Non-Intrusive Experimental Behavior Monitoring System Based on Pre-Trained Model

Traditional electronic classroom software is widely used in teaching university computer labs. When the teacher lectures in front of the classroom, he/she cannot see all the students' learning behaviors. Thus the students like playing computer games, which undoubtedly increases the teacher's workload. To address above problems, we designed a non-intrusive student behavior monitoring system. This system collects a lot of images of students' screens on the teacher's computer through a data acquisition device. The images are processed by the image processing techniques and the pre-trained model. If non-learning behaviors are detected, the system sends the results to the teacher, who can stop students from doing these behaviors. The system not only reduces the teacher's teaching workload but also improves the studying atmosphere.