{
    "pdf_info": [
        {
            "para_blocks": [
                {
                    "bbox": [
                        108,
                        103,
                        486,
                        121
                    ],
                    "type": "title",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                108,
                                103,
                                486,
                                121
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        108,
                                        103,
                                        486,
                                        121
                                    ],
                                    "type": "text",
                                    "content": "3D Packing for Self-Supervised Monocular Depth Estimation"
                                }
                            ]
                        }
                    ],
                    "index": 2
                },
                {
                    "bbox": [
                        91,
                        143,
                        502,
                        174
                    ],
                    "type": "text",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                91,
                                143,
                                502,
                                174
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        91,
                                        143,
                                        502,
                                        174
                                    ],
                                    "type": "text",
                                    "content": "Vitor Guizilini Rares Ambrus Sudeep Pillai Allan Raventos Adrien Gaidon Toyota Research Institute (TRI)"
                                }
                            ]
                        }
                    ],
                    "index": 3
                },
                {
                    "bbox": [
                        228,
                        176,
                        364,
                        186
                    ],
                    "type": "text",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                228,
                                176,
                                364,
                                186
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        228,
                                        176,
                                        364,
                                        186
                                    ],
                                    "type": "text",
                                    "content": "first这个名字@tri.global"
                                }
                            ]
                        }
                    ],
                    "index": 4
                },
                {
                    "bbox": [
                        143,
                        199,
                        192,
                        211
                    ],
                    "type": "title",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                143,
                                199,
                                192,
                                211
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        143,
                                        199,
                                        192,
                                        211
                                    ],
                                    "type": "text",
                                    "content": "Abstract"
                                }
                            ]
                        }
                    ],
                    "index": 5
                },
                {
                    "bbox": [
                        46,
                        219,
                        290,
                        470
                    ],
                    "type": "text",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                46,
                                219,
                                290,
                                470
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        46,
                                        219,
                                        290,
                                        470
                                    ],
                                    "type": "text",
                                    "content": "Although cameras are ubiquitous, robotic platforms typically rely on active sensors like LiDAR for direct 3D perception. In this work, we propose a novel self-supervised monocular depth estimation method combining geometry with a new deep network, PackNet, learned only from unlabeled monocular videos. Our architecture leverages novel symmetrical packing and unpacking blocks to jointly learn to compress and decompress detail-preserving representations using 3D convolutions. Although self-supervised, our method outperforms other self, semi, and fully supervised methods on the KITTI benchmark. The 3D inductive bias in PackNet enables it to scale with input resolution and number of parameters without overfitting, generalizing better on out-of-domain data such as the NuScenes dataset. Furthermore, it does not require large-scale supervised pretraining on ImageNet and can run in real-time. Finally, we release DDAD (Dense Depth for Automated Driving), a new urban driving dataset with more challenging and accurate depth evaluation, thanks to longer-range and denser ground-truth depth generated from high-density LiDARs mounted on a fleet of self-driving cars operating world-wide."
                                },
                                {
                                    "bbox": [
                                        46,
                                        219,
                                        290,
                                        470
                                    ],
                                    "type": "inline_equation",
                                    "content": "^{\\dagger}"
                                }
                            ]
                        }
                    ],
                    "index": 6
                },
                {
                    "bbox": [
                        47,
                        479,
                        128,
                        491
                    ],
                    "type": "title",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                47,
                                479,
                                128,
                                491
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        47,
                                        479,
                                        128,
                                        491
                                    ],
                                    "type": "text",
                                    "content": "1. Introduction"
                                }
                            ]
                        }
                    ],
                    "index": 7
                },
                {
                    "bbox": [
                        46,
                        496,
                        287,
                        662
                    ],
                    "type": "text",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                46,
                                496,
                                287,
                                662
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        46,
                                        496,
                                        287,
                                        662
                                    ],
                                    "type": "text",
                                    "content": "Accurate depth estimation is a key prerequisite in many robotics tasks, including perception, navigation, and planning. Depth from monocular camera configurations can provide useful cues for a wide array of tasks [23, 30, 34, 36], producing dense depth maps that could complement or eventually replace expensive range sensors. However, learning monocular depth via direct supervision requires ground-truth information from additional sensors and precise cross-calibration. Self-supervised methods do not suffer from these limitations, as they use geometrical constraints on image sequences as the sole source of supervision. In this work, we address the problem of jointly estimating scene structure and camera motion across RGB image sequences using a self-supervised deep network."
                                }
                            ]
                        }
                    ],
                    "index": 8
                },
                {
                    "bbox": [
                        59,
                        663,
                        287,
                        675
                    ],
                    "type": "text",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                59,
                                663,
                                287,
                                675
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        59,
                                        663,
                                        287,
                                        675
                                    ],
                                    "type": "text",
                                    "content": "While recent works in self-supervised monocular depth"
                                }
                            ]
                        }
                    ],
                    "index": 9
                },
                {
                    "type": "image",
                    "bbox": [
                        336,
                        202,
                        517,
                        407
                    ],
                    "blocks": [
                        {
                            "bbox": [
                                336,
                                202,
                                517,
                                407
                            ],
                            "lines": [
                                {
                                    "bbox": [
                                        336,
                                        202,
                                        517,
                                        407
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                336,
                                                202,
                                                517,
                                                407
                                            ],
                                            "type": "image",
                                            "image_path": "a2e623f975a8450811d159aa26df6e3056855de343f38fc1dd2a737649a44947.jpg"
                                        }
                                    ]
                                }
                            ],
                            "index": 10,
                            "angle": 0,
                            "type": "image_body"
                        },
                        {
                            "bbox": [
                                305,
                                415,
                                545,
                                441
                            ],
                            "lines": [
                                {
                                    "bbox": [
                                        305,
                                        415,
                                        545,
                                        441
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                305,
                                                415,
                                                545,
                                                441
                                            ],
                                            "type": "text",
                                            "content": "Figure 1: Example metrically accurate PackNet prediction (map and textured point cloud) on our DDAD dataset."
                                        }
                                    ]
                                }
                            ],
                            "index": 11,
                            "angle": 0,
                            "type": "image_caption"
                        }
                    ],
                    "index": 10
                },
                {
                    "bbox": [
                        304,
                        450,
                        546,
                        715
                    ],
                    "type": "text",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                304,
                                450,
                                546,
                                715
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        304,
                                        450,
                                        546,
                                        715
                                    ],
                                    "type": "text",
                                    "content": "estimation have mostly focused on engineering the loss function [5, 33, 47, 53], we show that performance critically depends on the model architecture, in line with the observations of [27] for other self-supervised tasks. Going beyond image classification models like ResNet [20], our main contribution is a new convolutional network architecture, called PackNet, for high-resolution self-supervised monocular depth estimation. We propose new packing and unpacking blocks that jointly leverage 3D convolutions to learn representations that maximally propagate dense appearance and geometric information while still being able to run in real time. Our second contribution is a novel loss that can optionally leverage the camera's velocity when available (e.g., from cars, robots, mobile phones) to solve the inherent scale ambiguity in monocular vision. Our third contribution is a new dataset: Dense Depth for Automated Driving (DDAD). It leverages diverse logs from a fleet of well-calibrated self-driving cars equipped with cameras and high-accuracy long-range LiDARs. Compared to existing benchmarks, DDAD enables much more accurate depth evaluation at range, which is key for high resolution monocular depth estimation methods (cf. Figure 1)."
                                }
                            ]
                        }
                    ],
                    "index": 12
                }
            ],
            "discarded_blocks": [
                {
                    "bbox": [
                        65,
                        2,
                        111,
                        34
                    ],
                    "type": "header",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                65,
                                2,
                                111,
                                34
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        65,
                                        2,
                                        111,
                                        34
                                    ],
                                    "type": "text",
                                    "content": "CVF"
                                }
                            ]
                        }
                    ],
                    "index": 0
                },
                {
                    "bbox": [
                        134,
                        0,
                        504,
                        37
                    ],
                    "type": "header",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                134,
                                0,
                                504,
                                37
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        134,
                                        0,
                                        504,
                                        37
                                    ],
                                    "type": "text",
                                    "content": "This CVPR 2020 paper is the Open Access version, provided by the Computer Vision Foundation. Except for this watermark, it is identical to the accepted version; the final published version of the proceedings is available on IEEE Xplore."
                                }
                            ]
                        }
                    ],
                    "index": 1
                },
                {
                    "bbox": [
                        61,
                        682,
                        255,
                        693
                    ],
                    "type": "page_footnote",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                61,
                                682,
                                255,
                                693
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        61,
                                        682,
                                        255,
                                        693
                                    ],
                                    "type": "text",
                                    "content": "†Video: https://www.youtube.com/watch?v=b62iDkLgGSI"
                                }
                            ]
                        }
                    ],
                    "index": 13
                },
                {
                    "bbox": [
                        63,
                        693,
                        211,
                        703
                    ],
                    "type": "page_footnote",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                63,
                                693,
                                211,
                                703
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        63,
                                        693,
                                        211,
                                        703
                                    ],
                                    "type": "inline_equation",
                                    "content": "\\dagger"
                                },
                                {
                                    "bbox": [
                                        63,
                                        693,
                                        211,
                                        703
                                    ],
                                    "type": "text",
                                    "content": "Dataset: https://github.com/TRI-ML/DDAD"
                                }
                            ]
                        }
                    ],
                    "index": 14
                },
                {
                    "bbox": [
                        63,
                        703,
                        220,
                        712
                    ],
                    "type": "page_footnote",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                63,
                                703,
                                220,
                                712
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        63,
                                        703,
                                        220,
                                        712
                                    ],
                                    "type": "text",
                                    "content": "†Code: https://github.com/TRI-ML/packnet-sfm"
                                }
                            ]
                        }
                    ],
                    "index": 15
                },
                {
                    "bbox": [
                        295,
                        734,
                        315,
                        742
                    ],
                    "type": "page_number",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                295,
                                734,
                                315,
                                742
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        295,
                                        734,
                                        315,
                                        742
                                    ],
                                    "type": "text",
                                    "content": "2485"
                                }
                            ]
                        }
                    ],
                    "index": 17
                }
            ],
            "page_size": [
                612,
                792
            ],
            "page_idx": 0
        },
        {
            "para_blocks": [
                {
                    "bbox": [
                        46,
                        72,
                        289,
                        205
                    ],
                    "type": "text",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                46,
                                72,
                                289,
                                205
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        46,
                                        72,
                                        289,
                                        205
                                    ],
                                    "type": "text",
                                    "content": "Our experiments on the standard KITTI benchmark [16], the recent NuScenes dataset [4], and our new proposed DDAD benchmark show that our self-supervised monocular approach "
                                },
                                {
                                    "bbox": [
                                        46,
                                        72,
                                        289,
                                        205
                                    ],
                                    "type": "inline_equation",
                                    "content": "i)"
                                },
                                {
                                    "bbox": [
                                        46,
                                        72,
                                        289,
                                        205
                                    ],
                                    "type": "text",
                                    "content": " improves on the state of the art, especially at longer ranges; "
                                },
                                {
                                    "bbox": [
                                        46,
                                        72,
                                        289,
                                        205
                                    ],
                                    "type": "inline_equation",
                                    "content": "ii)"
                                },
                                {
                                    "bbox": [
                                        46,
                                        72,
                                        289,
                                        205
                                    ],
                                    "type": "text",
                                    "content": " is competitive with fully supervised methods; "
                                },
                                {
                                    "bbox": [
                                        46,
                                        72,
                                        289,
                                        205
                                    ],
                                    "type": "inline_equation",
                                    "content": "iii)"
                                },
                                {
                                    "bbox": [
                                        46,
                                        72,
                                        289,
                                        205
                                    ],
                                    "type": "text",
                                    "content": " generalizes better on unseen data; "
                                },
                                {
                                    "bbox": [
                                        46,
                                        72,
                                        289,
                                        205
                                    ],
                                    "type": "inline_equation",
                                    "content": "iv)"
                                },
                                {
                                    "bbox": [
                                        46,
                                        72,
                                        289,
                                        205
                                    ],
                                    "type": "text",
                                    "content": " scales better with number of parameters, input resolution, and more unlabeled training data; "
                                },
                                {
                                    "bbox": [
                                        46,
                                        72,
                                        289,
                                        205
                                    ],
                                    "type": "inline_equation",
                                    "content": "v)"
                                },
                                {
                                    "bbox": [
                                        46,
                                        72,
                                        289,
                                        205
                                    ],
                                    "type": "text",
                                    "content": " can run in real time at high resolution; and "
                                },
                                {
                                    "bbox": [
                                        46,
                                        72,
                                        289,
                                        205
                                    ],
                                    "type": "inline_equation",
                                    "content": "vi)"
                                },
                                {
                                    "bbox": [
                                        46,
                                        72,
                                        289,
                                        205
                                    ],
                                    "type": "text",
                                    "content": " does not require supervised pretraining on ImageNet to achieve state-of-the-art results; or test-time ground-truth scaling if velocity information is available at training time."
                                }
                            ]
                        }
                    ],
                    "index": 0
                },
                {
                    "bbox": [
                        47,
                        216,
                        134,
                        228
                    ],
                    "type": "title",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                47,
                                216,
                                134,
                                228
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        47,
                                        216,
                                        134,
                                        228
                                    ],
                                    "type": "text",
                                    "content": "2. Related Work"
                                }
                            ]
                        }
                    ],
                    "index": 1
                },
                {
                    "bbox": [
                        46,
                        236,
                        288,
                        299
                    ],
                    "type": "text",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                46,
                                236,
                                288,
                                299
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        46,
                                        236,
                                        288,
                                        299
                                    ],
                                    "type": "text",
                                    "content": "Depth estimation from a single image poses several challenges due to its ill-posed and ambiguous nature. However, modern convolutional networks have shown that it is possible to successfully leverage appearance-based patterns in large scale datasets in order to make accurate predictions."
                                }
                            ]
                        }
                    ],
                    "index": 2
                },
                {
                    "bbox": [
                        46,
                        313,
                        287,
                        445
                    ],
                    "type": "text",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                46,
                                313,
                                287,
                                445
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        46,
                                        313,
                                        287,
                                        445
                                    ],
                                    "type": "text",
                                    "content": "Depth Network Architectures Eigen et al. [13] proposed one of the earliest works in convolutional-based depth estimation using a multi-scale deep network trained on RGB-D sensor data to regress the depth directly from single images. Subsequent works extended these network architectures to perform two-view stereo disparity estimation [35] using techniques developed in the flow estimation literature [12]. Following [12, 35], Umenhofer et al. [42] applied these concepts to simultaneously train a depth and pose network to predict depth and camera ego-motion between successive unconstrained image pairs."
                                }
                            ]
                        }
                    ],
                    "index": 3
                },
                {
                    "bbox": [
                        46,
                        445,
                        288,
                        650
                    ],
                    "type": "text",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                46,
                                445,
                                288,
                                650
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        46,
                                        445,
                                        288,
                                        650
                                    ],
                                    "type": "text",
                                    "content": "Independently, dense pixel-prediction networks [2, 31, 48] have made significant progress towards improving the flow of information between encoding and decoding layers. Fractional pooling [19] was introduced to amortize the rapid spatial reduction during downsampling. Lee et al. [29] generalized the pooling function to allow the learning of more complex patterns, including linear combinations and learnable pooling operations. Shi et al. [39] used sub-pixel convolutions to perform Single-Image-Super-Resolution, synthesizing and super-resolving images beyond their input resolutions, while still operating at lower resolutions. Recent works [38, 51] in self-supervised monocular depth estimation use this concept to super-resolve estimates and further improve performance. Here, we go one step further and introduce new operations relying on 3D convolutions for learning to preserve and process spatial information in the features of encoding and decoding layers."
                                }
                            ]
                        }
                    ],
                    "index": 4
                },
                {
                    "bbox": [
                        46,
                        665,
                        288,
                        715
                    ],
                    "type": "text",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                46,
                                665,
                                288,
                                715
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        46,
                                        665,
                                        288,
                                        715
                                    ],
                                    "type": "text",
                                    "content": "Self-Supervised Monocular Depth and Pose As supervised techniques for depth estimation advanced rapidly, the availability of target depth labels became challenging, especially for outdoor applications. To this end, [15, 17] pro"
                                }
                            ]
                        }
                    ],
                    "index": 5
                },
                {
                    "bbox": [
                        304,
                        72,
                        545,
                        204
                    ],
                    "type": "text",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                304,
                                72,
                                545,
                                204
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        304,
                                        72,
                                        545,
                                        204
                                    ],
                                    "type": "text",
                                    "content": "vided an alternative strategy involving training a monocular depth network with stereo cameras, without requiring ground-truth depth labels. By leveraging Spatial Transformer Networks [22], Godard et al [17] use stereo imagery to geometrically transform the right image plus a predicted depth of the left image into a synthesized left image. The loss between the resulting synthesized and original left images is then defined in a fully-differentiable manner, using a Structural Similarity [44] term and additional depth regularization terms, thus allowing the depth network to be self-supervised in an end-to-end fashion."
                                }
                            ]
                        }
                    ],
                    "index": 6
                },
                {
                    "bbox": [
                        304,
                        210,
                        546,
                        402
                    ],
                    "type": "text",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                304,
                                210,
                                546,
                                402
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        304,
                                        210,
                                        546,
                                        402
                                    ],
                                    "type": "text",
                                    "content": "Following [17] and [42], Zhou et al. [52] generalize this to self-supervised training in the purely monocular setting, where a depth and pose network are simultaneously learned from unlabeled monocular videos. Several methods [5, 26, 33, 43, 46, 47, 51, 53] have advanced this line terms, of work by incorporating these methods, ad,ditional loss and constraints. All, however, take advantage of constraints in monocular Structure-from-Motion (SfM) training that only allow the estimation of depth and pose up to an unknown scale factor, and rely on the ground-truth LiDAR measu, rements to scale their depth estimates appropriately for evaluation purposes [52]. Instead, in this work we show that, by simply using the instantaneous velocity of the camera during training, we are able to learn a scale-aware depth and pose model, alleviating the impractical need to use Li-DAR ground-truth depth measurements at test-time."
                                }
                            ]
                        }
                    ],
                    "index": 7
                },
                {
                    "bbox": [
                        305,
                        412,
                        493,
                        426
                    ],
                    "type": "title",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                305,
                                412,
                                493,
                                426
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        305,
                                        412,
                                        493,
                                        426
                                    ],
                                    "type": "text",
                                    "content": "3. Self-Supervised Scale-Aware SfM"
                                }
                            ]
                        }
                    ],
                    "index": 8
                },
                {
                    "bbox": [
                        304,
                        434,
                        547,
                        590
                    ],
                    "type": "text",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                304,
                                434,
                                547,
                                590
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        304,
                                        434,
                                        547,
                                        590
                                    ],
                                    "type": "text",
                                    "content": "In self-supervised monocular SfM training (Fig. 2), we aim to learn: (i) a monocular depth model "
                                },
                                {
                                    "bbox": [
                                        304,
                                        434,
                                        547,
                                        590
                                    ],
                                    "type": "inline_equation",
                                    "content": "f_{D}: I \\to D"
                                },
                                {
                                    "bbox": [
                                        304,
                                        434,
                                        547,
                                        590
                                    ],
                                    "type": "text",
                                    "content": ", that predicts the scale-ambiguous depth "
                                },
                                {
                                    "bbox": [
                                        304,
                                        434,
                                        547,
                                        590
                                    ],
                                    "type": "inline_equation",
                                    "content": "\\hat{D} = f_{D}(I(p))"
                                },
                                {
                                    "bbox": [
                                        304,
                                        434,
                                        547,
                                        590
                                    ],
                                    "type": "text",
                                    "content": " for every pixel "
                                },
                                {
                                    "bbox": [
                                        304,
                                        434,
                                        547,
                                        590
                                    ],
                                    "type": "inline_equation",
                                    "content": "p"
                                },
                                {
                                    "bbox": [
                                        304,
                                        434,
                                        547,
                                        590
                                    ],
                                    "type": "text",
                                    "content": " in the target image "
                                },
                                {
                                    "bbox": [
                                        304,
                                        434,
                                        547,
                                        590
                                    ],
                                    "type": "inline_equation",
                                    "content": "I"
                                },
                                {
                                    "bbox": [
                                        304,
                                        434,
                                        547,
                                        590
                                    ],
                                    "type": "text",
                                    "content": "; and (ii) a monocular ego-motion estimator "
                                },
                                {
                                    "bbox": [
                                        304,
                                        434,
                                        547,
                                        590
                                    ],
                                    "type": "inline_equation",
                                    "content": "f_{\\mathbf{x}}: (I_t, I_S) \\to \\mathbf{x}_{t \\to S}"
                                },
                                {
                                    "bbox": [
                                        304,
                                        434,
                                        547,
                                        590
                                    ],
                                    "type": "text",
                                    "content": ", that predicts the set of 6-DoF rigid transformations for all "
                                },
                                {
                                    "bbox": [
                                        304,
                                        434,
                                        547,
                                        590
                                    ],
                                    "type": "inline_equation",
                                    "content": "s \\in S"
                                },
                                {
                                    "bbox": [
                                        304,
                                        434,
                                        547,
                                        590
                                    ],
                                    "type": "text",
                                    "content": " given by "
                                },
                                {
                                    "bbox": [
                                        304,
                                        434,
                                        547,
                                        590
                                    ],
                                    "type": "inline_equation",
                                    "content": "\\mathbf{x}_{t \\to s} = \\begin{pmatrix} \\mathbf{R} & \\mathbf{t} \\\\ \\mathbf{0} & \\mathbf{1} \\end{pmatrix} \\in \\mathrm{SE}(3)"
                                },
                                {
                                    "bbox": [
                                        304,
                                        434,
                                        547,
                                        590
                                    ],
                                    "type": "text",
                                    "content": ", between the target image "
                                },
                                {
                                    "bbox": [
                                        304,
                                        434,
                                        547,
                                        590
                                    ],
                                    "type": "inline_equation",
                                    "content": "I_t"
                                },
                                {
                                    "bbox": [
                                        304,
                                        434,
                                        547,
                                        590
                                    ],
                                    "type": "text",
                                    "content": " and the set of source images "
                                },
                                {
                                    "bbox": [
                                        304,
                                        434,
                                        547,
                                        590
                                    ],
                                    "type": "inline_equation",
                                    "content": "I_s \\in I_S"
                                },
                                {
                                    "bbox": [
                                        304,
                                        434,
                                        547,
                                        590
                                    ],
                                    "type": "text",
                                    "content": " considered as part of the temporal context. In practice, we use the frames "
                                },
                                {
                                    "bbox": [
                                        304,
                                        434,
                                        547,
                                        590
                                    ],
                                    "type": "inline_equation",
                                    "content": "I_{t-1}"
                                },
                                {
                                    "bbox": [
                                        304,
                                        434,
                                        547,
                                        590
                                    ],
                                    "type": "text",
                                    "content": " and "
                                },
                                {
                                    "bbox": [
                                        304,
                                        434,
                                        547,
                                        590
                                    ],
                                    "type": "inline_equation",
                                    "content": "I_{t+1}"
                                },
                                {
                                    "bbox": [
                                        304,
                                        434,
                                        547,
                                        590
                                    ],
                                    "type": "text",
                                    "content": " as source images, although using a larger context is possible. Note that in the case of monocular SfM both depth and pose are estimated up to an unknown scale factor, due to the inherent ambiguity of the photometric loss."
                                }
                            ]
                        }
                    ],
                    "index": 9
                },
                {
                    "bbox": [
                        305,
                        597,
                        451,
                        611
                    ],
                    "type": "title",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                305,
                                597,
                                451,
                                611
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        305,
                                        597,
                                        451,
                                        611
                                    ],
                                    "type": "text",
                                    "content": "3.1. Self-Supervised Objective"
                                }
                            ]
                        }
                    ],
                    "index": 10
                },
                {
                    "bbox": [
                        304,
                        616,
                        547,
                        715
                    ],
                    "type": "text",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                304,
                                616,
                                547,
                                715
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        304,
                                        616,
                                        547,
                                        715
                                    ],
                                    "type": "text",
                                    "content": "Following the work of Zhou et al. [52], we train the depth and pose network simultaneously in a self-supervised manner. In this work, however, we learn to recover the inverse-depth "
                                },
                                {
                                    "bbox": [
                                        304,
                                        616,
                                        547,
                                        715
                                    ],
                                    "type": "inline_equation",
                                    "content": "f_{d}: I \\to f_{D}^{-1}(I)"
                                },
                                {
                                    "bbox": [
                                        304,
                                        616,
                                        547,
                                        715
                                    ],
                                    "type": "text",
                                    "content": " instead, along with the ego-motion estimator "
                                },
                                {
                                    "bbox": [
                                        304,
                                        616,
                                        547,
                                        715
                                    ],
                                    "type": "inline_equation",
                                    "content": "f_{\\mathbf{x}}"
                                },
                                {
                                    "bbox": [
                                        304,
                                        616,
                                        547,
                                        715
                                    ],
                                    "type": "text",
                                    "content": ". Similar to [52], the overall self-supervised objective consists of an appearance matching loss term "
                                },
                                {
                                    "bbox": [
                                        304,
                                        616,
                                        547,
                                        715
                                    ],
                                    "type": "inline_equation",
                                    "content": "\\mathcal{L}_p"
                                },
                                {
                                    "bbox": [
                                        304,
                                        616,
                                        547,
                                        715
                                    ],
                                    "type": "text",
                                    "content": " that is imposed between the synthesized target image "
                                },
                                {
                                    "bbox": [
                                        304,
                                        616,
                                        547,
                                        715
                                    ],
                                    "type": "inline_equation",
                                    "content": "\\hat{I}_t"
                                },
                                {
                                    "bbox": [
                                        304,
                                        616,
                                        547,
                                        715
                                    ],
                                    "type": "text",
                                    "content": " and the target image "
                                },
                                {
                                    "bbox": [
                                        304,
                                        616,
                                        547,
                                        715
                                    ],
                                    "type": "inline_equation",
                                    "content": "I_t"
                                },
                                {
                                    "bbox": [
                                        304,
                                        616,
                                        547,
                                        715
                                    ],
                                    "type": "text",
                                    "content": ", and a depth regularization term "
                                },
                                {
                                    "bbox": [
                                        304,
                                        616,
                                        547,
                                        715
                                    ],
                                    "type": "inline_equation",
                                    "content": "\\mathcal{L}_s"
                                },
                                {
                                    "bbox": [
                                        304,
                                        616,
                                        547,
                                        715
                                    ],
                                    "type": "text",
                                    "content": " that"
                                }
                            ]
                        }
                    ],
                    "index": 11
                }
            ],
            "discarded_blocks": [
                {
                    "bbox": [
                        295,
                        733,
                        316,
                        742
                    ],
                    "type": "page_number",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                295,
                                733,
                                316,
                                742
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        295,
                                        733,
                                        316,
                                        742
                                    ],
                                    "type": "text",
                                    "content": "2486"
                                }
                            ]
                        }
                    ],
                    "index": 12
                }
            ],
            "page_size": [
                612,
                792
            ],
            "page_idx": 1
        },
        {
            "para_blocks": [
                {
                    "bbox": [
                        47,
                        71,
                        286,
                        96
                    ],
                    "type": "text",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                47,
                                71,
                                286,
                                96
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        47,
                                        71,
                                        286,
                                        96
                                    ],
                                    "type": "text",
                                    "content": "ensures edge-aware smoothing in the depth estimates "
                                },
                                {
                                    "bbox": [
                                        47,
                                        71,
                                        286,
                                        96
                                    ],
                                    "type": "inline_equation",
                                    "content": "\\hat{D}_t"
                                },
                                {
                                    "bbox": [
                                        47,
                                        71,
                                        286,
                                        96
                                    ],
                                    "type": "text",
                                    "content": ". The objective takes the following form:"
                                }
                            ]
                        }
                    ],
                    "index": 0
                },
                {
                    "bbox": [
                        59,
                        102,
                        287,
                        118
                    ],
                    "type": "interline_equation",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                59,
                                102,
                                287,
                                118
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        59,
                                        102,
                                        287,
                                        118
                                    ],
                                    "type": "interline_equation",
                                    "content": "\\mathcal {L} \\left(I _ {t}, \\hat {I} _ {t}\\right) = \\mathcal {L} _ {p} \\left(I _ {t}, I _ {S}\\right) \\odot \\mathcal {M} _ {p} \\odot \\mathcal {M} _ {t} + \\lambda_ {1} \\mathcal {L} _ {s} (\\hat {D} _ {t}) \\tag {1}",
                                    "image_path": "ff7bb6a9df59da09d1b4b1c01f8f7e51c84ccb66c4d5de6e3899fc218b67d9f4.jpg"
                                }
                            ]
                        }
                    ],
                    "index": 1
                },
                {
                    "bbox": [
                        47,
                        128,
                        287,
                        212
                    ],
                    "type": "text",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                47,
                                128,
                                287,
                                212
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        47,
                                        128,
                                        287,
                                        212
                                    ],
                                    "type": "text",
                                    "content": "where "
                                },
                                {
                                    "bbox": [
                                        47,
                                        128,
                                        287,
                                        212
                                    ],
                                    "type": "inline_equation",
                                    "content": "\\mathcal{M}_t"
                                },
                                {
                                    "bbox": [
                                        47,
                                        128,
                                        287,
                                        212
                                    ],
                                    "type": "text",
                                    "content": " is a binary mask that avoids computing the photometric loss on the pixels that do not have a valid mapping, and "
                                },
                                {
                                    "bbox": [
                                        47,
                                        128,
                                        287,
                                        212
                                    ],
                                    "type": "inline_equation",
                                    "content": "\\odot"
                                },
                                {
                                    "bbox": [
                                        47,
                                        128,
                                        287,
                                        212
                                    ],
                                    "type": "text",
                                    "content": " denotes element-wise multiplication. Additionally, "
                                },
                                {
                                    "bbox": [
                                        47,
                                        128,
                                        287,
                                        212
                                    ],
                                    "type": "inline_equation",
                                    "content": "\\lambda_1"
                                },
                                {
                                    "bbox": [
                                        47,
                                        128,
                                        287,
                                        212
                                    ],
                                    "type": "text",
                                    "content": " enforces a weighted depth regularization on the objective. The overall loss in Equation 1 is averaged per-pixel, pyramid-scale and image batch during training. Fig. 2 shows a high-level overview of our training pipeline."
                                }
                            ]
                        }
                    ],
                    "index": 2
                },
                {
                    "bbox": [
                        47,
                        218,
                        287,
                        289
                    ],
                    "type": "text",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                47,
                                218,
                                287,
                                289
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        47,
                                        218,
                                        287,
                                        289
                                    ],
                                    "type": "text",
                                    "content": "Appearance Matching Loss. Following [17, 52] the pixel-level similarity between the target image "
                                },
                                {
                                    "bbox": [
                                        47,
                                        218,
                                        287,
                                        289
                                    ],
                                    "type": "inline_equation",
                                    "content": "I_{t}"
                                },
                                {
                                    "bbox": [
                                        47,
                                        218,
                                        287,
                                        289
                                    ],
                                    "type": "text",
                                    "content": " and the synthesized target image "
                                },
                                {
                                    "bbox": [
                                        47,
                                        218,
                                        287,
                                        289
                                    ],
                                    "type": "inline_equation",
                                    "content": "\\hat{I}_{t}"
                                },
                                {
                                    "bbox": [
                                        47,
                                        218,
                                        287,
                                        289
                                    ],
                                    "type": "text",
                                    "content": " is estimated using the Structural Similarity (SSIM) [44] term combined with an L1 pixel-wise loss term, inducing an overall photometric loss given by Equation 2 below."
                                }
                            ]
                        }
                    ],
                    "index": 3
                },
                {
                    "bbox": [
                        55,
                        297,
                        287,
                        316
                    ],
                    "type": "interline_equation",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                55,
                                297,
                                287,
                                316
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        55,
                                        297,
                                        287,
                                        316
                                    ],
                                    "type": "interline_equation",
                                    "content": "\\mathcal {L} _ {p} \\left(I _ {t}, \\hat {I} _ {t}\\right) = \\alpha \\frac {1 - \\operatorname {S S I M} \\left(I _ {t} , \\hat {I} _ {t}\\right)}{2} + (1 - \\alpha) \\| I _ {t} - \\hat {I} _ {t} \\| \\tag {2}",
                                    "image_path": "00b3666914656769673aaafc2cac65893d48dfbf3523f92d5b45cff07dad45e7.jpg"
                                }
                            ]
                        }
                    ],
                    "index": 4
                },
                {
                    "bbox": [
                        47,
                        324,
                        287,
                        396
                    ],
                    "type": "text",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                47,
                                324,
                                287,
                                396
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        47,
                                        324,
                                        287,
                                        396
                                    ],
                                    "type": "text",
                                    "content": "While multi-view projective geometry provides strong cues for self-supervision, errors due to parallax in the scene have an undesirable effect incurred on the photometric loss. We mitigate these undesirable effects by calculating the minimum photometric loss per pixel for each source image in the context "
                                },
                                {
                                    "bbox": [
                                        47,
                                        324,
                                        287,
                                        396
                                    ],
                                    "type": "inline_equation",
                                    "content": "I_{S}"
                                },
                                {
                                    "bbox": [
                                        47,
                                        324,
                                        287,
                                        396
                                    ],
                                    "type": "text",
                                    "content": ", as shown in [18], so that:"
                                }
                            ]
                        }
                    ],
                    "index": 5
                },
                {
                    "bbox": [
                        108,
                        407,
                        287,
                        425
                    ],
                    "type": "interline_equation",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                108,
                                407,
                                287,
                                425
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        108,
                                        407,
                                        287,
                                        425
                                    ],
                                    "type": "interline_equation",
                                    "content": "\\mathcal {L} _ {p} \\left(I _ {t}, I _ {S}\\right) = \\min  _ {I _ {S}} \\mathcal {L} _ {p} \\left(I _ {t}, \\hat {I} _ {t}\\right) \\tag {3}",
                                    "image_path": "6793faf702467a679736c903e1db32884dac09c004cacc69b01bf7c2ae8901f1.jpg"
                                }
                            ]
                        }
                    ],
                    "index": 6
                },
                {
                    "bbox": [
                        47,
                        435,
                        287,
                        579
                    ],
                    "type": "text",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                47,
                                435,
                                287,
                                579
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        47,
                                        435,
                                        287,
                                        579
                                    ],
                                    "type": "text",
                                    "content": "The intuition is that the same pixel will not be occluded or out-of-bounds in all context images, and that the association with minimal photometric loss should be the correct one. Furthermore, we also mask out static pixels by removing those which have a warped photometric loss "
                                },
                                {
                                    "bbox": [
                                        47,
                                        435,
                                        287,
                                        579
                                    ],
                                    "type": "inline_equation",
                                    "content": "\\mathcal{L}_p(I_t,\\hat{I}_t)"
                                },
                                {
                                    "bbox": [
                                        47,
                                        435,
                                        287,
                                        579
                                    ],
                                    "type": "text",
                                    "content": " higher than their corresponding unwarped photometric loss "
                                },
                                {
                                    "bbox": [
                                        47,
                                        435,
                                        287,
                                        579
                                    ],
                                    "type": "inline_equation",
                                    "content": "\\mathcal{L}_p(I_t,I_s)"
                                },
                                {
                                    "bbox": [
                                        47,
                                        435,
                                        287,
                                        579
                                    ],
                                    "type": "text",
                                    "content": ", calculated using the original source image without view synthesis. Introduced in [18], this auto-mask removes pixels whose appearance does not change between frames, which includes static scenes and dynamic objects with no relative motion, since these will have a smaller photometric loss when assuming no ego-motion."
                                }
                            ]
                        }
                    ],
                    "index": 7
                },
                {
                    "bbox": [
                        85,
                        589,
                        287,
                        608
                    ],
                    "type": "interline_equation",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                85,
                                589,
                                287,
                                608
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        85,
                                        589,
                                        287,
                                        608
                                    ],
                                    "type": "interline_equation",
                                    "content": "\\mathcal {M} _ {p} = \\min  _ {I _ {S}} \\mathcal {L} _ {p} \\left(I _ {t}, I _ {s}\\right) > \\min  _ {I _ {S}} \\mathcal {L} _ {p} \\left(I _ {t}, \\hat {I} _ {t}\\right) \\tag {4}",
                                    "image_path": "1b8ecbcf4550571b00617fb85eb543e1ca1a205284be455e8a00ddc12fc8c3b9.jpg"
                                }
                            ]
                        }
                    ],
                    "index": 8
                },
                {
                    "bbox": [
                        47,
                        618,
                        287,
                        689
                    ],
                    "type": "text",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                47,
                                618,
                                287,
                                689
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        47,
                                        618,
                                        287,
                                        689
                                    ],
                                    "type": "text",
                                    "content": "Depth Smoothness Loss. In order to regularize the depth in texture-less low-image gradient regions, we incorporate an edge-aware term (Equation 5), similar to [17]. The loss is weighted for each of the pyramid-levels, and is decayed by a factor of 2 on down-sampling, starting with a weight of 1 for the "
                                },
                                {
                                    "bbox": [
                                        47,
                                        618,
                                        287,
                                        689
                                    ],
                                    "type": "inline_equation",
                                    "content": "0^{\\text{th}}"
                                },
                                {
                                    "bbox": [
                                        47,
                                        618,
                                        287,
                                        689
                                    ],
                                    "type": "text",
                                    "content": " pyramid level."
                                }
                            ]
                        }
                    ],
                    "index": 9
                },
                {
                    "bbox": [
                        80,
                        699,
                        287,
                        715
                    ],
                    "type": "interline_equation",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                80,
                                699,
                                287,
                                715
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        80,
                                        699,
                                        287,
                                        715
                                    ],
                                    "type": "interline_equation",
                                    "content": "\\mathcal {L} _ {s} (\\hat {D} _ {t}) = \\left| \\delta_ {x} \\hat {D} _ {t} \\right| e ^ {- \\left| \\delta_ {x} I _ {t} \\right|} + \\left| \\delta_ {y} \\hat {D} _ {t} \\right| e ^ {- \\left| \\delta_ {y} I _ {t} \\right|} \\tag {5}",
                                    "image_path": "48cd304e9cd585f42e6e2eb687fb6f24b1474134beb3b750369310c51b84db96.jpg"
                                }
                            ]
                        }
                    ],
                    "index": 10
                },
                {
                    "type": "image",
                    "bbox": [
                        328,
                        69,
                        523,
                        209
                    ],
                    "blocks": [
                        {
                            "bbox": [
                                328,
                                69,
                                523,
                                209
                            ],
                            "lines": [
                                {
                                    "bbox": [
                                        328,
                                        69,
                                        523,
                                        209
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                328,
                                                69,
                                                523,
                                                209
                                            ],
                                            "type": "image",
                                            "image_path": "fe05f02d46e4415aa192fc280dce3c735566ef15139602b3608dc9414958cde2.jpg"
                                        }
                                    ]
                                }
                            ],
                            "index": 11,
                            "angle": 0,
                            "type": "image_body"
                        },
                        {
                            "bbox": [
                                305,
                                217,
                                545,
                                277
                            ],
                            "lines": [
                                {
                                    "bbox": [
                                        305,
                                        217,
                                        545,
                                        277
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                305,
                                                217,
                                                545,
                                                277
                                            ],
                                            "type": "text",
                                            "content": "Figure 2: PackNet-SfM: Our proposed scale-aware self-supervised monocular structure-from-motion architecture. We introduce PackNet as a novel depth network, and optionally include weak velocity supervision at training time to produce scale-aware depth and pose models."
                                        }
                                    ]
                                }
                            ],
                            "index": 12,
                            "angle": 0,
                            "type": "image_caption"
                        }
                    ],
                    "index": 11
                },
                {
                    "bbox": [
                        306,
                        287,
                        409,
                        298
                    ],
                    "type": "title",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                306,
                                287,
                                409,
                                298
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        306,
                                        287,
                                        409,
                                        298
                                    ],
                                    "type": "text",
                                    "content": "3.2. Scale-Aware SfM"
                                }
                            ]
                        }
                    ],
                    "index": 13
                },
                {
                    "bbox": [
                        305,
                        305,
                        545,
                        449
                    ],
                    "type": "text",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                305,
                                305,
                                545,
                                449
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        305,
                                        305,
                                        545,
                                        449
                                    ],
                                    "type": "text",
                                    "content": "As previously mentioned, both the monocular depth and ego-motion estimators "
                                },
                                {
                                    "bbox": [
                                        305,
                                        305,
                                        545,
                                        449
                                    ],
                                    "type": "inline_equation",
                                    "content": "f_{d}"
                                },
                                {
                                    "bbox": [
                                        305,
                                        305,
                                        545,
                                        449
                                    ],
                                    "type": "text",
                                    "content": " and "
                                },
                                {
                                    "bbox": [
                                        305,
                                        305,
                                        545,
                                        449
                                    ],
                                    "type": "inline_equation",
                                    "content": "f_{\\mathbf{x}}"
                                },
                                {
                                    "bbox": [
                                        305,
                                        305,
                                        545,
                                        449
                                    ],
                                    "type": "text",
                                    "content": " predict scale-ambiguous values, due to the limitations of the monocular SfM training objective. In other words, the scene depth and the camera ego-motion can only be estimated up to an unknown and ambiguous scale factor. This is also reflected in the overall learning objective, where the photometric loss is agnostic to the metric depth of the scene. Furthermore, we note that all previous approaches which operate in the self-supervised monocular regime [5, 15, 17, 33] suffer from this limitation, and resort to artificially incorporating this scale factor at test-time, using LiDAR measurements."
                                }
                            ]
                        }
                    ],
                    "index": 14
                },
                {
                    "bbox": [
                        305,
                        454,
                        545,
                        574
                    ],
                    "type": "text",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                305,
                                454,
                                545,
                                574
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        305,
                                        454,
                                        545,
                                        574
                                    ],
                                    "type": "text",
                                    "content": "Velocity Supervision Loss. Since instantaneous velocity measurements are ubiquitous in most mobile systems today, we show that they can be directly incorporated in our self-supervised objective to learn a metrically accurate and scale-aware monocular depth estimator. During training, we impose an additional loss "
                                },
                                {
                                    "bbox": [
                                        305,
                                        454,
                                        545,
                                        574
                                    ],
                                    "type": "inline_equation",
                                    "content": "\\mathcal{L}_v"
                                },
                                {
                                    "bbox": [
                                        305,
                                        454,
                                        545,
                                        574
                                    ],
                                    "type": "text",
                                    "content": " between the magnitude of the pose-translation component of the pose network prediction "
                                },
                                {
                                    "bbox": [
                                        305,
                                        454,
                                        545,
                                        574
                                    ],
                                    "type": "inline_equation",
                                    "content": "\\hat{\\mathbf{t}}"
                                },
                                {
                                    "bbox": [
                                        305,
                                        454,
                                        545,
                                        574
                                    ],
                                    "type": "text",
                                    "content": " and the measured instantaneous velocity scalar "
                                },
                                {
                                    "bbox": [
                                        305,
                                        454,
                                        545,
                                        574
                                    ],
                                    "type": "inline_equation",
                                    "content": "v"
                                },
                                {
                                    "bbox": [
                                        305,
                                        454,
                                        545,
                                        574
                                    ],
                                    "type": "text",
                                    "content": " multiplied by the time difference between target and source frames "
                                },
                                {
                                    "bbox": [
                                        305,
                                        454,
                                        545,
                                        574
                                    ],
                                    "type": "inline_equation",
                                    "content": "\\Delta T_{t\\rightarrow s}"
                                },
                                {
                                    "bbox": [
                                        305,
                                        454,
                                        545,
                                        574
                                    ],
                                    "type": "text",
                                    "content": ", as shown below:"
                                }
                            ]
                        }
                    ],
                    "index": 15
                },
                {
                    "bbox": [
                        348,
                        577,
                        545,
                        597
                    ],
                    "type": "interline_equation",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                348,
                                577,
                                545,
                                597
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        348,
                                        577,
                                        545,
                                        597
                                    ],
                                    "type": "interline_equation",
                                    "content": "\\mathcal {L} _ {v} \\left(\\hat {\\mathbf {t}} _ {t \\rightarrow s}, v\\right) = \\left| \\| \\hat {\\mathbf {t}} _ {t \\rightarrow s} \\| - | v | \\Delta T _ {t \\rightarrow s} \\right| \\tag {6}",
                                    "image_path": "f9a9d42d9a5904521cc2b2ce730b2c0b849b17e207e3e0a8bf2ad3ae75c7e6de.jpg"
                                }
                            ]
                        }
                    ],
                    "index": 16
                },
                {
                    "bbox": [
                        305,
                        603,
                        544,
                        626
                    ],
                    "type": "text",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                305,
                                603,
                                544,
                                626
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        305,
                                        603,
                                        544,
                                        626
                                    ],
                                    "type": "text",
                                    "content": "Our final scale-aware self-supervised objective loss "
                                },
                                {
                                    "bbox": [
                                        305,
                                        603,
                                        544,
                                        626
                                    ],
                                    "type": "inline_equation",
                                    "content": "\\mathcal{L}_{\\mathrm{scale}}"
                                },
                                {
                                    "bbox": [
                                        305,
                                        603,
                                        544,
                                        626
                                    ],
                                    "type": "text",
                                    "content": " from Equation 1 becomes:"
                                }
                            ]
                        }
                    ],
                    "index": 17
                },
                {
                    "bbox": [
                        333,
                        632,
                        545,
                        647
                    ],
                    "type": "interline_equation",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                333,
                                632,
                                545,
                                647
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        333,
                                        632,
                                        545,
                                        647
                                    ],
                                    "type": "interline_equation",
                                    "content": "\\mathcal {L} _ {\\text {s c a l e}} \\left(I _ {t}, \\hat {I} _ {t}, v\\right) = \\mathcal {L} \\left(I _ {t}, \\hat {I} _ {t}\\right) + \\lambda_ {2} \\mathcal {L} _ {v} \\left(\\hat {\\mathbf {t}} _ {t \\rightarrow s}, v\\right) \\tag {7}",
                                    "image_path": "d21cf4a60ea62f657c55a6472e972a2d74c749ac51be5c868ab56c5276682b5a.jpg"
                                }
                            ]
                        }
                    ],
                    "index": 18
                },
                {
                    "bbox": [
                        305,
                        653,
                        545,
                        713
                    ],
                    "type": "text",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                305,
                                653,
                                545,
                                713
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        305,
                                        653,
                                        545,
                                        713
                                    ],
                                    "type": "text",
                                    "content": "where "
                                },
                                {
                                    "bbox": [
                                        305,
                                        653,
                                        545,
                                        713
                                    ],
                                    "type": "inline_equation",
                                    "content": "\\lambda_{2}"
                                },
                                {
                                    "bbox": [
                                        305,
                                        653,
                                        545,
                                        713
                                    ],
                                    "type": "text",
                                    "content": " is a weight used to balance the different loss terms. This additional velocity loss allows the pose network to make metrically accurate predictions, subsequently resulting in the depth network also learning metrically accurate estimates to maintain consistency (cf. Section 5.4)."
                                }
                            ]
                        }
                    ],
                    "index": 19
                }
            ],
            "discarded_blocks": [
                {
                    "bbox": [
                        294,
                        734,
                        315,
                        742
                    ],
                    "type": "page_number",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                294,
                                734,
                                315,
                                742
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        294,
                                        734,
                                        315,
                                        742
                                    ],
                                    "type": "text",
                                    "content": "2487"
                                }
                            ]
                        }
                    ],
                    "index": 20
                }
            ],
            "page_size": [
                612,
                792
            ],
            "page_idx": 2
        },
        {
            "para_blocks": [
                {
                    "type": "image",
                    "bbox": [
                        75,
                        64,
                        263,
                        241
                    ],
                    "blocks": [
                        {
                            "bbox": [
                                75,
                                64,
                                263,
                                241
                            ],
                            "lines": [
                                {
                                    "bbox": [
                                        75,
                                        64,
                                        263,
                                        241
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                75,
                                                64,
                                                263,
                                                241
                                            ],
                                            "type": "image",
                                            "image_path": "bcf9d1e61771766367837af7fa0ac4bdf59bb3b41a3cb0a4e40a9e9857e8c7d7.jpg"
                                        }
                                    ]
                                }
                            ],
                            "index": 0,
                            "angle": 0,
                            "type": "image_body"
                        },
                        {
                            "bbox": [
                                46,
                                249,
                                288,
                                285
                            ],
                            "lines": [
                                {
                                    "bbox": [
                                        46,
                                        249,
                                        288,
                                        285
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                46,
                                                249,
                                                288,
                                                285
                                            ],
                                            "type": "text",
                                            "content": "Figure 3: Proposed 3D packing and unpacking blocks. Packing replaces striding and pooling, while unpacking is its symmetrical feature upsampling mechanism."
                                        }
                                    ]
                                }
                            ],
                            "index": 1,
                            "angle": 0,
                            "type": "image_caption"
                        }
                    ],
                    "index": 0
                },
                {
                    "bbox": [
                        46,
                        297,
                        284,
                        312
                    ],
                    "type": "title",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                46,
                                297,
                                284,
                                312
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        46,
                                        297,
                                        284,
                                        312
                                    ],
                                    "type": "text",
                                    "content": "4. PackNet: 3D Packing for Depth Estimation"
                                }
                            ]
                        }
                    ],
                    "index": 2
                },
                {
                    "bbox": [
                        46,
                        316,
                        289,
                        557
                    ],
                    "type": "text",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                46,
                                316,
                                289,
                                557
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        46,
                                        316,
                                        289,
                                        557
                                    ],
                                    "type": "text",
                                    "content": "Standard convolutional architectures use aggressive striding and pooling to increase their receptive field size. However, this potentially decreases model performance for tasks requiring fine-grained representations [19, 49]. Similarly, traditional upsampling strategies [6, 11] fail to propagate and preserve sufficient details at the decoder layers to recover accurate depth predictions. In contrast, we propose a novel encoder-decoder architecture, called PackNet, that introduces new 3D packing and unpacking blocks to learn to jointly preserve and recover important spatial information for depth estimation. This is in alignment with recent observations that information loss is not a necessary condition to learn representations capable of generalizing to different scenarios [21]. In fact, progressive expansion and contraction in a fully invertible manner, without discarding \"uninformative\" input variability, has been shown to increase performance in a wide variety of tasks [3, 10, 25]. We first describe the different blocks of our proposed architecture, and then proceed to show how they are integrated together in a single model for monocular depth estimation."
                                }
                            ]
                        }
                    ],
                    "index": 3
                },
                {
                    "bbox": [
                        47,
                        563,
                        138,
                        576
                    ],
                    "type": "title",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                47,
                                563,
                                138,
                                576
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        47,
                                        563,
                                        138,
                                        576
                                    ],
                                    "type": "text",
                                    "content": "4.1. Packing Block"
                                }
                            ]
                        }
                    ],
                    "index": 4
                },
                {
                    "bbox": [
                        46,
                        582,
                        287,
                        715
                    ],
                    "type": "text",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                46,
                                582,
                                287,
                                715
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        46,
                                        582,
                                        287,
                                        715
                                    ],
                                    "type": "text",
                                    "content": "The packing block (Fig. 3a) starts by folding the spatial dimensions of convolutional feature maps into extra feature channels via a Space2Depth operation [39]. The resulting tensor is at a reduced resolution, but in contrast to striding or pooling, this transformation is invertible and comes at no loss. Next, we learn to compress this concatenated feature space in order to reduce its dimensionality to a desired number of output channels. As we show in our experiments (cf. Section 5.6), 2D convolutions are not designed to directly leverage the tiled structure of this feature space. Instead, we propose to first learn to expand this structured"
                                }
                            ]
                        }
                    ],
                    "index": 5
                },
                {
                    "type": "table",
                    "bbox": [
                        308,
                        59,
                        547,
                        222
                    ],
                    "blocks": [
                        {
                            "bbox": [
                                308,
                                59,
                                547,
                                222
                            ],
                            "lines": [
                                {
                                    "bbox": [
                                        308,
                                        59,
                                        547,
                                        222
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                308,
                                                59,
                                                547,
                                                222
                                            ],
                                            "type": "table",
                                            "html": "<table><tr><td></td><td>Layer Description</td><td>K</td><td>Output Tensor Dim.</td></tr><tr><td>#0</td><td>Input RGB image</td><td></td><td>3×H×W</td></tr><tr><td colspan=\"4\">Encoding Layers</td></tr><tr><td>#1</td><td>Conv2d</td><td>5</td><td>64×H×W</td></tr><tr><td>#2</td><td>Conv2d → Packing</td><td>7</td><td>64×H/2×W/2</td></tr><tr><td>#3</td><td>ResidualBlock (x2) → Packing</td><td>3</td><td>64×H/4×W/4</td></tr><tr><td>#4</td><td>ResidualBlock (x2) → Packing</td><td>3</td><td>128×H/8×W/8</td></tr><tr><td>#5</td><td>ResidualBlock (x3) → Packing</td><td>3</td><td>256×H/16×W/16</td></tr><tr><td>#6</td><td>ResidualBlock (x3) → Packing</td><td>3</td><td>512×H/32×W/32</td></tr><tr><td colspan=\"4\">Decoding Layers</td></tr><tr><td>#7</td><td>Unpacking (#6) → Conv2d (⊕ #5)</td><td>3</td><td>512×H/16×W/16</td></tr><tr><td>#8</td><td>Unpacking (#7) → Conv2d (⊕ #4)</td><td>3</td><td>256×H/8×W/8</td></tr><tr><td>#9</td><td>InvDepth (#8)</td><td>3</td><td>1×H/8×W/8</td></tr><tr><td>#10</td><td>Unpacking (#8) → Conv2d (⊕ #3 ⊕ Upsample(#9))</td><td>3</td><td>128×H/4×W/4</td></tr><tr><td>#11</td><td>InvDepth (#10)</td><td>3</td><td>1×H/4×W/4</td></tr><tr><td>#12</td><td>Unpacking (#10) → Conv2d (⊕ #2 ⊕ Upsample(#11))</td><td>3</td><td>64×H/2×W/2</td></tr><tr><td>#13</td><td>InvDepth (#12)</td><td>3</td><td>1×H/2×W/2</td></tr><tr><td>#14</td><td>Unpacking (#12) → Conv2d (⊕ #1 ⊕ Upsample(#13))</td><td>3</td><td>64×H×W</td></tr><tr><td>#15</td><td>InvDepth (#14)</td><td>3</td><td>1×H×W</td></tr></table>",
                                            "image_path": "e7b6b83e5f42dbeeb05f554be17d9effa3529b887a7be23f2cd146847018bfa8.jpg"
                                        }
                                    ]
                                }
                            ],
                            "index": 6,
                            "angle": 0,
                            "type": "table_body"
                        }
                    ],
                    "index": 6
                },
                {
                    "bbox": [
                        305,
                        224,
                        547,
                        382
                    ],
                    "lines": [
                        {
                            "bbox": [
                                305,
                                224,
                                547,
                                382
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        305,
                                        224,
                                        547,
                                        382
                                    ],
                                    "type": "text",
                                    "content": "Table 1: Summary of our PackNet architecture for self-supervised monocular depth estimation. The Packing and Unpacking blocks are described in Fig. 3, with kernel size "
                                },
                                {
                                    "bbox": [
                                        305,
                                        224,
                                        547,
                                        382
                                    ],
                                    "type": "inline_equation",
                                    "content": "K = 3"
                                },
                                {
                                    "bbox": [
                                        305,
                                        224,
                                        547,
                                        382
                                    ],
                                    "type": "text",
                                    "content": " and "
                                },
                                {
                                    "bbox": [
                                        305,
                                        224,
                                        547,
                                        382
                                    ],
                                    "type": "inline_equation",
                                    "content": "D = 8"
                                },
                                {
                                    "bbox": [
                                        305,
                                        224,
                                        547,
                                        382
                                    ],
                                    "type": "text",
                                    "content": ". Conv2d blocks include Group-Norm [45] with "
                                },
                                {
                                    "bbox": [
                                        305,
                                        224,
                                        547,
                                        382
                                    ],
                                    "type": "inline_equation",
                                    "content": "G = 16"
                                },
                                {
                                    "bbox": [
                                        305,
                                        224,
                                        547,
                                        382
                                    ],
                                    "type": "text",
                                    "content": " and ELU non-linearities [7]. In-vDepth blocks include a 2D convolutional layer with "
                                },
                                {
                                    "bbox": [
                                        305,
                                        224,
                                        547,
                                        382
                                    ],
                                    "type": "inline_equation",
                                    "content": "K = 3"
                                },
                                {
                                    "bbox": [
                                        305,
                                        224,
                                        547,
                                        382
                                    ],
                                    "type": "text",
                                    "content": " and sigmoid non-linearities. Each ResidualBlock is a sequence of 3 2D convolutional layers with "
                                },
                                {
                                    "bbox": [
                                        305,
                                        224,
                                        547,
                                        382
                                    ],
                                    "type": "inline_equation",
                                    "content": "K = 3/3/1"
                                },
                                {
                                    "bbox": [
                                        305,
                                        224,
                                        547,
                                        382
                                    ],
                                    "type": "text",
                                    "content": " and ELU non-linearities, followed by GroupNorm with "
                                },
                                {
                                    "bbox": [
                                        305,
                                        224,
                                        547,
                                        382
                                    ],
                                    "type": "inline_equation",
                                    "content": "G = 16"
                                },
                                {
                                    "bbox": [
                                        305,
                                        224,
                                        547,
                                        382
                                    ],
                                    "type": "text",
                                    "content": " and Dropout [40] of 0.5 in the final layer. Upsample is a nearest-neighbor resizing operation. Numbers in parentheses indicate input layers, with "
                                },
                                {
                                    "bbox": [
                                        305,
                                        224,
                                        547,
                                        382
                                    ],
                                    "type": "inline_equation",
                                    "content": "\\oplus"
                                },
                                {
                                    "bbox": [
                                        305,
                                        224,
                                        547,
                                        382
                                    ],
                                    "type": "text",
                                    "content": " as channel concatenation. Bold numbers indicate the four inverse depth output scales."
                                }
                            ]
                        }
                    ],
                    "index": 7,
                    "angle": 0,
                    "type": "text"
                },
                {
                    "bbox": [
                        304,
                        396,
                        545,
                        492
                    ],
                    "type": "text",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                304,
                                396,
                                545,
                                492
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        304,
                                        396,
                                        545,
                                        492
                                    ],
                                    "type": "text",
                                    "content": "representation via a 3D convolutional layer. The resulting higher dimensional feature space is then flattened (by simple reshaping) before a final 2D convolutional contraction layer. This structured feature expansion-contraction, inspired by invertible networks [3, 21] although we do not ensure invertibility, allows our architecture to dedicate more parameters to learn how to compress key spatial details that need to be preserved for high resolution depth decoding."
                                }
                            ]
                        }
                    ],
                    "index": 8
                },
                {
                    "bbox": [
                        305,
                        502,
                        410,
                        515
                    ],
                    "type": "title",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                305,
                                502,
                                410,
                                515
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        305,
                                        502,
                                        410,
                                        515
                                    ],
                                    "type": "text",
                                    "content": "4.2. Unpacking Block"
                                }
                            ]
                        }
                    ],
                    "index": 9
                },
                {
                    "bbox": [
                        304,
                        522,
                        547,
                        713
                    ],
                    "type": "text",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                304,
                                522,
                                547,
                                713
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        304,
                                        522,
                                        547,
                                        713
                                    ],
                                    "type": "text",
                                    "content": "Symmetrically, the unpacking block (Fig. 3b) learns to decompress and unfold packed convolutional feature channels back into higher resolution spatial dimensions during the decoding process. The unpacking block replaces convolutional feature upsampling, typically performed via nearest-neighbor or with learnable transposed convolutional weights. It is inspired by sub-pixel convolutions [39], but adapted to reverse the 3D packing process that the features went through in the encoder. First, we use a 2D convolutional layer to produce the required number of feature channels for a following 3D convolutional layer. Second, this 3D convolution learns to expand back the compressed spatial features. Third, these unpacked features are converted back to spatial details via a reshape and Depth2Space operation [39] to obtain a tensor with the desired number of output channels and target higher resolution."
                                }
                            ]
                        }
                    ],
                    "index": 10
                }
            ],
            "discarded_blocks": [
                {
                    "bbox": [
                        295,
                        734,
                        315,
                        742
                    ],
                    "type": "page_number",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                295,
                                734,
                                315,
                                742
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        295,
                                        734,
                                        315,
                                        742
                                    ],
                                    "type": "text",
                                    "content": "2488"
                                }
                            ]
                        }
                    ],
                    "index": 11
                }
            ],
            "page_size": [
                612,
                792
            ],
            "page_idx": 3
        },
        {
            "para_blocks": [
                {
                    "type": "image",
                    "bbox": [
                        51,
                        63,
                        127,
                        119
                    ],
                    "blocks": [
                        {
                            "bbox": [
                                51,
                                63,
                                127,
                                119
                            ],
                            "lines": [
                                {
                                    "bbox": [
                                        51,
                                        63,
                                        127,
                                        119
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                51,
                                                63,
                                                127,
                                                119
                                            ],
                                            "type": "image",
                                            "image_path": "5a5f34eb8f7fd6db80b96a0cb2aafd65c325cfb71f7d3db15354f5c9fc7e09a2.jpg"
                                        }
                                    ]
                                }
                            ],
                            "index": 0,
                            "angle": 0,
                            "type": "image_body"
                        },
                        {
                            "bbox": [
                                63,
                                121,
                                114,
                                130
                            ],
                            "lines": [
                                {
                                    "bbox": [
                                        63,
                                        121,
                                        114,
                                        130
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                63,
                                                121,
                                                114,
                                                130
                                            ],
                                            "type": "text",
                                            "content": "(a) Input Image"
                                        }
                                    ]
                                }
                            ],
                            "index": 1,
                            "angle": 0,
                            "type": "image_caption"
                        }
                    ],
                    "index": 0
                },
                {
                    "type": "image",
                    "bbox": [
                        129,
                        63,
                        205,
                        118
                    ],
                    "blocks": [
                        {
                            "bbox": [
                                129,
                                63,
                                205,
                                118
                            ],
                            "lines": [
                                {
                                    "bbox": [
                                        129,
                                        63,
                                        205,
                                        118
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                129,
                                                63,
                                                205,
                                                118
                                            ],
                                            "type": "image",
                                            "image_path": "c542c13c61e7404fda04aa60915f6fb39e7445f0c7f9cd4038e7ea386ca2cfb1.jpg"
                                        }
                                    ]
                                }
                            ],
                            "index": 2,
                            "angle": 0,
                            "type": "image_body"
                        },
                        {
                            "bbox": [
                                135,
                                121,
                                197,
                                140
                            ],
                            "lines": [
                                {
                                    "bbox": [
                                        135,
                                        121,
                                        197,
                                        140
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                135,
                                                121,
                                                197,
                                                140
                                            ],
                                            "type": "text",
                                            "content": "(b) Max Pooling + Bilinear Upsample"
                                        }
                                    ]
                                }
                            ],
                            "index": 3,
                            "angle": 0,
                            "type": "image_caption"
                        },
                        {
                            "bbox": [
                                46,
                                144,
                                287,
                                204
                            ],
                            "lines": [
                                {
                                    "bbox": [
                                        46,
                                        144,
                                        287,
                                        204
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                46,
                                                144,
                                                287,
                                                204
                                            ],
                                            "type": "text",
                                            "content": "Figure 4: Image reconstruction using different encoder-decoders: (b) standard max pooling and bilinear upsampling, each followed by 2D convolutions; (c) one packing-unpacking combination (cf. Fig. 3) with "
                                        },
                                        {
                                            "bbox": [
                                                46,
                                                144,
                                                287,
                                                204
                                            ],
                                            "type": "inline_equation",
                                            "content": "D = 2"
                                        },
                                        {
                                            "bbox": [
                                                46,
                                                144,
                                                287,
                                                204
                                            ],
                                            "type": "text",
                                            "content": ". All kernel sizes are "
                                        },
                                        {
                                            "bbox": [
                                                46,
                                                144,
                                                287,
                                                204
                                            ],
                                            "type": "inline_equation",
                                            "content": "K = 3"
                                        },
                                        {
                                            "bbox": [
                                                46,
                                                144,
                                                287,
                                                204
                                            ],
                                            "type": "text",
                                            "content": " and "
                                        },
                                        {
                                            "bbox": [
                                                46,
                                                144,
                                                287,
                                                204
                                            ],
                                            "type": "inline_equation",
                                            "content": "C = 4"
                                        },
                                        {
                                            "bbox": [
                                                46,
                                                144,
                                                287,
                                                204
                                            ],
                                            "type": "text",
                                            "content": " for intermediate channels."
                                        }
                                    ]
                                }
                            ],
                            "index": 6,
                            "angle": 0,
                            "type": "image_caption"
                        }
                    ],
                    "index": 2
                },
                {
                    "type": "image",
                    "bbox": [
                        207,
                        63,
                        283,
                        118
                    ],
                    "blocks": [
                        {
                            "bbox": [
                                207,
                                63,
                                283,
                                118
                            ],
                            "lines": [
                                {
                                    "bbox": [
                                        207,
                                        63,
                                        283,
                                        118
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                207,
                                                63,
                                                283,
                                                118
                                            ],
                                            "type": "image",
                                            "image_path": "6a38de17772036cac79816a20c9c8dae8e6a423bf9352006513e56fbd4bdfe00.jpg"
                                        }
                                    ]
                                }
                            ],
                            "index": 4,
                            "angle": 0,
                            "type": "image_body"
                        },
                        {
                            "bbox": [
                                214,
                                121,
                                274,
                                130
                            ],
                            "lines": [
                                {
                                    "bbox": [
                                        214,
                                        121,
                                        274,
                                        130
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                214,
                                                121,
                                                274,
                                                130
                                            ],
                                            "type": "text",
                                            "content": "(c) Pack + Unpack"
                                        }
                                    ]
                                }
                            ],
                            "index": 5,
                            "angle": 0,
                            "type": "image_caption"
                        }
                    ],
                    "index": 4
                },
                {
                    "bbox": [
                        47,
                        215,
                        203,
                        227
                    ],
                    "type": "title",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                47,
                                215,
                                203,
                                227
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        47,
                                        215,
                                        203,
                                        227
                                    ],
                                    "type": "text",
                                    "content": "4.3. Detail-Preserving Properties"
                                }
                            ]
                        }
                    ],
                    "index": 7
                },
                {
                    "bbox": [
                        46,
                        233,
                        288,
                        400
                    ],
                    "type": "text",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                46,
                                233,
                                288,
                                400
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        46,
                                        233,
                                        288,
                                        400
                                    ],
                                    "type": "text",
                                    "content": "In Fig. 4, we illustrate the detail-preserving properties of our packing / unpacking combination, showing we can get a near-lossless encoder-decoder for single image reconstruction by minimizing the L1 loss. We train a simple network composed of one packing layer followed by a symmetrical unpacking one and show it is able to almost exactly reconstruct the input image (final loss of 0.0079), including sharp edges and finer details. In contrast, a comparable baseline replacing packing / unpacking with max pooling / bilinear upsampling (and keeping the 2D convolutions) is only able to learn a blurry reconstruction (final loss of 0.063). This highlights how PackNet is able to learn more complex features by preserving spatial and appearance information end-to-end throughout the network."
                                }
                            ]
                        }
                    ],
                    "index": 8
                },
                {
                    "bbox": [
                        47,
                        407,
                        162,
                        418
                    ],
                    "type": "title",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                47,
                                407,
                                162,
                                418
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        47,
                                        407,
                                        162,
                                        418
                                    ],
                                    "type": "text",
                                    "content": "4.4. Model Architecture"
                                }
                            ]
                        }
                    ],
                    "index": 9
                },
                {
                    "bbox": [
                        46,
                        426,
                        287,
                        570
                    ],
                    "type": "text",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                46,
                                426,
                                287,
                                570
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        46,
                                        426,
                                        287,
                                        570
                                    ],
                                    "type": "text",
                                    "content": "Our PackNet architecture for self-supervised monocular depth estimation is detailed in Table 1. Our symmetrical encoder-decoder architecture incorporates several packing and unpacking blocks, and is supplemented with skip connections [35] to facilitate the flow of information and gradients throughout the network. The decoder produces intermediate inverse depth maps that are upsampled before being concatenated with their corresponding skip connections and unpacked feature maps. These intermediate inverse depth maps are also used at training time in the loss calculation, after being upsampled to the full output resolution using nearest neighbors interpolation."
                                }
                            ]
                        }
                    ],
                    "index": 10
                },
                {
                    "bbox": [
                        47,
                        579,
                        127,
                        593
                    ],
                    "type": "title",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                47,
                                579,
                                127,
                                593
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        47,
                                        579,
                                        127,
                                        593
                                    ],
                                    "type": "text",
                                    "content": "5. Experiments"
                                }
                            ]
                        }
                    ],
                    "index": 11
                },
                {
                    "bbox": [
                        47,
                        599,
                        110,
                        610
                    ],
                    "type": "title",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                47,
                                599,
                                110,
                                610
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        47,
                                        599,
                                        110,
                                        610
                                    ],
                                    "type": "text",
                                    "content": "5.1. Datasets"
                                }
                            ]
                        }
                    ],
                    "index": 12
                },
                {
                    "bbox": [
                        46,
                        617,
                        286,
                        713
                    ],
                    "type": "text",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                46,
                                617,
                                286,
                                713
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        46,
                                        617,
                                        286,
                                        713
                                    ],
                                    "type": "text",
                                    "content": "KITTI [16]. The KITTI benchmark is the de facto standard for depth evaluation. More specifically, we adopt the training protocol used in Eigen et al. [13], with Zhou et al.'s [52] pre-processing to remove static frames. This results in 39810 images for training, 4424 for validation and 697 for evaluation. We also consider the improved ground-truth depth maps from [41] for evaluation, which uses 5 consecutive frames to accumulate LiDAR points and stereo"
                                }
                            ]
                        }
                    ],
                    "index": 13
                },
                {
                    "bbox": [
                        305,
                        72,
                        544,
                        96
                    ],
                    "type": "text",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                305,
                                72,
                                544,
                                96
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        305,
                                        72,
                                        544,
                                        96
                                    ],
                                    "type": "text",
                                    "content": "information to handle moving objects, resulting in 652 high-quality depth maps."
                                }
                            ]
                        }
                    ],
                    "index": 14
                },
                {
                    "bbox": [
                        304,
                        97,
                        545,
                        228
                    ],
                    "type": "text",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                304,
                                97,
                                545,
                                228
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        304,
                                        97,
                                        545,
                                        228
                                    ],
                                    "type": "text",
                                    "content": "DDAD (Dense Depth for Automated Driving). As one of our contributions, we release a diverse dataset of urban, highway, and residential scenes curated from a global fleet of self-driving cars. It contains 17,050 training and 4,150 evaluation frames with ground-truth depth maps generated from dense LiDAR measurements using the Luminar-H2 sensor. This new dataset is a more realistic and challenging benchmark for depth estimation, as it is diverse and captures precise structure across images (30k points per frame) at longer ranges (up to "
                                },
                                {
                                    "bbox": [
                                        304,
                                        97,
                                        545,
                                        228
                                    ],
                                    "type": "inline_equation",
                                    "content": "200m"
                                },
                                {
                                    "bbox": [
                                        304,
                                        97,
                                        545,
                                        228
                                    ],
                                    "type": "text",
                                    "content": " vs "
                                },
                                {
                                    "bbox": [
                                        304,
                                        97,
                                        545,
                                        228
                                    ],
                                    "type": "inline_equation",
                                    "content": "80m"
                                },
                                {
                                    "bbox": [
                                        304,
                                        97,
                                        545,
                                        228
                                    ],
                                    "type": "text",
                                    "content": " for previous datasets). See supplementary material for more details."
                                }
                            ]
                        }
                    ],
                    "index": 15
                },
                {
                    "bbox": [
                        304,
                        229,
                        546,
                        289
                    ],
                    "type": "text",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                304,
                                229,
                                546,
                                289
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        304,
                                        229,
                                        546,
                                        289
                                    ],
                                    "type": "text",
                                    "content": "NuScenes [4]. To assess the generalization capability of our approach w.r.t. previous ones, we evaluate KITTI models (without fine-tuning) on the official NuScenes validation dataset of 6019 front-facing images with ground-truth depth maps generated by LiDAR reprojection."
                                }
                            ]
                        }
                    ],
                    "index": 16
                },
                {
                    "bbox": [
                        304,
                        289,
                        545,
                        385
                    ],
                    "type": "text",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                304,
                                289,
                                545,
                                385
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        304,
                                        289,
                                        545,
                                        385
                                    ],
                                    "type": "text",
                                    "content": "CityScapes [8]. We also experiment with pretraining our monocular networks on the CityScapes dataset, before finetuning on the KITTI dataset. This also allows us to explore the scalability and generalization performance of different models, as they are trained with increasing amounts of unlabeled data. A total of 88250 images were considered as the training split for the CityScapes dataset, using the same training parameters as KITTI for 20 epochs."
                                }
                            ]
                        }
                    ],
                    "index": 17
                },
                {
                    "bbox": [
                        306,
                        394,
                        438,
                        407
                    ],
                    "type": "title",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                306,
                                394,
                                438,
                                407
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        306,
                                        394,
                                        438,
                                        407
                                    ],
                                    "type": "text",
                                    "content": "5.2. Implementation Details"
                                }
                            ]
                        }
                    ],
                    "index": 18
                },
                {
                    "bbox": [
                        304,
                        413,
                        545,
                        557
                    ],
                    "type": "text",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                304,
                                413,
                                545,
                                557
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        304,
                                        413,
                                        545,
                                        557
                                    ],
                                    "type": "text",
                                    "content": "We use PyTorch [37] with all models trained across 8 Titan V100 GPUs. We use the Adam optimizer [24], with "
                                },
                                {
                                    "bbox": [
                                        304,
                                        413,
                                        545,
                                        557
                                    ],
                                    "type": "inline_equation",
                                    "content": "\\beta_{1} = 0.9"
                                },
                                {
                                    "bbox": [
                                        304,
                                        413,
                                        545,
                                        557
                                    ],
                                    "type": "text",
                                    "content": " and "
                                },
                                {
                                    "bbox": [
                                        304,
                                        413,
                                        545,
                                        557
                                    ],
                                    "type": "inline_equation",
                                    "content": "\\beta_{2} = 0.999"
                                },
                                {
                                    "bbox": [
                                        304,
                                        413,
                                        545,
                                        557
                                    ],
                                    "type": "text",
                                    "content": ". The monocular depth and pose networks are trained for 100 epochs, with a batch size of 4 and initial depth and pose learning rates of "
                                },
                                {
                                    "bbox": [
                                        304,
                                        413,
                                        545,
                                        557
                                    ],
                                    "type": "inline_equation",
                                    "content": "2 \\cdot 10^{-4}"
                                },
                                {
                                    "bbox": [
                                        304,
                                        413,
                                        545,
                                        557
                                    ],
                                    "type": "text",
                                    "content": " and "
                                },
                                {
                                    "bbox": [
                                        304,
                                        413,
                                        545,
                                        557
                                    ],
                                    "type": "inline_equation",
                                    "content": "5 \\cdot 10^{-4}"
                                },
                                {
                                    "bbox": [
                                        304,
                                        413,
                                        545,
                                        557
                                    ],
                                    "type": "text",
                                    "content": " respectively. Training sequences are generated using a stride of 1, meaning that the previous "
                                },
                                {
                                    "bbox": [
                                        304,
                                        413,
                                        545,
                                        557
                                    ],
                                    "type": "inline_equation",
                                    "content": "t - 1"
                                },
                                {
                                    "bbox": [
                                        304,
                                        413,
                                        545,
                                        557
                                    ],
                                    "type": "text",
                                    "content": ", current "
                                },
                                {
                                    "bbox": [
                                        304,
                                        413,
                                        545,
                                        557
                                    ],
                                    "type": "inline_equation",
                                    "content": "t"
                                },
                                {
                                    "bbox": [
                                        304,
                                        413,
                                        545,
                                        557
                                    ],
                                    "type": "text",
                                    "content": ", and posterior "
                                },
                                {
                                    "bbox": [
                                        304,
                                        413,
                                        545,
                                        557
                                    ],
                                    "type": "inline_equation",
                                    "content": "t + 1"
                                },
                                {
                                    "bbox": [
                                        304,
                                        413,
                                        545,
                                        557
                                    ],
                                    "type": "text",
                                    "content": " images are used in the loss calculation. As training proceeds, the learning rate is decayed every 40 epochs by a factor of 2. We set the SSIM weight to "
                                },
                                {
                                    "bbox": [
                                        304,
                                        413,
                                        545,
                                        557
                                    ],
                                    "type": "inline_equation",
                                    "content": "\\alpha = 0.85"
                                },
                                {
                                    "bbox": [
                                        304,
                                        413,
                                        545,
                                        557
                                    ],
                                    "type": "text",
                                    "content": ", the depth regularization weight to "
                                },
                                {
                                    "bbox": [
                                        304,
                                        413,
                                        545,
                                        557
                                    ],
                                    "type": "inline_equation",
                                    "content": "\\lambda_{1} = 0.001"
                                },
                                {
                                    "bbox": [
                                        304,
                                        413,
                                        545,
                                        557
                                    ],
                                    "type": "text",
                                    "content": " and, where applicable, the velocity-scaling weight to "
                                },
                                {
                                    "bbox": [
                                        304,
                                        413,
                                        545,
                                        557
                                    ],
                                    "type": "inline_equation",
                                    "content": "\\lambda_{2} = 0.05"
                                },
                                {
                                    "bbox": [
                                        304,
                                        413,
                                        545,
                                        557
                                    ],
                                    "type": "text",
                                    "content": "."
                                }
                            ]
                        }
                    ],
                    "index": 19
                },
                {
                    "bbox": [
                        304,
                        557,
                        545,
                        629
                    ],
                    "type": "text",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                304,
                                557,
                                545,
                                629
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        304,
                                        557,
                                        545,
                                        629
                                    ],
                                    "type": "text",
                                    "content": "Depth Network. Unless noted otherwise, we use our PackNet architecture as specified in Table 1. During training, all four inverse depth output scales are used in the loss calculation, and at test-time only the final output scale is used, after being resized to the full ground-truth depth map resolution using nearest neighbor interpolation."
                                }
                            ]
                        }
                    ],
                    "index": 20
                },
                {
                    "bbox": [
                        304,
                        630,
                        545,
                        713
                    ],
                    "type": "text",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                304,
                                630,
                                545,
                                713
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        304,
                                        630,
                                        545,
                                        713
                                    ],
                                    "type": "text",
                                    "content": "Pose Network. We use the architecture proposed by [52] without the explainability mask, which we found not to improve results. The pose network consists of 7 convolutional layers followed by a final "
                                },
                                {
                                    "bbox": [
                                        304,
                                        630,
                                        545,
                                        713
                                    ],
                                    "type": "inline_equation",
                                    "content": "1 \\times 1"
                                },
                                {
                                    "bbox": [
                                        304,
                                        630,
                                        545,
                                        713
                                    ],
                                    "type": "text",
                                    "content": " convolutional layer. The input to the network consists of the target view "
                                },
                                {
                                    "bbox": [
                                        304,
                                        630,
                                        545,
                                        713
                                    ],
                                    "type": "inline_equation",
                                    "content": "I_{t}"
                                },
                                {
                                    "bbox": [
                                        304,
                                        630,
                                        545,
                                        713
                                    ],
                                    "type": "text",
                                    "content": " and the context views "
                                },
                                {
                                    "bbox": [
                                        304,
                                        630,
                                        545,
                                        713
                                    ],
                                    "type": "inline_equation",
                                    "content": "I_{S}"
                                },
                                {
                                    "bbox": [
                                        304,
                                        630,
                                        545,
                                        713
                                    ],
                                    "type": "text",
                                    "content": ", and the output is the set of 6 DOF transformations between "
                                },
                                {
                                    "bbox": [
                                        304,
                                        630,
                                        545,
                                        713
                                    ],
                                    "type": "inline_equation",
                                    "content": "I_{t}"
                                },
                                {
                                    "bbox": [
                                        304,
                                        630,
                                        545,
                                        713
                                    ],
                                    "type": "text",
                                    "content": " and "
                                },
                                {
                                    "bbox": [
                                        304,
                                        630,
                                        545,
                                        713
                                    ],
                                    "type": "inline_equation",
                                    "content": "I_{s}"
                                },
                                {
                                    "bbox": [
                                        304,
                                        630,
                                        545,
                                        713
                                    ],
                                    "type": "text",
                                    "content": ", for "
                                },
                                {
                                    "bbox": [
                                        304,
                                        630,
                                        545,
                                        713
                                    ],
                                    "type": "inline_equation",
                                    "content": "s \\in S"
                                },
                                {
                                    "bbox": [
                                        304,
                                        630,
                                        545,
                                        713
                                    ],
                                    "type": "text",
                                    "content": "."
                                }
                            ]
                        }
                    ],
                    "index": 21
                }
            ],
            "discarded_blocks": [
                {
                    "bbox": [
                        294,
                        733,
                        315,
                        742
                    ],
                    "type": "page_number",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                294,
                                733,
                                315,
                                742
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        294,
                                        733,
                                        315,
                                        742
                                    ],
                                    "type": "text",
                                    "content": "2489"
                                }
                            ]
                        }
                    ],
                    "index": 22
                }
            ],
            "page_size": [
                612,
                792
            ],
            "page_idx": 4
        },
        {
            "para_blocks": [
                {
                    "bbox": [
                        47,
                        72,
                        217,
                        85
                    ],
                    "type": "title",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                47,
                                72,
                                217,
                                85
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        47,
                                        72,
                                        217,
                                        85
                                    ],
                                    "type": "text",
                                    "content": "5.3. Depth Estimation Performance"
                                }
                            ]
                        }
                    ],
                    "index": 0
                },
                {
                    "bbox": [
                        46,
                        91,
                        288,
                        235
                    ],
                    "type": "text",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                46,
                                91,
                                288,
                                235
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        46,
                                        91,
                                        288,
                                        235
                                    ],
                                    "type": "text",
                                    "content": "First, we report the performance of our proposed monocular depth estimation method when considering longer distances, which is now possible due to the introduction of our new DDAD dataset. Depth estimation results using this dataset for training and evaluation, considering cumulative distances up to "
                                },
                                {
                                    "bbox": [
                                        46,
                                        91,
                                        288,
                                        235
                                    ],
                                    "type": "inline_equation",
                                    "content": "200\\mathrm{m}"
                                },
                                {
                                    "bbox": [
                                        46,
                                        91,
                                        288,
                                        235
                                    ],
                                    "type": "text",
                                    "content": ", can be found in Fig. 5 and Table 2. Additionally, in Fig. 6 we present results for different depth intervals calculated independently. From these results we can see that our PackNet-SfM approach significantly outperforms the state-of-the-art [18], based on the ResNet family, the performance gap consistently increasing when larger distances are considered."
                                }
                            ]
                        }
                    ],
                    "index": 1
                },
                {
                    "bbox": [
                        46,
                        236,
                        287,
                        357
                    ],
                    "type": "text",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                46,
                                236,
                                287,
                                357
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        46,
                                        236,
                                        287,
                                        357
                                    ],
                                    "type": "text",
                                    "content": "Second, we evaluate depth predictions on KITTI using the metrics described in Eigen et al. [13]. We summarize our results in Table 3, for the original depth maps from [13] and the accumulated depth maps from [41], and illustrate their performance qualitatively in Fig. 7. In contrast to previous methods [5, 18] that predominantly focus on modifying the training objective, we show that our proposed Pack-Net architecture can by itself bolster performance and establish a new state of the art for the task of monocular depth estimation, trained in the self-supervised monocular setting."
                                }
                            ]
                        }
                    ],
                    "index": 2
                },
                {
                    "bbox": [
                        46,
                        358,
                        288,
                        502
                    ],
                    "type": "text",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                46,
                                358,
                                288,
                                502
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        46,
                                        358,
                                        288,
                                        502
                                    ],
                                    "type": "text",
                                    "content": "Furthermore, we show that by simply introducing an additional source of unlabeled videos, such as the publicly available CityScapes dataset (CS+K) [8], we are able to further improve monocular depth estimation performance. As indicated by Pillai et al. [38], we also observe an improvement in performance at higher image resolutions, which we attribute to the proposed network's ability to properly preserve and process spatial information end-to-end. Our best results are achieved when injecting both more unlabeled data at training time and processing higher resolution input images, achieving performance comparable to semi-supervised [28] and fully supervised [14] methods."
                                }
                            ]
                        }
                    ],
                    "index": 3
                },
                {
                    "bbox": [
                        47,
                        514,
                        277,
                        528
                    ],
                    "type": "title",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                47,
                                514,
                                277,
                                528
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        47,
                                        514,
                                        277,
                                        528
                                    ],
                                    "type": "text",
                                    "content": "5.4. Scale-Aware Depth Estimation Performance"
                                }
                            ]
                        }
                    ],
                    "index": 4
                },
                {
                    "bbox": [
                        46,
                        534,
                        288,
                        715
                    ],
                    "type": "text",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                46,
                                534,
                                288,
                                715
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        46,
                                        534,
                                        288,
                                        715
                                    ],
                                    "type": "text",
                                    "content": "Due to their inherent scale ambiguity, self-supervised monocular methods [18, 33, 52] evaluate depth by scaling their estimates to the median ground-truth as measured via LiDAR. In Section 3.2 we propose to also recover the metric scale of the scene from a single image by imposing a loss on the magnitude of the translation for the pose network output. Table 3 shows that introducing this weak velocity supervision at training time allows the generation of scale-aware depth models with similar performance as their unscaled counterparts, with the added benefit of not requiring ground-truth depth scaling (or even velocity information) at test-time. Another benefit of scale-awareness is that we can compose metrically accurate trajectories directly from the output of the pose network. Due to space constraints, we report pose estimation results in supplementary material."
                                }
                            ]
                        }
                    ],
                    "index": 5
                },
                {
                    "type": "image",
                    "bbox": [
                        312,
                        61,
                        541,
                        175
                    ],
                    "blocks": [
                        {
                            "bbox": [
                                312,
                                61,
                                541,
                                175
                            ],
                            "lines": [
                                {
                                    "bbox": [
                                        312,
                                        61,
                                        541,
                                        175
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                312,
                                                61,
                                                541,
                                                175
                                            ],
                                            "type": "image",
                                            "image_path": "e329bdb0953d24246e48cf62f2fffea3935999cf8aa4d3f71b3973d6bdf6f393.jpg"
                                        }
                                    ]
                                }
                            ],
                            "index": 6,
                            "angle": 0,
                            "type": "image_body"
                        },
                        {
                            "bbox": [
                                307,
                                177,
                                543,
                                190
                            ],
                            "lines": [
                                {
                                    "bbox": [
                                        307,
                                        177,
                                        543,
                                        190
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                307,
                                                177,
                                                543,
                                                190
                                            ],
                                            "type": "text",
                                            "content": "Figure 5: PackNet pointcloud reconstructions on DDAD."
                                        }
                                    ]
                                }
                            ],
                            "index": 7,
                            "angle": 0,
                            "type": "image_caption"
                        }
                    ],
                    "index": 6
                },
                {
                    "type": "table",
                    "bbox": [
                        306,
                        198,
                        547,
                        274
                    ],
                    "blocks": [
                        {
                            "bbox": [
                                306,
                                198,
                                547,
                                274
                            ],
                            "lines": [
                                {
                                    "bbox": [
                                        306,
                                        198,
                                        547,
                                        274
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                306,
                                                198,
                                                547,
                                                274
                                            ],
                                            "type": "table",
                                            "html": "<table><tr><td>Method</td><td>Abs Rel</td><td>Sq Rel</td><td>RMSE</td><td>\\( RMSE_{log} \\)</td><td>\\( \\delta_{1.25} \\)</td></tr><tr><td>Monodepth2 (R18)</td><td>0.381</td><td>8.387</td><td>21.277</td><td>0.371</td><td>0.587</td></tr><tr><td>Monodepth2‡ (R18)</td><td>0.213</td><td>4.975</td><td>18.051</td><td>0.340</td><td>0.761</td></tr><tr><td>Monodepth2 (R50)</td><td>0.324</td><td>7.348</td><td>20.538</td><td>0.344</td><td>0.615</td></tr><tr><td>Monodepth2‡ (R50)</td><td>0.198</td><td>4.504</td><td>16.641</td><td>0.318</td><td>0.781</td></tr><tr><td>PackNet-SfM</td><td>0.162</td><td>3.917</td><td>13.452</td><td>0.269</td><td>0.823</td></tr></table>",
                                            "image_path": "8d1abc2d64ceeabc93c4ddcaf59b65cde0f5dfa2ac7fe250df540b43e0789e23.jpg"
                                        }
                                    ]
                                }
                            ],
                            "index": 8,
                            "angle": 0,
                            "type": "table_body"
                        }
                    ],
                    "index": 8
                },
                {
                    "bbox": [
                        304,
                        277,
                        547,
                        338
                    ],
                    "lines": [
                        {
                            "bbox": [
                                304,
                                277,
                                547,
                                338
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        304,
                                        277,
                                        547,
                                        338
                                    ],
                                    "type": "text",
                                    "content": "Table 2: Depth Evaluation on DDAD, for "
                                },
                                {
                                    "bbox": [
                                        304,
                                        277,
                                        547,
                                        338
                                    ],
                                    "type": "inline_equation",
                                    "content": "640 \\times 384"
                                },
                                {
                                    "bbox": [
                                        304,
                                        277,
                                        547,
                                        338
                                    ],
                                    "type": "text",
                                    "content": " resolution and distances up to "
                                },
                                {
                                    "bbox": [
                                        304,
                                        277,
                                        547,
                                        338
                                    ],
                                    "type": "inline_equation",
                                    "content": "200\\mathrm{m}"
                                },
                                {
                                    "bbox": [
                                        304,
                                        277,
                                        547,
                                        338
                                    ],
                                    "type": "text",
                                    "content": ". While the ResNet family heavily relies on large-scale supervised ImageNet [9] pretraining (denoted by "
                                },
                                {
                                    "bbox": [
                                        304,
                                        277,
                                        547,
                                        338
                                    ],
                                    "type": "inline_equation",
                                    "content": "\\ddagger"
                                },
                                {
                                    "bbox": [
                                        304,
                                        277,
                                        547,
                                        338
                                    ],
                                    "type": "text",
                                    "content": "), PackNet achieves significantly better results despite being trained from scratch."
                                }
                            ]
                        }
                    ],
                    "index": 9,
                    "angle": 0,
                    "type": "text"
                },
                {
                    "type": "image",
                    "bbox": [
                        324,
                        345,
                        525,
                        437
                    ],
                    "blocks": [
                        {
                            "bbox": [
                                324,
                                345,
                                525,
                                437
                            ],
                            "lines": [
                                {
                                    "bbox": [
                                        324,
                                        345,
                                        525,
                                        437
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                324,
                                                345,
                                                525,
                                                437
                                            ],
                                            "type": "image",
                                            "image_path": "e11cf925d4273221a2b12b1a7de064b116b9c5f7f6f8cfb3a5fe5d0abf47f6ef.jpg"
                                        }
                                    ]
                                }
                            ],
                            "index": 10,
                            "angle": 0,
                            "type": "image_body"
                        },
                        {
                            "bbox": [
                                304,
                                442,
                                547,
                                479
                            ],
                            "lines": [
                                {
                                    "bbox": [
                                        304,
                                        442,
                                        547,
                                        479
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                304,
                                                442,
                                                547,
                                                479
                                            ],
                                            "type": "text",
                                            "content": "Figure 6: Depth Evaluation on DDAD binned at different intervals, calculated independently by only considering ground-truth depth pixels in that range (0-20m, 20-40m, ...)."
                                        }
                                    ]
                                }
                            ],
                            "index": 11,
                            "angle": 0,
                            "type": "image_caption"
                        }
                    ],
                    "index": 10
                },
                {
                    "bbox": [
                        306,
                        491,
                        426,
                        504
                    ],
                    "type": "title",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                306,
                                491,
                                426,
                                504
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        306,
                                        491,
                                        426,
                                        504
                                    ],
                                    "type": "text",
                                    "content": "5.5. Network Complexity"
                                }
                            ]
                        }
                    ],
                    "index": 12
                },
                {
                    "bbox": [
                        304,
                        510,
                        547,
                        715
                    ],
                    "type": "text",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                304,
                                510,
                                547,
                                715
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        304,
                                        510,
                                        547,
                                        715
                                    ],
                                    "type": "text",
                                    "content": "The introduction of packing and unpacking as alternatives to standard downsampling and upsampling operations increases the complexity of the network, due to the number of added parameters. To ensure that the gain in performance shown in our experiments is not only due to an increase in model capacity, we compare different variations of our PackNet architecture (obtained by modifying the number of layers and feature channels) against available ResNet architectures. These results are depicted in Fig. 8 and show that, while the ResNet family stabilizes with diminishing returns as the number of parameters increase, the PackNet family matches its performance at around 70M parameters and further improves as more complexity is added. Finally, the proposed architecture (Table 1) reaches around 128M parameters with an inference time of 60ms on a Titan V100 GPU, which can be further improved to "
                                },
                                {
                                    "bbox": [
                                        304,
                                        510,
                                        547,
                                        715
                                    ],
                                    "type": "inline_equation",
                                    "content": "< 30"
                                },
                                {
                                    "bbox": [
                                        304,
                                        510,
                                        547,
                                        715
                                    ],
                                    "type": "text",
                                    "content": " ms using TensorRT [1], making it suitable for real-time applications."
                                }
                            ]
                        }
                    ],
                    "index": 13
                }
            ],
            "discarded_blocks": [
                {
                    "bbox": [
                        295,
                        733,
                        316,
                        742
                    ],
                    "type": "page_number",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                295,
                                733,
                                316,
                                742
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        295,
                                        733,
                                        316,
                                        742
                                    ],
                                    "type": "text",
                                    "content": "2490"
                                }
                            ]
                        }
                    ],
                    "index": 14
                }
            ],
            "page_size": [
                612,
                792
            ],
            "page_idx": 5
        },
        {
            "para_blocks": [
                {
                    "type": "table",
                    "bbox": [
                        53,
                        71,
                        538,
                        374
                    ],
                    "blocks": [
                        {
                            "bbox": [
                                53,
                                71,
                                538,
                                374
                            ],
                            "lines": [
                                {
                                    "bbox": [
                                        53,
                                        71,
                                        538,
                                        374
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                53,
                                                71,
                                                538,
                                                374
                                            ],
                                            "type": "table",
                                            "html": "<table><tr><td></td><td>Method</td><td>Supervision</td><td>Resolution</td><td>Dataset</td><td>Abs Rel</td><td>Sq Rel</td><td>RMSE</td><td>RMSElog</td><td>δ &lt; 1.25</td><td>δ &lt; 1.252</td><td>δ &lt; 1.253</td></tr><tr><td rowspan=\"15\">Original [13]</td><td>SfMLearner [52]</td><td>M</td><td>416 x 128</td><td>CS + K</td><td>0.198</td><td>1.836</td><td>6.565</td><td>0.275</td><td>0.718</td><td>0.901</td><td>0.960</td></tr><tr><td>Vid2Depth [33]</td><td>M</td><td>416 x 128</td><td>CS + K</td><td>0.159</td><td>1.231</td><td>5.912</td><td>0.243</td><td>0.784</td><td>0.923</td><td>0.970</td></tr><tr><td>DF-Net [53]</td><td>M</td><td>576 x 160</td><td>CS + K</td><td>0.146</td><td>1.182</td><td>5.215</td><td>0.213</td><td>0.818</td><td>0.943</td><td>0.978</td></tr><tr><td>Struct2Depth [5]</td><td>M</td><td>416 x 128</td><td>K</td><td>0.141</td><td>1.026</td><td>5.291</td><td>0.215</td><td>0.816</td><td>0.945</td><td>0.979</td></tr><tr><td>Zhou et al.‡ [50]</td><td>M</td><td>1248 x 384</td><td>K</td><td>0.121</td><td>0.837</td><td>4.945</td><td>0.197</td><td>0.853</td><td>0.955</td><td>0.982</td></tr><tr><td>Monodepth2‡ [18]</td><td>M</td><td>640 x 192</td><td>K</td><td>0.115</td><td>0.903</td><td>4.863</td><td>0.193</td><td>0.877</td><td>0.959</td><td>0.981</td></tr><tr><td>Monodepth2‡ [18]</td><td>M</td><td>1024 x 320</td><td>K</td><td>0.115</td><td>0.882</td><td>4.701</td><td>0.190</td><td>0.879</td><td>0.961</td><td>0.982</td></tr><tr><td>PackNet-SfM</td><td>M</td><td>640 x 192</td><td>K</td><td>0.111</td><td>0.785</td><td>4.601</td><td>0.189</td><td>0.878</td><td>0.960</td><td>0.982</td></tr><tr><td>PackNet-SfM</td><td>M+v</td><td>640 x 192</td><td>K</td><td>0.111</td><td>0.829</td><td>4.788</td><td>0.199</td><td>0.864</td><td>0.954</td><td>0.980</td></tr><tr><td>PackNet-SfM</td><td>M</td><td>640 x 192</td><td>CS + K</td><td>0.108</td><td>0.727</td><td>4.426</td><td>0.184</td><td>0.885</td><td>0.963</td><td>0.983</td></tr><tr><td>PackNet-SfM</td><td>M+v</td><td>640 x 192</td><td>CS + K</td><td>0.108</td><td>0.803</td><td>4.642</td><td>0.195</td><td>0.875</td><td>0.958</td><td>0.980</td></tr><tr><td>PackNet-SfM</td><td>M</td><td>1280 x 384</td><td>K</td><td>0.107</td><td>0.802</td><td>4.538</td><td>0.186</td><td>0.889</td><td>0.962</td><td>0.981</td></tr><tr><td>PackNet-SfM</td><td>M+v</td><td>1280 x 384</td><td>K</td><td>0.107</td><td>0.803</td><td>4.566</td><td>0.197</td><td>0.876</td><td>0.957</td><td>0.979</td></tr><tr><td>PackNet-SfM</td><td>M</td><td>1280 x 384</td><td>CS + K</td><td>0.104</td><td>0.758</td><td>4.386</td><td>0.182</td><td>0.895</td><td>0.964</td><td>0.982</td></tr><tr><td>PackNet-SfM</td><td>M+v</td><td>1280 x 384</td><td>CS + K</td><td>0.103</td><td>0.796</td><td>4.404</td><td>0.189</td><td>0.881</td><td>0.959</td><td>0.980</td></tr><tr><td rowspan=\"11\">Improved [41]</td><td>SfMLeaner [52]</td><td>M</td><td>416 x 128</td><td>CS + K</td><td>0.176</td><td>1.532</td><td>6.129</td><td>0.244</td><td>0.758</td><td>0.921</td><td>0.971</td></tr><tr><td>Vid2Depth [33]</td><td>M</td><td>416 x 128</td><td>CS + K</td><td>0.134</td><td>0.983</td><td>5.501</td><td>0.203</td><td>0.827</td><td>0.944</td><td>0.981</td></tr><tr><td>GeoNet [47]</td><td>M</td><td>416 x 128</td><td>CS + K</td><td>0.132</td><td>0.994</td><td>5.240</td><td>0.193</td><td>0.883</td><td>0.953</td><td>0.985</td></tr><tr><td>DDVO [43]</td><td>M</td><td>416 x 128</td><td>CS + K</td><td>0.126</td><td>0.866</td><td>4.932</td><td>0.185</td><td>0.851</td><td>0.958</td><td>0.986</td></tr><tr><td>EPC++ [32]</td><td>M</td><td>640 x 192</td><td>K</td><td>0.120</td><td>0.789</td><td>4.755</td><td>0.177</td><td>0.856</td><td>0.961</td><td>0.987</td></tr><tr><td>Monodepth2‡ [18]</td><td>M</td><td>640 x 192</td><td>K</td><td>0.090</td><td>0.545</td><td>3.942</td><td>0.137</td><td>0.914</td><td>0.983</td><td>0.995</td></tr><tr><td>Kuznietsov et al.‡ [28]</td><td>D</td><td>621 x 187</td><td>K</td><td>0.089</td><td>0.478</td><td>3.610</td><td>0.138</td><td>0.906</td><td>0.980</td><td>0.995</td></tr><tr><td>DORN‡ [14]</td><td>D</td><td>513 x 385</td><td>K</td><td>0.072</td><td>0.307</td><td>2.727</td><td>0.120</td><td>0.932</td><td>0.984</td><td>0.995</td></tr><tr><td>PackNet-SfM</td><td>M</td><td>640 x 192</td><td>K</td><td>0.078</td><td>0.420</td><td>3.485</td><td>0.121</td><td>0.931</td><td>0.986</td><td>0.996</td></tr><tr><td>PackNet-SfM</td><td>M</td><td>1280 x 384</td><td>CS + K</td><td>0.071</td><td>0.359</td><td>3.153</td><td>0.109</td><td>0.944</td><td>0.990</td><td>0.997</td></tr><tr><td>PackNet-SfM</td><td>M+v</td><td>1280 x 384</td><td>CS + K</td><td>0.075</td><td>0.384</td><td>3.293</td><td>0.114</td><td>0.938</td><td>0.984</td><td>0.995</td></tr></table>",
                                            "image_path": "a5271919ad4d007e3bedcb6a549119292e71937e5d6e52994c0882a140c79020.jpg"
                                        }
                                    ]
                                }
                            ],
                            "index": 0,
                            "angle": 0,
                            "type": "table_body"
                        }
                    ],
                    "index": 0
                },
                {
                    "bbox": [
                        46,
                        383,
                        545,
                        468
                    ],
                    "lines": [
                        {
                            "bbox": [
                                46,
                                383,
                                545,
                                468
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        46,
                                        383,
                                        545,
                                        468
                                    ],
                                    "type": "text",
                                    "content": "Table 3: Quantitative performance comparison of PackNet-SfM on the KITTI dataset for distances up to "
                                },
                                {
                                    "bbox": [
                                        46,
                                        383,
                                        545,
                                        468
                                    ],
                                    "type": "inline_equation",
                                    "content": "80\\mathrm{m}"
                                },
                                {
                                    "bbox": [
                                        46,
                                        383,
                                        545,
                                        468
                                    ],
                                    "type": "text",
                                    "content": ". For Abs Rel, Sq Rel, RMSE and "
                                },
                                {
                                    "bbox": [
                                        46,
                                        383,
                                        545,
                                        468
                                    ],
                                    "type": "inline_equation",
                                    "content": "\\mathrm{RMSE}_{log}"
                                },
                                {
                                    "bbox": [
                                        46,
                                        383,
                                        545,
                                        468
                                    ],
                                    "type": "text",
                                    "content": " lower is better, and for "
                                },
                                {
                                    "bbox": [
                                        46,
                                        383,
                                        545,
                                        468
                                    ],
                                    "type": "inline_equation",
                                    "content": "\\delta < 1.25"
                                },
                                {
                                    "bbox": [
                                        46,
                                        383,
                                        545,
                                        468
                                    ],
                                    "type": "text",
                                    "content": ", "
                                },
                                {
                                    "bbox": [
                                        46,
                                        383,
                                        545,
                                        468
                                    ],
                                    "type": "inline_equation",
                                    "content": "\\delta < 1.25^2"
                                },
                                {
                                    "bbox": [
                                        46,
                                        383,
                                        545,
                                        468
                                    ],
                                    "type": "text",
                                    "content": " and "
                                },
                                {
                                    "bbox": [
                                        46,
                                        383,
                                        545,
                                        468
                                    ],
                                    "type": "inline_equation",
                                    "content": "\\delta < 1.25^3"
                                },
                                {
                                    "bbox": [
                                        46,
                                        383,
                                        545,
                                        468
                                    ],
                                    "type": "text",
                                    "content": " higher is better. In the Dataset column, CS+K refers to pretraining on CityScapes (CS) and fine-tuning on KITTI (K). M refers to methods that train using monocular (M) images, and M+v refers to added velocity weak supervision (v), as shown in Section 3.2. "
                                },
                                {
                                    "bbox": [
                                        46,
                                        383,
                                        545,
                                        468
                                    ],
                                    "type": "inline_equation",
                                    "content": "\\ddagger"
                                },
                                {
                                    "bbox": [
                                        46,
                                        383,
                                        545,
                                        468
                                    ],
                                    "type": "text",
                                    "content": " indicates ImageNet [9] pretraining. Original uses raw depth maps from [13] for evaluation, and Improved uses annotated depth maps from [41]. At test-time, all monocular methods (M) scale estimated depths with median ground-truth LiDAR information. Velocity-scaled (M+v) and supervised (D) methods are not scaled in such way, since they are already metrically accurate."
                                }
                            ]
                        }
                    ],
                    "index": 1,
                    "angle": 0,
                    "type": "text"
                },
                {
                    "type": "image",
                    "bbox": [
                        56,
                        476,
                        537,
                        681
                    ],
                    "blocks": [
                        {
                            "bbox": [
                                56,
                                476,
                                537,
                                681
                            ],
                            "lines": [
                                {
                                    "bbox": [
                                        56,
                                        476,
                                        537,
                                        681
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                56,
                                                476,
                                                537,
                                                681
                                            ],
                                            "type": "image",
                                            "image_path": "dc80df5c4772bac7fd59ee9aa150408c4a4b1275fe69c3d769d47417ae475413.jpg"
                                        }
                                    ]
                                }
                            ],
                            "index": 2,
                            "angle": 0,
                            "type": "image_body"
                        },
                        {
                            "bbox": [
                                46,
                                689,
                                546,
                                726
                            ],
                            "lines": [
                                {
                                    "bbox": [
                                        46,
                                        689,
                                        546,
                                        726
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                46,
                                                689,
                                                546,
                                                726
                                            ],
                                            "type": "text",
                                            "content": "Figure 7: Qualitative monocular depth estimation performance comparing PackNet with previous methods, on frames from the KITTI dataset (Eigen test split). Our method is able to capture sharper details and structure (e.g., on vehicles, pedestrians, and thin poles) thanks to the learned preservation of spatial information."
                                        }
                                    ]
                                }
                            ],
                            "index": 3,
                            "angle": 0,
                            "type": "image_caption"
                        }
                    ],
                    "index": 2
                }
            ],
            "discarded_blocks": [
                {
                    "bbox": [
                        294,
                        754,
                        315,
                        764
                    ],
                    "type": "page_number",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                294,
                                754,
                                315,
                                764
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        294,
                                        754,
                                        315,
                                        764
                                    ],
                                    "type": "text",
                                    "content": "2491"
                                }
                            ]
                        }
                    ],
                    "index": 4
                }
            ],
            "page_size": [
                612,
                792
            ],
            "page_idx": 6
        },
        {
            "para_blocks": [
                {
                    "type": "image",
                    "bbox": [
                        62,
                        62,
                        272,
                        159
                    ],
                    "blocks": [
                        {
                            "bbox": [
                                62,
                                62,
                                272,
                                159
                            ],
                            "lines": [
                                {
                                    "bbox": [
                                        62,
                                        62,
                                        272,
                                        159
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                62,
                                                62,
                                                272,
                                                159
                                            ],
                                            "type": "image",
                                            "image_path": "1e30ca174d9bfc19e852603f7dd6bd66fd575a031af90cb6c70b45286ce1bbec.jpg"
                                        }
                                    ]
                                }
                            ],
                            "index": 0,
                            "angle": 0,
                            "type": "image_body"
                        },
                        {
                            "bbox": [
                                46,
                                170,
                                287,
                                266
                            ],
                            "lines": [
                                {
                                    "bbox": [
                                        46,
                                        170,
                                        287,
                                        266
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                46,
                                                170,
                                                287,
                                                266
                                            ],
                                            "type": "text",
                                            "content": "Figure 8: Performance of different depth network architectures for varying numbers of parameters on the original KITTI Eigen split [13] with resolutions of "
                                        },
                                        {
                                            "bbox": [
                                                46,
                                                170,
                                                287,
                                                266
                                            ],
                                            "type": "inline_equation",
                                            "content": "640 \\times 192"
                                        },
                                        {
                                            "bbox": [
                                                46,
                                                170,
                                                287,
                                                266
                                            ],
                                            "type": "text",
                                            "content": " (MR) and "
                                        },
                                        {
                                            "bbox": [
                                                46,
                                                170,
                                                287,
                                                266
                                            ],
                                            "type": "inline_equation",
                                            "content": "1280 \\times 384"
                                        },
                                        {
                                            "bbox": [
                                                46,
                                                170,
                                                287,
                                                266
                                            ],
                                            "type": "text",
                                            "content": " (HR). While the ResNet family plateaus at 70M parameters, the PackNet family matches its performance at the same number of parameters for MR, outperforms it clearly for HR, and improves significantly with more parameters in both cases without overfitting."
                                        }
                                    ]
                                }
                            ],
                            "index": 1,
                            "angle": 0,
                            "type": "image_caption"
                        }
                    ],
                    "index": 0
                },
                {
                    "bbox": [
                        46,
                        268,
                        287,
                        329
                    ],
                    "type": "text",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                46,
                                268,
                                287,
                                329
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        46,
                                        268,
                                        287,
                                        329
                                    ],
                                    "type": "text",
                                    "content": "The PackNet family is also consistently better at higher resolution, as it properly preserves and propagates spatial information between layers. In contrast, as reported in prior works [18], ResNet architectures do not scale well, with only minor improvements at higher resolution."
                                }
                            ]
                        }
                    ],
                    "index": 2
                },
                {
                    "bbox": [
                        47,
                        335,
                        148,
                        346
                    ],
                    "type": "title",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                47,
                                335,
                                148,
                                346
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        47,
                                        335,
                                        148,
                                        346
                                    ],
                                    "type": "text",
                                    "content": "5.6. Ablation Studies"
                                }
                            ]
                        }
                    ],
                    "index": 3
                },
                {
                    "bbox": [
                        46,
                        354,
                        287,
                        472
                    ],
                    "type": "text",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                46,
                                354,
                                287,
                                472
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        46,
                                        354,
                                        287,
                                        472
                                    ],
                                    "type": "text",
                                    "content": "To further study the performance improvements that PackNet provides, we perform an ablative analysis on the different architectural components introduced, as depicted in Table 4. We show that the base architecture, without the proposed packing and unpacking blocks, already produces a strong baseline for the monocular depth estimation task. The introduction of packing and unpacking boosts depth estimation performance, especially as more 3D convolutional filters are added, with new state-of-the-art results being achieved by the architecture described in Table 1."
                                }
                            ]
                        }
                    ],
                    "index": 4
                },
                {
                    "bbox": [
                        46,
                        473,
                        287,
                        605
                    ],
                    "type": "text",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                46,
                                473,
                                287,
                                605
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        46,
                                        473,
                                        287,
                                        605
                                    ],
                                    "type": "text",
                                    "content": "As mentioned in [14, 18], ResNet architectures highly benefit from ImageNet pretraining, since they were originally developed for classification tasks. Interestingly, we also noticed that the performance of pretrained ResNet architectures degrades in longer training periods, due to catastrophic forgetting that leads to overfitting. The proposed PackNet architecture, on the other hand, achieves state-of-the-art results from randomly initialized weights, and can be further improved by self-supervised pretraining on other datasets, thus properly leveraging the large-scale availability of unlabeled information thanks to its structure."
                                }
                            ]
                        }
                    ],
                    "index": 5
                },
                {
                    "bbox": [
                        47,
                        611,
                        190,
                        624
                    ],
                    "type": "title",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                47,
                                611,
                                190,
                                624
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        47,
                                        611,
                                        190,
                                        624
                                    ],
                                    "type": "text",
                                    "content": "5.7. Generalization Capability"
                                }
                            ]
                        }
                    ],
                    "index": 6
                },
                {
                    "bbox": [
                        46,
                        629,
                        288,
                        715
                    ],
                    "type": "text",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                46,
                                629,
                                288,
                                715
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        46,
                                        629,
                                        288,
                                        715
                                    ],
                                    "type": "text",
                                    "content": "We also investigate the generalization performance of PackNet, as evidence that it does not simply memorize training data but learns transferable discriminative features. To assess this, we evaluate on the recent NuScenes dataset [4] models trained on a combination of CityScapes and KITTI "
                                },
                                {
                                    "bbox": [
                                        46,
                                        629,
                                        288,
                                        715
                                    ],
                                    "type": "inline_equation",
                                    "content": "(\\mathrm{CS} + \\mathrm{K})"
                                },
                                {
                                    "bbox": [
                                        46,
                                        629,
                                        288,
                                        715
                                    ],
                                    "type": "text",
                                    "content": ", without any fine-tuning. Results in Table 5 show PackNet indeed generalizes better across a large spectrum of"
                                }
                            ]
                        }
                    ],
                    "index": 7
                },
                {
                    "type": "table",
                    "bbox": [
                        312,
                        59,
                        537,
                        184
                    ],
                    "blocks": [
                        {
                            "bbox": [
                                312,
                                59,
                                537,
                                184
                            ],
                            "lines": [
                                {
                                    "bbox": [
                                        312,
                                        59,
                                        537,
                                        184
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                312,
                                                59,
                                                537,
                                                184
                                            ],
                                            "type": "table",
                                            "html": "<table><tr><td>Depth Network</td><td>Abs Rel</td><td>Sq Rel</td><td>RMSE</td><td>\\( RMSE_{log} \\)</td><td>\\( \\delta_{1.25} \\)</td></tr><tr><td>ResNet18</td><td>0.133</td><td>1.023</td><td>5.123</td><td>0.211</td><td>0.845</td></tr><tr><td>\\( ResNet18^{\\ddagger} \\)</td><td>0.120</td><td>0.896</td><td>4.869</td><td>0.198</td><td>0.868</td></tr><tr><td>ResNet50</td><td>0.127</td><td>0.977</td><td>5.023</td><td>0.205</td><td>0.856</td></tr><tr><td>\\( ResNet50^{\\ddagger} \\)</td><td>0.117</td><td>0.900</td><td>4.826</td><td>0.196</td><td>0.873</td></tr><tr><td>PackNet (w/o pack/unpack)</td><td>0.122</td><td>0.880</td><td>4.816</td><td>0.198</td><td>0.864</td></tr><tr><td>PackNet (D=0)</td><td>0.121</td><td>0.922</td><td>4.831</td><td>0.195</td><td>0.869</td></tr><tr><td>PackNet (D=2)</td><td>0.118</td><td>0.802</td><td>4.656</td><td>0.194</td><td>0.868</td></tr><tr><td>PackNet (D=4)</td><td>0.113</td><td>0.818</td><td>4.621</td><td>0.190</td><td>0.875</td></tr><tr><td>PackNet (D=8)</td><td>0.111</td><td>0.785</td><td>4.601</td><td>0.189</td><td>0.878</td></tr></table>",
                                            "image_path": "499937867aa8d5e7a38837979e46b53d3e510b42cde845e815bf01e84972bb3c.jpg"
                                        }
                                    ]
                                }
                            ],
                            "index": 8,
                            "angle": 0,
                            "type": "table_body"
                        }
                    ],
                    "index": 8
                },
                {
                    "type": "table",
                    "bbox": [
                        320,
                        316,
                        527,
                        389
                    ],
                    "blocks": [
                        {
                            "bbox": [
                                304,
                                189,
                                545,
                                309
                            ],
                            "lines": [
                                {
                                    "bbox": [
                                        304,
                                        189,
                                        545,
                                        309
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                304,
                                                189,
                                                545,
                                                309
                                            ],
                                            "type": "text",
                                            "content": "Table 4: Ablation study on the PackNet architecture, on the standard KITTI benchmark for "
                                        },
                                        {
                                            "bbox": [
                                                304,
                                                189,
                                                545,
                                                309
                                            ],
                                            "type": "inline_equation",
                                            "content": "640 \\times 192"
                                        },
                                        {
                                            "bbox": [
                                                304,
                                                189,
                                                545,
                                                309
                                            ],
                                            "type": "text",
                                            "content": " resolution. ResNetXX indicates that specific architecture [20] as encoder, with and without ImageNet [9] pretraining (denoted with "
                                        },
                                        {
                                            "bbox": [
                                                304,
                                                189,
                                                545,
                                                309
                                            ],
                                            "type": "inline_equation",
                                            "content": "\\ddagger"
                                        },
                                        {
                                            "bbox": [
                                                304,
                                                189,
                                                545,
                                                309
                                            ],
                                            "type": "text",
                                            "content": "). We also show results with the proposed PackNet architecture, first without packing and unpacking (replaced respectively with convolutional striding and bilinear upsampling) and then with increasing numbers of 3D convolutional filters ("
                                        },
                                        {
                                            "bbox": [
                                                304,
                                                189,
                                                545,
                                                309
                                            ],
                                            "type": "inline_equation",
                                            "content": "D = 0"
                                        },
                                        {
                                            "bbox": [
                                                304,
                                                189,
                                                545,
                                                309
                                            ],
                                            "type": "text",
                                            "content": " indicates no 3D convolutions and the corresponding reshape operations)."
                                        }
                                    ]
                                }
                            ],
                            "index": 9,
                            "angle": 0,
                            "type": "table_caption"
                        },
                        {
                            "bbox": [
                                320,
                                316,
                                527,
                                389
                            ],
                            "lines": [
                                {
                                    "bbox": [
                                        320,
                                        316,
                                        527,
                                        389
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                320,
                                                316,
                                                527,
                                                389
                                            ],
                                            "type": "table",
                                            "html": "<table><tr><td>Method</td><td>Abs Rel</td><td>Sq Rel</td><td>RMSE</td><td>RMSElog</td><td>δ1.25</td></tr><tr><td>ResNet18</td><td>0.218</td><td>2.053</td><td>8.154</td><td>0.355</td><td>0.650</td></tr><tr><td>ResNet18‡</td><td>0.212</td><td>1.918</td><td>7.958</td><td>0.323</td><td>0.674</td></tr><tr><td>ResNet50</td><td>0.216</td><td>2.165</td><td>8.477</td><td>0.371</td><td>0.637</td></tr><tr><td>ResNet50‡</td><td>0.210</td><td>2.017</td><td>8.111</td><td>0.328</td><td>0.697</td></tr><tr><td>PackNet</td><td>0.187</td><td>1.852</td><td>7.636</td><td>0.289</td><td>0.742</td></tr></table>",
                                            "image_path": "5be66bf8ae8a17ddd21cf5853176699deffe6388e7356d9c555617b0fd7c70bc.jpg"
                                        }
                                    ]
                                }
                            ],
                            "index": 10,
                            "angle": 0,
                            "type": "table_body"
                        }
                    ],
                    "index": 10
                },
                {
                    "bbox": [
                        304,
                        395,
                        545,
                        444
                    ],
                    "lines": [
                        {
                            "bbox": [
                                304,
                                395,
                                545,
                                444
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        304,
                                        395,
                                        545,
                                        444
                                    ],
                                    "type": "text",
                                    "content": "Table 5: Generalization capability of different depth networks, trained on both KITTI and CityScapes and evaluated on NuScenes [4], for "
                                },
                                {
                                    "bbox": [
                                        304,
                                        395,
                                        545,
                                        444
                                    ],
                                    "type": "inline_equation",
                                    "content": "640 \\times 192"
                                },
                                {
                                    "bbox": [
                                        304,
                                        395,
                                        545,
                                        444
                                    ],
                                    "type": "text",
                                    "content": " resolution and distances up to "
                                },
                                {
                                    "bbox": [
                                        304,
                                        395,
                                        545,
                                        444
                                    ],
                                    "type": "inline_equation",
                                    "content": "80\\mathrm{m}"
                                },
                                {
                                    "bbox": [
                                        304,
                                        395,
                                        545,
                                        444
                                    ],
                                    "type": "text",
                                    "content": ". "
                                },
                                {
                                    "bbox": [
                                        304,
                                        395,
                                        545,
                                        444
                                    ],
                                    "type": "inline_equation",
                                    "content": "\\ddagger"
                                },
                                {
                                    "bbox": [
                                        304,
                                        395,
                                        545,
                                        444
                                    ],
                                    "type": "text",
                                    "content": " denotes ImageNet [9] pretraining."
                                }
                            ]
                        }
                    ],
                    "index": 11,
                    "angle": 0,
                    "type": "text"
                },
                {
                    "bbox": [
                        304,
                        449,
                        545,
                        498
                    ],
                    "type": "text",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                304,
                                449,
                                545,
                                498
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        304,
                                        449,
                                        545,
                                        498
                                    ],
                                    "type": "text",
                                    "content": "vehicles and countries (Germany for "
                                },
                                {
                                    "bbox": [
                                        304,
                                        449,
                                        545,
                                        498
                                    ],
                                    "type": "inline_equation",
                                    "content": "\\mathrm{CS + K}"
                                },
                                {
                                    "bbox": [
                                        304,
                                        449,
                                        545,
                                        498
                                    ],
                                    "type": "text",
                                    "content": ", USA + Singapore for NuScenes), outperforming standard architectures in all considered metrics without the need for large-scale supervised pretraining on ImageNet."
                                }
                            ]
                        }
                    ],
                    "index": 12
                },
                {
                    "bbox": [
                        306,
                        502,
                        378,
                        514
                    ],
                    "type": "title",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                306,
                                502,
                                378,
                                514
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        306,
                                        502,
                                        378,
                                        514
                                    ],
                                    "type": "text",
                                    "content": "6. Conclusion"
                                }
                            ]
                        }
                    ],
                    "index": 13
                },
                {
                    "bbox": [
                        304,
                        521,
                        545,
                        715
                    ],
                    "type": "text",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                304,
                                521,
                                545,
                                715
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        304,
                                        521,
                                        545,
                                        715
                                    ],
                                    "type": "text",
                                    "content": "We propose a new convolutional network architecture for self-supervised monocular depth estimation: PackNet. It leverages novel, symmetrical, detail-preserving packing and unpacking blocks that jointly learn to compress and decompress high resolution visual information for fine-grained predictions. Although purely trained on unlabeled monocular videos, our approach outperforms other existing self- and semi-supervised methods and is even competitive with fully-supervised methods while able to run in real-time. It also generalizes better to different datasets and unseen environments without the need for ImageNet pretraining, especially when considering longer depth ranges, as assessed up to "
                                },
                                {
                                    "bbox": [
                                        304,
                                        521,
                                        545,
                                        715
                                    ],
                                    "type": "inline_equation",
                                    "content": "200\\mathrm{m}"
                                },
                                {
                                    "bbox": [
                                        304,
                                        521,
                                        545,
                                        715
                                    ],
                                    "type": "text",
                                    "content": " on our new DDAD dataset. Additionally, by leveraging during training only weak velocity information, we are able to make our model scale-aware, i.e. producing metrically accurate depth maps from a single image."
                                }
                            ]
                        }
                    ],
                    "index": 14
                }
            ],
            "discarded_blocks": [
                {
                    "bbox": [
                        295,
                        753,
                        315,
                        763
                    ],
                    "type": "page_number",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                295,
                                753,
                                315,
                                763
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        295,
                                        753,
                                        315,
                                        763
                                    ],
                                    "type": "text",
                                    "content": "2492"
                                }
                            ]
                        }
                    ],
                    "index": 15
                }
            ],
            "page_size": [
                612,
                792
            ],
            "page_idx": 7
        },
        {
            "para_blocks": [
                {
                    "bbox": [
                        48,
                        71,
                        106,
                        83
                    ],
                    "type": "title",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                48,
                                71,
                                106,
                                83
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        48,
                                        71,
                                        106,
                                        83
                                    ],
                                    "type": "text",
                                    "content": "References"
                                }
                            ]
                        }
                    ],
                    "index": 0
                },
                {
                    "bbox": [
                        48,
                        91,
                        287,
                        713
                    ],
                    "type": "list",
                    "angle": 0,
                    "index": 15,
                    "blocks": [
                        {
                            "bbox": [
                                53,
                                91,
                                285,
                                112
                            ],
                            "type": "ref_text",
                            "angle": 0,
                            "lines": [
                                {
                                    "bbox": [
                                        53,
                                        91,
                                        285,
                                        112
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                53,
                                                91,
                                                285,
                                                112
                                            ],
                                            "type": "text",
                                            "content": "[1] TensorRT python library. https://developer. nvidia.com/tensorrt. Accessed: 2019-11-09. 6"
                                        }
                                    ]
                                }
                            ],
                            "index": 1
                        },
                        {
                            "bbox": [
                                53,
                                114,
                                287,
                                156
                            ],
                            "type": "ref_text",
                            "angle": 0,
                            "lines": [
                                {
                                    "bbox": [
                                        53,
                                        114,
                                        287,
                                        156
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                53,
                                                114,
                                                287,
                                                156
                                            ],
                                            "type": "text",
                                            "content": "[2] Aayush Bansal, Xinlei Chen, Bryan Russell, Abhinav Gupta, and Deva Ramanan. Pixelnet: Representation of the pixels, by the pixels, and for the pixels. arXiv preprint arXiv:1702.06506, 2017. 2"
                                        }
                                    ]
                                }
                            ],
                            "index": 2
                        },
                        {
                            "bbox": [
                                53,
                                159,
                                287,
                                191
                            ],
                            "type": "ref_text",
                            "angle": 0,
                            "lines": [
                                {
                                    "bbox": [
                                        53,
                                        159,
                                        287,
                                        191
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                53,
                                                159,
                                                287,
                                                191
                                            ],
                                            "type": "text",
                                            "content": "[3] Jens Behrmann, Will Grathwohl, Ricky TQ Chen, David Duvenaud, and Jörn-Henrik Jacobsen. Invertible residual networks. arXiv preprint arXiv:1811.00995, 2018. 4"
                                        }
                                    ]
                                }
                            ],
                            "index": 3
                        },
                        {
                            "bbox": [
                                53,
                                192,
                                287,
                                245
                            ],
                            "type": "ref_text",
                            "angle": 0,
                            "lines": [
                                {
                                    "bbox": [
                                        53,
                                        192,
                                        287,
                                        245
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                53,
                                                192,
                                                287,
                                                245
                                            ],
                                            "type": "text",
                                            "content": "[4] Holger Caesar, Varun Bankiti, Alex H. Lang, Sourabh Vora, Venice Erin Liong, Qiang Xu, Anush Krishnan, Yu Pan, Giancarlo Baldan, and Oscar Beijbom. nuscenes: A multimodal dataset for autonomous driving. CoRR, 2019. 2, 5, 8"
                                        }
                                    ]
                                }
                            ],
                            "index": 4
                        },
                        {
                            "bbox": [
                                53,
                                247,
                                287,
                                290
                            ],
                            "type": "ref_text",
                            "angle": 0,
                            "lines": [
                                {
                                    "bbox": [
                                        53,
                                        247,
                                        287,
                                        290
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                53,
                                                247,
                                                287,
                                                290
                                            ],
                                            "type": "text",
                                            "content": "[5] Vincent Casser, Soeren Pirk, Reza Mahjourian, and Anelia Angelova. Depth prediction without the sensors: Leveraging structure for unsupervised learning from monocular videos. In AAAI, 2019. 1, 2, 3, 6, 7"
                                        }
                                    ]
                                }
                            ],
                            "index": 5
                        },
                        {
                            "bbox": [
                                53,
                                292,
                                287,
                                335
                            ],
                            "type": "ref_text",
                            "angle": 0,
                            "lines": [
                                {
                                    "bbox": [
                                        53,
                                        292,
                                        287,
                                        335
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                53,
                                                292,
                                                287,
                                                335
                                            ],
                                            "type": "text",
                                            "content": "[6] Yunjin Chen and Thomas Pock. Trainable nonlinear reaction diffusion: A flexible framework for fast and effective image restoration. IEEE Transactions on Pattern Analysis and Machine Intelligence, 39:1256-1272, 2017. 4"
                                        }
                                    ]
                                }
                            ],
                            "index": 6
                        },
                        {
                            "bbox": [
                                53,
                                337,
                                287,
                                368
                            ],
                            "type": "ref_text",
                            "angle": 0,
                            "lines": [
                                {
                                    "bbox": [
                                        53,
                                        337,
                                        287,
                                        368
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                53,
                                                337,
                                                287,
                                                368
                                            ],
                                            "type": "text",
                                            "content": "[7] Djork-Arné Clevert, Thomas Unterthiner, and Sepp Hochreiter. Fast and accurate deep network learning by exponential linear units (elus). In ICLR, 2016. 4"
                                        }
                                    ]
                                }
                            ],
                            "index": 7
                        },
                        {
                            "bbox": [
                                53,
                                370,
                                287,
                                434
                            ],
                            "type": "ref_text",
                            "angle": 0,
                            "lines": [
                                {
                                    "bbox": [
                                        53,
                                        370,
                                        287,
                                        434
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                53,
                                                370,
                                                287,
                                                434
                                            ],
                                            "type": "text",
                                            "content": "[8] Marius Cordts, Mohamed Omran, Sebastian Ramos, Timo Rehfeld, Markus Enzweiler, Rodrigo Benenson, Uwe Franke, Stefan Roth, and Bernt Schiele. The cityscapes dataset for semantic urban scene understanding. In IEEE conference on computer vision and pattern recognition, pages 3213-3223, 2016. 5, 6"
                                        }
                                    ]
                                }
                            ],
                            "index": 8
                        },
                        {
                            "bbox": [
                                53,
                                436,
                                287,
                                479
                            ],
                            "type": "ref_text",
                            "angle": 0,
                            "lines": [
                                {
                                    "bbox": [
                                        53,
                                        436,
                                        287,
                                        479
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                53,
                                                436,
                                                287,
                                                479
                                            ],
                                            "type": "text",
                                            "content": "[9] Jia Deng, Wei Dong, Richard Socher, Li jia Li, Kai Li, and Li Fei-fei. Imagenet: A large-scale hierarchical image database. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2009. 6, 7, 8"
                                        }
                                    ]
                                }
                            ],
                            "index": 9
                        },
                        {
                            "bbox": [
                                48,
                                481,
                                286,
                                502
                            ],
                            "type": "ref_text",
                            "angle": 0,
                            "lines": [
                                {
                                    "bbox": [
                                        48,
                                        481,
                                        286,
                                        502
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                48,
                                                481,
                                                286,
                                                502
                                            ],
                                            "type": "text",
                                            "content": "[10] Laurent Dinh, Jascha Sohl-Dickstein, and Samy Bengio. Density estimation using real nvp. In *ICLR*, 2017. 4"
                                        }
                                    ]
                                }
                            ],
                            "index": 10
                        },
                        {
                            "bbox": [
                                48,
                                504,
                                287,
                                545
                            ],
                            "type": "ref_text",
                            "angle": 0,
                            "lines": [
                                {
                                    "bbox": [
                                        48,
                                        504,
                                        287,
                                        545
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                48,
                                                504,
                                                287,
                                                545
                                            ],
                                            "type": "text",
                                            "content": "[11] Chao Dong, Chen Change Loy, Kaiming He, and Xiaou Tang. Image super-resolution using deep convolutional networks. IEEE Trans. Pattern Anal. Mach. Intell., 38(2):295-307, Feb. 2016. 4"
                                        }
                                    ]
                                }
                            ],
                            "index": 11
                        },
                        {
                            "bbox": [
                                48,
                                548,
                                287,
                                612
                            ],
                            "type": "ref_text",
                            "angle": 0,
                            "lines": [
                                {
                                    "bbox": [
                                        48,
                                        548,
                                        287,
                                        612
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                48,
                                                548,
                                                287,
                                                612
                                            ],
                                            "type": "text",
                                            "content": "[12] Alexey Dosovitskiy, Philipp Fischer, Eddy Ilg, Philip Hausser, Caner Hazirbas, Vladimir Golkov, Patrick Van Der Smagt, Daniel Cremers, and Thomas Brox. Flownet: Learning optical flow with convolutional networks. In Proceedings of the IEEE international conference on computer vision, pages 2758-2766, 2015. 2"
                                        }
                                    ]
                                }
                            ],
                            "index": 12
                        },
                        {
                            "bbox": [
                                48,
                                614,
                                287,
                                658
                            ],
                            "type": "ref_text",
                            "angle": 0,
                            "lines": [
                                {
                                    "bbox": [
                                        48,
                                        614,
                                        287,
                                        658
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                48,
                                                614,
                                                287,
                                                658
                                            ],
                                            "type": "text",
                                            "content": "[13] David Eigen, Christian Puhrsch, and Rob Fergus. Depth map prediction from a single image using a multi-scale deep network. In Advances in neural information processing systems, pages 2366-2374, 2014. 2, 5, 6, 7, 8"
                                        }
                                    ]
                                }
                            ],
                            "index": 13
                        },
                        {
                            "bbox": [
                                48,
                                659,
                                287,
                                713
                            ],
                            "type": "ref_text",
                            "angle": 0,
                            "lines": [
                                {
                                    "bbox": [
                                        48,
                                        659,
                                        287,
                                        713
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                48,
                                                659,
                                                287,
                                                713
                                            ],
                                            "type": "text",
                                            "content": "[14] Huan Fu, Mingming Gong, Chaohui Wang, Kayhan Bat-manghelich, and Dacheng Tao. Deep ordinal regression network for monocular depth estimation. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 2002-2011, 2018. 6, 7, 8"
                                        }
                                    ]
                                }
                            ],
                            "index": 14
                        }
                    ],
                    "sub_type": "ref_text"
                },
                {
                    "bbox": [
                        307,
                        73,
                        545,
                        713
                    ],
                    "type": "list",
                    "angle": 0,
                    "index": 33,
                    "blocks": [
                        {
                            "bbox": [
                                307,
                                73,
                                545,
                                117
                            ],
                            "type": "ref_text",
                            "angle": 0,
                            "lines": [
                                {
                                    "bbox": [
                                        307,
                                        73,
                                        545,
                                        117
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                307,
                                                73,
                                                545,
                                                117
                                            ],
                                            "type": "text",
                                            "content": "[15] Ravi Garg, Vijay Kumar BG, Gustavo Carneiro, and Ian Reid. Unsupervised cnn for single view depth estimation: Geometry to the rescue. In European Conference on Computer Vision, pages 740-756. Springer, 2016. 2, 3"
                                        }
                                    ]
                                }
                            ],
                            "index": 16
                        },
                        {
                            "bbox": [
                                308,
                                118,
                                545,
                                160
                            ],
                            "type": "ref_text",
                            "angle": 0,
                            "lines": [
                                {
                                    "bbox": [
                                        308,
                                        118,
                                        545,
                                        160
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                308,
                                                118,
                                                545,
                                                160
                                            ],
                                            "type": "text",
                                            "content": "[16] Andreas Geiger, Philip Lenz, Christoph Stiller, and Raquel Urtasun. Vision meets robotics: The kitti dataset. The International Journal of Robotics Research, 32(11):1231-1237, 2013. 2, 5"
                                        }
                                    ]
                                }
                            ],
                            "index": 17
                        },
                        {
                            "bbox": [
                                308,
                                163,
                                545,
                                205
                            ],
                            "type": "ref_text",
                            "angle": 0,
                            "lines": [
                                {
                                    "bbox": [
                                        308,
                                        163,
                                        545,
                                        205
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                308,
                                                163,
                                                545,
                                                205
                                            ],
                                            "type": "text",
                                            "content": "[17] Clément Godard, Oisin Mac Aodha, and Gabriel J Brosstow. Unsupervised monocular depth estimation with left-right consistency. In CVPR, volume 2, page 7, 2017. 2, 3"
                                        }
                                    ]
                                }
                            ],
                            "index": 18
                        },
                        {
                            "bbox": [
                                308,
                                208,
                                545,
                                240
                            ],
                            "type": "ref_text",
                            "angle": 0,
                            "lines": [
                                {
                                    "bbox": [
                                        308,
                                        208,
                                        545,
                                        240
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                308,
                                                208,
                                                545,
                                                240
                                            ],
                                            "type": "text",
                                            "content": "[18] Clément Godard, Oisin Mac Aodha, Michael Firman, and Gabriel J. Brostow. Digging into self-supervised monocular depth prediction. In ICCV, 2019. 3, 6, 7, 8"
                                        }
                                    ]
                                }
                            ],
                            "index": 19
                        },
                        {
                            "bbox": [
                                308,
                                241,
                                545,
                                262
                            ],
                            "type": "ref_text",
                            "angle": 0,
                            "lines": [
                                {
                                    "bbox": [
                                        308,
                                        241,
                                        545,
                                        262
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                308,
                                                241,
                                                545,
                                                262
                                            ],
                                            "type": "text",
                                            "content": "[19] Benjamin Graham. Fractional max-pooling. arXiv:1412.607, 2015. 2, 4"
                                        }
                                    ]
                                }
                            ],
                            "index": 20
                        },
                        {
                            "bbox": [
                                308,
                                264,
                                545,
                                308
                            ],
                            "type": "ref_text",
                            "angle": 0,
                            "lines": [
                                {
                                    "bbox": [
                                        308,
                                        264,
                                        545,
                                        308
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                308,
                                                264,
                                                545,
                                                308
                                            ],
                                            "type": "text",
                                            "content": "[20] Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Deep residual learning for image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 770-778, 2016. 1, 8"
                                        }
                                    ]
                                }
                            ],
                            "index": 21
                        },
                        {
                            "bbox": [
                                308,
                                309,
                                545,
                                342
                            ],
                            "type": "ref_text",
                            "angle": 0,
                            "lines": [
                                {
                                    "bbox": [
                                        308,
                                        309,
                                        545,
                                        342
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                308,
                                                309,
                                                545,
                                                342
                                            ],
                                            "type": "text",
                                            "content": "[21] Jrn-Henrik Jacobsen, Arnold W.M. Smeulders, and Edouard Oyallon. i-revnet: Deep invertible networks. In International Conference on Learning Representations, 2018. 4"
                                        }
                                    ]
                                }
                            ],
                            "index": 22
                        },
                        {
                            "bbox": [
                                308,
                                343,
                                545,
                                376
                            ],
                            "type": "ref_text",
                            "angle": 0,
                            "lines": [
                                {
                                    "bbox": [
                                        308,
                                        343,
                                        545,
                                        376
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                308,
                                                343,
                                                545,
                                                376
                                            ],
                                            "type": "text",
                                            "content": "[22] Max Jaderberg, Karen Simonyan, Andrew Zisserman, et al. Spatial transformer networks. In Advances in neural information processing systems, pages 2017-2025, 2015. 2"
                                        }
                                    ]
                                }
                            ],
                            "index": 23
                        },
                        {
                            "bbox": [
                                308,
                                377,
                                545,
                                430
                            ],
                            "type": "ref_text",
                            "angle": 0,
                            "lines": [
                                {
                                    "bbox": [
                                        308,
                                        377,
                                        545,
                                        430
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                308,
                                                377,
                                                545,
                                                430
                                            ],
                                            "type": "text",
                                            "content": "[23] Alex Kendall, Yarin Gal, and Roberto Cipolla. Multi-task learning using uncertainty to weigh losses for scene geometry and semantics. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 7482-7491, 2018. 1"
                                        }
                                    ]
                                }
                            ],
                            "index": 24
                        },
                        {
                            "bbox": [
                                308,
                                433,
                                545,
                                464
                            ],
                            "type": "ref_text",
                            "angle": 0,
                            "lines": [
                                {
                                    "bbox": [
                                        308,
                                        433,
                                        545,
                                        464
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                308,
                                                433,
                                                545,
                                                464
                                            ],
                                            "type": "text",
                                            "content": "[24] Diederik P Kingma and Jimmy Ba. Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980, 2014.5"
                                        }
                                    ]
                                }
                            ],
                            "index": 25
                        },
                        {
                            "bbox": [
                                308,
                                467,
                                545,
                                499
                            ],
                            "type": "ref_text",
                            "angle": 0,
                            "lines": [
                                {
                                    "bbox": [
                                        308,
                                        467,
                                        545,
                                        499
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                308,
                                                467,
                                                545,
                                                499
                                            ],
                                            "type": "text",
                                            "content": "[25] Durk P Kingma and Prafulla Dhariwal. Glow: Generative flow with invertible 1x1 convolutions. In Advances in Neural Information Processing Systems, 2018. 4"
                                        }
                                    ]
                                }
                            ],
                            "index": 26
                        },
                        {
                            "bbox": [
                                308,
                                501,
                                545,
                                533
                            ],
                            "type": "ref_text",
                            "angle": 0,
                            "lines": [
                                {
                                    "bbox": [
                                        308,
                                        501,
                                        545,
                                        533
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                308,
                                                501,
                                                545,
                                                533
                                            ],
                                            "type": "text",
                                            "content": "[26] Maria Klodt and Andrea Vedaldi. Supervising the new with the old: Learning sfm from sfm. In European Conference on Computer Vision, pages 713-728. Springer, 2018. 2"
                                        }
                                    ]
                                }
                            ],
                            "index": 27
                        },
                        {
                            "bbox": [
                                308,
                                535,
                                545,
                                567
                            ],
                            "type": "ref_text",
                            "angle": 0,
                            "lines": [
                                {
                                    "bbox": [
                                        308,
                                        535,
                                        545,
                                        567
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                308,
                                                535,
                                                545,
                                                567
                                            ],
                                            "type": "text",
                                            "content": "[27] Alexander Kolesnikov, Xiaohua Zhai, and Lucas Beyer. Revisiting self-supervised visual representation learning. arXiv preprint arXiv:1901.09005, 2019. 1"
                                        }
                                    ]
                                }
                            ],
                            "index": 28
                        },
                        {
                            "bbox": [
                                308,
                                568,
                                545,
                                611
                            ],
                            "type": "ref_text",
                            "angle": 0,
                            "lines": [
                                {
                                    "bbox": [
                                        308,
                                        568,
                                        545,
                                        611
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                308,
                                                568,
                                                545,
                                                611
                                            ],
                                            "type": "text",
                                            "content": "[28] Yevhen Kuznietsov, Jörg Stuckler, and Bastian Leibe. Semi-supervised deep learning for monocular depth map prediction. In IEEE Conference on Computer Vision and Pattern Recognition, pages 6647-6655, 2017. 6, 7"
                                        }
                                    ]
                                }
                            ],
                            "index": 29
                        },
                        {
                            "bbox": [
                                308,
                                613,
                                545,
                                656
                            ],
                            "type": "ref_text",
                            "angle": 0,
                            "lines": [
                                {
                                    "bbox": [
                                        308,
                                        613,
                                        545,
                                        656
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                308,
                                                613,
                                                545,
                                                656
                                            ],
                                            "type": "text",
                                            "content": "[29] Chen-Yu Lee, Patrick Gallagher, and Zhuowen Tu. Generalizing pooling functions in convolutional neural networks: Mixed, gated, and tree. In International Conference on Artificial Intelligence and Statistics (AISTATS), 2016. 2"
                                        }
                                    ]
                                }
                            ],
                            "index": 30
                        },
                        {
                            "bbox": [
                                308,
                                658,
                                545,
                                689
                            ],
                            "type": "ref_text",
                            "angle": 0,
                            "lines": [
                                {
                                    "bbox": [
                                        308,
                                        658,
                                        545,
                                        689
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                308,
                                                658,
                                                545,
                                                689
                                            ],
                                            "type": "text",
                                            "content": "[30] Kuan-Hui Lee, German Ros, Jie Li, and Adrien Gaidon. Spigan: Privileged adversarial learning from simulation. In ICLR, 2019. 1"
                                        }
                                    ]
                                }
                            ],
                            "index": 31
                        },
                        {
                            "bbox": [
                                308,
                                691,
                                545,
                                713
                            ],
                            "type": "ref_text",
                            "angle": 0,
                            "lines": [
                                {
                                    "bbox": [
                                        308,
                                        691,
                                        545,
                                        713
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                308,
                                                691,
                                                545,
                                                713
                                            ],
                                            "type": "text",
                                            "content": "[31] Jonathan Long, Evan Shelhamer, and Trevor Darrell. Fully convolutional networks for semantic segmentation. In Pro"
                                        }
                                    ]
                                }
                            ],
                            "index": 32
                        }
                    ],
                    "sub_type": "ref_text"
                }
            ],
            "discarded_blocks": [
                {
                    "bbox": [
                        295,
                        754,
                        315,
                        763
                    ],
                    "type": "page_number",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                295,
                                754,
                                315,
                                763
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        295,
                                        754,
                                        315,
                                        763
                                    ],
                                    "type": "text",
                                    "content": "2493"
                                }
                            ]
                        }
                    ],
                    "index": 34
                }
            ],
            "page_size": [
                612,
                792
            ],
            "page_idx": 8
        },
        {
            "para_blocks": [
                {
                    "bbox": [
                        48,
                        73,
                        287,
                        576
                    ],
                    "type": "list",
                    "angle": 0,
                    "index": 11,
                    "blocks": [
                        {
                            "bbox": [
                                67,
                                73,
                                286,
                                95
                            ],
                            "type": "ref_text",
                            "angle": 0,
                            "lines": [
                                {
                                    "bbox": [
                                        67,
                                        73,
                                        286,
                                        95
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                67,
                                                73,
                                                286,
                                                95
                                            ],
                                            "type": "text",
                                            "content": "ceedings of the IEEE conference on computer vision and pattern recognition, pages 3431-3440, 2015. 2"
                                        }
                                    ]
                                }
                            ],
                            "index": 0
                        },
                        {
                            "bbox": [
                                48,
                                96,
                                287,
                                139
                            ],
                            "type": "ref_text",
                            "angle": 0,
                            "lines": [
                                {
                                    "bbox": [
                                        48,
                                        96,
                                        287,
                                        139
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                48,
                                                96,
                                                287,
                                                139
                                            ],
                                            "type": "text",
                                            "content": "[32] C. Luo, Z. Yang, P. Wang, Y. Wang, W. Xu, R. Nevatia, and A. Yuille. Every pixel counts++: Joint learning of geometry and motion with 3d holistic understanding. arXiv preprint arXiv:1810.06125, 2018. 7"
                                        }
                                    ]
                                }
                            ],
                            "index": 1
                        },
                        {
                            "bbox": [
                                48,
                                141,
                                287,
                                196
                            ],
                            "type": "ref_text",
                            "angle": 0,
                            "lines": [
                                {
                                    "bbox": [
                                        48,
                                        141,
                                        287,
                                        196
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                48,
                                                141,
                                                287,
                                                196
                                            ],
                                            "type": "text",
                                            "content": "[33] Reza Mahjourian, Martin Wicke, and Anelia Angelova. Unsupervised learning of depth and ego-motion from monocular video using 3d geometric constraints. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 5667-5675, 2018. 1, 2, 3, 6, 7"
                                        }
                                    ]
                                }
                            ],
                            "index": 2
                        },
                        {
                            "bbox": [
                                48,
                                198,
                                287,
                                239
                            ],
                            "type": "ref_text",
                            "angle": 0,
                            "lines": [
                                {
                                    "bbox": [
                                        48,
                                        198,
                                        287,
                                        239
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                48,
                                                198,
                                                287,
                                                239
                                            ],
                                            "type": "text",
                                            "content": "[34] Fabian Manhardt, Wadim Kehl, and Adrien Gaidon. Roi10d: Monocular lifting of 2d detection to 6d pose and metric shape. IEEE Conference on Computer Vision and Pattern Recognition, 2018. 1"
                                        }
                                    ]
                                }
                            ],
                            "index": 3
                        },
                        {
                            "bbox": [
                                48,
                                241,
                                287,
                                307
                            ],
                            "type": "ref_text",
                            "angle": 0,
                            "lines": [
                                {
                                    "bbox": [
                                        48,
                                        241,
                                        287,
                                        307
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                48,
                                                241,
                                                287,
                                                307
                                            ],
                                            "type": "text",
                                            "content": "[35] Nikolaus Mayer, Eddy Ilg, Philip Hausser, Philipp Fischer, Daniel Cremers, Alexey Dosovitskiy, and Thomas Brox. A large dataset to train convolutional networks for disparity, optical flow, and scene flow estimation. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 4040-4048, 2016. 2, 5"
                                        }
                                    ]
                                }
                            ],
                            "index": 4
                        },
                        {
                            "bbox": [
                                48,
                                308,
                                287,
                                352
                            ],
                            "type": "ref_text",
                            "angle": 0,
                            "lines": [
                                {
                                    "bbox": [
                                        48,
                                        308,
                                        287,
                                        352
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                48,
                                                308,
                                                287,
                                                352
                                            ],
                                            "type": "text",
                                            "content": "[36] Jeff Michels, Ashutosh Saxena, and Andrew Y Ng. High speed obstacle avoidance using monocular vision and reinforcement learning. In 22nd international conference on Machine learning, pages 593-600. ACM, 2005. 1"
                                        }
                                    ]
                                }
                            ],
                            "index": 5
                        },
                        {
                            "bbox": [
                                48,
                                354,
                                287,
                                396
                            ],
                            "type": "ref_text",
                            "angle": 0,
                            "lines": [
                                {
                                    "bbox": [
                                        48,
                                        354,
                                        287,
                                        396
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                48,
                                                354,
                                                287,
                                                396
                                            ],
                                            "type": "text",
                                            "content": "[37] Adam Paszke, Sam Gross, Soumith Chintala, Gregory Chanan, Edward Yang, Zachary DeVito, Zeming Lin, Alban Desmaison, Luca Antiga, and Adam Lerer. Automatic differentiation in pytorch. In NIPS-W, 2017. 5"
                                        }
                                    ]
                                }
                            ],
                            "index": 6
                        },
                        {
                            "bbox": [
                                48,
                                398,
                                287,
                                441
                            ],
                            "type": "ref_text",
                            "angle": 0,
                            "lines": [
                                {
                                    "bbox": [
                                        48,
                                        398,
                                        287,
                                        441
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                48,
                                                398,
                                                287,
                                                441
                                            ],
                                            "type": "text",
                                            "content": "[38] Sudeep Pillai, Rares Ambrus, and Adrien Gaidon. Superdepth: Self-supervised, super-resolved monocular depth estimation. In Robotics and Automation (ICRA), 2019 IEEE International Conference on, 2018. 2, 6"
                                        }
                                    ]
                                }
                            ],
                            "index": 7
                        },
                        {
                            "bbox": [
                                48,
                                443,
                                287,
                                508
                            ],
                            "type": "ref_text",
                            "angle": 0,
                            "lines": [
                                {
                                    "bbox": [
                                        48,
                                        443,
                                        287,
                                        508
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                48,
                                                443,
                                                287,
                                                508
                                            ],
                                            "type": "text",
                                            "content": "[39] Wenzhe Shi, Jose Caballero, Ferenc Huszár, Johannes Totz, Andrew P Aitken, Rob Bishop, Daniel Rueckert, and Zehan Wang. Real-time single image and video super-resolution using an efficient sub-pixel convolutional neural network. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 1874-1883, 2016. 2, 4"
                                        }
                                    ]
                                }
                            ],
                            "index": 8
                        },
                        {
                            "bbox": [
                                48,
                                510,
                                287,
                                553
                            ],
                            "type": "ref_text",
                            "angle": 0,
                            "lines": [
                                {
                                    "bbox": [
                                        48,
                                        510,
                                        287,
                                        553
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                48,
                                                510,
                                                287,
                                                553
                                            ],
                                            "type": "text",
                                            "content": "[40] Nitish Srivastava, Geoffrey Hinton, Alex Krizhevsky, Ilya Sutskever, and Ruslan Salakhutdinov. Dropout: A simple way to prevent neural networks from overfitting. Journal of Machine Learning Research, 15:1929-1958, 2014. 4"
                                        }
                                    ]
                                }
                            ],
                            "index": 9
                        },
                        {
                            "bbox": [
                                48,
                                555,
                                287,
                                576
                            ],
                            "type": "ref_text",
                            "angle": 0,
                            "lines": [
                                {
                                    "bbox": [
                                        48,
                                        555,
                                        287,
                                        576
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                48,
                                                555,
                                                287,
                                                576
                                            ],
                                            "type": "text",
                                            "content": "[41] J. Uhrig, N. Schneider, L. Schneider, U. Franke, T. Brox, and A. Geiger. Sparsity invariant cnns. 3DV, 2017. 5, 6, 7"
                                        }
                                    ]
                                }
                            ],
                            "index": 10
                        }
                    ],
                    "sub_type": "ref_text"
                },
                {
                    "bbox": [
                        307,
                        73,
                        545,
                        576
                    ],
                    "type": "list",
                    "angle": 0,
                    "index": 24,
                    "blocks": [
                        {
                            "bbox": [
                                307,
                                73,
                                545,
                                128
                            ],
                            "type": "ref_text",
                            "angle": 0,
                            "lines": [
                                {
                                    "bbox": [
                                        307,
                                        73,
                                        545,
                                        128
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                307,
                                                73,
                                                545,
                                                128
                                            ],
                                            "type": "text",
                                            "content": "[42] Benjamin Ummenhofer, Huizhong Zhou, Jonas Uhrig, Nikolaus Mayer, Eddy Ilg, Alexey Dosovitskiy, and Thomas Brox. Demon: Depth and motion network for learning monocular stereo. In IEEE Conference on computer vision and pattern recognition (CVPR), volume 5, page 6, 2017. 2"
                                        }
                                    ]
                                }
                            ],
                            "index": 12
                        },
                        {
                            "bbox": [
                                307,
                                129,
                                545,
                                173
                            ],
                            "type": "ref_text",
                            "angle": 0,
                            "lines": [
                                {
                                    "bbox": [
                                        307,
                                        129,
                                        545,
                                        173
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                307,
                                                129,
                                                545,
                                                173
                                            ],
                                            "type": "text",
                                            "content": "[43] Chaoyang Wang, José Miguel Buenaposada, Rui Zhu, and Simon Lucey. Learning depth from monocular videos using direct methods. In IEEE Conference on Computer Vision and Pattern Recognition, pages 2022–2030, 2018. 2, 7"
                                        }
                                    ]
                                }
                            ],
                            "index": 13
                        },
                        {
                            "bbox": [
                                307,
                                175,
                                545,
                                217
                            ],
                            "type": "ref_text",
                            "angle": 0,
                            "lines": [
                                {
                                    "bbox": [
                                        307,
                                        175,
                                        545,
                                        217
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                307,
                                                175,
                                                545,
                                                217
                                            ],
                                            "type": "text",
                                            "content": "[44] Zhou Wang, Alan C Bovik, Hamid R Sheikh, and Eero P Simoncelli. Image quality assessment: from error visibility to structural similarity. IEEE transactions on image processing, 13(4):600-612, 2004. 2, 3"
                                        }
                                    ]
                                }
                            ],
                            "index": 14
                        },
                        {
                            "bbox": [
                                307,
                                217,
                                545,
                                261
                            ],
                            "type": "ref_text",
                            "angle": 0,
                            "lines": [
                                {
                                    "bbox": [
                                        307,
                                        217,
                                        545,
                                        261
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                307,
                                                217,
                                                545,
                                                261
                                            ],
                                            "type": "text",
                                            "content": "[45] Yuxin Wu and Kaiming He. Group normalization. In Computer Vision - ECCV 2018 - 15th European Conference, Munich, Germany, September 8-14, 2018, Proceedings, Part XIII, pages 3-19, 2018. 4"
                                        }
                                    ]
                                }
                            ],
                            "index": 15
                        },
                        {
                            "bbox": [
                                307,
                                262,
                                545,
                                305
                            ],
                            "type": "ref_text",
                            "angle": 0,
                            "lines": [
                                {
                                    "bbox": [
                                        307,
                                        262,
                                        545,
                                        305
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                307,
                                                262,
                                                545,
                                                305
                                            ],
                                            "type": "text",
                                            "content": "[46] Nan Yang, Rui Wang, Jörg Stuckler, and Daniel Cremers. Deep virtual stereo odometry: Leveraging deep depth prediction for monocular direct sparse odometry. arXiv preprint arXiv:1807.02570, 2018. 2"
                                        }
                                    ]
                                }
                            ],
                            "index": 16
                        },
                        {
                            "bbox": [
                                307,
                                307,
                                545,
                                350
                            ],
                            "type": "ref_text",
                            "angle": 0,
                            "lines": [
                                {
                                    "bbox": [
                                        307,
                                        307,
                                        545,
                                        350
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                307,
                                                307,
                                                545,
                                                350
                                            ],
                                            "type": "text",
                                            "content": "[47] Zhichao Yin and Jianping Shi. Geonet: Unsupervised learning of dense depth, optical flow and camera pose. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), volume 2, 2018. 1, 2, 7"
                                        }
                                    ]
                                }
                            ],
                            "index": 17
                        },
                        {
                            "bbox": [
                                307,
                                352,
                                545,
                                385
                            ],
                            "type": "ref_text",
                            "angle": 0,
                            "lines": [
                                {
                                    "bbox": [
                                        307,
                                        352,
                                        545,
                                        385
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                307,
                                                352,
                                                545,
                                                385
                                            ],
                                            "type": "text",
                                            "content": "[48] Fisher Yu, Vladlen Koltun, and Thomas Funkhouser. Dilated residual networks. In The IEEE Conference on Computer Vision and Pattern Recognition (CVPR), July 2017. 2"
                                        }
                                    ]
                                }
                            ],
                            "index": 18
                        },
                        {
                            "bbox": [
                                307,
                                386,
                                545,
                                418
                            ],
                            "type": "ref_text",
                            "angle": 0,
                            "lines": [
                                {
                                    "bbox": [
                                        307,
                                        386,
                                        545,
                                        418
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                307,
                                                386,
                                                545,
                                                418
                                            ],
                                            "type": "text",
                                            "content": "[49] Hao Zhang and Jianwei Ma. Hartley spectral pooling for deep learning. Computing Research Repository, abs/1810.04028, 2018. 4"
                                        }
                                    ]
                                }
                            ],
                            "index": 19
                        },
                        {
                            "bbox": [
                                307,
                                419,
                                545,
                                463
                            ],
                            "type": "ref_text",
                            "angle": 0,
                            "lines": [
                                {
                                    "bbox": [
                                        307,
                                        419,
                                        545,
                                        463
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                307,
                                                419,
                                                545,
                                                463
                                            ],
                                            "type": "text",
                                            "content": "[50] Junsheng Zhou, Yuwang Wang, Naiyan Wang, and Wenjun Zeng. Unsupervised high-resolution depth learning from videos with dual networks. In Inter. Conf. on Computer Vision. IEEE, IEEE, 2019. 7"
                                        }
                                    ]
                                }
                            ],
                            "index": 20
                        },
                        {
                            "bbox": [
                                307,
                                464,
                                545,
                                508
                            ],
                            "type": "ref_text",
                            "angle": 0,
                            "lines": [
                                {
                                    "bbox": [
                                        307,
                                        464,
                                        545,
                                        508
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                307,
                                                464,
                                                545,
                                                508
                                            ],
                                            "type": "text",
                                            "content": "[51] Lipu Zhou, Jiamin Ye, Montiel Abello, Shengze Wang, and Michael Kaess. Unsupervised learning of monocular depth estimation with bundle adjustment, super-resolution and clip loss. arXiv preprint arXiv:1812.03368, 2018. 2"
                                        }
                                    ]
                                }
                            ],
                            "index": 21
                        },
                        {
                            "bbox": [
                                307,
                                510,
                                545,
                                542
                            ],
                            "type": "ref_text",
                            "angle": 0,
                            "lines": [
                                {
                                    "bbox": [
                                        307,
                                        510,
                                        545,
                                        542
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                307,
                                                510,
                                                545,
                                                542
                                            ],
                                            "type": "text",
                                            "content": "[52] Tinghui Zhou, Matthew Brown, Noah Snavely, and David G Lowe. Unsupervised learning of depth and ego-motion from video. In CVPR, volume 2, page 7, 2017. 2, 3, 5, 6, 7"
                                        }
                                    ]
                                }
                            ],
                            "index": 22
                        },
                        {
                            "bbox": [
                                307,
                                544,
                                545,
                                576
                            ],
                            "type": "ref_text",
                            "angle": 0,
                            "lines": [
                                {
                                    "bbox": [
                                        307,
                                        544,
                                        545,
                                        576
                                    ],
                                    "spans": [
                                        {
                                            "bbox": [
                                                307,
                                                544,
                                                545,
                                                576
                                            ],
                                            "type": "text",
                                            "content": "[53] Yuliang Zou, Zelun Luo, and Jia-Bin Huang. Df-net: Unsupervised joint learning of depth and flow using cross-task consistency. In ECCV, 2018. 1, 2, 7"
                                        }
                                    ]
                                }
                            ],
                            "index": 23
                        }
                    ],
                    "sub_type": "ref_text"
                }
            ],
            "discarded_blocks": [
                {
                    "bbox": [
                        295,
                        754,
                        315,
                        763
                    ],
                    "type": "page_number",
                    "angle": 0,
                    "lines": [
                        {
                            "bbox": [
                                295,
                                754,
                                315,
                                763
                            ],
                            "spans": [
                                {
                                    "bbox": [
                                        295,
                                        754,
                                        315,
                                        763
                                    ],
                                    "type": "text",
                                    "content": "2494"
                                }
                            ]
                        }
                    ],
                    "index": 25
                }
            ],
            "page_size": [
                612,
                792
            ],
            "page_idx": 9
        }
    ],
    "_backend": "vlm",
    "_version_name": "2.6.4"
}