Research map / 3D vision / Face and shape
Reduced Representation of Deformation Fields for Effective Non-rigid Shape Matching
Read the paperCheck 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 |
|---|---|---|---|
| Represent deformation with sparse nodal parameters and mesh-free approximation instead of a point-wise MLP | Non-rigid shape correspondence · SCAPE (PC+N), SHREC19 | correspondence error: 13.7 -> 6.6; 7.4 -> 4.8 | Helped |
| Replace mesh-free approximation with RBF interpolation (ablation) | Reduced deformation representation · SCAPE (PC+N), SHREC19 | correspondence error: 6.6 -> 9.3; 4.8 -> 8.1 | Hurt |
| Add training volume and ARAP regularization (ablation) | Ours · SHREC’19, SCAPE | correspondence error: 7.7 -> 5.2; 9.8 -> 7.0 | Helped |
| Use training ARAP regularization alone (ablation) | Ours · SHREC’19, SCAPE | correspondence error: 7.7 -> 6.9; 9.8 -> 10.2 | Mixed |
| Use training volume regularization alone (ablation) | Ours · SHREC’19, SCAPE | correspondence error: 7.7 -> 7.3; 9.8 -> 8.7 | Helped |
| Add test-time Chamfer-based regularization (ablation) | Ours with training regularization · SHREC’19, SCAPE | correspondence error: 7.9 -> 5.0; 12.5 -> 6.7 | Helped |
| Remove training regularization (ablation) | Ours · SHREC’19, SCAPE | correspondence error: 4.8 -> 7.7; 6.6 -> 9.8 | Hurt |
| Increase node count from 300 to 900 or 2700 at the listed radii (ablation) | Ours · SHREC’19, SCAPE-PC | correspondence error: 4.8 -> 5.4/6.4/6.5; 6.6 -> 7.0/8.7/9.8/9.6 | Hurt |
| Increase node radius from 0.2 to 0.5 with 300 nodes (ablation) | Ours · SHREC’19, SCAPE-PC | correspondence error: 4.8 -> 5.4; 6.6 -> 7.6 | Hurt |
| Use template segmentation or geodesic filtering instead of uniform node sampling (ablation) | Ours · SHREC’19, SCAPE | correspondence error: A-pose SHREC’19: 5.9 -> 5.1; A-pose SCAPE: 9.5 -> 7.3; T-pose changes are marginal | Mixed |
| Use learned latent-space constraints with the reduced representation (ablation) | Ours · SCAPE (PC+N), SHREC19 | correspondence error: 9.8 -> 6.6; 7.7 -> 4.8 | Helped |
| Add first-order and latent constraints to PW-MLP (ablation) | PW-MLP · SCAPE (PC+N), SHREC19 | training iteration time (ms): 32.0 -> 4098.3 | Hurt |
Closest papers by meaning
- Neural Shape Deformation Priors (NeurIPS 2022 · similarity 0.90)
- Deformed Implicit Field: Modeling 3D Shapes With Learned Dense Correspondence (CVPR 2021 · similarity 0.90)
- 4Deform: Neural Surface Deformation for Robust Shape Interpolation (CVPR 2025 · similarity 0.89)
- Implicit Neural Surface Deformation with Explicit Velocity Fields (ICLR 2025 · similarity 0.88)
- Implicit-ARAP: Efficient Handle-Guided Neural Field Deformation via Local Patch Meshing (NeurIPS 2025 · similarity 0.87)
- Neural Pose Representation Learning for Generating and Transferring Non-Rigid Object Poses (NeurIPS 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.