AutoProc - AI Based Automated Exam Proctoring System with Test Score Analysis

Exam Proctoring or invigilation is no longer restricted to a particular time, place or infrastructure. The COVID pandemic caused a fast change in how assessments or evaluations are performed. Proctoring systems currently in use are bulky, cumbersome, and expensive and cannot be used freely in the education sector. In this scenario, an artificial intelligence-assisted proctoring system is needed to monitor online tests that would be most convenient to academic institutions and students. In our work, we developed an AI that enables students to be proctored remotely using the webcam and microphone of their laptops. The proposed framework is separated into three main functional groups - Audio, Visual, and Safe Browser. The remarkable feature of our proctoring system is Test score analysis, which summarises students’ malpractices during the entire exam duration. This particular metric gives a quick output of every student’s activity, and the professor can gauge activity through this metric rather than checking the student log for the entire class. It reports multiple face detection with a picture of it, switching of tab to another application for a specific duration or noise detection with its recording. Our efficient approach satisfies the need for an online exam as a trustworthy, helpful tool and supports the expanding online and distance learning sectors/ industries.