Deep learning enabled exercise monitoring system for sustainable online education of future teacher-trainers
Nurlan Omarov, Bakһytzһan Omarov, Quwanishbay Mamutov, Zhanibek Kissebayev, Almas Anarbayev, Adilbay Tastanov, Zhandos Yessirkepov · Frontiers in Education · 2024
Introduction In recent years, the importance of effective training methods for future physical education teacher-trainers has grown significantly, particularly in the context of online education. Methods This research introduces a pioneering Deep Learning Enabled Exercise Monitoring System aimed at enhancing the online education experience for these trainers. The study employs a skeleton-based approach utilizing the PoseNet model to monitor and provide real-time feedback on physical exercises such as pull-ups, push-ups, sit-ups, squats, and bicep workouts. The system achieves a remarkable accuracy rate of 99.8% by analyzing key skeletal points extracted from video frames, addressing the challenge of ensuring correct exercise execution without physical supervision–a common issue in remote learning environments. Results To validate the system’s effectiveness, data was collected through a series of controlled experiments involving various exercises. The system’s design focuses on low-resource requirements, making it accessible and sustainable for diverse educational contexts. Discussion The findings demonstrate the system’s potential to revolutionize online physical education by offering a balance of technological innovation and educational utility. This research not only elevates the quality of training for future educators but also contributes to the broader field of sustainable digital education technologies.