Research map / Point clouds / Registration and matching
Robust Multiview Point Cloud Registration With Reliable Pose Graph Initialization and History Reweighting
Haiping Wang, Yuan Liu, Zhen Dong, Yulan Guo, Yu-Shen Liu, Wenping Wang, Bisheng Yang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 the proposed sparse graph and history-reweighted IRLS | Multiview point cloud registration · 3DMatch | registration recall (%): 83.1 -> 96.2 | Improved |
| Use the proposed sparse graph and history-reweighted IRLS | Multiview point cloud registration · 3DLoMatch | registration recall (%): 68.7 -> 81.6 | Improved |
| Use the sparse graph construction | Multiview registration on ScanNet · ScanNet | pairwise registrations: 13920 -> 6004 | Improved |
| Add overlap score s_ij to weight initialization (ablation) | History-reweighted IRLS · 3DMatch | registration recall (%): 95.5 -> 96.2 | Improved |
| Add overlap score s_ij to weight initialization (ablation) | History-reweighted IRLS · 3DLoMatch | registration recall (%): 79.9 -> 81.6 | Improved |
| Add inlier correspondence count r_ij to weight initialization (ablation) | History-reweighted IRLS · 3DMatch | registration recall (%): 76.9 -> 96.2 | Improved |
| Add inlier correspondence count r_ij to weight initialization (ablation) | History-reweighted IRLS · 3DLoMatch | registration recall (%): 63.4 -> 81.6 | Improved |
| Use history of residuals instead of current-iteration residuals only (ablation) | IRLS reweighting · 3DMatch | registration recall (%): 83.1 -> 96.2 | Improved |
| Use history of residuals instead of current-iteration residuals only (ablation) | IRLS reweighting · 3DLoMatch | registration recall (%): 68.9 -> 81.6 | Improved |
| Increase g(m) with iteration instead of giving iterations equal weight (ablation) | History-reweighted IRLS · 3DMatch | registration recall (%): 94.1 -> 96.2 | Improved |
| Increase g(m) with iteration instead of giving iterations equal weight (ablation) | History-reweighted IRLS · 3DLoMatch | registration recall (%): 79.8 -> 81.6 | Improved |
| Set the sparse-graph top-k to 12 instead of 10 (ablation) | Sparse pose graph construction · 3DMatch | registration recall (%): 96.2 -> 96.6 | Improved |
Closest papers by meaning
- Learning Multiview 3D Point Cloud Registration (CVPR 2020 · similarity 0.92)
- Partial Point Cloud Registration with Multi-view 2D Image Learning (AAAI 2025 · similarity 0.89)
- Register Any Point: Scaling 3D Point Cloud Registration by Flow Matching (ECCV 2026 · similarity 0.89)
- PointMC: Multi-instance Point Cloud Registration based on Maximal Cliques (ICML 2024 · similarity 0.88)
- 3D Focusing-and-Matching Network for Multi-Instance Point Cloud Registration (NeurIPS 2024 · similarity 0.88)
- PSReg: Prior-guided Sparse Mixture of Experts for Point Cloud Registration (AAAI 2025 · similarity 0.87)
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.
