Research map / 3D vision / Depth and stereo
Augmenting Depth Estimation With Geospatial Context
Scott Workman, Hunter Blanton
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 geospatial context (overhead-derived synthetic depth) to ground-level depth estimation | depth network · HoliCity (cap 80m) | RMSE (m): 5.626 -> 4.895 | Improved |
| add geospatial context | depth network · HoliCity (cap 80m) | Abs Rel: 0.503 -> 0.501 | No clear effect |
| fuse synthetic depth in the decoder instead of concatenating it as an input channel (ablation) | ours · HoliCity (cap 80m) | RMSE (m): 5.057 -> 4.895 | Improved |
| fuse in decoder instead of concatenating as input channel (ablation) | ours · HoliCity (cap 80m) | Abs Rel: 0.515 -> 0.501 | Improved |
| train end-to-end with overhead-derived scale instead of using the overhead estimate for median scaling | ground only baseline · HoliCity (cap 80m) | RMSE (m): 8.635 -> 4.895 | Improved |
| add median scaling with ground truth to our method | ours · HoliCity (cap 80m) | RMSE (m): 4.895 -> 4.782 | Improved |
| learn height maps from overhead images instead of using known height maps (ablation) | ours · HoliCity (cap 80m) | RMSE (m): 4.895 -> 4.935 | No clear effect |
| learn height maps instead of using known height maps (ablation) | ours · HoliCity (cap 80m) | RMSE log: 0.279 -> 0.287 | No clear effect |
| use lower zoom level overhead data (greater spatial coverage) (ablation) | ours · HoliCity long range | RMSE at larger depth cap | Improved |
| add increasing heading noise to geo-orientation at inference (ablation) | ours · HoliCity | depth error | Hurt |
| disable median scaling of Monodepth2 on KITTI (ablation) | Monodepth2 · KITTI depth (cap 80m) | RMSE (m): 4.863 -> 19.176 | Hurt |
| use overhead-derived scale for median scaling of Monodepth2 | Monodepth2 · HoliCity | RMSE (m): 17.555 -> 15.743 | Improved |
Closest papers by meaning
- Depth2Pose: A Pose-Based Benchmark for Monocular Depth Estimation without Ground-Truth Depth (NeurIPS 2026 · similarity 0.88)
- GeoDepth: From Point-to-Depth to Plane-to-Depth Modeling for Self-Supervised Monocular Depth Estimation (CVPR 2025 · similarity 0.87)
- Depth Pro: Sharp Monocular Metric Depth in Less Than a Second (ICLR 2025 · similarity 0.85)
- Depth Prompting for Sensor-Agnostic Depth Estimation (CVPR 2024 · similarity 0.85)
- Mind The Edge: Refining Depth Edges in Sparsely-Supervised Monocular Depth Estimation (CVPR 2024 · similarity 0.85)
- RSA: Resolving Scale Ambiguities in Monocular Depth Estimators through Language Descriptions (NeurIPS 2024 · 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.
