VeTrack leverages transformer-based appearance modeling and confidence-guided multi-level association for robust multi-vehicle tracking
Haichao Liu, Jiangwei Qin, Zhiguang Zhang, Haoyu Liang, Haotian Deng, Shuangjie Xiao, Miao Yu · 2025
Read VeTrack leverages transformer-based appearance modeling and confidence-guided multi-level association for robust multi-vehicle tracking by Haichao Liu, Jiangwei Qin, Zhiguang Zhang, Haoyu Liang, Haotian Deng, Shuangjie Xiao, Miao Yu (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/pmc-12824284.txt.
Description
Liu and co-authors present VeTrack, a system for following many vehicles at once in traffic video. Tracking fails most often when similar-looking cars swap identities or when vehicles hide behind each other in dense traffic, both common in real road scenes.
VeTrack pairs a YOLOX detector with a transformer-based module that studies each vehicle's appearance in fine detail, making look-alikes easier to tell apart. A multi-level association strategy then chooses how to match detections to tracks based on confidence, which reduces breaks during occlusion. On standard benchmarks the system reached state-of-the-art scores, including a strong result on MOT17, and ran fast enough for real traffic monitoring.
Catalog record
| CREATOR | Haichao Liu, Jiangwei Qin, Zhiguang Zhang, Haoyu Liang, Haotian Deng, Shuangjie Xiao, Miao Yu |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 58 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |