An Integrated AI System for Exam Proctoring Using YOLOv8, Face Recognition, and Noise Detection

Abstract

Online examinations have become increasingly 
prevalent, demanding robust proctoring solutions to 
ensure integrity and prevent unauthorized activities. 
This paper presents an integrated AI-based proctoring 
system that combines real-time face recognition, 
object detection using YOLOv8, and audio anomaly 
detection to enhance the reliability and fairness of 
online 
assessments. 
The proposed system 
authenticates candidates through webcam-based face 
recognition using facial encodings, preventing 
impersonation. 
Once verified, the system 
continuously monitors the candidate throughout the 
exam session.