Online Exam Proctoring System Using Machine Learning

A web-based exam monitoring system enhances the integrity and fairness of online examinations by simulating the role of an invigilator through advanced AI technologies, ensuring secure, fair, and regulated conditions. In a virtual setting where direct supervision is not possible, this system serves as an automated proctor, detecting and flagging suspicious behavior that may indicate potential cheating. One of its key features is Face Detection and Recognition, which verifies the student’s identity to eliminate the risk of impersonation. Additionally, People Counting enhances security by scanning the environment for additional faces, preventing unauthorized assistance. The system also employs Head and Eye Tracking, which monitors the student’s gaze and head movements to ensure their attention remains on the exam screen. Any prolonged distraction is logged and flagged for review. To counter fraudulent attempts, Face Spoofing Detection uses liveness detection techniques to differentiate between a real person and fake representations like photographs, videos, or 3D models, ensuring only authentic users take the exam. Furthermore, Object Recognition identifies and flags unauthorized items such as mobile phones, calculators, and notes, which are typically restricted in exam settings. By integrating object detection algorithms, the system can automatically recognize such items, alert proctors, and even pause the exam if necessary. Together, these features create a comprehensive, AI-driven monitoring solution that closely mimics in-person invigilation, making online exams more secure and fair. This application not only upholds academic integrity but also ensures students are evaluated purely on their knowledge and efforts, providing a virtual invigilation process that is reliable, effective, and as close as possible to traditional exam supervision.