A Kidnapping Detection Scheme Using Frame-based Classification for Intelligent Video Surveillance

초록

The purpose of this study is to develop a kidnapping event detection scheme for intelligent video surveillance by frame-based classification which is able to assort each frame into a kidnapping or normally accompanying situation. In this study, for generating training data from videos, a semi-automatic video annotation tool named INHA-VAT is used. Also, we developed a frame-based event classifier using Bayesian network model to distinguish the frame of kidnapping situations from one of accompanying ones. When a video has more frames of kidnapping situation than the threshold ratio after two people meet in the video, the proposed scheme detects and notifies the occurrence of kidnapping event. To check the feasibility of the proposed scheme, we also performed the accuracy evaluation against test videos. According to the experiment results, the proposed scheme could detect kidnapping situations appropriately according to the threshold ratio.