ABSTRACT

The use of IoT devices for behavior monitoring has become increasingly common across various industries. However, observing student behavior in the classroom still presents many challenges in teaching practice. Teachers’ direct observation alone cannot capture all the details of a dynamic classroom or monitor each student’s performance consistently. As a result, classroom observation urgently requires new methods and technological support to help teachers assess student behavior and make real-time adjustments to improve teaching efficiency. To address this, there is a critical need for a behavior detection management platform. This study aims to combine target detection with behavior classification to develop a new behavior analysis platform. Through in-depth analysis of multi-dimensional data, including IT expert interviews and user feedback, the platform will provide teachers and administrators with accurate and real-time information. This study analyzed student classroom behavior detection algorithm based on IoT devices and realized behavior detection by combining target detection with behavior recognition. Additionally, both teachers and students can review the behavior analysis results after class, which help guide students to listen more attentively. This not only enhances the effectiveness of the teachers’ instruction but also improves students’ learning efficiency.

Keywords: Action Recognition, Behavior Analysis System, Student Classroom Behavior, YOLO_V3
https://doi.org/10.57180/nlyi7719