Research map / 3D vision / Face and shape
Shape Non-rigid Kinematics (SNK): A Zero-Shot Method for Non-Rigid Shape Matching via Unsupervised Functional Map Regularized Reconstruction
Souhaib Attaiki, Maks Ovsjanikov
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 |
|---|---|---|---|
| replace proposed decoder with MLP decoder (ablation) | SNK · FAUST | geodesic error (x100): 1.9 -> 35.9 | Hurt |
| replace proposed decoder with DiffusionNet decoder (ablation) | SNK · FAUST | geodesic error (x100): 1.9 -> 13.4 | Hurt |
| add ARAP loss to DiffusionNet decoder (ablation) | DiffusionNet decoder · FAUST | geodesic error (x100): 13.4 -> 14.8 | Hurt |
| replace shape encoder with directly learned latent code (ablation) | SNK · FAUST | geodesic error (x100): 1.9 -> 4.0 | Hurt |
| replace MSE loss with Chamfer loss (ablation) | SNK · FAUST | geodesic error (x100): 1.9 -> 80.7 | Hurt |
| use only spectral loss (ablation) | SNK · FAUST | geodesic error (x100): 1.9 -> 8.7 | Hurt |
| remove PriMo energy loss (ablation) | SNK · FAUST | geodesic error (x100): 1.9 -> 6.5 | Hurt |
| remove cycle loss (ablation) | SNK · FAUST | geodesic error (x100): 1.9 -> 2.7 | Hurt |
| replace neural feature extractor with HKS features (ablation) | SNK · FAUST | geodesic error (x100): 1.9 -> 20.3 | Hurt |
| replace network with free variables for features (ablation) | SNK · FAUST | geodesic error (x100): 1.9 -> 64.0 | Hurt |
| use SNK instead of Smooth Shells on SHREC'19 | SNK · SHREC'19 | geodesic error (x100): 12.2 -> 5.8 | Improved |
| use SNK instead of runtime-heavy BCICP | SNK · FAUST | runtime (s): 780 -> 70 | Improved |
Closest papers by meaning
- Non-Rigid Shape Registration via Deep Functional Maps Prior (NeurIPS 2023 · similarity 0.88)
- Probabilistic Deformation Consistency for Unsupervised Shape Matching (AAAI 2026 · similarity 0.88)
- KillingFusion: Non-Rigid 3D Reconstruction Without Correspondences (CVPR 2017 · similarity 0.88)
- Hybrid Functional Maps for Crease-Aware Non-Isometric Shape Matching (CVPR 2024 · similarity 0.88)
- Spectral Meets Spatial: Harmonising 3D Shape Matching and Interpolation (CVPR 2024 · similarity 0.87)
- Integrating Efficient Optimal Transport and Functional Maps For Unsupervised Shape Correspondence Learning (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.
