An Automated Proctor Assistant in Online Exams Using Computer Vision

Abstract

Cheating or attempting to cheat in education has had the chance to increase in both number and complexity since the outbreak of the COVID-19 pandemic. With teaching and testing conducted online, learners can easily access prohibited materials without notice of a human proctor. Such problems raise the need for an automated intelligent system to help proctors in supervising test takers. Therefore, this work proposes a system that can automatically examine students’ behaviors through two main cameras. The first camera takes images of a student's frontal face and use them as input for a facial landmark model, detecting anomalies in student’s face movements. The second camera captures a student’s whole body and the surrounding environment, and by using a trained pose recognition model, it can efficiently classify student actions as suspicion or not. Results of this research show good remarks and can be applied in schools, universities experimentally in the future.