The Implementation of Artificial Intelligence Based Body Tracking for the Assessment of Orientation and Mobility Skills in Visual Impaired Individuals

Visually impaired individuals face immense challenges during navigation. A comprehensive understanding of orientation and mobility (O &M) skills would shed light on the difficulties they face and expand the scope of therapeutic interventions. Here, we aim to complement the existing metrics of performance measurement in O &M assessments by presenting the methodological implementation of body tracking using artificial intelligence (AI) tools. We video-recorded a participant navigating a line marked on the floor within an indoor environment designed to replicate real-world conditions. Utilizing the YOLOv8 neural network for human detection and tracking, we transformed raw data to determine participant locations. To assess the efficacy of our body-tracking system, we examined deviations from the designated route. We found minimal disparities and a strong positive correlation between the marked path and the tracked route. Hence, our YOLOv8-based body-tracking system accurately captures participant locations. Furthermore, we provided two practical applications of the body tracking data: first, the real-time estimation of participant speed throughout the task, and second, the precise measurement of the total path length covered. Here, we presented a body tracking implementation which can capture the precise locations of participants in complex environments. This data analysis offers insights into the dynamic interactions between individuals with vision impairments and their surroundings, complementing existing measures of Orientation and Mobility (O &M) performance with additional outcome metrics.