Searching for just a few words should be enough to get started. If you need to make more complex queries, use the tips below to guide you.
Article type: Research Article
Authors: Li, Fanshua | Yao, Dengfenga; b; * | Jiang, Minghub | Kang, Xinchena
Affiliations: [a] Beijing Key Laboratory of Information Service Engineering, Beijing Union University, Beijing, China | [b] Lab of Computational Linguistics, School of Humanities, Tsinghua University, Beijing, China
Correspondence: [*] Corresponding author. Dengfeng Yao, E-mail: yaodengfeng@gmail.com.
Abstract: A new smoking behavior recognition algorithm based on a weak supervision fine-grained structure and the EficientDet network is proposed in this study to solve the poor recognition effect and lack of data samples of smoking behavior in complex situations. The proposed algorithm uses the framework of a fine-grained two-level attention model with weak supervision. First, the feature edge of the image block is detected by a structured method, and the edge is screened by non-maximum suppression to form a candidate region block. Smoking behavior can then be recognized effectively by combining the results of the object-level filter for specific objects and the local-level filter for locating discriminant parts. Second, the object-level filter uses an improved EfficientDet network to classify prospective objects and candidate regions with strong features. The present smoking behavior recognition algorithm and coarse- and fine-grained algorithms are compared to verify the effectiveness of the algorithm. Experimental results show that the accuracy of the proposed algorithm is 93.10%, which is higher than that of the optimal smoking behavior detection algorithm by 1.7%, and the error detection rate is 3.6%.
Keywords: Smoking, EfficienDet network, weakly supervised fine-grained target detection, attention mechanism, behavior recognition
DOI: 10.3233/JIFS-213042
Journal: Journal of Intelligent & Fuzzy Systems, vol. 43, no. 5, pp. 5733-5747, 2022
IOS Press, Inc.
6751 Tepper Drive
Clifton, VA 20124
USA
Tel: +1 703 830 6300
Fax: +1 703 830 2300
sales@iospress.com
For editorial issues, like the status of your submitted paper or proposals, write to editorial@iospress.nl
IOS Press
Nieuwe Hemweg 6B
1013 BG Amsterdam
The Netherlands
Tel: +31 20 688 3355
Fax: +31 20 687 0091
info@iospress.nl
For editorial issues, permissions, book requests, submissions and proceedings, contact the Amsterdam office info@iospress.nl
Inspirees International (China Office)
Ciyunsi Beili 207(CapitaLand), Bld 1, 7-901
100025, Beijing
China
Free service line: 400 661 8717
Fax: +86 10 8446 7947
china@iospress.cn
For editorial issues, like the status of your submitted paper or proposals, write to editorial@iospress.nl
如果您在出版方面需要帮助或有任何建, 件至: editorial@iospress.nl