CVPR 2021 accepted papers on 3D vision
297 of the 1,661 papers accepted at CVPR 2021 (17.9%) are on 3D vision, compared with 18.0% at CVPR 2020. They fall into 5 topics.
See CVPR 2021 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
Face and shape · 93 papers
- Learning Continuous Image Representation With Local Implicit Image Function (770 citations)
- Neural Body: Implicit Neural Representations With Structured Latent Codes for Novel View Synthesis of Dynamic Humans (633 citations)
- Neural Geometric Level of Detail: Real-Time Rendering With Implicit 3D Shapes (439 citations)
- Pose-Controllable Talking Face Generation by Implicitly Modularized Audio-Visual Representation (358 citations)
- Function4D: Real-Time Human Volumetric Capture From Very Sparse Consumer RGBD Sensors (351 citations)
- PhySG: Inverse Rendering With Spherical Gaussians for Physics-Based Material Editing and Relighting (302 citations)
- Learned Initializations for Optimizing Coordinate-Based Neural Representations (223 citations)
- SCANimate: Weakly Supervised Learning of Skinned Clothed Avatar Networks (223 citations)
- SMPLicit: Topology-Aware Generative Model for Clothed People (189 citations)
- Learning 3D Shape Feature for Texture-Insensitive Person Re-Identification (158 citations)
- Pulsar: Efficient Sphere-Based Neural Rendering (155 citations)
- Deformed Implicit Field: Modeling 3D Shapes With Learned Dense Correspondence (145 citations)
and 81 more in the interactive map.
Depth and stereo · 92 papers
- PatchmatchNet: Learned Multi-View Patchmatch Stereo (424 citations)
- NeuralRecon: Real-Time Coherent 3D Reconstruction From Monocular Video (313 citations)
- HITNet: Hierarchical Iterative Tile Refinement Network for Real-time Stereo Matching (310 citations)
- Learning To Recover 3D Scene Shape From a Single Image (212 citations)
- Stable View Synthesis (188 citations)
- Boosting Monocular Depth Estimation Models to High-Resolution via Content-Adaptive Multi-Resolution Merging (180 citations)
- Deep Gaussian Scale Mixture Prior for Spectral Compressive Imaging (170 citations)
- Robust Consistent Video Depth Estimation (165 citations)
- Back to Event Basics: Self-Supervised Learning of Image Reconstruction for Event Cameras via Photometric Constancy (147 citations)
- Learning Accurate Dense Correspondences and When To Trust Them (127 citations)
- RAFT-3D: Scene Flow Using Rigid-Motion Embeddings (122 citations)
- Depth Completion With Twin Surface Extrapolation at Occlusion Boundaries (111 citations)
and 80 more in the interactive map.
Human pose estimation · 64 papers
- End-to-End Human Pose and Mesh Reconstruction with Transformers (707 citations)
- GDR-Net: Geometry-Guided Direct Regression Network for Monocular 6D Object Pose Estimation (453 citations)
- HybrIK: A Hybrid Analytical-Neural Inverse Kinematics Solution for 3D Human Pose and Shape Estimation (399 citations)
- Bottom-Up Human Pose Estimation via Disentangled Keypoint Regression (393 citations)
- FFB6D: A Full Flow Bidirectional Fusion Network for 6D Pose Estimation (379 citations)
- Graph Stacked Hourglass Networks for 3D Human Pose Estimation (298 citations)
- Back to the Feature: Learning Robust Camera Localization From Pixels To Pose (295 citations)
- DexYCB: A Benchmark for Capturing Hand Grasping of Objects (281 citations)
- Beyond Static Features for Temporally Consistent 3D Human Pose and Shape From a Video (257 citations)
- Motion Representations for Articulated Animation (242 citations)
- AGORA: Avatars in Geography Optimized for Regression Analysis (197 citations)
- Rethinking the Heatmap Regression for Bottom-Up Human Pose Estimation (182 citations)
and 52 more in the interactive map.
3D generation · 31 papers
- 3D Human Action Representation Learning via Cross-View Consistency Pursuit (233 citations)
- Coming Down to Earth: Satellite-to-Street View Synthesis for Geo-Localization (171 citations)
- SceneGraphFusion: Incremental 3D Scene Graph Prediction From RGB-D Sequences (155 citations)
- Holistic 3D Scene Understanding From a Single Image With Implicit Representation (113 citations)
- SceneGen: Learning To Generate Realistic Traffic Scenes (96 citations)
- Dual Attention Guided Gaze Target Detection in the Wild (95 citations)
- Sketch2Model: View-Aware 3D Modeling From Single Free-Hand Sketches (93 citations)
- GeoSim: Realistic Video Simulation via Geometry-Aware Composition for Self-Driving (91 citations)
- Weakly Supervised Learning of Rigid 3D Scene Flow (91 citations)
- Projecting Your View Attentively: Monocular Road Scene Layout Estimation via Cross-View Transformation (88 citations)
- Introvert: Human Trajectory Prediction via Conditional 3D Attention (75 citations)
- OpenRooms: An Open Framework for Photorealistic Indoor Scene Datasets (74 citations)
and 19 more in the interactive map.
NeRF and radiance fields · 17 papers
- pixelNeRF: Neural Radiance Fields From One or Few Images (1,530 citations)
- NeRF in the Wild: Neural Radiance Fields for Unconstrained Photo Collections (1,399 citations)
- D-NeRF: Neural Radiance Fields for Dynamic Scenes (1,356 citations)
- GIRAFFE: Representing Scenes As Compositional Generative Neural Feature Fields (742 citations)
- IBRNet: Learning Multi-View Image-Based Rendering (722 citations)
- Pi-GAN: Periodic Implicit Generative Adversarial Networks for 3D-Aware Image Synthesis (698 citations)
- Neural Scene Flow Fields for Space-Time View Synthesis of Dynamic Scenes (682 citations)
- NeRV: Neural Reflectance and Visibility Fields for Relighting and View Synthesis (522 citations)
- Dynamic Neural Radiance Fields for Monocular 4D Facial Avatar Reconstruction (507 citations)
- Space-Time Neural Irradiance Fields for Free-Viewpoint Video (364 citations)
- Neural Scene Graphs for Dynamic Scenes (265 citations)
- NeX: Real-Time View Synthesis With Neural Basis Expansion (246 citations)
and 5 more in the interactive map.
