Research map / 3D vision / Human pose estimation
PVN3D: A Deep Point-Wise 3D Keypoints Voting Network for 6DoF Pose Estimation
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use 3D keypoint voting rather than directly regressing pose parameters (ablation) | PVN3D on YCB-Video · YCB-Video | ADD(S): 87.3 -> 91.8 | Helped |
| Use 3D keypoint voting rather than projected 2D keypoints and PnP (ablation) | PVN3D on YCB-Video · YCB-Video | ADD-S: 81.8 -> 95.5 | Helped |
| Use 3D keypoint voting rather than dense object-coordinate regression (ablation) | PVN3D on YCB-Video · YCB-Video | ADD(S): 88.1 -> 91.8 | Helped |
| Select mesh-surface points with FPS instead of bounding-box corners (ablation) | PVN3D on YCB-Video · YCB-Video | ADD(S): 90.2 -> 91.8 | Helped |
| Use 8 FPS keypoints rather than 4 FPS keypoints (ablation) | PVN3D on YCB-Video · YCB-Video | ADD(S): 90.5 -> 91.8 | Helped |
| Use 8 FPS keypoints rather than 12 FPS keypoints (ablation) | PVN3D on YCB-Video · YCB-Video | ADD(S): 90.7 -> 91.8 | Helped |
| Jointly train semantic segmentation with keypoint voting (ablation) | PVN3D on YCB-Video · YCB-Video | ADD(S): 90.6 -> 91.3 | Helped |
| Add center voting to jointly trained keypoint and semantic modules (ablation) | PVN3D instance segmentation on YCB-Video · YCB-Video, extra-large clamp | mIoU: 50.7 -> 69.0 | Helped |
| Add iterative ICP refinement (ablation) | PVN3D on YCB-Video · YCB-Video | ADD(S) AUC: 91.8 -> 92.3 | Helped |
| Use PVN3D instead of the strongest listed baseline | 6DoF pose estimation · LineMOD | ADD(S): 94.3 -> 99.4 | Helped |
| Use PVN3D instead of the strongest listed baseline | 6DoF pose estimation · Occlusion LineMOD | ADD(S): 40.8 -> 63.2 | Helped |
Closest papers by meaning
- Pos3R: 6D Pose Estimation for Unseen Objects Made Easy (CVPR 2025 · similarity 0.88)
- SingRef6D: Monocular Novel Object Pose Estimation with a Single RGB Reference (NeurIPS 2025 · similarity 0.88)
- One2Any: One-Reference 6D Pose Estimation for Any Object (CVPR 2025 · similarity 0.87)
- iG-6DoF: Model-free 6DoF Pose Estimation for Unseen Object via Iterative 3D Gaussian Splatting (CVPR 2025 · similarity 0.86)
- UA-Pose: Uncertainty-Aware 6D Object Pose Estimation and Online Object Completion with Partial References (CVPR 2025 · similarity 0.86)
- Any6D: Model-free 6D Pose Estimation of Novel Objects (CVPR 2025 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.