Research map / 3D vision / Depth and stereo
GeoRefine: Self-Supervised Online Depth Refinement for Accurate Dense Mapping
Pan Ji, Qingan Yan, Yuxin Ma, Yi 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 basic self-supervised online refinement (ablation) | DPT · EuRoC V2_03 (monocular) | AbsRel: 9.9% -> 9.0% | Improved |
| add RAFT-flow front-end to SLAM (ablation) | BaseSystem · EuRoC V2_03 (monocular) | AbsRel: 0.090 -> 0.083 | Improved |
| add scale alignment between SLAM and depth model (ablation) | BaseSystem + RAFT-flow · EuRoC V2_03 (monocular) | AbsRel: 8.3% -> 6.4% | Improved |
| add scale alignment (ablation) | BaseSystem + RAFT-flow · EuRoC V2_03 (monocular) | delta1: 91.5% -> 95.2% | Improved |
| add occlusion-aware depth consistency loss (ablation) | previous system · EuRoC V2_03 (monocular) | AbsRel: 0.064 -> 0.053 | Improved |
| use refined depth in pRGBD SLAM (ablation) | BaseSystem (pRGBD) · EuRoC V2_03 (pRGBD) | AbsRel: 0.076 -> 0.065 | Improved |
| add RAFT-flow in pRGBD mode (ablation) | pRGBD system · EuRoC V2_03 (pRGBD) | RMSE ATE (m): 0.133 -> 0.069 | Improved |
| remove the extra reprojection term in bundle adjustment (ablation) | pRGBD system · EuRoC V2_03 (pRGBD) | RMSE ATE (m): 0.069 -> 0.034 | Improved |
| GeoRefine online refinement over pretrained DPT | DPT · EuRoC V1_01 (monocular) | AbsRel: 14.0% -> 5.0% | Improved |
| GeoRefine online refinement over pretrained DPT | DPT · TUM-RGBD | AbsRel: 2-4 times reduction | Improved |
| visual-inertial mode instead of monocular mode | GeoRefine · EuRoC | AbsRel | No clear effect |
| GeoRefine over Monodepth2 on KITTI | Monodepth2 · KITTI Eigen | AbsRel: 0.106 -> 0.096 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 1 on most shared results.
- DROID-SLAM — DROID-SLAM: Deep Visual SLAM for Monocular, Stereo, and RGB-D Cameras (NeurIPS 2021) · better on 0 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- GeoDepth: From Point-to-Depth to Plane-to-Depth Modeling for Self-Supervised Monocular Depth Estimation (CVPR 2025 · similarity 0.85)
- Self-Distilled Depth Refinement with Noisy Poisson Fusion (NeurIPS 2024 · similarity 0.85)
- Loopy-SLAM: Dense Neural SLAM with Loop Closures (CVPR 2024 · similarity 0.84)
- Depth2Pose: A Pose-Based Benchmark for Monocular Depth Estimation without Ground-Truth Depth (NeurIPS 2026 · similarity 0.84)
- Photo-SLAM: Real-time Simultaneous Localization and Photorealistic Mapping for Monocular Stereo and RGB-D Cameras (CVPR 2024 · similarity 0.84)
- A Simple yet Universal Framework for Depth Completion (NeurIPS 2024 · similarity 0.84)
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.
