Research map / 3D vision / Face and shape
Unsupervised Learning of Dense Shape Correspondence
Check 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 |
|---|---|---|---|
| Train FMNet with geodesic-distance distortion instead of ground-truth correspondence supervision | FMNet on synthetic FAUST · FAUST synthetic intra-subject pairs | correspondence accuracy | No clear effect |
| Train with the unsupervised geodesic-distance distortion loss | FMNet · FAUST synthetic | supervised loss | Helped |
| Train on real scans rather than synthetic FAUST shapes | Unsupervised network · SCAPE | generalization performance | Helped |
| Apply PMF refinement to the network prediction | Unsupervised network · TOSCA | matching accuracy | Helped |
| Train on human scans and transfer to animal shapes without animal training examples | Unsupervised network · TOSCA | matching performance | Helped |
| Optimize on the input pair without correspondence labels rather than use the axiomatic matching pipeline | Unsupervised network · Single artistic shape pair | runtime: about 15 minutes vs. over one hour | Helped |
| Use the unsupervised network trained on the single pair instead of a supervised network trained on FAUST | Shape correspondence · Deadpool pose pairs | matching performance | Helped |
| Use the unsupervised network without adaptation for partial shapes | Unsupervised network · SHREC'16 dog with holes | matching performance | Mixed |
Closest papers by meaning
- Continuous Surface Embeddings (NeurIPS 2020 · similarity 0.91)
- CorrNet3D: Unsupervised End-to-End Learning of Dense Correspondence for 3D Point Clouds (CVPR 2021 · similarity 0.91)
- Probabilistic Deformation Consistency for Unsupervised Shape Matching (AAAI 2026 · similarity 0.90)
- Deep Shells: Unsupervised Shape Correspondence with Optimal Transport (NeurIPS 2020 · similarity 0.90)
- Reduced Representation of Deformation Fields for Effective Non-rigid Shape Matching (NeurIPS 2022 · similarity 0.90)
- SGSoft: Learning Fused Semantic-Geometric Features for 3D Shape Correspondence via Template-Guided Soft Signals (CVPR 2026 · similarity 0.90)
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.