摘要
针对基于监控视频的人体异常行为识别问题,提出了基于主题隐马尔科夫模型的人体异常行为识别方法,即通过无任何人工标注的视频训练集自动学习人体行为模型,并能够应用学到的人体行为模型实时检测异常行为和识别正常行为。这一方法主要围绕"低层视频表示-中层语义行为建模-高层语义分类"3个方面进行:1)基于时-空间兴趣点构建了一种紧凑的和有效的视频表示方法。2)提出一种新颖的语义主题模型(Topic Model,TM)——主题隐马尔科夫模型(Topic Hidden Markov Model,THMM),它能够自然地分组视频中检测到的人体行为。主题隐马尔科夫模型基于已有的马尔科夫模型和主题模型构造,不但聚类运动词汇成简单动作,而且聚类简单动作成全局行为,同时建模了行为时间上的相关性。THMM是一个4层贝叶斯主题模型,它将视频序列建模为行为的马尔科夫链,同时行为是视频序列中某些视频剪辑(Clip)的概率分布;将视频剪辑建模为动作的随机组合,同时动作是视频剪辑中运动词汇的概率分布。克服了传统隐马尔科夫模型和主题模型在人体复杂行为建模过程中精度、鲁棒性和计算效率上的不足。3)提出运行时累积的异常性测度及其在线异常行为检测方法和基于在线似然比检验(Likelihood Ratio Test,LRT)的实时正常行为分类方法,从而克服了实时行为识别过程中由于缺乏充分的视觉证据而引发的行为类型歧义,能完较好地完成监控场景中实时异常行为检测和在线正常行为识别的任务。取自实际监控场景的实验数据集上的实验结果证明了本方法的有效性。
This paper aimed to address the problem of modeling human behavior patterns captured in surveillance videos for the application of online normal behavior recognition and anomaly detection. From the perspective of cognitive psychology, a novel method was developed for automatic behavior modeling and online anomaly detection without the need for manual labeling of the training data set. The work has been done with the hierarchical structure,following the routine of Video Representation-Semantic Behavior （Topic） Model-Behavior Classification. 1） A compact and effective behavior representation method is developed based on spatial-temporal interest point detection. 2） The natural grouping of behavior patterns is determined through a novel clustering algorithm, topic hidden Markov model （THMM） built up- on the existing hidden Markov model （HMM） and latent Dirichlet allocation （LDA）, which overcomes the current limi- tations in accuracy, robustness, and computational efficiency. The new model is a four-level hierarchical Bayesian model, in which each video is modeled as a Markov chain of behavior patterns where each behavior pattern is a distribution over some segments of the video. Each of these segments in the video can be modeled as a mixture of actions where each ac- tion is a distribution over spatial-temporal words. 3） An online anomaly measure is introduced to detect abnormal be- havior, whereas normal behavior is recognized by runtime accumulative visual evidence using likelihood ratio test （LRT） method. Experimental results demonstrate the effectiveness and robustness of our approach using noisy and sparse data sets collected from a real surveillance scenario.
出处
《计算机科学》
CSCD
北大核心
2012年第3期251-255,275,共6页
Computer Science
基金
国家自然科学基金(60573139)资助
关键词
计算机视觉
语义主题模型
异常检测
运动词包
行为聚类
Computer vision, Topic model, Anomaly detection, Bag of motion word, Behavior clustering
