Research map / Vision-language and multimodal / Cross-modal learning
XVO: Generalized Visual Odometry via Cross-Modal Self-Training
Lei Lai, Zhongkai Shangguan, Jimuyang Zhang, Eshed Ohn-Bar
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 semi-supervised self-training on pseudo-labeled YouTube video (student) to nuScenes teacher | XVO · KITTI seqs 03-10 average | t_rel / r_rel: 25.27 -> 15.17 t_rel; 8.17 -> 4.08 r_rel | Improved |
| add uncertainty-based pseudo-label filtering (ablation) | Student · KITTI seqs 03-10 average | t_rel: 18.73 -> 15.17 | Improved |
| add uncertainty-based filtering (ablation) | Student · nuScenes unseen regions | t_err: 9.32 -> 10.54 | Mixed |
| add audio prediction auxiliary task (ablation) | Student · KITTI 00-10 | t_err: 17.04 -> 14.37 | Improved |
| add audio prediction auxiliary task (ablation) | Student · Argoverse 2 | t_err: 9.16 -> 8.00 | Improved |
| add audio prediction auxiliary task (ablation) | Student · nuScenes unseen regions | t_err: 10.54 -> 9.26 | Improved |
| add segmentation auxiliary task (ablation) | Student · nuScenes unseen regions | t_err: 10.54 -> 11.35 | Hurt |
| add depth auxiliary task (ablation) | Student · KITTI 00-10 / nuScenes | t_err: 17.04 -> 17.49 (KITTI); 10.54 -> 11.86 (nuScenes) | Mixed |
| add flow auxiliary task (ablation) | Student · KITTI 00-10 | t_err: 17.04 -> 15.60 | Improved |
| combine all four auxiliary tasks instead of audio alone (ablation) | Student+Audio · KITTI 00-10 | t_err: 14.37 -> 16.74 | Hurt |
| add noise to camera intrinsics for TartanVO | TartanVO · KITTI seqs 03-10 average | t_rel: 5.93 -> 11.72 (10% noise) -> 24.51 (30% noise) | Hurt |
Closest papers by meaning
- VSO: Visual Semantic Odometry (ECCV 2018 · similarity 0.85)
- Adaptive VIO: Deep Visual-Inertial Odometry with Online Continual Learning (CVPR 2024 · similarity 0.85)
- Modality-Invariant Visual Odometry for Embodied Vision (CVPR 2023 · similarity 0.85)
- Generalizing to the Open World: Deep Visual Odometry With Online Adaptation (CVPR 2021 · similarity 0.85)
- LEAP-VO: Long-term Effective Any Point Tracking for Visual Odometry (CVPR 2024 · similarity 0.84)
- ZeroVO: Visual Odometry with Minimal Assumptions (CVPR 2025 · similarity 0.83)
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.
