This project presents an automated exam monitoring system that leverages machine learning and computer vision to enhance academic integrity in both online and offline examinations. The system integrates facial recognition, gaze tracking, and hand movement detection to identify and prevent cheating behaviors in real time.

Using InsightFace for accurate face recognition and MediaPipe for behavioral analysis, the system ensures that only registered students are present during exams while continuously monitoring their actions. Any suspicious behavior, such as abnormal eye movement or unusual hand gestures, is instantly detected and flagged for review. The model was trained on diverse, well-labelled datasets collected under varying environmental conditions to improve robustness and reliability.

The system also automates attendance marking and generates real-time alerts, significantly reducing the need for manual invigilation. Experimental results show high accuracy in student identification and behavior tracking, even in crowded settings, with minimal false positives. Overall, the system demonstrates how artificial intelligence can enhance fairness, efficiency, and scalability in modern assessment environments.

7_Electronic Exam monitoring using Machine Learning and Computer Vision