Research map / 3D vision / Depth and stereo
PlaneMatch: Patch Coplanarity Prediction for Robust RGB-D Reconstruction
Yifei Shi, Kai Xu, Matthias Niessner, Szymon Rusinkiewicz, Thomas Funkhouser
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 learned patch descriptor in RANSAC voting (ablation) | PlaneMatch registration · TUM RGB-D fr1/desk | ATE RMSE (cm): 9.6 -> 1.4 | Improved |
| add learned patch descriptor in RANSAC voting (ablation) | PlaneMatch registration · TUM RGB-D fr3/office | ATE RMSE (cm): 12.6 -> 1.7 | Improved |
| add learned patch descriptor in RANSAC voting (ablation) | PlaneMatch registration · TUM RGB-D fr3/nst | ATE RMSE (cm): 2.3 -> 1.5 | Improved |
| weight coplanarity objective by predicted matching confidence (ablation) | PlaneMatch registration · TUM RGB-D fr1/desk | ATE RMSE (cm): 4.8 -> 1.4 | Improved |
| weight coplanarity objective by predicted matching confidence (ablation) | PlaneMatch registration · TUM RGB-D fr3/office | ATE RMSE (cm): 2.5 -> 1.7 | Improved |
| use learned descriptor in both RANSAC and optimization (ablation) | PlaneMatch registration · TUM RGB-D fr1/desk | ATE RMSE (cm): 18.9 -> 1.4 | Improved |
| add coplanarity constraints to key-point matching (ablation) | PlaneMatch registration · TUM RGB-D fr1/desk | ATE RMSE (cm): 5.6 -> 1.4 | Improved |
| add coplanarity constraints to key-point matching (ablation) | PlaneMatch registration · TUM RGB-D fr3/office | ATE RMSE (cm): 5.2 -> 1.7 | Improved |
| add key-point matching to coplanarity constraints (ablation) | PlaneMatch registration · TUM RGB-D fr1/desk | ATE RMSE (cm): 2.5 -> 1.4 | Improved |
| add key-point matching to coplanarity constraints (ablation) | PlaneMatch registration · TUM RGB-D fr3/office | ATE RMSE (cm): 3.7 -> 1.7 | Improved |
| pre-filter key-point matches with BundleFusion | PlaneMatch registration · TUM RGB-D fr3/nst | ATE RMSE (cm): 1.5 -> 0.9 | Improved |
| use coplanarity only instead of key-points only | PlaneMatch registration · TUM RGB-D fr1/desk | ATE RMSE (cm): 5.6 -> 2.5 | Improved |
Closest papers by meaning
- 3DMatch: Learning Local Geometric Descriptors From RGB-D Reconstructions (CVPR 2017 · similarity 0.87)
- GGPT: Geometry-Grounded Point Transformer (CVPR 2026 · similarity 0.85)
- Recovering 3D Planes from a Single Image via Convolutional Neural Networks (ECCV 2018 · similarity 0.85)
- Interacted Planes Reveal 3D Line Mapping (TPAMI 2026 · similarity 0.85)
- VGGT-SLAM: Dense RGB SLAM Optimized on the SL(4) Manifold (NeurIPS 2025 · similarity 0.85)
- SLAM3R: Real-Time Dense Scene Reconstruction from Monocular RGB Videos (CVPR 2025 · similarity 0.84)
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.
