Research map / 3D vision / Depth and stereo
PlaneRCNN: 3D Plane Detection and Reconstruction From a Single Image
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 PlaneRCNN instead of PlaneNet without fine-tuning | Plane reconstruction on NYUv2 · NYUv2, w/o fine-tuning | Rel: 0.220 -> 0.164 | Helped |
| Use PlaneRCNN instead of PlaneNet after fine-tuning | Plane reconstruction on NYUv2 · NYUv2, w/ fine-tuning | RMSE: 0.397 -> 0.395 | Helped |
| Add pixel-wise depth estimation (ablation) | PlaneRCNN · ScanNet | AP@0.6m: 0.329 -> 0.376 | Helped |
| Add anchor-based plane normal regression (ablation) | PlaneRCNN · ScanNet | AP@0.6m: 0.376 -> 0.378 | Helped |
| Add the warping loss module (ablation) | PlaneRCNN · ScanNet | AP@0.6m: 0.378 -> 0.384 | Helped |
| Add the segmentation refinement network (ablation) | PlaneRCNN · ScanNet | SC: 0.766 -> 0.810 | Helped |
| Add the segmentation refinement network (ablation) | PlaneRCNN · ScanNet | AP@0.6m: 0.384 -> 0.386 | Helped |
| Add plane offset supervision (ablation) | PlaneRCNN · Plane offset estimation | Performance | No clear effect |
| Use a detection-based PlaneRCNN rather than PlaneNet | Plane detection and reconstruction · ScanNet | Plane recall | Helped |
| Use MWS with ground-truth depth rather than PlaneRCNN | Plane detection · ScanNet, small depth-error thresholds | Plane recall | Mixed |
Closest papers by meaning
- PlaneNet: Piece-Wise Planar Reconstruction From a Single RGB Image (CVPR 2018 · similarity 0.89)
- Single-Image Piece-Wise Planar 3D Reconstruction via Associative Embedding (CVPR 2019 · similarity 0.86)
- PlaneMVS: 3D Plane Reconstruction From Multi-View Stereo (CVPR 2022 · similarity 0.86)
- DPSNet: End-to-end Deep Plane Sweep Stereo (ICLR 2019 · similarity 0.85)
- AirPlanes: Accurate Plane Estimation via 3D-Consistent Embeddings (CVPR 2024 · similarity 0.85)
- Towards In-the-wild 3D Plane Reconstruction from a Single Image (CVPR 2025 · similarity 0.85)
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.