Research map / Point clouds / Registration and matching
(Just) A Spoonful of Refinements Helps the Registration Error Go Down
Sérgio Agostinho, Aljoša Ošep, Alessio Del Bue, Laura Leal-Taixé
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 |
|---|---|---|---|
| add iterative linearized refinement layer during training | DCP-v2 · ModelNet40, DCP, noise on source only (unseen instances) | RMSE(R) deg: 12.974750 -> 8.938098 | Improved |
| add iterative linearized refinement layer during training | DCP-v2 · ModelNet40, DCP, noise on source only (unseen instances) | RMSE(t): 0.003941 -> 0.004221 | Hurt |
| add iterative linearized refinement layer during training | DCP-v2 · ModelNet40, DCP, unseen categories | RMSE(R) deg: 3.150191 -> 2.051713 | Improved |
| add iterative linearized refinement layer during training | DCP-v2 · ModelNet40, DCP, unseen categories | RMSE(t): 0.005039 -> 0.004543 | Improved |
| add iterative linearized refinement layer during training | DCP-v2 · 3DMatch, DCP, 30% min overlap | RMSE(R) deg: 11.211800 -> 10.232300 | Improved |
| add iterative linearized refinement layer during training | DCP-v2 · 3DMatch, DCP, 30% min overlap | RMSE(t): 0.137432 -> 0.128846 | Improved |
| add iterative linearized refinement layer during training | RPM-Net · ModelNet40, RPM-Net, partial visible data with noise | Anisotropic rot. err (deg): 0.893 -> 0.826 | Improved |
| add iterative linearized refinement layer during training | RPM-Net · ModelNet40, RPM-Net, partial visible data with noise | Anisotropic trans. err: 0.0087 -> 0.0081 | Improved |
| add inlier loss term to RPM-Net training (ablation) | RPM-Net · ModelNet40, RPM-Net, partial visible data with noise | Anisotropic rot. err (deg): 0.993 -> 0.893 | Improved |
| add inlier loss term to RPM-Net + Ours training (ablation) | RPM-Net + Ours · ModelNet40, RPM-Net, partial visible data with noise | Anisotropic rot. err (deg): 0.872 -> 0.826 | Mixed |
| add iterative refinement layer, 26 training runs (unseen categories) | DCP-v2 · ModelNet40, DCP, unseen categories, 26 runs | RMSE(R) deg: 4.264873 -> 3.311812 | Improved |
| add iterative refinement layer, 26 training runs (unseen categories) | DCP-v2 · ModelNet40, DCP, unseen categories, 26 runs | RMSE(t): 0.006250 -> 0.006532 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- RPM-Net — RPM-Net: Robust Point Matching Using Learned Features (CVPR 2020) · better on 4 of 4 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Learning-Based Point Cloud Registration for 6D Object Pose Estimation in the Real World (ECCV 2022 · similarity 0.90)
- Scalable and Differentiable Point-Cloud Registration Using Maximum Mean Discrepancy (ICML 2026 · similarity 0.89)
- Partial Point Cloud Registration with Multi-view 2D Image Learning (AAAI 2025 · similarity 0.87)
- A robust inlier identification algorithm for point cloud registration via $\mathbf{\ell_0}$-minimization (NeurIPS 2024 · similarity 0.87)
- Where Precision Meets Efficiency: Transformation Diffusion Model for Point Cloud Registration (AAAI 2025 · similarity 0.87)
- Implicit Correspondence Learning for Image-to-Point Cloud Registration (CVPR 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.
