Enhancing AI Proctoring System Using Various ML Models

Online exams have been the most popular in all educational sectors during the past year because of COVID-19. But proctoring methods are proving to be quite problematic for the universities. Artificial intelligence (AI)-based proctoring technologies automate the monitoring process in this project, doing away with the necessity for human superintendence. When several tests need to be given at the same time, this can increase efficiency and scalability. The project uses ArcFace for accurate facial recognition and YOLO for real-time object detection to provide a complete and effective online test monitoring solution. This multi-modal strategy improves the accuracy of identity verification, identifies things that are not permitted, and adjusts to new security risks. The system provides a strong solution for preserving the integrity of online assessments, prioritizing user privacy, integrating easily with learning management systems, and surpassing current models in terms of speed, scalability, and overall effectiveness. Exam proctors and exam candidate will find it easy to use the user-friendly interfaces of the proctoring system. It will allow the proctor to intervene if needed and provide the test-taker with immediate feedback, including alerts for questionable behavior. The suggested model outperforms other current models by 2.26% when tested using a variety of assessment matrices.