Research map / Point clouds / Shape and surface completion
CorrNet3D: Unsupervised End-to-End Learning of Dense Correspondence for 3D Point Clouds
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 |
|---|---|---|---|
| replace the symmetric deformer with fully connected layers (ablation) | CorrNet3D · Surreal non-rigid, tolerance 20% | Corr (%): 95.61 -> 26.21 | Hurt |
| remove global features from the symmetric deformer (ablation) | CorrNet3D · Surreal non-rigid, tolerance 20% | Corr (%): 95.61 -> 25.24 | Hurt |
| use shared rather than unshared symmetric-deformer parameters (ablation) | CorrNet3D · Surreal non-rigid, tolerance 20% | Corr (%): 95.97 -> 95.61 | No clear effect |
| share the symmetric-deformer parameters (ablation) | CorrNet3D · Surreal non-rigid | number of parameters: reduces by half | Helped |
| add permutation and manifold regularization to the reconstruction loss (ablation) | CorrNet3D · Surreal non-rigid, tolerance 20% | Corr (%): 31.49 -> 95.61 | Helped |
| use Euclidean reconstruction loss instead of Chamfer Distance, with regularization (ablation) | CorrNet3D · Surreal non-rigid, tolerance 20% | Corr (%): 25.74 -> 95.61 | Helped |
| replace DeSmooth with a Sinkhorn layer (ablation) | CorrNet3D · Surreal non-rigid | efficiency: up to 8× | Hurt |
Closest papers by meaning
- Unsupervised Learning of Dense Shape Correspondence (CVPR 2019 · similarity 0.91)
- Correspondence learning via linearly-invariant embedding (NeurIPS 2020 · similarity 0.89)
- Topology-aware Feature Propagation for Unsupervised Non-rigid Point Cloud Correspondence (CVPR 2026 · similarity 0.89)
- Learning SO(3)-Invariant Semantic Correspondence via Local Shape Transform (CVPR 2024 · similarity 0.88)
- DV-Matcher: Deformation-based Non-rigid Point Cloud Matching Guided by Pre-trained Visual Features (CVPR 2025 · similarity 0.88)
- Occlusion-aware Non-Rigid Point Cloud Registration via Unsupervised Neural Deformation Correntropy (ICLR 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.