Research map / Point clouds / Registration and matching
Multiway Non-Rigid Point Cloud Registration via Learned Functional Map Synchronization
Jiahui Huang, Tolga Birdal, Žan Gojčič, Leonidas J. Guibas, Shi‐Min Hu
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 functional map synchronization (ablation) | SyNoRiM · MPC-CAPE (self-supervised) | L2 error: 3.07 -> 2.95 | Improved |
| add functional map synchronization (ablation) | SyNoRiM · MPC-DT4D (full point cloud) | L2 error: 3.73 -> 3.53 | Improved |
| add functional map synchronization (ablation) | SyNoRiM · MPC-SAPIEN | L2 error: 3.48 -> 3.09 | Improved |
| add learned basis functions on top of descriptor network only (ablation) | SyNoRiM · MPC-DT4D (full point cloud) | L2 error: 5.40 -> 3.73 | Improved |
| add learned basis functions on top of descriptor network only (ablation) | SyNoRiM · MPC-SAPIEN | L2 error: 4.66 -> 3.48 | Improved |
| increase number of bases M from 24 to 32 (ablation) | SyNoRiM · MPC-CAPE | L2 error: 1.08 -> 1.05 | No clear effect |
| increase number of bases M from 24 to 80 (ablation) | SyNoRiM · MPC-CAPE | L2 error: 1.08 -> 1.55 | Hurt |
| use learned bases instead of hand-crafted bases (24 bases) (ablation) | SyNoRiM · MPC-CAPE | L2 error: 2.96 -> 1.08 | Improved |
| add robust term rho, consistency loss and mixed g for basis net training (ablation) | SyNoRiM basis net · MPC-DT4D | L2 error (non-occluded): 3.64 -> 3.14 | Improved |
| feed refinement net with F^n, F^f and X (ablation) | SyNoRiM refinement · MPC-DT4D | L2 error (full): 4.59 -> 3.73 | Improved |
| add robust term rho in synchronization (ablation) | SyNoRiM · MPC-SAPIEN | L2 error (fixed iter.): 3.31 -> 3.09 | Improved |
| use direct joint manifold optimizer instead of alternating optimization (ablation) | SyNoRiM · MPC-SAPIEN | L2 error (converged): 3.03 -> 3.00 | No clear effect |
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.
- PointPWC-Net — PointPWC-Net: Cost Volume on Point Clouds for (Self-)Supervised Scene Flow Estimation (ECCV 2020) · better on 1 of 1 shared result
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
- Non-Rigid Shape Registration via Deep Functional Maps Prior (NeurIPS 2023 · similarity 0.92)
- Occlusion-aware Non-Rigid Point Cloud Registration via Unsupervised Neural Deformation Correntropy (ICLR 2025 · similarity 0.90)
- Deterministic Point Cloud Registration via Novel Transformation Decomposition (CVPR 2022 · similarity 0.90)
- Scalable 3D Registration via Truncated Entry-wise Absolute Residuals (CVPR 2024 · similarity 0.87)
- Rectified Point Flow: Generic Point Cloud Pose Estimation (NeurIPS 2025 · similarity 0.87)
- Correspondence-Free Non-Rigid Point Set Registration Using Unsupervised Clustering Analysis (CVPR 2024 · 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.
