<![CDATA[The adoption of online tests has introduced significant challenges in maintaining academic integrity, particularly in real-time detection of cheating behaviors. This study proposes an intelligent proctoring system that automatically detects suspicious participant behavior during an online test by integrating image processing and computer vision techniques. The system integrates a YOLOv8s model based on the YOLO neural network algorithm to localize and classify facial states and suspicious objects in each video frame. This detection layer is complemented by an InsightFace face recognition module, which extracts deep facial embedding features and performs similarity matching against a registered reference image to continuously verify the identity of the participant and detect attempts at impersonation. In parallel, the GazeTracking module analyzes eye landmarks and pupil dynamics to monitor eye behavior, including blinking and significant gaze deviation, providing additional behavioral cues related to attention and potential cheating. The system consists of three detection layers: (1) YOLOv8s for object and behavior detection, (2) InsightFace for identity verification, and (3) GazeTracking for eye behavior analysis. Together, these components form a synchronized computer vision module that performs real-time analysis from live video streams, allowing the system to classify behavioral states such as abnormal head orientation, multiple faces, foreign objects, no face detected, identity mismatch, and eye closure. The experimental results show that the YOLOv8s model achieves an mAP@50 of 0.9918, a precision of 0.9856, and a recall of 0.9903 on the validation set while maintaining real-time performance at an average of 10 frames per second. The findings demonstrate that deep learning-based visual monitoring can effectively support automated online exam supervision, offering a viable computer vision-based proctoring approach.]]>
Copyrights © 2026
