Research map / 3D vision / Human pose estimation
PlaneRecTR++: Unified Query Learning for Joint 3D Planar Reconstruction and Pose Estimation
Jingjia Shi, Shuaifeng Zhi, Kai Xu
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 cross-placed embeddings (CE) in plane-aware cross attention (ablation) | PlaneRecTR++ · ScanNetv2 pose | rotation error median (deg): 5.24 -> 4.30 | Improved |
| add cross-placed embeddings (CE) (ablation) | PlaneRecTR++ · MatterPort3D pose | rotation error median (deg): 4.39 -> 2.60 | Improved |
| keep query/key unsplit (1) instead of 4 heads (ablation) | PlaneRecTR++ · ScanNetv2 pose | rotation error median (deg): 4.45 -> 4.30 | Improved |
| keep query/key unsplit (1) instead of 4 heads (ablation) | PlaneRecTR++ · MatterPort3D pose | rotation error median (deg): 3.06 -> 2.60 | Improved |
| use 4 value segments instead of 1 (ablation) | PlaneRecTR++ · ScanNetv2 pose | rotation error median (deg): 5.13 -> 4.30 | Improved |
| add cross-placed embeddings (CE) (ablation) | PlaneRecTR++ · ScanNetv2 3D plane AP (0.2m, 5deg) | AP: 15.32 -> 18.19 | Improved |
| add cross-placed embeddings (CE) (ablation) | PlaneRecTR++ · ScanNetv2 plane correspondence | F-score: 0.529 -> 0.564 | Improved |
| add plane-level depth prediction task (ablation) | PlaneRecTR · ScanNetv1 | plane offset error (mm): 166.85 -> 165.13 | Improved |
| add plane-level depth prediction task (ablation) | PlaneRecTR · ScanNetv1 per-pixel depth recall @0.10m | recall: 50.28 -> 53.07 | Improved |
| add dense plane-level mask prediction (ablation) | PlaneRecTR · ScanNetv1 | plane offset error (mm): 177.69 -> 166.85 | Improved |
| add plane geometry tasks (parameters and depth) to segmentation (ablation) | PlaneRecTR · ScanNetv1 plane segmentation | VI/RI/SC | No clear effect |
| replace ResNet-50 backbone with Swin-B (ablation) | PlaneRecTR · ScanNetv1 plane segmentation | VI: 0.698 -> 0.651 | Improved |
Closest papers by meaning
- Towards In-the-wild 3D Plane Reconstruction from a Single Image (CVPR 2025 · similarity 0.89)
- AirPlanes: Accurate Plane Estimation via 3D-Consistent Embeddings (CVPR 2024 · similarity 0.88)
- PLANA3R: Zero-shot Metric Planar 3D Reconstruction via Feed-forward Planar Splatting (NeurIPS 2025 · similarity 0.87)
- PF3plat: Pose-Free Feed-Forward 3D Gaussian Splatting for Novel View Synthesis (ICML 2025 · similarity 0.87)
- Know Your Neighbors: Improving Single-View Reconstruction via Spatial Vision-Language Reasoning (CVPR 2024 · similarity 0.87)
- Test3R: Learning to Reconstruct 3D at Test Time (NeurIPS 2025 · 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.
