Validation of a Markerless Motion Capture System for Human Movement Analysis
Three-dimensional human movement analysis is a widely used tool in clinical and research biomechanics to provide comprehensive 3D representations and quantification of individuals’ movement patterns, particularly gait. Marker-based optical motion capture is the current ‘gold standard’ for performing human movement analyses; however, these systems have several inherent issues that affect the accuracy and reliability of their data and limit the environments in which data can be collected. Markerless motion capture is a quickly evolving technology that has the potential to eliminate many of the issues associated with marker-based motion capture. This research aims to validate a deep neural network-based markerless motion capture system, Theia3D, against a current field-accepted marker-based motion capture system for human gait. Three studies were undertaken towards the validation of this technology: (i) a comparison of time- and distance-based gait parameter measurements obtained simultaneously by the markerless and marker-based motion capture systems; (ii) a comparison of kinematic measurements obtained simultaneously by both systems; and (iii) a multi-session study of the repeatability of kinematic measurements obtained by the markerless motion capture system. The results of these studies indicate that this markerless motion capture system can measure time- and distance-based gait parameters and gait kinematics with sufficient accuracy for use in research and clinical applications, and the kinematic measurements were more reliable than those previously reported for ‘gold standard’ marker-based motion capture systems. These findings indicate the markerless motion capture system is sufficiently accurate and reliable for use in clinical and research biomechanics and can potentially reduce the limitations previously associated with performing human movement analysis.