Research map / 3D vision / Depth and stereo
"Towards Scale-Aware, Robust, and Generalizable Unsupervised Monocular Depth Estimation by Integrating IMU Motion Dynamics"
Sen Zhang, Jing Zhang, Dacheng Tao
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 IMU and cross-sensor losses to get absolute scale | Monodepth2-style network · KITTI Eigen | scale ratio: 29.128 -> 1.013 | Improved |
| add cross-sensor photometric consistency loss to IMU photometric loss (ablation) | DynaDepth (EKF) · KITTI Eigen ablation | scale ratio: 1.130 -> 1.076 | Improved |
| add IMU photometric loss to consistency loss (ablation) | DynaDepth (EKF) · KITTI Eigen ablation | scale ratio: 4.271 -> 1.076 | Improved |
| add Lvg velocity-gravity loss (ablation) | DynaDepth (EKF) · KITTI Eigen ablation | AbsRel: 0.113 -> 0.111 | Improved |
| add Lvg velocity-gravity loss (ablation) | DynaDepth · Make3D (rescaled) | AbsRel: 0.313 -> 0.334 | Hurt |
| add EKF fusion (ablation) | DynaDepth (IMU+cons) · KITTI Eigen ablation | AbsRel: 0.115 -> 0.113 | Improved |
| add EKF fusion (ablation) | DynaDepth (IMU+cons) · KITTI Eigen ablation | scale ratio: 0.968 -> 1.076 | Mixed |
| add EKF fusion (ablation) | DynaDepth · Make3D (rescaled) | AbsRel: 0.316 -> 0.313 | Improved |
| add EKF fusion (ablation) | DynaDepth · Make3D (unscaled) | AbsRel: 0.360 -> 0.337 | Improved |
| add EKF fusion (ablation) | DynaDepth · KITTI simulated moving objects | AbsRel: 0.169 -> 0.126 | Improved |
| add EKF fusion (ablation) | DynaDepth · KITTI simulated illumination change | AbsRel: 0.124 -> 0.123 | Improved |
| add Lvg on top of EKF (ablation) | DynaDepth · KITTI simulated moving objects | scale ratio: 0.087 -> 0.956 | Improved |
Closest papers by meaning
- Unsupervised Scale-consistent Depth and Ego-motion Learning from Monocular Video (NeurIPS 2019 · similarity 0.89)
- SynDeMo: Synergistic Deep Feature Alignment for Joint Learning of Depth and Ego-Motion (ICCV 2019 · similarity 0.89)
- Unsupervised Learning of Depth and Ego-Motion From Monocular Video Using 3D Geometric Constraints (CVPR 2018 · similarity 0.88)
- Learning Monocular Depth in Dynamic Scenes via Instance-Aware Projection Consistency (AAAI 2021 · similarity 0.88)
- Seurat: From Moving Points to Depth (CVPR 2025 · similarity 0.87)
- UniDepth: Universal Monocular Metric Depth Estimation (CVPR 2024 · 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.
