Lower Extremity Function Evaluation System at the Workplace Using Azure Kinect

In this paper, we developed a quantitative and detailed evaluation system for the single-leg standing test and clarified the relationship between evaluation results and lower extremity function. The proposed method uses the Azure Kinect to estimate joint coordinates, calculate joint angles, and represent standing motion as a state transition model. The center of mass was estimated and used as an evaluation index. In the evaluation experiment, 41 subjects were measured, and it was confirmed that the proposed system could calculate the same evaluation values as an athletic trainer with high accuracy. It was also confirmed that the shape of the radar chart consisting of the 14 indicator groups was different for each subject. Furthermore, the results were compared with the physical fitness test data, and it was found that these results were related to three lower extremity functions: lower-extremity muscle strength, quadriceps muscle and hip flexibility, and hamstring flexibility. By reconstructing the radar chart with only the most relevant indicators, we were able to provide more specific feedback on lower extremity function. Cluster analysis suggests the possibility of detecting clusters with similar physical functions. The proposed system is capable of quantitative and detailed evaluations, even in the absence of skilled evaluators.