Aberrant Behavior Detection Using Three-Dimensional Stereo Video
Jui‐Feng Yeh, Shu-Po Hsu, Kai-Siang You · 2024
Surveillance technology has become fundamental to the security of living spaces. Police agencies, manufacturers, and individuals have deployed many video monitoring equipment in various environments. In this study, we used stereoscopic image technology for malicious behavior detection. The system framework proposed in this study consists of active capture, visual focus retrieval, and event detection. We used a depth sensor for three-dimensional stereo video. We developed a method by using the three-dimensional stereo video to detect the behavior in the aberrant event. We experimented with collected datasets in this study, and an accuracy of 95.8% and an F1 Score of 79% were obtained in predicting aberrant behavior detection.