Image Recognition and Extraction of Students' Human Motion Features Based on Graph Neural Network

In order to improve students' overall subhealth behavior, teenagers' physical health problems have attracted more and more attention. The state clearly requires students to increase the number and frequency of exercise in school. In order to study the physical changes in the process of students' sports and the impact on their health caused by a sports injury, a student human motion feature image recognition based on a graph neural network is proposed in this paper. This paper combines image recognition technology with graphic neural network management and uses image recognition technology to detect and track targets. It also analyzes the health changes of students in sports and the influencing factors of physical subhealth in classroom learning. The results show that image recognition technology can accurately analyze the process of cervical spine injury and sports injury in students' classroom activities. It provides accurate experimental data for analyzing students' physical health and effective suggestions for promoting students' healthy development. Compared with the traditional image recognition and analysis results, the advantage of using a graph neural network to manage the detection and tracking results is that a graph neural network is used to manage the detection and tracking results, and the visual expression of students' physical health test data is completed.