Performance Evaluation of Automatic Suspicious Activity Detection Method

The video monitoring is hectic task when large dataset of surveillance video is to be monitored. The errors involved in monitoring task and suspicious activity detection may lead to missed detection performance. The automation of suspicious activity detection is the need of the time. In this paper, the performance evaluation of suspicious activity detection of video data is shown. The proposed method of combination of GLCM, harries corner detection, speeded up robust features, shows average 96% percent accuracy of suspicious activity detection on self-designed dataset.