Analysis on AI Proctoring System Using Various ML Models

The increase in online examinations as a result of the COVID-19 pandemic has emphasized the need for effective proctoring methods in educational institutions. This work presents an automated AI-based proctoring system that tackles these challenges by eliminating the requirement for human supervision. Utilizing YOLO for real-time object detection and FaceNet for accurate facial recognition, the system provides a comprehensive solution for online exam monitoring. Utilizing a multi-modal approach significantly improves the accuracy of identity verification, enables the detection of unauthorized objects, and ensures adaptability to emerging security threats. Performance evaluation shows that our models outperform existing ones in terms of speed, scalability, and overall effectiveness. Significantly, the proposed model demonstrates a 2.28% increase in accuracy. The software offers intuitive interfaces for both test-takers and exam proctors, making it easy to use and providing instant feedback. Proctors have the ability to intervene if needed, and the system will alert them if any suspicious behavior is detected. With meticulous testing and evaluation, this proctoring system provides a strong solution for preserving the integrity of online assessments in today's educational environment.