Research map / 3D vision / Depth and stereo
PlaneMVS: 3D Plane Reconstruction From Multi-View Stereo
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 single-view plane geometry regression with multi-view plane reconstruction | PlaneMVS · ScanNet | AbsRel: 0.164 -> 0.088 | Helped |
| Add soft-pooling loss to couple plane detection and geometry (ablation) | PlaneMVS · ScanNet | AbsRel: 0.170 -> 0.119 | Helped |
| Add soft-pooling loss to couple plane detection and geometry (ablation) | PlaneMVS · ScanNet | AP: 0.288 -> 0.380 | Helped |
| Weight loss terms with learned uncertainty (ablation) | PlaneMVS · ScanNet | AbsRel: 0.119 -> 0.089 | Helped |
| Weight loss terms with learned uncertainty (ablation) | PlaneMVS · ScanNet | AP: 0.520 -> 0.535 | Helped |
| Use convex upsampling instead of bilinear upsampling (ablation) | PlaneMVS · ScanNet | AbsRel: 0.089 -> 0.088 | Helped |
| Use convex upsampling instead of bilinear upsampling (ablation) | PlaneMVS · ScanNet | AP: 0.562 -> 0.564 | Helped |
| Use slanted rather than fronto-parallel plane hypotheses (ablation) | PlaneMVS · ScanNet | AbsRel: 0.094 -> 0.088 | Helped |
| Use slanted rather than fronto-parallel plane hypotheses (ablation) | PlaneMVS · ScanNet | AP: 0.548 -> 0.564 | Helped |
| Fine-tune the ScanNet-pretrained model on 7-Scenes | PlaneMVS · 7-Scenes | AbsRel: 0.158 -> 0.113 | Helped |
Closest papers by meaning
- MVS2D: Efficient Multi-View Stereo via Attention-Driven 2D Convolutions (CVPR 2022 · similarity 0.90)
- Towards In-the-wild 3D Plane Reconstruction from a Single Image (CVPR 2025 · similarity 0.90)
- AirPlanes: Accurate Plane Estimation via 3D-Consistent Embeddings (CVPR 2024 · similarity 0.89)
- MUSt3R: Multi-view Network for Stereo 3D Reconstruction (CVPR 2025 · similarity 0.87)
- DUSt3R: Geometric 3D Vision Made Easy (CVPR 2024 · similarity 0.87)
- AlphaTablets: A Generic Plane Representation for 3D Planar Reconstruction from Monocular Videos (NeurIPS 2024 · 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.