Research map / 3D vision / Depth and stereo
Dynamo-Depth: Fixing Unsupervised Depth Estimation for Dynamical Scenes
Read the paperCheck 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 |
|---|---|---|---|
| add motion consistency loss L_c (ablation) | Dynamo-Depth · Waymo Open | Abs Rel (all): 0.525 -> 0.116 | Helped |
| add motion initialization (ablation) | Dynamo-Depth · Waymo Open | Abs Rel (all): 0.522 -> 0.116 | Helped |
| add above-ground loss L_g (ablation) | Dynamo-Depth · Waymo Open | Abs Rel (all): 0.266 -> 0.116 | Helped |
| add above-ground loss L_g (ablation) | Dynamo-Depth · Waymo Open moving objects | Abs Rel: 0.211 -> 0.194 | No clear effect |
| add ImageNet pretraining of encoder (ablation) | Dynamo-Depth · Waymo Open | Abs Rel (all): 0.136 -> 0.116 | Helped |
| model independent motion (Dynamo-Depth vs LiteMono) | LiteMono depth network · Waymo Open moving objects | Abs Rel: 0.599 -> 0.194 | Helped |
| model independent motion (Dynamo-Depth vs LiteMono) | LiteMono depth network · nuScenes | Abs Rel (all): 0.419 -> 0.179 | Helped |
| model independent motion (Dynamo-Depth vs LiteMono) | LiteMono depth network · KITTI Eigen | Abs Rel: 0.101 -> 0.112 | Hurt |
Closest papers by meaning
- Mining Supervision for Dynamic Regions in Self-Supervised Monocular Depth Estimation (CVPR 2024 · similarity 0.88)
- Seurat: From Moving Points to Depth (CVPR 2025 · similarity 0.88)
- Depth2Pose: A Pose-Based Benchmark for Monocular Depth Estimation without Ground-Truth Depth (NeurIPS 2026 · similarity 0.87)
- Stereo4D: Learning How Things Move in 3D from Internet Stereo Videos (CVPR 2025 · similarity 0.86)
- Layered Motion Fusion: Lifting Motion Segmentation to 3D in Egocentric Videos (CVPR 2025 · similarity 0.86)
- Video Depth Anything: Consistent Depth Estimation for Super-Long Videos (CVPR 2025 · similarity 0.86)
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.