[an error occurred while processing this directive]

Journal of Systems Engineering and Electronics ›› 2020, Vol. 31 ›› Issue (1): 45-55.doi: 10.21629/JSEE.2020.01.06

• • 上一篇    下一篇

  

  • 收稿日期:2019-04-02 出版日期:2020-02-20 发布日期:2020-02-25

Attributes-based person re-identification via CNNs with coupled clusters loss

Rui SUN1,2,*(), Qiheng HUANG1,2(), Wei FANG1,2(), Xudong ZHANG1,2()   

  1. 1 Key Laboratory of Knowledge Engineering with Big Data (Ministry of Education), Hefei University of Technology, Hefei 230601, China
    2 School of Computer and Information, Hefei University of Technology, Hefei 230601, China
  • Received:2019-04-02 Online:2020-02-20 Published:2020-02-25
  • Contact: Rui SUN E-mail:sunrui@hfut.edu.cn;jchqh123@163.com;1204764020@qq.com;xudong@hfut.edu.cn
  • About author:SUN Rui was born in 1976. He is a Ph.D. and a professor in Hefei University of Technology. His research interests are computer vision, intelligent information processing and machine learning. E-mail: sunrui@hfut.edu.cn|HUANG Qiheng was born in 1994. He received his master degree from Hefei University of Technology. He is a R & D engineer in the China Mobile Anhui Branch. His research interests are computer vision and deep learning. E-mail: jchqh123@163.com|FANG Wei was born in 1993. He received his master degree from Hefei University of Technology. He is a R & D engineer in the ArcSoft Corporation Limited now. His research interests are computer vision and deep learning. E-mail: 1204764020@qq.com|ZHANG Xudong was born in 1966. He is a Ph.D. and a professor in Hefei University of Technology. His research interests are pattern recognition and intelligent information processing. E-mail: xudong@hfut.edu.cn
  • Supported by:
    the National Natural Science Foundation of China(61471154);the National Natural Science Foundation of China(61876057);the Fundamental Research Funds for Central Universities(JZ2018YYPY0287);This work was supported by the National Natural Science Foundation of China (61471154; 61876057) and the Fundamental Research Funds for Central Universities (JZ2018YYPY0287)

Abstract:

Person re-identification (re-id) involves matching a person across nonoverlapping views, with different poses, illuminations and conditions. Visual attributes are understandable semantic information to help improve the issues including illumination changes, viewpoint variations and occlusions. This paper proposes an end-to-end framework of deep learning for attribute-based person re-id. In the feature representation stage of framework, the improved convolutional neural network (CNN) model is designed to leverage the information contained in automatically detected attributes and learned low-dimensional CNN features. Moreover, an attribute classifier is trained on separate data and includes its responses into the training process of our person re-id model. The coupled clusters loss function is used in the training stage of the framework, which enhances the discriminability of both types of features. The combined features are mapped into the Euclidean space. The L2 distance can be used to calculate the distance between any two pedestrians to determine whether they are the same. Extensive experiments validate the superiority and advantages of our proposed framework over state-of-the-art competitors on contemporary challenging person re-id datasets.

Key words: person re-identification (re-id), convolutions neural network (CNN), attributes, coupled clusters loss (CCL)