Research map / 3D vision / Depth and stereo
Planar Surface Reconstruction From Sparse Views
Linyi Jin, Shengyi Qian, Andrew Owens, David F. Fouhey
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 geometric reasoning to merge planes (instead of appearance embedding only) (ablation) | Proposed · Relative camera pose (Matterport3D pairs) | AP (All): 33.04 -> 36.02 | Improved |
| add the proposed optimization instead of Associative3D heuristic (ablation) | Proposed · Relative camera pose (Matterport3D pairs) | AP (All): 33.01 -> 36.02 | Improved |
| add continuous optimization (ablation) | Proposed · Relative camera pose (Matterport3D pairs) | AP (All): 35.87 -> 36.02 | No clear effect |
| add continuous optimization (ablation) | Proposed · Relative camera pose (Matterport3D pairs) | translation median error (m): 0.88 -> 0.63 | Improved |
| add continuous optimization (ablation) | Proposed · Relative camera pose (Matterport3D pairs) | translation error <= 1m (%): 56.5 -> 66.6 | Improved |
| add continuous optimization (ablation) | Proposed · Relative camera pose (Matterport3D pairs) | rotation error <= 30 deg (%): 83.7 -> 83.4 | Mixed |
| add discrete optimization over cameras and planes (ablation) | Proposed · Relative camera pose (Matterport3D pairs) | translation median error (m): 0.90 -> 0.88 | Improved |
| add geometry to appearance for plane correspondence (ablation) | Proposed · Plane correspondence with GT boxes | IPAA-100 (%): 6.8 -> 16.2 | Improved |
| add geometry to appearance for plane correspondence (ablation) | Proposed · Plane correspondence with GT boxes | IPAA-80 (%): 55.7 -> 55.3 | No clear effect |
| use proposed optimization instead of Associative3D heuristic (ablation) | Proposed · Plane correspondence with GT boxes | IPAA-100 (%): 0.9 -> 16.2 | Improved |
| use attention features instead of concatenation in camera module (ablation) | ResNet50 camera branch · Camera pose module ablation (standalone backbone) | translation median error (m): 1.55 -> 0.98 | Improved |
| run attention module on the plane detection FPN backbone (ablation) | ResNet50-Attention · Camera pose module ablation (standalone backbone) | translation median error (m): 0.98 -> 0.63 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 1 on most shared results.
- SuperGlue — SuperGlue: Learning Feature Matching With Graph Neural Networks (CVPR 2020) · better on 0 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- PlanarSplatting: Accurate Planar Surface Reconstruction in 3 Minutes (CVPR 2025 · similarity 0.90)
- PLANA3R: Zero-shot Metric Planar 3D Reconstruction via Feed-forward Planar Splatting (NeurIPS 2025 · similarity 0.89)
- Sparis: Neural Implicit Surface Reconstruction of Indoor Scenes from Sparse Views (AAAI 2025 · similarity 0.88)
- Behind the Veil: Enhanced Indoor 3D Scene Reconstruction with Occluded Surfaces Completion (CVPR 2024 · similarity 0.86)
- SparseGNV: Generating Novel Views of Indoor Scenes with Sparse RGB-D Images (AAAI 2024 · similarity 0.86)
- 3D Gaussian Flats: Hybrid 2D/3D Photometric Scene Reconstruction (NeurIPS 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.
