Validation of a markerless motion capture system (OpenCap) for jump-landing kinematic in the landing error scoring system (LESS)

Abstract

OpenCap, a markerless motion capture system, offers a practical alternative for movement assessment; however, additional validation is needed to confirm its applicability. This study evaluated how closely lower extremity kinematics measured by OpenCap matched those obtained from a conventional marker-based system during jump-landing tasks. We also examined whether OpenCap's accuracy varied between normal and simulated abnormal jump-landing conditions. Kinematic data were simultaneously recorded by both systems while 21 male participants performed jump-landing tasks under both normal and simulated abnormal conditions. Waveform similarity was assessed using the coefficient of multiple correlation (CMC), and error was quantified using root mean square error (RMSE). Wilcoxon signed-rank tests with Bonferroni correction were used to compare RMSE across conditions. OpenCap showed moderate to strong waveform similarity with the marker-based system during jump-landing tasks, with a grand mean CMC of 0.78 across all joint angles and conditions. Agreement was highest in the sagittal plane (0.84–0.98) and lowest in the transverse plane (0.51–0.60). A grand mean RMSE of 6.62° was observed across all joint angles and movement conditions. Wilcoxon signed-rank tests revealed significant RMSE differences only for ankle flexion (p < 0.001) and inversion (p < 0.001) under the simulated impaired postural control and restricted ankle motion conditions, respectively. These findings suggest the potential of OpenCap as a practical tool for assessing lower extremity kinematics in clinical and sports settings, especially for sagittal plane movements. While certain limitations remain in capturing movement in the transverse and frontal planes, continued algorithmic advancement may enhance its accuracy and broaden its applicability.