Abstract
BACKGROUND: Gait analysis is traditionally conducted using marker-based methods yet markerless motion capture is emerging as an alternative. Initial studies have begun to evaluate the reliability of markerless motion capture yet the evaluation of different clothing conditions across sessions and complete evaluation of the lower limb and pelvis reliability have yet to be considered. The aim of this study was to evaluate the inter-trial, inter-session and inter-session-clothing variation and root mean square differences between tight- or loose-fitting clothing during walking. METHOD: Twenty-two healthy adult participants walked along an indoor walkway whilst eight video cameras recorded their gait in either tight- or loose-fitting clothing. A commercial markerless motion capture system (Theia3D) provided gait kinematics for evaluation. RESULTS: Reliability results showed average inter-trial variation of <2°, inter-session variation of <3° and inter-session-clothing variation <3.5°. Root mean square differences (RMSD) between clothing conditions were <2°. DISCUSSION: Pelvis variations were smaller than those at the hip, knee and ankle. Our results showed smaller variation than in previous studies which may be due to updates to software. The demonstration of the reliability of markerless motion capture for gait analysis in healthy adults should prompt further evaluation in clinical conditions and reconsideration of multi-assessor marker-based gait analysis protocols, where variation is highest.