ProTracker: Probabilistic Integration for Robust and Accurate Point Tracking
Tingyang Zhang, Chen Wang, Zhiyang Dou, Qingzhe Gao, Jiahui Lei, Baoquan Chen, Lingjie Liu · 2025
Read ProTracker: Probabilistic Integration for Robust and Accurate Point Tracking by Tingyang Zhang, Chen Wang, Zhiyang Dou, Qingzhe Gao, Jiahui Lei, Baoquan Chen, Lingjie Liu (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/arxiv-2501-03220.txt.
Description
Following a single point through a long video is hard when the point hides behind objects or the scene cuts. Tingyang Zhang and co-authors propose ProTracker, which mixes two older strategies. Global methods search the whole frame and survive big changes but land imprecisely; local methods refine step by step and stay accurate until they drift. ProTracker uses local optical flow to predict where a point goes and global heat maps, sharpened by probability reasoning, to correct the guess.
The mix brings global scene knowledge together with fine, time-aware detail. In tests the system matched the best optimisation-based trackers and beat supervised feed-forward rivals on several benchmarks. The authors plan to release code and models after publication, aiming at long, dense tracking of any chosen point.
Catalog record
| CREATOR | Tingyang Zhang, Chen Wang, Zhiyang Dou, Qingzhe Gao, Jiahui Lei, Baoquan Chen, Lingjie Liu |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 39 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2501-03220