Validation of 5G-Enabled Mobile Gait Assessment System in a Clinical Setting

This paper presents a non-wearable, vision-based human gait analysis system that enables accurate spatiotemporal gait assessment using a standard smartphone camera. The system employs a pretrained BlazePose model to extract 3D skeletal joint trajectories, from which clinically relevant gait parameters, including step length, gait speed and cadence, are derived. Unlike conventional gait analysis approaches that rely on expensive motion capture (MoCap) systems or wearable inertial sensors, the proposed framework offers a cost-effective and easily deployable alternative, eliminating the need for body-mounted devices. The system was validated on a cohort of 56 subjects, comprising 30 participants evaluated against a gold-standard MoCap system and 26 elderly patients assessed in a hospital setting. The results indicate that the proposed system achieves reliable gait measurement accuracy while maintaining practical feasibility for real-world clinical environments. This study demonstrates the potential of combining markerless computer vision with 5G connectivity to support objective gait assessment for remote monitoring.