Overview of behavior recognition based on deep learning

被引：54
作者：
Hu, Kai [1 ,2 ]
Jin, Junlan [1 ,2 ]
Zheng, Fei [2 ,3 ]
Weng, Liguo [1 ,2 ]
Ding, Yiwu [1 ,2 ]
机构：
[1] Nanjing Univ Informat Sci & Technol, Sch Automat, 219 Ningliu Rd, Nanjing 210044, Jiangsu, Peoples R China
[2] Nanjing Univ Informat Sci & Technol, Jiangsu Collaborat Innovat Ctr Atmospher Environm, 219 Ningliu Rd, Nanjing 210044, Jiangsu, Peoples R China
[3] Innovat Dept Ind Internet, China Telecom Ningbo Branch, 96 HeYi Rd, Ningbo 315000, Zhejiang, Peoples R China
来源：
ARTIFICIAL INTELLIGENCE REVIEW | 2023年 / 56卷 / 03期
基金：
中国国家自然科学基金;
关键词：
Behavior recognition; Deep learning; Skeleton data; NETWORK; LSTM;
D O I：
10.1007/s10462-022-10210-8
中图分类号：
TP18 [人工智能理论];
学科分类号：
081104 ; 0812 ; 0835 ; 1405 ;
摘要：
Human behavior recognition has always been a hot spot for research in computer vision. With the wide application of behavior recognition in virtual reality and short video in recent years and the rapid development of deep learning algorithms, behavior recognition algorithms based on deep learning have emerged. Compared with traditional methods, behavior recognition algorithms based on deep learning have the advantages of strong robustness and high accuracy. This paper systemizes and introduces behavior recognition algorithms based on deep learning proposed in recent years, then focuses on a series of behavior recognition algorithms based on image and bone data; deeply analyzes their theories and performance, and finally, puts forward further prospects.
引用
收藏
页码：1833 / 1865
页数：33
相关论文
共 76 条
[51]   SUACDNet: Attentional change detection network based on siamese U-shaped structure [J].
Song, Lei ;
Xia, Min ;
Jin, Junlan ;
Qian, Ming ;
Zhang, Yonghong .
INTERNATIONAL JOURNAL OF APPLIED EARTH OBSERVATION AND GEOINFORMATION, 2021, 105
[52]   Action Recognition with Improved Trajectories [J].
Wang, Heng ;
Schmid, Cordelia .
2013 IEEE INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV), 2013, :3551-3558
[53]   Dense Trajectories and Motion Boundary Descriptors for Action Recognition [J].
Wang, Heng ;
Klaeser, Alexander ;
Schmid, Cordelia ;
Liu, Cheng-Lin .
INTERNATIONAL JOURNAL OF COMPUTER VISION, 2013, 103 (01) :60-79
[54]   Temporal Segment Networks: Towards Good Practices for Deep Action Recognition [J].
Wang, Limin ;
Xiong, Yuanjun ;
Wang, Zhe ;
Qiao, Yu ;
Lin, Dahua ;
Tang, Xiaoou ;
Van Gool, Luc .
COMPUTER VISION - ECCV 2016, PT VIII, 2016, 9912 :20-36
[55]   Compressed Video Action Recognition [J].
Wu, Chao-Yuan ;
Zaheer, Manzil ;
Hu, Hexiang ;
Manmatha, R. ;
Smola, Alexander J. ;
Krahenbuhl, Philipp .
2018 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR), 2018, :6026-6035
[56]   DAU-Net: a novel water areas segmentation structure for remote sensing image [J].
Xia, Min ;
Cui, Yichen ;
Zhang, Yonghong ;
Xu, Yiming ;
Liu, Jia ;
Xu, Yiqing .
INTERNATIONAL JOURNAL OF REMOTE SENSING, 2021, 42 (07) :2594-2621
[57]   Cloud/shadow segmentation based on global attention feature fusion residual network for remote sensing imagery [J].
Xia, Min ;
Wang, Tao ;
Zhang, Yonghong ;
Liu, Jia ;
Xu, Yiqing .
INTERNATIONAL JOURNAL OF REMOTE SENSING, 2021, 42 (06) :2022-2045
[58]   Non-intrusive load disaggregation based on composite deep long short-term memory network [J].
Xia, Min ;
Liu, Wan'an ;
Wang, Ke ;
Song, Wenzhu ;
Chen, Chunling ;
Li, Yaping .
EXPERT SYSTEMS WITH APPLICATIONS, 2020, 160
[59]   Multi-Stage Feature Constraints Learning for Age Estimation [J].
Xia, Min ;
Zhang, Xu ;
Liu, Wan'an ;
Weng, Liguo ;
Xu, Yiqing .
IEEE TRANSACTIONS ON INFORMATION FORENSICS AND SECURITY, 2020, 15 :2417-2428
[60]  
Xianyuan Wang, 2019, IOP Conference Series: Materials Science and Engineering, V569, DOI 10.1088/1757-899X/569/3/032035
12345678