Student Engagement and Cheating Detection Using Transfer Learning
摘要
This paper presents a dual-purpose system for Student Engagement Detection and Cheating Detection in educational environments. The first module classifies classroom activities—hand-raising, reading, and writing based on the SCB dataset using YOLOv8, YOLOv9, and YOLOv10 models. The second module detects cheating behavior during examinations using the Roboflow Exam Cheating dataset, categorizing actions as cheating or not-cheating. Our models aim to enhance classroom engagement and examination monitoring, providing real-time, automated solutions for improved educational integrity.