Research map / 3D vision / Face and shape
MDND: Unsupervised Learning Guided by Non-Differentiable Refinement for Shape Correspondence
Qinsong Li, Jing Meng, Haibo Wang, Shengjun Liu
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 |
|---|---|---|---|
| use non-differentiable iterative refinement as supervisory oracle instead of a differentiable solver (ablation) | MDND · TOPKIDS | mean geodesic error x100: 9.4 -> 3.5 | Improved |
| use non-differentiable iterative refinement instead of differentiable solver (ablation) | MDND · SMAL | mean geodesic error x100: 4.2 -> 3.1 | Improved |
| use non-differentiable iterative refinement instead of differentiable solver (ablation) | MDND · SCAPE | mean geodesic error x100: 2.0 -> 1.9 | Improved |
| use hybrid LBO+ELA basis in the refiner instead of LBO only (ablation) | MDND oracle (HWF) · TOPKIDS | mean geodesic error x100: 14.7 -> 3.5 | Improved |
| use hybrid LBO+ELA basis instead of ELA only (ablation) | MDND oracle (HWF) · TOPKIDS | mean geodesic error x100: 5.5 -> 3.5 | Improved |
| use hybrid basis instead of LBO only (ablation) | MDND oracle (HWF) · SMAL | mean geodesic error x100: 4.9 -> 3.1 | Improved |
| use hybrid basis instead of LBO only (ablation) | MDND oracle (HWF) · SCAPE | mean geodesic error x100: 2.2 -> 1.9 | Improved |
| MDND versus state of the art on near-isometric FAUST and SCAPE | MDND · FAUST and SCAPE | mean geodesic error x100: FAUST 1.6, SCAPE 1.9 vs HybridFMaps 1.5, ULRSSM 1.8 | No clear effect |
| MDND on cross-dataset generalization | MDND · FAUST to SCAPE | mean geodesic error x100: 6.4 -> 2.1 (vs ULRSSM) | Improved |
| MDND on TOPKIDS compared with next best method | MDND · TOPKIDS | mean geodesic error x100: 5.0 -> 3.5 | Improved |
Closest papers by meaning
- Deep Geometric Functional Maps: Robust Feature Learning for Shape Correspondence (CVPR 2020 · similarity 0.91)
- Memory-Scalable and Simplified Functional Map Learning (CVPR 2024 · similarity 0.89)
- Denoising Functional Maps: Diffusion Models for Shape Correspondence (CVPR 2025 · similarity 0.89)
- Hybrid Functional Maps for Crease-Aware Non-Isometric Shape Matching (CVPR 2024 · similarity 0.88)
- Integrating Efficient Optimal Transport and Functional Maps For Unsupervised Shape Correspondence Learning (CVPR 2024 · similarity 0.88)
- 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.
