CVPR 2022 accepted papers on 3D vision
370 of the 2,077 papers accepted at CVPR 2022 (17.8%) are on 3D vision, compared with 17.9% at CVPR 2021. They fall into 5 topics.
See CVPR 2022 in the interactive map
Submitting to CVPR? Checklist →
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
Depth and stereo · 93 papers
- Neural Window Fully-Connected CRFs for Monocular Depth Estimation (358 citations)
- TransMVSNet: Global Context-Aware Multi-View Stereo Network With Transformers (264 citations)
- Rethinking Depth Estimation for Multi-View Stereo: A Unified Representation (173 citations)
- RayMVSNet: Learning Ray-Based 1D Implicit Fields for Accurate Multi-View Stereo (160 citations)
- P3Depth: Monocular Depth Estimation With a Piecewise Planarity Prior (159 citations)
- Neural 3D Scene Reconstruction With the Manhattan-World Assumption (158 citations)
- IterMVS: Iterative Probability Estimation for Efficient Multi-View Stereo (148 citations)
- Event-Aided Direct Sparse Odometry (126 citations)
- Raw High-Definition Radar for Multi-Task Learning (120 citations)
- Light Field Neural Rendering (112 citations)
- Occlusion-Aware Cost Constructor for Light Field Depth Estimation (107 citations)
- Beyond Cross-View Image Retrieval: Highly Accurate Vehicle Localization Using Satellite Image (94 citations)
and 81 more in the interactive map.
Face and shape · 89 papers
- Efficient Geometry-Aware 3D Generative Adversarial Networks (1,071 citations)
- Extracting Triangular 3D Models, Materials, and Lighting From Images (342 citations)
- ICON: Implicit Clothed Humans Obtained From Normals (313 citations)
- Text2Mesh: Text-Driven Neural Stylization for Meshes (289 citations)
- Gait Recognition in the Wild With Dense 3D Representations and a Benchmark (256 citations)
- StyleSDF: High-Resolution 3D-Consistent Image and Geometry Generation (240 citations)
- EMOCA: Emotion Driven Monocular Face Capture and Animation (221 citations)
- Neural Head Avatars From Monocular RGB Videos (197 citations)
- I M Avatar: Implicit Morphable Head Avatars From Videos (192 citations)
- AutoSDF: Shape Priors for 3D Completion, Reconstruction and Generation (187 citations)
- Depth-Aware Generative Adversarial Network for Talking Head Video Generation (180 citations)
- SelfRecon: Self Reconstruction Your Digital Avatar From Monocular Video (151 citations)
and 77 more in the interactive map.
Human pose estimation · 80 papers
- MHFormer: Multi-Hypothesis Transformer for 3D Human Pose Estimation (453 citations)
- HumanNeRF: Free-Viewpoint Rendering of Moving People From Monocular Video (450 citations)
- MixSTE: Seq2seq Mixed Spatio-Temporal Encoder for 3D Human Pose Estimation in Video (420 citations)
- Physical Inertial Poser (PIP): Physics-Aware Real-Time Human Motion Tracking From Sparse Inertial Sensors (214 citations)
- EPro-PnP: Generalized End-to-End Probabilistic Perspective-N-Points for Monocular Object Pose Estimation (189 citations)
- ZebraPose: Coarse To Fine Surface Encoding for 6DoF Object Pose Estimation (183 citations)
- GraFormer: Graph-Oriented Transformer for 3D Pose Estimation (176 citations)
- OnePose: One-Shot Object Pose Estimation Without CAD Models (169 citations)
- End-to-End Multi-Person Pose Estimation With Transformers (164 citations)
- Keypoint Transformer: Solving Joint Identification in Challenging Hands and Object Interactions for Accurate 3D Pose Estimation (162 citations)
- Putting People in Their Place: Monocular Regression of 3D People in Depth (162 citations)
- Progressively Generating Better Initial Guesses Towards Next Stages for High-Quality Human Motion Prediction (160 citations)
and 68 more in the interactive map.
3D generation · 54 papers
- NICE-SLAM: Neural Implicit Scalable Encoding for SLAM (812 citations)
- Generating Diverse and Natural 3D Human Motions From Text (553 citations)
- Zero-Shot Text-Guided Object Generation With Dream Fields (381 citations)
- MonoScene: Monocular 3D Semantic Scene Completion (288 citations)
- Panoptic Neural Fields: A Semantic Object-Aware Neural Scene Representation (219 citations)
- Bailando: 3D Dance Generation by Actor-Critic GPT With Choreographic Memory (198 citations)
- ScanQA: 3D Question Answering for Spatial Scene Understanding (186 citations)
- BEHAVE: Dataset and Method for Tracking Human Object Interactions (166 citations)
- ABO: Dataset and Benchmarks for Real-World 3D Object Understanding (154 citations)
- Rope3D: The Roadside Perception Dataset for Autonomous Driving and Monocular 3D Object Detection Task (154 citations)
- M3T: Three-Dimensional Medical Image Classifier Using Multi-Plane and Multi-Slice Transformer (131 citations)
- Multi-View Transformer for 3D Visual Grounding (111 citations)
and 42 more in the interactive map.
NeRF and radiance fields · 54 papers
- Mip-NeRF 360: Unbounded Anti-Aliased Neural Radiance Fields (1,901 citations)
- Plenoxels: Radiance Fields Without Neural Networks (1,375 citations)
- Direct Voxel Grid Optimization: Super-Fast Convergence for Radiance Fields Reconstruction (932 citations)
- Depth-Supervised NeRF: Fewer Views and Faster Training for Free (825 citations)
- Block-NeRF: Scalable Large Scene Neural View Synthesis (765 citations)
- RegNeRF: Regularizing Neural Radiance Fields for View Synthesis From Sparse Inputs (597 citations)
- Ref-NeRF: Structured View-Dependent Appearance for Neural Radiance Fields (517 citations)
- Mega-NERF: Scalable Construction of Large-Scale NeRFs for Virtual Fly-Throughs (410 citations)
- Neural 3D Video Synthesis From Multi-View Video (380 citations)
- Dense Depth Priors for Neural Radiance Fields From Sparse Input Views (352 citations)
- NeRF in the Dark: High Dynamic Range View Synthesis From Noisy Raw Images (342 citations)
- CLIP-NeRF: Text-and-Image Driven Manipulation of Neural Radiance Fields (337 citations)
and 42 more in the interactive map.
