Research map / 3D vision / Face and shape
Learning 3D Keypoint Descriptors for Non-Rigid Shape Matching
Hanyu Wang, Jianwei Guo, Dong-Ming Yan, Weize Quan, Xiaopeng Zhang
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 Min-CV triplet loss instead of classic triplet loss (ablation) | triplet CNN descriptor · FAUST | P1% (CMC): 25.93 -> 40.42 | Improved |
| use Min-CV triplet loss instead of classic triplet loss (ablation) | triplet CNN descriptor · FAUST | P20% (CMC): 57.82 -> 71.29 | Improved |
| use Min-CV triplet loss instead of Kumar's loss (ablation) | triplet CNN descriptor · FAUST | P1% (CMC): 33.39 -> 40.42 | Improved |
| use Min-CV triplet loss instead of Kumar's loss (ablation) | triplet CNN descriptor · FAUST | P20% (CMC): 66.25 -> 71.29 | Improved |
| use Min-CV triplet loss instead of Jan's loss (ablation) | triplet CNN descriptor · FAUST | P1% (CMC): 12.28 -> 40.42 | Improved |
| use learned descriptor instead of MoNet when transferring across datasets | descriptor · FAUST | P1% (CMC): 56.93 -> 49.14 | Hurt |
| use learned descriptor instead of FMNet | descriptor · SPRING (trained on FAUST) | P1% (CMC): 13.40 -> 63.30 | Improved |
| use learned descriptor instead of FMNet | descriptor · FACE (trained on FAUST) | P1% (CMC): 12.00 -> 35.22 | Improved |
| use learned descriptor instead of RoPS | descriptor · FACE (trained on FAUST) | AP: 0.629 -> 0.619 | Mixed |
| use learned descriptor instead of LSCNN | descriptor · FAUST | P20% (CMC): 58.61 -> 81.70 | Improved |
Closest papers by meaning
- Hierarchical Metric Learning and Matching for 2D and 3D Geometric Correspondences (ECCV 2018 · similarity 0.91)
- Deep Shape Matching (ECCV 2018 · similarity 0.89)
- PPFNet: Global Context Aware Local Features for Robust 3D Point Matching (CVPR 2018 · similarity 0.89)
- End-to-End Learning of Keypoint Detector and Descriptor for Pose Invariant 3D Matching (CVPR 2018 · similarity 0.89)
- Unsupervised Deep Shape Descriptor With Point Distribution Learning (CVPR 2020 · similarity 0.89)
- DGC-GNN: Leveraging Geometry and Color Cues for Visual Descriptor-Free 2D-3D Matching (CVPR 2024 · similarity 0.88)
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.
