Real-Time Classroom Behaviour Recognition based YOLOv8 for Enhancing Higher Vocational Education Quality

Identification and analysis of higher vocational college students' learning behavior in the classroom are an essential part of enhancing educational quality and learning outcomes. This paper suggests a deep learning-based method based on YOLOv8 to identify and analyze higher vocational college students' classroom learning behaviors to improve educational quality and outcomes. Conventional methods based on manual observation are time-consuming and susceptible to bias. YOLOv8, one of the cutting-edge object detection models, provides real-time detection, effective feature extraction, and rapid processing to classify automatically behaviors like attentiveness, engagement, interaction, and distractions. Having been trained on a classroom images and videos dataset, the model solves low accuracy and scalability problems of previous systems. Comparative studies show remarkable enhancements, with 99.3% accuracy, a 98.1% F1-score, and processing speed compared to conventional methods. The capability of the system to process vast amounts of data and offer real-time responses makes it efficient and scalable for ongoing behavior monitoring in dynamic classroom settings. The innovative approach allows teachers to make informed decisions based on data, promoting increased student engagement and customized teaching practices.