Complex Event Detection in Video Streams

Abstract

With the development of monitoring camera and the update of hardware, the processing of video data is becoming more and more. It is an important topic to detect the event in the video data and analyze its intrinsic relationship in order to form a semantic description. Complex event detection in video includes detection of video events, description of target features, and finding the concept of semantic. Based on the hypergraph theory, this paper put forward the characteristics of the construction of track and multi label hypergraph description of the moving target, and classification of video events. The experimental results show that, in comparison with other methods including ordinary graph based method and hypergraph based multi-label semi-supervised learning method, our method achieves better average precision and average recall when detecting complex events.