作者投稿和查稿 主编审稿 专家审稿 编委审稿 远程编辑

计算机工程 ›› 2025, Vol. 51 ›› Issue (7): 12-30. doi: 10.19678/j.issn.1000-3428.0069698

• 热点与综述 • 上一篇    下一篇

无监督行人重识别研究综述

田青1,2,3,*(), 王斌1, 周子枭1   

  1. 1. 南京信息工程大学软件学院, 江苏 南京 210044
    2. 南京信息工程大学无锡研究院, 江苏 无锡 214000
    3. 南京大学计算机软件新技术国家重点实验室, 江苏 南京 210023
  • 收稿日期:2024-04-07 出版日期:2025-07-15 发布日期:2024-08-07
  • 通讯作者: 田青
  • 基金资助:
    国家自然科学基金(62176128); 江苏省自然科学基金(BK20231143); 南京大学计算机软件新技术国家重点实验室开放课题(KFKT2022B06); 中央高校基本科研业务费专项资金(NJ2022028); 江苏省“青蓝工程”人才项目

Survey on Unsupervised Person Re-Identification

TIAN Qing1,2,3,*(), WANG Bin1, ZHOU Zixiao1   

  1. 1. School of Software, Nanjing University of Information Science and Technology, Nanjing 210044, Jiangsu, China
    2. Wuxi Institute of Technology, Nanjing University of Information Science and Technology, Wuxi 214000, Jiangsu, China
    3. State Key Laboratory for Novel Software Technology, Nanjing University, Nanjing 210023, Jiangsu, China
  • Received:2024-04-07 Online:2025-07-15 Published:2024-08-07
  • Contact: TIAN Qing

摘要:

从多个不同的摄像机中检索出一个特定的行人是行人重识别(ReID)的主要任务。随着深度神经网络的发展和智能视频监控需求的增加,行人ReID逐渐受到研究人员的关注。现有的行人ReID方法大多利用带有标签的数据集进行有监督训练,但该方式的数据标注成本高昂,使得有监督行人ReID难以扩展到大型的未标记数据集场景。无监督行人ReID可以有效地改善行人ReID模型的扩展问题,更符合现实场景的应用,并逐渐成为研究热点。尽管已有行人ReID的相关综述,但它们主要聚焦于有监督学习领域的方法和应用。为此,对现有无监督行人ReID研究工作进行系统归纳、分析和总结,以便为该领域研究人员提供参考。首先,全面回顾无监督场景的行人ReID方法,根据模型训练是否使用源域监督信息,将无监督行人ReID研究划分为无监督域适应方法和完全无监督方法,并对这2类方法进行分析和总结。然后,介绍和归纳总结行人ReID研究相关的经典数据集,并展示和讨论不同行人ReID方法在这些数据集上的性能和优劣。最后,指出当前无监督行人ReID研究所面临的问题,并提出未来发展方向。

关键词: 行人重识别, 深度神经网络, 智能视频监控, 无监督域适应方法, 完全无监督方法

Abstract:

The primary task of person Re-IDentification (ReID) is to identify and track a specific pedestrian across multiple non-overlapping cameras. With the development of deep neural networks and owing to the increasing demand for intelligent video surveillance, ReID has gradually attracted research attention. Most existing ReID methods primarily adopt labeled data for supervised training; however, the high annotation cost makes the scaling supervised ReID to large unlabeled datasets challenging. The paradigm of unsupervised ReID can significantly alleviate such issues. This can improve its applicability to real-life scenarios, enhancing its research potential. Although several ReID surveys have been published, they have primarily focused on supervised methods and their applications. This survey systematically reviews, analyzes, and summarizes existing ReID studies to provide a reference for researchers in this field. First, the ReID methods are comprehensively reviewed in an unsupervised setting. Based on the availability of source domain labels, the unsupervised ReID methods are categorized into unsupervised domain adaptation methods and fully unsupervised methods. Additionally, their merits and drawbacks are discussed. Subsequently, the benchmark datasets widely evaluated in ReID research are summarized, and the performance of different ReID methods on these datasets is compared. Finally, the current challenges in this field are discussed and potential future directions are proposed.

Key words: person Re-Identification (ReID), deep neural network, intelligent video surveillance, unsupervised domain adaptation method, fully unsupervised method