Development and realization of home online teaching system based on video data analysis

With the outbreak of the COVID-19, most students can only take classes online at home. However, many students cannot consciously control their learning behaviors when they are in class at home, and some negative learning behaviors such as sleeping on the table or wandering have appeared. These negative learning behaviors greatly affect the effectiveness of learning. In order to deal with these problems, a online teaching system based on human action recognition is developed to assist teachers in real-time capturing the class status of classmates and improve the efficiency of home-based online teaching. Here, the system uses the OpenPose human body gesture recognition algorithm to obtain the key points of the student's body in class in front of the camera, and recognizes the student's class behavior through the analysis of the coordinates. The RTMP protocol is used to solve the problem of audio and video transmission during the live broadcast. We have conducted experiments to show the effectiveness of our system for analyzing the status of students and evaluating the online class.