• 中国计算机学会会刊
  • 中国科技核心期刊
  • 中文核心期刊

计算机工程与科学 ›› 2023, Vol. 45 ›› Issue (09): 1578-1592.

• 图形与图像 • 上一篇    下一篇

基于孪生网络的目标跟踪算法综述

马玉民1,2,3,钱育蓉1,2,3,周伟航1,2,3,公维军1,2,3,帕力旦·吐尔逊4   

  1. (1.新疆大学软件学院,新疆 乌鲁木齐830000;2.新疆维吾尔自治区信号检测处理重点实验室,新疆 乌鲁木齐 830046;
    3.新疆大学软件工程重点实验室,新疆 乌鲁木齐 830000;4.新疆师范大学,新疆 乌鲁木齐 830000)
  • 收稿日期:2022-01-07 修回日期:2022-06-06 接受日期:2023-09-25 出版日期:2023-09-25 发布日期:2023-09-12
  • 基金资助:
    国家自然科学基金(61966035,U1803261)

A survey of target tracking algorithms based on Siamese network

MA Yu-min1,2,3,QIAN Yu-rong1,2,3,ZHOU Wei-hang1,2,3,GONG Wei-jun1,2,3,Palladium Turson4   

  1. (1.School of Software,Xinjiang University,Urumqi 830000;
    2.Key Laboratory of Signal Detection & Processing in Xinjiang Autonomous Region,Urumqi 830046;
    3.Key Laboratory of Software Engineering,Xinjiang University,Urumqi 830000;
    4.Xinjiang Normal University,Urumqi 830000,China)
  • Received:2022-01-07 Revised:2022-06-06 Accepted:2023-09-25 Online:2023-09-25 Published:2023-09-12

摘要: 孪生网络是由2个或多个人工神经网络建立的耦合框架,因其将回归问题转换为相似度匹配问题,备受计算机视觉领域的研究人员关注。随着深度学习理论的快速发展,目标跟踪技术在生活中得到了广泛的应用。基于孪生网络的目标跟踪算法以其相对优越的准确率和实时性逐渐代替了传统的目标跟踪算法,成为目标跟踪的主流算法。首先,介绍了目标跟踪任务面对的挑战和传统方法;然后,介绍了孪生网络的基础结构及其发展,汇总了近年来基于孪生网络的目标跟踪算法与相应设计原理;另外,介绍多个用于目标跟踪测试的主流数据集,并基于这些数据集对比了基于孪生网络的目标跟踪算法的性能;最后,提出基于孪生网络目标跟踪算法目前存在的问题及对未来的展望。

关键词: 孪生网络, 人工智能, 计算机视觉, 视觉目标跟踪

Abstract: Siamese network is a coupled framework established by two or more artificial neural networks, which turns the regression problem into a similarity matching problem and has attracted much attention from researchers in the computer vision field. With the rapid development of deep learning theory, target tracking technology has been widely used in daily life. Siamese network-based target tracking algorithms have gradually replaced traditional target tracking algorithms with their relatively superior accuracy and real-time performance, becoming the mainstream algorithm for target tracking. Firstly, the challenges and traditional methods faced by target tracking tasks are introduced. Then, the basic structure and development of Siamese network are introduced, and the design principles of Siamese network-based target tracking algorithms in recent years are summarized. In addition, the performance of Siamese network-based target tracking algorithms is compared using multiple mainstream datasets for target tracking testing. Finally, the problems and prospects of Siamese network-based target tracking algorithms are proposed.

Key words: Siamese network, artificial intelligence, computer vision, visual target tracking