文献类型：期刊文献

英文题名：Guided Attention and Joint Loss for Infrared Dim Small Target Detection

作者：Tong, Yunfei[1,2];Liu, Jing[1,2];Fu, Zhiling[1,2];Wang, Zhe[1,2];Yang, Hai[1,2];Niu, Saisai[3,4];Tan, Qinyan[3,4]

机构：[1]East China Univ Sci & Technol, Minist Educ, Key Lab Smart Mfg Energy Chem Proc, Shanghai 200237, Peoples R China;[2]East China Univ Sci & Technol, Dept Comp Sci & Engn, Shanghai 200237, Peoples R China;[3]China Aerosp Sci & Technol Corp, Shanghai Aerosp Control Technol Inst, Shanghai 201109, Peoples R China;[4]China Aerosp Sci & Technol Corp, Res & Dev Ctr Infrared Detect Technol, Shanghai 201109, Peoples R China

年份：2024

卷号：62

外文期刊名：IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING

收录：;EI(收录号：20244417298414);WOS:【SCI-EXPANDED(收录号:WOS:001342538500017)】；

基金：This work was supported in part by the Natural Science Foundation of China under Grant 62476087, in part by the National Key Research and Development Program of China under Grant 2022YFB3203500, and in part by the Chinese Defense Program of Science and Technology under Grant 2021-JCJQ-JJ-0041.

语种：英文

外文关键词：Feature extraction; Object detection; Accuracy; Attention mechanisms; Sensitivity; Geoscience and remote sensing; Fitting; Detection algorithms; Deep learning; Windows; Guided-attention (GA) mechanism; infrared dim small target (IDST) detection; joint loss (JL); positional bias; YOLO series

摘要：Infrared dim small target (IDST) detection is of great significance in security surveillance and disaster relief. However, the complex background interference and tiny targets in infrared images keep it still a long-term challenge. Existing deep learning models stack network layers to expand the model fitting capability, but this operation also increases redundant features which reduce model speed and accuracy. Meanwhile, small targets are more susceptible to positional bias, with this dramatically reducing the model's localization accuracy. In this article, we propose a guided attention and joint loss (GA-JL) network for infrared small target detection. More specifically, the method visualizes the feature maps at each resolution through a two-branch detection head (TDH) module, filters out the features that are strongly related to the task, and cuts out the redundant features. On this basis, the guided attention (GA) module guides the prediction layer features using the features that are associated closely with the task and combines spatial and channel bidirectional attention to make the prediction layer feature maps embedded with effective messages. Finally, through the joint loss (JL) module, the target position regression is performed with multiangle metrics for enhancing the target detection accuracy. Experimental results of our method on the SIATD, SIRST, and IRSTD_1k datasets reveal that it is capable of accurately identifying IDSTs, remarkably reduces the false alarm rate, and outperforms other methods.