SiamCAF: Complementary Attention Fusion-Based Siamese Network for RGBT Tracking

The tracking community is increasingly focused on RGBT tracking, which leverages the complementary strengths of corresponding visible light and thermal infrared images. most well-known trackers, however, are unable to balance performance speed at same time for UAV tracking. In this paper, an innovative Siamese tracker named SiamCAF proposed, utilizes multi-modal features with a beyond-real-time running speed. Specifically, we used dual-modal subnetwork extract features. addition, similar reduce modality differences fusing efficiently, designed Complementary Coupling Feature fusion module (CCF). Simultaneously, Residual Channel Attention Enhanced (RCAE) was enhance extracted representational power. Furthermore, Maximum Fusion Prediction (MFP) constructed boost in response map stage. Finally, comprehensive experiments three real datasets one visible–thermal dataset showed that outperforms other methods, remarkable over 105 frames per second.