Research map / 3D vision / Depth and stereo
Learned Monocular Depth Priors in Visual-Inertial Initialization
Yunwen Zhou, Abhishek Kar, Eric Turner, Adarsh Kowdle, Chao X. Guo, Ryan C. DuToit, Konstantine Tsotsos
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 monocular depth constraints to VI-SFM initialization (vs Baseline) | VI-BA initialization · EuRoC exhaustive init (5 KFs, mean over 11 sequences) | scale error (%): 45.87 -> 30.55 | Improved |
| add monocular depth constraints to VI-SFM initialization (vs Baseline) | VI-BA initialization · EuRoC exhaustive init (5 KFs, mean over 11 sequences) | position RMSE (m): 0.053 -> 0.036 | Improved |
| add monocular depth constraints to VI-SFM initialization (vs Baseline) | VI-BA initialization · EuRoC exhaustive init (5 KFs, mean over 11 sequences) | gravity RMSE (deg): 2.92 -> 2.26 | Improved |
| add monocular depth constraints to VI-SFM initialization | VI-BA initialization · EuRoC exhaustive init (5 KFs, mean over 11 sequences) | log condition number: 13.27 -> 12.14 | Improved |
| add scale-shift prior (ablation) | Mono-depth VI initialization · EuRoC exhaustive init ablation (5 KFs, all sequences) | scale error (%): 35.47 -> 31.23 | Improved |
| add scale-shift prior (ablation) | Mono-depth VI initialization · EuRoC exhaustive init ablation (5 KFs, all sequences) | log condition number: 13.24 -> 12.14 | Improved |
| add edge-awareness weight (ablation) | Mono-depth VI initialization · EuRoC exhaustive init ablation (5 KFs, all sequences) | scale error (%): 34.22 -> 31.23 | Improved |
| add edge-awareness weight (ablation) | Mono-depth VI initialization · EuRoC exhaustive init ablation (5 KFs, all sequences) | position RMSE (m): 0.041 -> 0.036 | Improved |
| add outlier rejection (ablation) | Mono-depth VI initialization · EuRoC exhaustive init ablation (5 KFs, all sequences) | scale error (%): 36.59 -> 31.23 | Improved |
| add outlier rejection (ablation) | Mono-depth VI initialization · EuRoC exhaustive init ablation (5 KFs, all sequences) | log condition number: 13.18 -> 12.14 | Improved |
| add prior, weight and outlier rejection together (ablation) | Mono-depth VI initialization · EuRoC exhaustive init ablation (5 KFs, all sequences) | position RMSE (m): 0.044 -> 0.036 | Improved |
| add mono-depth initialization to VIO tracking | OpenVINS-like VIO · EuRoC VIO tracking (5 KF init, 10 s segments, mean) | position RMSE (m): 2.243 -> 0.351 | Improved |
Closest papers by meaning
- A Rotation-Translation-Decoupled Solution for Robust and Efficient Visual-Inertial Initialization (CVPR 2023 · similarity 0.91)
- Adaptive VIO: Deep Visual-Inertial Odometry with Online Continual Learning (CVPR 2024 · similarity 0.90)
- Depth2Pose: A Pose-Based Benchmark for Monocular Depth Estimation without Ground-Truth Depth (NeurIPS 2026 · similarity 0.87)
- EqNIO: Subequivariant Neural Inertial Odometry (ICLR 2025 · similarity 0.87)
- LEAP-VO: Long-term Effective Any Point Tracking for Visual Odometry (CVPR 2024 · similarity 0.85)
- Seurat: From Moving Points to Depth (CVPR 2025 · 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.
