SPA<SUP>2</SUP>Net: Structure-Preserved Attention Activated Network for Weakly Supervised Object Localization

doi:10.1109/TIP.2023.3323793

	SPA²Net: Structure-Preserved Attention Activated Network for Weakly Supervised Object Localization
	Chen, Dong 1; Pan, Xingjia 2; Tang, Fan 3; Dong, Weiming 4; Xu, Changsheng 4
	2023
发表期刊	IEEE TRANSACTIONS ON IMAGE PROCESSING
ISSN	1057-7149
卷号	32 页码:5779-5793
摘要	By exploring the localizable representations in deep CNN, weakly supervised object localization (WSOL) methods could determine the position of the object in each image just trained by the classification task. However, the partial activation problem caused by the discriminant function makes the network unable to locate objects accurately. To alleviate this problem, we propose Structure-Preserved Attention Activated Network (SPA(2)Net), a simple and effective one-stage WSOL framework to explore the ability of structure preservation of deep features. Different from traditional WSOL approaches, we decouple the object localization task from the classification branch to reduce their mutual influence by involving a localization branch which is online refined by a self-supervised structural-preserved localization mask. Specifically, we employ the high-order self-correlation as structural prior to enhance the perception of spatial interaction within convolutional features. By succinctly combining the structural prior with spatial attention, activations by SPA(2)Net will spread from part to the whole object during training. To avoid the structure-missing issue caused by the classification network, we furthermore utilize the restricted activation loss (RAL) to distinguish the difference between foreground and background in the channel dimension. In conjunction with the self-supervised localization branch, SPA(2)Net can directly predict the class-irrelevant localization map while prompting the network to pay more attention to the target region for accurate localization. Extensive experiments on two publicly available benchmarks, including CUB-200-2011 and ILSVRC, show that our SPA(2)Net achieves substantial and consistent performance gains compared with baseline approaches. The code and models are available at https://github.com/MsterDC/SPA2Net.
关键词	High-order self-correlation class activation map structure preservation weakly supervised object localization
DOI	10.1109/TIP.2023.3323793
收录类别	SCI
语种	英语
资助项目	National Natural Science Foundation of China[62102162] ; National Natural Science Foundation of China[U20B2070] ; National Natural Science Foundation of China[61832016] ; National Natural Science Foundation of China[61832002] ; Beijing Natural Science Foundation[L221013]
WOS研究方向	Computer Science ; Engineering
WOS类目	Computer Science, Artificial Intelligence ; Engineering, Electrical & Electronic
WOS记录号	WOS:001104979200002
出版者	IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
引用统计	被引频次：2[WOS] [WOS记录] [WOS相关记录]
文献类型	期刊论文
条目标识符	http://119.78.100.204/handle/2XEOYT63/38072
专题	中国科学院计算技术研究所期刊论文_英文
通讯作者	Tang, Fan
作者单位	1.Jilin Univ, Sch Artificial Intelligence, Changchun 130012, Peoples R China 2.Momenta, Beijing 100018, Peoples R China 3.Chinese Acad Sci, Inst Comp Technol, Beijing 100190, Peoples R China 4.Chinese Acad Sci, Inst Automat, Beijing 100190, Peoples R China
推荐引用方式 GB/T 7714	Chen, Dong,Pan, Xingjia,Tang, Fan,et al. SPA²Net: Structure-Preserved Attention Activated Network for Weakly Supervised Object Localization[J]. IEEE TRANSACTIONS ON IMAGE PROCESSING,2023,32:5779-5793.
APA	Chen, Dong,Pan, Xingjia,Tang, Fan,Dong, Weiming,&Xu, Changsheng.(2023).SPA²Net: Structure-Preserved Attention Activated Network for Weakly Supervised Object Localization.IEEE TRANSACTIONS ON IMAGE PROCESSING,32,5779-5793.
MLA	Chen, Dong,et al."SPA²Net: Structure-Preserved Attention Activated Network for Weakly Supervised Object Localization".IEEE TRANSACTIONS ON IMAGE PROCESSING 32(2023):5779-5793.