BACKGROUND: In the field of human movement sciences, a marker-less motion capture system that is user-friendly and sufficiently accurate to measure human performance is needed. The feasibility study aims to validate a three-dimensional (3D) marker-less motion capture technique, while using a professional grade motion analysis system as the reference. METHODS: The marker based camera systems were positioned on the perimeter of the workspace. Three iOS cameras coupled with the marker-less system were arranged according to the system’s recommendations. A test subject performed 6 vertical jumps on 2 different flooring systems, and joint kinematics were measured using both marker-based optical motion capture and the open-source marker-less motion capture. The differences in corresponding joint positions, estimated from the three different camera systems throughout the analysis, were presented as means and standard deviations. The 3D model reconstruction accuracy between systems was analyzed using Pearson’s correlation coefficient, along with a Bland-Altman agreement analysis. Reliability of accuracy across trials was accomplished by comparing means using an analysis of variance technique. RESULTS: During landing, maximum knee range of motion (ROM) from the reference source (M = 99.1 degrees, SD = 0.52), the ASC system (M = 99.8 degrees, SD = 0.91), and Open-Cap (M = 100.2 degrees, SD = 1.05) were compared, where p = .01. The relationship between the reference system and the marker-based ASC system (r = .98, p = .001) and (r = .97, p = .001) between the reference system and the open-source marker-less system (r = .97, p = .001) was validated for accuracy. CONCLUSIONS: Experimental camera system accuracy was based on the output measures compared to the output measures of the reference system. The results demonstrated that, quantitatively, 3D reconstruction estimation using marker-less motion capture correctly reproduced the movements of the participant. The studies preliminary findings demonstrate that the open-source marker-less motion capture can be used for human movement science with an error of less than 1.5 degrees, and the budget friendly ASC system demonstrated less than 1 degree of error. Future work in this area should investigate the validity of marker-less 3D motion analysis in real-world settings.

This document is currently not available here.