Research map / 3D vision / Depth and stereo
Relationship Spatialization for Depth Estimation
Xiaoyu Xu, Jiayan Qiu, Xinchao Wang, Zhou Wang
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 equally weighted relationship representations to image features (ablation) | Baseline · KITTI | abs rel: 0.0775 -> 0.0738 | Improved |
| add equally weighted relationship representations to image features (ablation) | Baseline · NYU Depth v2 | abs rel: 0.1390 -> 0.1295 | Improved |
| keep only identified effective relationships instead of all (ablation) | Baseline+R · KITTI | abs rel: 0.0738 -> 0.0725 | Improved |
| keep only identified effective relationships instead of all (ablation) | Baseline+R · NYU Depth v2 | abs rel: 0.1295 -> 0.1107 | Improved |
| weight relationships by quantified contribution via attention (ablation) | Baseline+ER · KITTI | abs rel: 0.0725 -> 0.0690 | Improved |
| weight relationships by quantified contribution via attention (ablation) | Baseline+ER · NYU Depth v2 | abs rel: 0.1107 -> 0.1071 | Improved |
| add redundancy elimination to relationship representations (ablation) | Baseline · KITTI | abs rel: 0.0775 -> 0.0640 | Improved |
| add redundancy elimination to relationship representations (ablation) | Baseline · NYU Depth v2 | abs rel: 0.1390 -> 0.1205 | Improved |
| add relationship spatialization to Bts | Bts · KITTI | silog: 11.8858 -> 13.5749 | Hurt |
| add relationship spatialization to Bts | Bts · NYU Depth v2 | silog: 18.5223 -> 21.9191 | Hurt |
| add relationship spatialization to Adabins | Adabins · KITTI | abs rel: 0.0775 -> 0.0758 | Improved |
| add relationship spatialization to Adabins | Adabins · NYU Depth v2 | abs rel: 0.1537 -> 0.1203 | Improved |
Closest papers by meaning
- RelationVGGT : Visual Geometry Transformers for 3D Spatial Relation Segmentation (NeurIPS 2026 · similarity 0.85)
- Mind The Edge: Refining Depth Edges in Sparsely-Supervised Monocular Depth Estimation (CVPR 2024 · similarity 0.84)
- DME: Unveiling the Bias for Better Generalized Monocular Depth Estimation (AAAI 2024 · similarity 0.84)
- Depth2Pose: A Pose-Based Benchmark for Monocular Depth Estimation without Ground-Truth Depth (NeurIPS 2026 · similarity 0.84)
- Can Transformers Capture Spatial Relations between Objects? (ICLR 2024 · similarity 0.84)
- SQLdepth: Generalizable Self-Supervised Fine-Structured Monocular Depth Estimation (AAAI 2024 · similarity 0.83)
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.
