Weakly Supervised Local-Global Attention Network for Facial Expression Recognition

doi:10.1109/ACCESS.2020.2975913

	Weakly Supervised Local-Global Attention Network for Facial Expression Recognition
	Zhang, Haifeng 1; Su, Wen 3; Wang, Zengfu 1,2
刊名	IEEE ACCESS
	2020
卷号	8
关键词	Feature extraction Face recognition Mouth Measurement Generators Licenses Fuses Facial expression recognition weak supervision attention mechanism local features global features
ISSN号	2169-3536
DOI	10.1109/ACCESS.2020.2975913
通讯作者	Wang, Zengfu(zfwang@ustc.edu.cn)
英文摘要	Combining global and local features is an essential solution to improve discriminative performances in facial expression recognition tasks. The limitations of existing methods are that they cannot extract crucial local features and ignore the complementary effects of local and global features. To address these problems, this paper proposes a Weakly Supervised Local-Global Attention Network (WS-LGAN), which uses the attention mechanism to deal with part location and feature fusion problems. Firstly, an Attention Map Generator is designed to get a set of attention maps under weak supervision. It mimics the attention mechanism of human brain and quickly finds the local regions-of-interest. Secondly, bilinear attention pooling is employed to generate and refine local features based on attention maps. Thirdly, a building block called Selective Feature Unit is designed. It allows adaptive weighted fusion of global and local features before making classification. In WS-LGAN, global and local features represent expressions from different aspects. Compared with methods relying on single type of feature, it benefits from local-global complementary advantages. Additionally, contrastive loss is introduced for both local and global features to increase inter-class dispersion and intra-class compactness under different granularities. Experiments on three popular facial expression datasets, including two lab-controlled facial expression datasets and one real-world facial expression dataset show that WS-LGAN achieves state-of-the-art performance, which demonstrates our superiority in facial expression recognition.
资助项目	National Natural Science Foundation of China[61472393]
WOS研究方向	Computer Science ; Engineering ; Telecommunications
语种	英语
出版者	IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
WOS记录号	WOS:000525545900041
资助机构	National Natural Science Foundation of China
内容类型	期刊论文
源URL	[http://ir.hfcas.ac.cn:8080/handle/334002/103333]
专题	中国科学院合肥物质科学研究院
通讯作者	Wang, Zengfu
作者单位	1.Univ Sci & Technol China, Dept Automat, Hefei 230022, Peoples R China 2.Chinese Acad Sci, Inst Intelligent Machines, Hefei 230031, Peoples R China 3.Zhejiang Sci Tech Univ, Fac Mech Engn & Automat, Hangzhou 310018, Peoples R China
推荐引用方式 GB/T 7714	Zhang, Haifeng,Su, Wen,Wang, Zengfu. Weakly Supervised Local-Global Attention Network for Facial Expression Recognition[J]. IEEE ACCESS,2020,8.
APA	Zhang, Haifeng,Su, Wen,&Wang, Zengfu.(2020).Weakly Supervised Local-Global Attention Network for Facial Expression Recognition.IEEE ACCESS,8.
MLA	Zhang, Haifeng,et al."Weakly Supervised Local-Global Attention Network for Facial Expression Recognition".IEEE ACCESS 8(2020).