An In-Browser Proctoring System Using YOLO-Based Object Detection and Gaze Analysis
Published in Journal of Artificial Intelligence and Capsule Networks, 2025
This paper proposes a web-based automated proctoring system designed to improve online exam integrity by detecting cheating behaviors in real time. It combines YOLOv8 object detection to recognize faces and restricted objects with gaze-based analysis to monitor eye movement and head orientation. A rule-based decision layer integrates these signals with browser activity patterns to identify suspicious behavior and generate alerts. The system is lightweight, works within a browser environment, and demonstrates reliable performance in monitoring remote examinations.
Recommended citation: Kaundinya, A. S., Adhikari, P., KC, M. B., & Shrestha, P. (2025). "An In-Browser Proctoring System Using YOLO-Based Object Detection and Gaze Analysis." *Journal of Artificial Intelligence and Capsule Networks*, 7(4), 388–412. https://doi.org/10.36548/jaicn.2025.4.005
Download Paper