Research map / Point clouds / Shape and surface completion
Self-Supervised Learning for Multimodal Non-Rigid 3D Shape Matching
Dongliang Cao, Florian Bernard
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 |
|---|---|---|---|
| Add mesh-to-point-cloud alignment loss Ealign (ablation) | Multimodal shape matching · FAUST, SCAPE, and SHREC’19 point clouds | mean geodesic error (×100): 22.7 -> 2.4; 20.9 -> 4.1; 30.8 -> 4.5 | Improved |
| Add mesh-point contrastive loss Ence (ablation) | Multimodal shape matching · FAUST, SCAPE, and SHREC’19 meshes | mean geodesic error (×100): 2.3 -> 2.0; 4.8 -> 3.1; 4.3 -> 4.0 | Improved |
| Add mesh-point contrastive loss Ence (ablation) | Multimodal shape matching · FAUST, SCAPE, and SHREC’19 point clouds | mean geodesic error (×100): 5.2 -> 2.4; 9.3 -> 4.1; 11.2 -> 4.5 | Mixed |
| Remove Ence while retaining Ealign (ablation) | Multimodal shape matching · FAUST, SCAPE, and SHREC’19 point clouds | mean geodesic error (×100): 2.4 -> 2.7; 4.1 -> 5.6; 4.5 -> 5.1 | Hurt |
| Use multimodal training with mesh functional-map regularization and point-cloud feature matching | Ours · FAUST, SCAPE, and SHREC’19 point clouds | mean geodesic error (×100): 6.0 -> 2.4; 7.8 -> 4.1; 11.7 -> 4.5 | Improved |
| Train on SURREAL and evaluate on another dataset | Ours · FAUST, SCAPE, and SHREC’19 cross-dataset point clouds | mean geodesic error (×100): 12.2 -> 3.5; 14.1 -> 3.8; 15.9 -> 6.6 | Improved |
| Use the proposed self-supervised method rather than DPFM unsupervised | SHREC’16 partial shape matching · CUTS and HOLES point clouds | mean geodesic error (×100): 9.0 -> 7.6; 20.5 -> 15.9 | Improved |
| Use the proposed method rather than supervised DPFM | SURREAL-PV partial-view matching · SURREAL-PV test | mean geodesic error (×100): 7.8 -> 6.3 | Improved |
Closest papers by meaning
- Deep Geometric Functional Maps: Robust Feature Learning for Shape Correspondence (CVPR 2020 · similarity 0.90)
- Correspondence learning via linearly-invariant embedding (NeurIPS 2020 · similarity 0.89)
- Neural Intrinsic Embedding for Non-Rigid Point Cloud Matching (CVPR 2023 · similarity 0.88)
- Spectral Meets Spatial: Harmonising 3D Shape Matching and Interpolation (CVPR 2024 · similarity 0.88)
- Learning SO(3)-Invariant Semantic Correspondence via Local Shape Transform (CVPR 2024 · similarity 0.87)
- Denoising Functional Maps: Diffusion Models for Shape Correspondence (CVPR 2025 · 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.
