Multi-domain universal representation learning for hyperspectral object tracking

Zhuanfeng Li, Fengchao Xiong, Jianfeng Lu, Jing Wang, Diqi Chen, Jun Zhou, Yuntao Qian

Abstract: Highlights•This work learns universal representation for multi-domain hyperspectral tracking.•The spatial–spectral prompt guides foundation model to learn universal features.•The domain adapter learns domain-specific feature to address distribution difference.•The effectiveness of our method is validated on hyperspectral tracking datasets.