Research map / Point clouds / Registration and matching
Accurate Point Cloud Registration with Robust Optimal Transport
Zhengyang Shen, Jean Feydy, Peirong Liu, Ariel Hernán Curiale, Ruben San Jose Estepar, Raul San Jose Estepar, Marc Niethammer
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 spline smoothing to D-RobOT matching (ablation) | D-RobOT · DirLab lung landmarks | average error (mm): 3.40 -> 2.95 | Improved |
| use LDDMM deformation model instead of spline (ablation) | D-RobOT · DirLab lung landmarks | average error (mm): 2.95 -> 2.86 | Improved |
| add deep registration module on top of S-RobOT (D-RobOT vs S-RobOT) (ablation) | D-RobOT (spline) · DirLab lung landmarks | average error (mm): 5.72 -> 2.95 | Improved |
| replace nearest-neighbor/affine ICP with raw RobOT matching | RobOT · DirLab lung landmarks | average error (mm): 15.05 -> 9.41 | Improved |
| RobOT raw vs non-rigid CPD | RobOT raw · DirLab lung landmarks | average error (mm): 9.30 -> 9.41 | No clear effect |
| LDDMM optimization instead of spline for S-RobOT (ablation) | S-RobOT · DirLab lung landmarks | runtime (s): 2.77 -> 42.30 | Hurt |
| rigid S-RobOT vs FGR/Go-ICP on partial registration (unseen objects) | S-RobOT rigid · ModelNet40 | rotation RMSE (degrees): 11.238 -> 2.989 | Improved |
| rigid S-RobOT on partial registration, unseen categories vs PRNet | S-RobOT rigid · ModelNet40 | rotation RMSE (degrees): 4.986 -> 7.142 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 1 on most shared results.
- PRNet — PRNet: Self-Supervised Learning for Partial-to-Partial Registration (NeurIPS 2019) · better on 1 of 2 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
- Efficient and Robust Point Cloud Registration via Heuristics-Guided Parameter Search (TPAMI 2024 · similarity 0.88)
- Register Any Point: Scaling 3D Point Cloud Registration by Flow Matching (ECCV 2026 · similarity 0.88)
- Integrating Efficient Optimal Transport and Functional Maps For Unsupervised Shape Correspondence Learning (CVPR 2024 · similarity 0.88)
- Rectified Point Flow: Generic Point Cloud Pose Estimation (NeurIPS 2025 · similarity 0.87)
- Multi-scale Consistency for Robust 3D Registration via Hierarchical Sinkhorn Tree (NeurIPS 2024 · similarity 0.87)
- Scalable 3D Registration via Truncated Entry-wise Absolute Residuals (CVPR 2024 · 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.
