Advanced Human Motion Detection and Precision Movement Measurement via Mobile Devices Using YOLOv8, R-CNN, and Augmented Reality
Oras F. Baker, Wei Li, Qing Yuan · 2024
Human locomotion detection is a significant area in machine learning, especially in computer vision. Methodologies have advanced with the rising computational power of mobile devices and large datasets, enabling mobile adoption of machine learning. This research employs object detection to identify human movement, using facial recognition and augmented reality for measurement via handheld devices. An iOS mobile application was developed, detecting human figures and using YOLOv8 and R-CNN methods for facial recognition to acquire bodily parameters. These parameters are combined with augmented reality metrics for final calculations. The system has two main components: a web service for facial recognition and a mobile application for movement detection and measurement. Rigorous testing, including unit and black-box testing, demonstrated outstanding results. The system detected human movement with over 95% precision and achieved a 98% identification rate in facial recognition. Augmented reality enabled superior motion quantification, surpassing conventional methods. With an intuitive interface, the system is ready for widespread deployment, revolutionising human motion analysis.