Detection of Violent Crowd Behavior Based on Statistical Characteristics of the Optical Flow
Violent crowd behavior detection is an important research subject in crowd surveillance.By observing that in crowded scenes,the variation of optical flows is large when violent behavior occurs.In this paper,we propose an optical flow based method to detect violent behaviors in crowd scenes.Our method considers the statistical characteristics of optical flow field and extracts a statistical characteristic of the optical flow(SCOF)descriptor from these characteristics to represent a short frame sequences.The SCOF descriptors are then classified as either non-violent or violent using linear Support Vector Machine.The experiments are conducted on Crowd ViolenceNon-violence Database and Hockey dataset.Experimental results show the effectiveness of our method,and it is competitive with the state-of-art methods.
component violence detection linear support vector machine SCOF descriptor
Jian-Feng Huang Shui-Li Chen
College of Mathematics and Computer Science,Fuzhou University Fuzhou,Fujian 350002,China School of Science,Jimei University Xiamen,Fujian 361021,China
国际会议
厦门
英文
574-578
2014-08-19（万方平台首次上网日期，不代表论文的发表时间）