Research map / 3D vision / Face and shape
3DMatch: Learning Local Geometric Descriptors From RGB-D Reconstructions
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 |
|---|---|---|---|
| Use a learned metric network instead of the contrastive-loss descriptor (ablation) | 3DMatch · Keypoint matching benchmark | error at 95% recall: 35.3 -> 33.1 | Helped |
| Replace the 3D TDF volume with a 2D depth patch (ablation) | 3DMatch · Keypoint matching benchmark | error at 95% recall: 35.3 -> 38.5 | Hurt |
| Add pretraining on RGB-D reconstructions (ablation) | 3DMatch + RANSAC · Amazon Picking Challenge Shelf & Tote test split | correct rotation predictions: 53.8 -> 61.0 | Helped |
| Add pretraining on RGB-D reconstructions (ablation) | 3DMatch + RANSAC · Amazon Picking Challenge Shelf & Tote test split | correct translation predictions: 69.1 -> 71.7 | Helped |
| Use 3DMatch descriptors rather than the strongest listed baseline | 3DMatch + RANSAC · Synthetic scan fragment registration | recall: 59.2 -> 65.1 | Helped |
| Use 3DMatch descriptors rather than the strongest listed baseline | 3DMatch + RANSAC · Synthetic scan fragment registration | precision: 23.2 -> 25.2 | Helped |
| Use 3DMatch descriptors rather than Johnson et al. descriptors | 3DMatch + RANSAC · Real-world scan fragment registration | recall: 51.8 -> 66.8 | Helped |
| Use 3DMatch descriptors rather than Johnson et al. descriptors | 3DMatch + RANSAC · Real-world scan fragment registration | precision: 31.6 -> 40.1 | Helped |
Closest papers by meaning
- Test3R: Learning to Reconstruct 3D at Test Time (NeurIPS 2025 · similarity 0.87)
- SemAlign3D: Semantic Correspondence between RGB-Images through Aligning 3D Object-Class Representations (CVPR 2025 · similarity 0.87)
- Partial-to-Partial Shape Matching with Geometric Consistency (CVPR 2024 · similarity 0.87)
- DGC-GNN: Leveraging Geometry and Color Cues for Visual Descriptor-Free 2D-3D Matching (CVPR 2024 · similarity 0.87)
- Common3D: Self-Supervised Learning of 3D Morphable Models for Common Objects in Neural Feature Space (CVPR 2025 · similarity 0.86)
- Sensing Surface Patches in Volume Rendering for Inferring Signed Distance Functions (AAAI 2025 · similarity 0.86)
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.