Optimizing Multi-Modal Trackers via Sensitivity-aware Regularized Tuning
Zhiwen Chen, Jinjian Wu, Zhiyu Zhu, Yifan Zhang, Guangming Shi · 2025
Read Optimizing Multi-Modal Trackers via Sensitivity-aware Regularized Tuning by Zhiwen Chen, Jinjian Wu, Zhiyu Zhu, Yifan Zhang, Guangming Shi (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-2508-17488.txt.
Description
Zhiwen Chen and colleagues improve fine tuning for trackers that use more than one data type, in work from 2025. Trackers pretrained on colour video must adapt to added signals such as depth or heat views. Standard tuning can be too loose or too tight, and both extremes hurt the balance of change and stability.
Their framework scores how much each parameter matters before and during tuning. Early scores protect core patterns learned in pretraining. Later scores track what helps the shift to new data types while keeping results stable. Both scores act as one shared penalty during training. Tests across multi-modal tracking benchmarks beat leading methods. Code and models are shared in public for reuse.
Catalog record
| CREATOR | Zhiwen Chen, Jinjian Wu, Zhiyu Zhu, Yifan Zhang, Guangming Shi |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 70 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2508-17488