CVPR 2022 accepted papers on Point clouds
98 of the 2,077 papers accepted at CVPR 2022 (4.7%) are on Point clouds, compared with 5.1% at CVPR 2021. They fall into 4 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
Point cloud classification · 42 papers
- Point-BERT: Pre-Training 3D Point Cloud Transformers With Masked Point Modeling (780 citations)
- Point-NeRF: Point-Based Neural Radiance Fields (561 citations)
- Stratified Transformer for 3D Point Cloud Segmentation (557 citations)
- PointCLIP: Point Cloud Understanding by CLIP (435 citations)
- CrossPoint: Self-Supervised Cross-Modal Contrastive Learning for 3D Point Cloud Understanding (300 citations)
- A Keypoint-Based Global Association Network for Lane Detection (172 citations)
- PTTR: Relational 3D Point Cloud Object Tracking With Transformer (123 citations)
- No-Reference Point Cloud Quality Assessment via Domain Adaptation (119 citations)
- Beyond 3D Siamese Tracking: A Motion-Centric Paradigm for 3D Single Object Tracking in Point Clouds (108 citations)
- Shape-Invariant 3D Adversarial Point Clouds (95 citations)
- 3DJCG: A Unified Framework for Joint Dense Captioning and Visual Grounding on 3D Point Clouds (91 citations)
- PatchFormer: An Efficient Point Transformer With Patch Attention (88 citations)
and 30 more in the interactive map.
Segmentation and detection · 26 papers
- Not All Points Are Equal: Learning Highly Efficient Point-Based Detectors for 3D LiDAR Point Clouds (413 citations)
- Embracing Single Stride 3D Object Detector With Sparse Transformer (289 citations)
- SoftGroup for 3D Instance Segmentation on Point Clouds (278 citations)
- Sparse Fuse Dense: Towards High Quality 3D Detection With Depth Completion (275 citations)
- Voxel Set Transformer: A Set-to-Set Approach to 3D Object Detection From Point Clouds (234 citations)
- Fast Point Transformer (217 citations)
- Contrastive Boundary Learning for Point Cloud Segmentation (202 citations)
- Point-to-Voxel Knowledge Distillation for LiDAR Semantic Segmentation (198 citations)
- Point Density-Aware Voxels for LiDAR 3D Object Detection (176 citations)
- LiDAR Snowfall Simulation for Robust 3D Object Detection (165 citations)
- Image-to-Lidar Self-Supervised Distillation for Autonomous Driving Data (112 citations)
- HybridCR: Weakly-Supervised 3D Point Cloud Semantic Segmentation via Hybrid Contrastive Regularization (104 citations)
and 14 more in the interactive map.
Shape and surface completion · 20 papers
- Surface Representation for Point Clouds (247 citations)
- POCO: Point Convolution for Surface Reconstruction (146 citations)
- ShapeFormer: Transformer-Based Shape Completion via Sparse Representation (134 citations)
- Neural Points: Point Cloud Representation With Neural Fields for Arbitrary Upsampling (91 citations)
- DiGS: Divergence Guided Shape Implicit Neural Representation for Unoriented Point Clouds (81 citations)
- LAKe-Net: Topology-Aware Point Cloud Completion by Localizing Aligned Keypoints (80 citations)
- Surface Reconstruction From Point Clouds by Learning Predictive Context Priors (76 citations)
- Reconstructing Surfaces for Sparse Point Clouds With On-Surface Priors (74 citations)
- Self-Supervised Arbitrary-Scale Point Clouds Upsampling via Implicit Neural Representation (68 citations)
- Point2Cyl: Reverse Engineering 3D Objects From Point Clouds to Extrusion Cylinders (59 citations)
- Robust and Accurate Superquadric Recovery: A Probabilistic Approach (48 citations)
- Learning a Structured Latent Space for Unsupervised Point Cloud Completion (45 citations)
and 8 more in the interactive map.
Registration and matching · 10 papers
- Geometric Transformer for Fast and Robust Point Cloud Registration (501 citations)
- REGTR: End-to-End Point Cloud Correspondences With Transformers (272 citations)
- Lepard: Learning Partial Point Cloud Matching in Rigid and Deformable Scenes (183 citations)
- SC2-PCR: A Second Order Spatial Compatibility for Efficient and Robust Point Cloud Registration (181 citations)
- RCP: Recurrent Closest Point for Point Cloud (32 citations)
- Coherent Point Drift Revisited for Non-Rigid Shape Matching and Registration (30 citations)
- Deterministic Point Cloud Registration via Novel Transformation Decomposition (25 citations)
- Multi-Instance Point Cloud Registration by Efficient Correspondence Clustering (16 citations)
- An Iterative Quantum Approach for Transformation Estimation From Point Sets (15 citations)
- ARCS: Accurate Rotation and Correspondence Search (15 citations)
