Research map / 3D vision / Face and shape
Learning Multi-resolution Functional Maps with Spectral Attention for Robust 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 |
|---|---|---|---|
| add intermediate multi-resolution map loss L_inter (ablation) | Ours · SMAL unsup | geodesic error x100: 13.4 -> 5.4 | Helped |
| add learned spectral attention network instead of uniform weights (ablation) | Ours-Fast · SMAL unsup | geodesic error x100: 56.2 -> 5.8 | Helped |
| add learned spectral attention network instead of uniform weights (ablation) | Ours · SMAL unsup | geodesic error x100: 5.9 -> 5.4 | Helped |
| add learned attention instead of mean residual softmax (ablation) | Ours-Fast · SMAL unsup | geodesic error x100: 11.1 -> 5.8 | Helped |
| add learned attention instead of mean residual softmax (ablation) | Ours · SMAL unsup | geodesic error x100: 5.6 -> 5.4 | Helped |
| spectral step size tau 5 or 20 instead of 10 (ablation) | Ours / Ours-Fast · SMAL unsup | geodesic error x100: Fast: 5.1 (tau=5), 6.4 (tau=20) vs 5.8; Ours: 7.2, 6.9 vs 5.4 | Mixed |
| replace DiffusionNet with KPConv backbone (ablation) | Ours-Fast vs GeomFmaps · SMAL sup | geodesic error x100: GeomFmaps 9.4 vs Ours-Fast 6.8 | Helped |
| multi-resolution maps with spectral attention vs single-resolution GeomFmaps | GeomFmaps · SMAL sup | geodesic error x100: 8.4 -> 5.3 | Helped |
| multi-resolution maps with spectral attention vs GeomFmaps | GeomFmaps · DT4D-H inter-class sup | geodesic error x100: 4.1 -> 4.6 | No clear effect |
| multi-resolution maps with spectral attention vs GeomFmaps | GeomFmaps · DT4D-H inter-class unsup | geodesic error x100: 22.6 -> 11.6 | Helped |
| acceleration via principal submatrices (Ours-Fast) vs full | Ours · FAUST/SCAPE/SHREC'19 | geodesic error x100 | No clear effect |
Closest papers by meaning
- Memory-Scalable and Simplified Functional Map Learning (CVPR 2024 · similarity 0.90)
- Latent Functional Maps: a spectral framework for representation alignment (NeurIPS 2024 · 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 0.88)
- Denoising Functional Maps: Diffusion Models for Shape Correspondence (CVPR 2025 · similarity 0.87)
- Locality Preserving Refinement for Shape Matching with Functional Maps (AAAI 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.