With the rapid development of national fitness and competitive sports, the prevention and rehabilitation management of sports-related injuries have become important topics in the field of sports health and wellness. To address the problems of traditional injury prediction methods, which rely on empirical judgment and statistical models and suffer from strong subjectivity, poor real-time performance, and insufficient accuracy, this study proposes an Artificial Intelligence Injury Prediction Model based on Convolutional Neural Network (AI-CNN-IPM). It aims to achieve high-precision and real-time injury prediction through multimodal physiological and sports behavior data. The model integrates time-series signals collected by wearable devices, including Heart Rate Variability (HRV), Electromyography (EMG), Accelerometer (ACC), and Gyroscope (GYRO). It uses CNN to automatically extract deep features and construct an injury prediction system. Experiments are conducted on a longitudinal dataset containing 120 athletes and fitness participants. The results show that the proposed model achieves an accuracy of 94.7% in predicting common sports injuries such as muscle strains and joint strains, which is better than the traditional Support Vector Machine (SVM) and Long Short-Term Memory (LSTM) models. This study provides a practical technical path for the intellectualization of sports health and wellness, and contributes to the realization of personalized training intervention and active health management.
Zhao et al. (Thu,) studied this question.