Exam Sentinel: A Real-Time Visual Intelligence System for Secure Online Examinations

1 J Raghunath, 2 PolikiKomala Sai, 3 Mondikalla Lakshmi, 4 Kothalaiah Gari Chandrashekar, 5 Peddapothula Hari Krishna, 6 Medikurthi Chandra Sekhar

doi:10.62643/

Authors

1 J Raghunath, 2 PolikiKomala Sai, 3 Mondikalla Lakshmi, 4 Kothalaiah Gari Chandrashekar, 5 Peddapothula Hari Krishna, 6 Medikurthi Chandra Sekhar Author

DOI:

https://doi.org/10.62643/

Keywords:

Face Recognition, YOLO Algorithm,Open CV, Head Pose Estimation, Online Proctoring, Artificial Intelligence

Abstract

The rapid growth of online education has significantly increased the demand for secure and reliable remote examination systems. However, ensuring academic integrity in virtual environments remains a major challenge due to the absence of physical supervision. This paper presents Exam Sentinel, a real-time AIbased proctoring system designed to monitor candidates during online examinations using computer vision techniques. The system integrates face recognition for identity verification, head pose estimation for behavioral analysis, and object detection using YOLO for identifying prohibited items such as mobile phones. A Flask-based web application enables real-time video streaming and interaction. The system automatically generates alerts and terminates sessions upon detecting suspicious activities. Experimental results demonstrate that the proposed system enhances accuracy, reduces human dependency, and provides a scalable solution for secure online examinations.