ECCV 2020 accepted papers on 3D vision
245 of the 1,358 papers accepted at ECCV 2020 (18.0%) are on 3D vision, compared with 21.9% at ECCV 2018. They fall into 6 topics.
See ECCV 2020 in the interactive map
Working on something in this area? Describe your idea in scime atlas to see the closest papers by meaning and how crowded the spot has become.
Topics and papers
Human pose estimation · 80 papers
- CosyPose: Consistent multi-view multi-object 6D pose estimation (466 citations)
- Pose2Mesh: Graph Convolutional Network for 3D Human Pose and Mesh Recovery from a 2D Human Pose (411 citations)
- I2L-MeshNet: Image-to-Lixel Prediction Network for Accurate 3D Human Pose and Mesh Estimation from a Single RGB Image (398 citations)
- ETH-XGaze: A Large Scale Dataset for Gaze Estimation under Extreme Head Pose and Gaze Variation (369 citations)
- History Repeats Itself: Human Motion Prediction via Motion Attention (340 citations)
- InterHand2.6M: A Dataset and Baseline for 3D Interacting Hand Pose Estimation from a Single RGB Image (322 citations)
- Whole-Body Human Pose Estimation in the Wild (304 citations)
- STAR: Sparse Trained Articulated Human Body Regressor (268 citations)
- Learning Delicate Local Representations for Multi-Person Pose Estimation (253 citations)
- Monocular Expressive Body Regression through Body-Driven Attention (249 citations)
- Shape Prior Deformation for Categorical 6D Object Pose and Size Estimation (240 citations)
- Motion Guided 3D Pose Estimation from Videos (230 citations)
and 68 more in the interactive map.
Depth and stereo · 72 papers
- Self-Supervised Monocular Depth Estimation: Solving the Dynamic Object Problem by Semantic Guidance (399 citations)
- Non-Local Spatial Propagation Network for Depth Completion (396 citations)
- Dense Hybrid Recurrent Multi-view Stereo Net with Dynamic Consistency Checking (222 citations)
- Spike-FlowNet: Event-based Optical Flow Estimation with Energy-Efficient Hybrid Neural Networks (214 citations)
- Guiding Monocular Depth Estimation Using Depth-Attention Volume (163 citations)
- Pyramid Multi-view Stereo Net with Self-adaptive View Aggregation (144 citations)
- Deep Shape from Polarization (135 citations)
- RANSAC-Flow: Generic Two-stage Image Alignment (105 citations)
- Learning Monocular Visual Odometry via Self-Supervised Long-Term Modeling (84 citations)
- Pseudo RGB-D for Self-Improving Monocular SLAM and Depth Prediction (71 citations)
- CLIFFNet for Monocular Depth Estimation with Hierarchical Embedding Loss (67 citations)
- Occlusion-Aware Depth Estimation with Adaptive Normal Constraints (65 citations)
and 60 more in the interactive map.
Face and shape · 61 papers
- Convolutional Occupancy Networks (902 citations)
- Towards Fast, Accurate and Stable 3D Dense Face Alignment (482 citations)
- Deep Local Shapes: Learning Local SDF Priors for Detailed 3D Reconstruction (412 citations)
- NASA Neural Articulated Shape Approximation (208 citations)
- BCNet: Learning Body and Cloth Shape from A Single Image (181 citations)
- CLOTH3D: Clothed 3D Humans (167 citations)
- On Disentangling Spoof Trace for Generic Face Anti-Spoofing (159 citations)
- Self-supervised Single-view 3D Reconstruction via Semantic Consistency (151 citations)
- Fast Bi-layer Neural Synthesis of One-Shot Realistic Head Avatars (143 citations)
- Self-Supervised Monocular 3D Face Reconstruction by Occlusion-Aware Multi-view Geometry Consistency (138 citations)
- Who Left the Dogs Out? 3D Animal Reconstruction with Expectation Maximization in the Loop (111 citations)
- Shape and Viewpoint without Keypoints (102 citations)
and 49 more in the interactive map.
3D generation · 23 papers
- Lift, Splat, Shoot: Encoding Images from Arbitrary Camera Rigs by Implicitly Unprojecting to 3D (1,385 citations)
- GRAB: A Dataset of Whole-Body Human Grasping of Objects (372 citations)
- Structured3D: A Large Photo-realistic Dataset for Structured 3D Modeling (309 citations)
- Atlas: End-to-End 3D Scene Reconstruction from Posed Images (294 citations)
- Virtual Multi-view Fusion for 3D Semantic Segmentation (180 citations)
- Gen-LaneNet: A Generalized and Scalable Approach for 3D Lane Detection (130 citations)
- Perceiving 3D Human-Object Spatial Arrangements from a Single Image in the Wild (126 citations)
- SimAug: Learning Robust Representations from Simulation for Trajectory Prediction (107 citations)
- Meta-Sim2: Unsupervised Learning of Scene Structure for Synthetic Data Generation (79 citations)
- SceneCAD: Predicting Object Alignments and Layouts in RGB-D Scans (67 citations)
- Malleable 2.5D Convolution: Learning Receptive Fields along the Depth-axis for RGB-D Scene Parsing (47 citations)
- GeoLayout: Geometry Driven Room Layout Estimation Based on Depth Maps of Planes (37 citations)
and 11 more in the interactive map.
NeRF and radiance fields · 8 papers
- Free View Synthesis (281 citations)
- Deep Reflectance Volumes: Relightable Reconstructions from Multi-View Photometric Images (114 citations)
- Learning to Factorize and Relight a City (39 citations)
- Deep Multi Depth Panoramas for View Synthesis (35 citations)
- Semantic View Synthesis (30 citations)
- Acquiring Dynamic Light Fields through Coded Aperture Camera (18 citations)
- Crowdsampling the Plenoptic Function
- NeRF: Representing Scenes as Neural Radiance Fields for View Synthesis
