Infrared Camera Region of Interest Tracking Based on Siamese Networks and Attention Mechanisms
Bo Li, Bingsheng Cao · 2024
In infrared target tracking, the robustness of the tracking algorithm may be affected due to the blurring of the target features in the infrared image and the obvious background noise.For this reason, we introduce two attention mechanism modules, CAFE (Channel Attention Feature Enhancement Module) and SAFE (Spatial Attention Feature Enhancement Module), into the Siamese network for enhancing IR image features and suppressing background noise.Specifically, the CAFÉ module aims to enhance attention to IR image features, while the SAFE module utilizes the spatial relationship of features to generate a spatial attention map to suppress the effect of background noise.Experimental results show that our proposed algorithm can significantly improve the tracking accuracy.