A Supervisory System to Detect Suspicious Behavior in Online Testing System

Computer vision is one of the most emerging fields in Information Technology. We propose a unique application of computer vision in this paper. In the proposed method, we have detected the suspicious behaviour of a student in an online or computer based testing system. The work is valuable especially for test centers, schools, colleges and universities. For the development of such system, we present a three stage process to detect mysterious actions. The experiment has shown the impressive results for warning generation and activity registration on successful detection. The most occurring inexplicable actions are listed: arm movement and flipping, head shaking dual sides, wobbling around premises, revolving side-by-side or back to back.