Research map / 3D vision / Depth and stereo
SC-DepthV3: Robust Self-Supervised Monocular Depth Estimation for Dynamic Scenes
Libo Sun, Jia-Wang Bian, Huangying Zhan, Wei Yin, Ian D. Reid, Chunhua Shen
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 dynamic region refinement (DRR) (ablation) | SC-DepthV3 on DDAD · DDAD full image | AbsRel: 0.153 -> 0.142 | Improved |
| add local structure refinement (LSR) (ablation) | SC-DepthV3 on DDAD · DDAD full image | AbsRel: 0.149 -> 0.142 | Improved |
| add dynamic region refinement (DRR) (ablation) | SC-DepthV3 on BONN · BONN full image | AbsRel: 0.138 -> 0.126 | Improved |
| add local structure refinement (LSR) (ablation) | SC-DepthV3 on BONN · BONN full image | AbsRel: 0.130 -> 0.126 | Improved |
| add dynamic region refinement (DRR) (ablation) | SC-DepthV3 on TUM · TUM full image | AbsRel: 0.185 -> 0.163 | Improved |
| add local structure refinement (LSR) (ablation) | SC-DepthV3 on TUM · TUM full image | AbsRel: 0.195 -> 0.163 | Improved |
| add local structure refinement (LSR) (ablation) | SC-DepthV3 on iBims-1 · iBims-1 | AbsRel: 0.152 -> 0.146 | Improved |
| use dynamic-focused sampling instead of random sampling in DRR (ablation) | DRR · DDAD full image | AbsRel: 0.154 -> 0.149 | Improved |
| use confident depth ranking loss instead of original ranking loss (ablation) | DRR · DDAD full image | AbsRel: 0.159 -> 0.149 | Improved |
| use normal matching loss instead of edge-aware depth smoothness (ablation) | LSR · DDAD full image | AbsRel: 0.148 -> 0.142 | Improved |
| do not add random sampling to edge-guided sampling (ablation) | LSR · DDAD full image | AbsRel: 0.146 -> 0.142 | Improved |
| use LeReS ResNet-101 pseudo-depth instead of DPT (ablation) | SC-DepthV3 · DDAD full image | AbsRel: 0.151 -> 0.142 | Improved |
Closest papers by meaning
- Digging Into Self-Supervised Monocular Depth Estimation (ICCV 2019 · similarity 0.92)
- Mining Supervision for Dynamic Regions in Self-Supervised Monocular Depth Estimation (CVPR 2024 · similarity 0.92)
- Self-Supervised Monocular Trained Depth Estimation Using Self-Attention and Discrete Disparity Volume (CVPR 2020 · similarity 0.91)
- The Temporal Opportunist: Self-Supervised Multi-Frame Monocular Depth (CVPR 2021 · similarity 0.91)
- From-Ground-To-Objects: Coarse-to-Fine Self-supervised Monocular Depth Estimation of Dynamic Objects with Ground Contact Prior (CVPR 2024 · similarity 0.89)
- Align3R: Aligned Monocular Depth Estimation for Dynamic Videos (CVPR 2025 · similarity 0.88)
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.
