Research map / 3D vision / Depth and stereo
Learning To Recover 3D Scene Shape 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 |
|---|---|---|---|
| Recover depth shift from the predicted point cloud | 3D scene reconstruction · ETH3D | AbsRel: 23.7 -> 15.9 | Helped |
| Recover depth shift from the predicted point cloud | 3D scene reconstruction · NYU, KITTI, Sintel, DIODE | AbsRel: NYU: 25.8 -> 15.1; KITTI: 23.3 -> 17.5; Sintel: 47.4 -> 40.3; DIODE: 46.8 -> 36.9 | Helped |
| Use the combined DPM and PCM with pinhole projection | 3D scene reconstruction · OASIS | LSIV: 0.63 -> 0.52 | Helped |
| Use the combined DPM and PCM with pinhole projection | 3D scene reconstruction · 2D-3D-S | LSIV: 2.65 -> 0.80 | Helped |
| Use the proposed PCM instead of Hold-Geoffroy focal length with MiDaS | MiDaS 3D reconstruction · OASIS | LSIV: 1.60 -> 1.32 | Helped |
| Add the PWN loss (ablation) | Ours depth predictor, small training subset · iBims-1 | εplanPE: 3.91 -> 2.95 | Helped |
| Add the PWN loss (ablation) | Ours depth predictor, small training subset · iBims-1 | AbsRel: 0.106 -> 0.101 | Helped |
| Use full training data rather than the small training subset (ablation) | Ours depth predictor · iBims-1 | AbsRel: 0.101 -> 0.079 | Helped |
| Use ILNR instead of SSMAE | Depth prediction · NYU | AbsRel: 14.4 -> 13.9 | Helped |
| Use ILNR instead of ILNR-MinMax (ablation) | Depth prediction · DIODE | AbsRel: 46.1 -> 34.2 | Helped |
| Use ILNR instead of ILNR-MAD (ablation) | Depth prediction · RedWeb | WHDR: 18.8 -> 18.7 | Helped |
| Use ResNeXt101 rather than ResNet50 backbone | Ours depth predictor · NYU | AbsRel: 9.1 -> 9.0 | Helped |
Closest papers by meaning
- Depth Pro: Sharp Monocular Metric Depth in Less Than a Second (ICLR 2025 · similarity 0.88)
- Know Your Neighbors: Improving Single-View Reconstruction via Spatial Vision-Language Reasoning (CVPR 2024 · similarity 0.87)
- A Simple yet Universal Framework for Depth Completion (NeurIPS 2024 · similarity 0.87)
- Depth2Pose: A Pose-Based Benchmark for Monocular Depth Estimation without Ground-Truth Depth (NeurIPS 2026 · similarity 0.87)
- Depth Anything: Unleashing the Power of Large-Scale Unlabeled Data (CVPR 2024 · similarity 0.87)
- Seurat: From Moving Points to Depth (CVPR 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.