Self-attention mechanism in person re-identification models

被引:0
作者
Wenbai Chen
Yue Lu
Hang Ma
Qili Chen
Xibao Wu
Peiliang Wu
机构
[1] Beijing Information Science & Technology University,School of Automation
[2] Yanshan University,School of Information and Engineering
[3] Chinese Academy of Sciences,State Key Laboratory of Management and Control for Complex Systems, Institute of Automation
来源
Multimedia Tools and Applications | 2022年 / 81卷
关键词
Person re-identification; Deep neural network; Self-attention; Computer vision;
D O I
暂无
中图分类号
学科分类号
摘要
In recent years, person re-identification based on video has become a hot topic in the field of person re-identification. The self-attention mechanism can improve the ability of deep neural networks in computer vision tasks such as image classification, image segmentation and natural language processing tasks. In order to verify whether the self-attention can improve the performance or not in person re-identification tasks, this paper applies two self-attention mechanisms, non-local attention and recurrent criss-cross attention to person re-identification model, and experiments are conducted on Market-1501, DukeMTMC-reID and MSMT17 person re-identification datasets. The results show that the self-attention mechanism can improve the accuracy of the person re-identification model. The accuracy is higher when the self-attention module is inserted into the convolutional layers of the re-identification network.
引用
收藏
页码:4649 / 4667
页数:18
相关论文
共 18 条
  • [1] Hao L(2019)A survey on deep learning based person re-identification ACTA AUTOMATICA SINICA 45 2032-2049
  • [2] Wei J(2020)Research progress on batch normalization of deep learning and its related algorithms Acta Automatica Sinica 46 1090-1120
  • [3] Jian-Wei L(2018)Applying the convolutional neural network deep learning technology to behavioural recognition in intelligent video Tehnicki Vjesnik 25 528-535
  • [4] Hui-Dan Z(2017)Survey on pedestrian re-identification research CAAI Trans Intell Syst 12 770-780
  • [5] Xiong-Lin L(2018)A survey of person re-identification ACTA AUTOMATICA SINICA 44 1554-1568
  • [6] Jun X(2017)A discriminatively learned cnn embedding for person reidentification ACM Trans Multimed Comput Commun Applic (TOMM) 14 1-20
  • [7] Qin L(2020)Pedestrian re-identification feature extraction method based on attention mechanism J Comput Applic 40 672-undefined
  • [8] Yu N(undefined)undefined undefined undefined undefined-undefined
  • [9] Zhao D(undefined)undefined undefined undefined undefined-undefined
  • [10] SONG W(undefined)undefined undefined undefined undefined-undefined