Journal of Computer Applications ›› 2019, Vol. 39 ›› Issue (8): 2223-2229.DOI: 10.11772/j.issn.1001-9081.2018122505

• Artificial intelligence • Previous Articles     Next Articles

Person re-identification based on deep multi-view feature distance learning

DENG Xuan1, LIAO Kaiyang1,2, ZHENG Yuanlin1,3, YUAN Hui1, LEI Hao1, CHEN Bing1   

  1. 1. College of Printing, Packaging Engineering and Digital Media Technology, Xi'an University of Technology, Xi'an Shaanxi 710048, China;
    2. Printing and Packaging Engineering Technology Research Centre of Shaanxi Province, Xi'an Shaanxi 710048, China;
    3. Key Laboratory of Printing and Packaging Engineering of Shaanxi Province, Xi'an Shaanxi 710048, China
  • Received:2018-12-19 Revised:2019-02-28 Online:2019-08-10 Published:2019-08-14
  • Supported by:
    This work is partially supported by the National Natural Science Foundation of China (61671376, 61771386), the Scientific Research Project of Shaanxi Provincial Department of Education (18JK0556).

基于深度多视图特征距离学习的行人重识别

邓轩1, 廖开阳1,2, 郑元林1,3, 袁晖1, 雷浩1, 陈兵1   

  1. 1. 西安理工大学 印刷包装与数字媒体学院, 西安 710048;
    2. 陕西省印刷包装工程技术研究中心, 西安 710048;
    3. 陕西省印刷包装工程重点实验室, 西安 710048
  • 通讯作者: 廖开阳
  • 作者简介:邓轩(1995-),女,山东菏泽人,硕士研究生,主要研究方向:深度学习、图像处理;廖开阳(1976-),男,湖北荆州人,讲师,博士,主要研究方向:机器视觉、人工智能;郑元林(1975-),男,山东泰安人,副教授,博士,主要研究方向:色彩管理、彩色图像质量评估、颜色科学;袁晖(1993-),男,江西赣州人,硕士研究生,主要研究方向:深度学习、图像处理;雷浩(1995-),男,湖南岳阳人,硕士研究生,主要研究方向:深度学习、图像处理;陈兵(1997-),男,陕西安康人,主要研究方向:图像处理。
  • 基金资助:
    国家自然科学基金资助项目(61671376,61771386);陕西省教育厅科学研究项目(18JK0556)。

Abstract: The traditional handcrafted features rely heavily on the appearance characteristics of pedestrians and the deep convolution feature is a high-dimensional feature, so, it will consume a lot of time and memory when the feature is directly used to match the image. Moreover, features from higher levels are easily affected by human pose or background clutter. Aiming at these problems, a method based on deep multi-view feature distance learning was proposed. Firstly, a new feature to improve and integrate the convolution feature of the deep region was proposed. The convolution feature was processed by the sliding frame technique, and the integration feature of low-dimensional deep region with the dimension equal to the number of convolution layer channels was obtained. Secondly, from the perspectives of the deep regional integration feature and the handcrafted feature, a multi-view feature distance learning algorithm was proposed by utilizing the cross-view quadratic discriminant analysis method. Finally, the weighted fusion strategy was used to accomplish the collaboration between handcrafted features and deep convolution features. Experimental results show that the Rank1 value of the proposed method reaches 80.17% and 75.32% respectively on the Market-1501 and VIPeR datasets; under the new classification rules of CHUK03 dataset, the Rank1 value of the proposed method reaches 33.5%. The results show that the accuracy of pedestrian re-identification after distance-weighted fusion is significantly higher than that of the separate feature distance metric, and the effectiveness of the proposed deep region features and algorithm model are proved.

Key words: person re-identification, Convolutional Neural Network (CNN), regional integration feature, weighted fusion strategy, distance metric

摘要: 传统手工特征很大程度上依赖于行人的外观特征,而深度卷积特征作为高维特征,直接用来匹配图像会消耗大量的时间和内存,并且来自较高层的特征很容易受到行人姿势背景杂波影响。针对这些问题,提出一种基于深度多视图特征距离学习的方法。首先,提出一种新的整合和改善深度区域的卷积特征,利用滑框技术对卷积特征进行处理,得到低维的深度区域聚合特征并使其维数等于卷积层通道数;其次,通过交叉视图二次判别分析方法,从深度区域聚合特征和手工特征两个角度出发,提出一种多视图特征距离学习算法;最后,利用加权融合策略来完成传统特征和卷积特征之间的协作。在Market-1501和VIPeR数据集上的实验结果显示,所提融合模型的Rank1值在两个数据集上分别达到80.17%和75.32%;在CUHK03数据集新分类规则下,所提方法的Rank1值达到33.5%。实验结果表明,通过距离加权融合之后的行人重识别的精度明显高于单独的特征距离度量取得的精度,验证了所提的深度区域特征和算法模型的有效性。

关键词: 行人重识别, 卷积神经网络, 区域聚合特征, 加权融合策略, 距离度量

CLC Number: