计算机科学与探索 ›› 2022, Vol. 16 ›› Issue (3): 529-540.DOI: 10.3778/j.issn.1673-9418.2106117

• 综述·探索 • 上一篇    下一篇

视频异常检测技术研究进展

邬开俊1, 黄涛1,+(), 王迪聪1,2, 白晨帅1, 陶小苗1   

  1. 1.兰州交通大学 电子与信息工程学院,兰州 730070
    2.天津大学 智能与计算学部,天津 300350
  • 收稿日期:2021-06-21 修回日期:2021-08-18 出版日期:2022-03-01 发布日期:2021-08-19
  • 通讯作者: + E-mail: 1479987020@qq.com
  • 作者简介:邬开俊(1978—),男,山东莒南人,博士,教授,博士生导师,主要研究方向为视频检测与神经元的非线性动力学。
    黄涛(1996—),男,重庆彭水人,硕士研究生,CCF 学生会员,主要研究方向为视频的异常检测。
    王迪聪(1992—),男,甘肃张掖人,博士研究生,CCF 学生会员,主要研究方向为视频的目标检测、视频的异常检测。
    白晨帅(1998—),男,陕西渭南人,硕士研究生,CCF学生会员,主要研究方向为视频的目标检测。
    陶小苗(1985—),女,山西运城人,博士研究生,讲师,主要研究方向为视频的目标检测和追踪。
  • 基金资助:
    国家自然科学基金(61966022)

Research Progress of Video Anomaly Detection Technology

WU Kaijun1, HUANG Tao1,+(), WANG Dicong1,2, BAI Chenshuai1, TAO Xiaomiao1   

  1. 1. School of Electronic and Information Engineering, Lanzhou Jiaotong University, Lanzhou 730070, China
    2. College of Intelligence and Computing, Tianjin University, Tianjin 300350, China
  • Received:2021-06-21 Revised:2021-08-18 Online:2022-03-01 Published:2021-08-19
  • About author:WU Kaijun, born in 1978, Ph.D., professor, Ph.D. supervisor. His research interests include video detection and nonlinear dynamics of neurons.
    HUANG Tao, born in 1996, M.S. candidate, student member of CCF. His research interest is video anomaly detection.
    WANG Dicong, born in 1992, Ph.D. candidate, student member of CCF. His research interests include video object detection and video anomaly detection.
    BAI Chenshuai, born in 1998, M.S. candidate, student member of CCF. His research interest is video object detection.
    TAO Xiaomiao, born in 1985, Ph.D. candidate, lecturer. Her research interests include video object detection and tracking.
  • Supported by:
    National Natural Science Foundation of China(61966022)

摘要:

视频异常检测是指对偏离正常行为事件的检测识别,在监控视频中有着广泛的应用。对基于深度学习的视频异常检测算法进行了深入的调查研究和全面的梳理与总结。首先,对视频异常检测相关内容以及异常检测面临的挑战进行了分析;然后,从有监督、半监督和无监督三方面对视频异常检测的相关算法进行了介绍和分析。对三种不同场景下的算法进一步细化分类,将监督场景下的算法划分为二分类和多分类两种方式,将半监督场景下的算法划分为计算异常得分和聚类判别两种方式,将无监督场景下的算法划分为重构判别和预测判别两种方式,并且分析了不同技术的特点和优缺点。介绍了目前在视频异常检测领域常用的数据集,以及检测性能的评估标准,对目前主流的视频异常检测算法性能进行了对比分析。最后,对视频异常检测算法的未来研究方向进行了讨论和展望。

关键词: 深度学习, 异常检测, 有监督, 半监督, 无监督

Abstract:

Video anomaly detection refers to the detection and identification of events that deviate from normal behavior, which has a wide range of applications in surveillance video. In this paper, the video anomaly detection algorithm based on deep learning is investigated in depth and summarized comprehensively. Firstly, this paper analyzes the related content of video anomaly detection and the challenges faced by anomaly detection, then introduces and analyzes the related algorithms of video anomaly detection from three aspects: supervised, semi-supervised and unsupervised. The algorithms in three different scenarios are further refined and classified. The algorithms in the supervised scenario are divided into two types: binary classification and multi-classification. The algorithms in the semi-supervised scenario are divided into two types: calculating anomaly scores and clustering discrimination. The algorithms in the unsupervised scenario are divided into two types: reconstruction discrimination and prediction discrimination. The characteristics, advantages and disadvantages of different technologies are analyzed. The commonly used datasets in the field of video anomaly detection and the evaluation criteria of detection performance are introduced, and the performance of current mainstream video anomaly detection algorithms is compared and analyzed. Finally, the future research direction of video anomaly detection algorithm is discussed and prospected.

Key words: deep learning, anomaly detection, supervised, semi-supervised, unsupervised

中图分类号: