![]() | Only 14 pages are availabe for public view |
Abstract When administering an e-exam, human observation stands out as a prevalent approach to prevent students from engaging in any unforeseen actions. This method involves placing the student in a designated laboratory or hall at the university and closely monitoring them both visually and audibly throughout the entire examination duration. However, this method is costly and labor-intensive. Moreover, students’ behavior is not entirely under control, and some of them attempt to act dishonestly to gain a competitive advantage. This makes monitoring students uncomfortable and challenging. This work introduces a novel approach to monitor students during online exams automatically based on face and voice recognition using a deep learning method. The approach tracks a student’s pupils, lips, and facial movements using a camera on their computer, keeping track of their behavior throughout the test and intervening if anything unusual happens. The suggested approach initially validates the students’ identities prior to the commencement of the exam and continues to be employed throughout the test duration to prevent any instances of impersonation. Furthermore, this approach effectively detects and addresses any unexpected or suspicious behavior that may arise during the examination |