Validation of a markerless motion capture system for centre of mass kinematic analysis

Abstract

In recent years, markerless optical systems for biomechanical movement analysis in sports, gait and balance assessments are being used as an alternative to conventional marker based measuring systems. This study compares the performance of the Zed 2i stereoscopic camera against a VICON system in a standing position under three conditions: quiet standing and two movements simulating disturbances in two directions, anteroposterior and mediolateral. This study originates from a collaborative project with a medical team that aims to objectively evaluate balance function in patients recovering from stroke. The displacement and velocities of the centre of mass were calculated and compared in two directions, x and y. A Bland–Altman analysis for non-parametric data, along with the coefficient of determination and mean square error, were used for statistical evaluation. The results demonstrate that the limits of agreement in both sway tasks were greater than those observed in static conditions. However, the coefficient of determination of the sway tasks indicates a significant degree of agreement between the two systems. In contrast, in the static condition, it appears that noise may have a greater influence on the signal than the centre of mass estimate, due to the limitation of the depth algorithm used to estimate the joint positions.