The Influence of Occlusion on Markerless Motion Capture using Opencap
C.E. Dekker (TU Delft - Mechanical Engineering)
A.H.A. Stienen – Mentor (TU Delft - Mechanical Engineering)
J.K. Moore – Graduation committee member (TU Delft - Mechanical Engineering)
More Info
expand_more
Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons.
Abstract
Markerless motion capture systems such as OpenCap provide a low-cost and accessible alternative to traditional laboratory-based motion capture. This makes them promising for in-the-field movement analysis, where ideal recording conditions cannot always be achieved. However, reduced visibility of body segments due to self-occlusion, object occlusion, or unfavorable camera angles may affect the reliability of the estimated kinematic measurements. This study investigated the influence of occlusion on OpenCap kinematic output using a single-subject repeated-measures design. Six standardized movement tasks were recorded under three visual conditions: an optimal reference condition, a self-occlusion condition, and an object-occlusion condition. The exported OpenCap joint-angle trajectories were processed and compared using RMSE with respect to the reference condition. The results showed that lower-extremity sagittal-plane tasks remained relatively stable across conditions, while upper-extremity, rotational, and supine tasks showed larger deviations and more processing errors. These findings suggest that OpenCap can be useful for in-the-field movement analysis, but that task selection, camera placement, and body-segment visibility should be carefully considered when interpreting kinematic output.