Real-time measurement of pelvis and trunk kinematics during treadmill locomotion using a low-cost depth-sensing camera: A concurrent validity study

There is currently no suitable kinematic system for a large-scale prospective trial assessing risk factors of musculoskeletal disorders. A practical kinematic system is described which involves the use of a single low-cost depth-sensing camera for the real-time measurement of 3-dimensional linear and angular pelvic and trunk range-of-movement (ROM). The method is based on the creation and processing of dynamic point clouds taken from the posterior surface of the pelvis and trunk. Nine healthy participants performed 3 trials of treadmill locomotion when walking at self-selected speed (3.6-5.6 km/h), running at 70% (10.9-14.0 km/h) and 90% of maximal speed (14.0-18.0 km/h). Stride-by-stride linear and angular ROM data were captured concurrently using the single depth-sensing camera running at 30Hz (KinectTM for Windows, Microsoft, USA) and a six-camera motion capture system at 100Hz (Vicon MX13, Vicon Motion Systems, United Kingdom). Within subject correlation coefficients between the practical and criterion method ranged from very large to nearly perfect (r = 0.87–1.00) for the linear ROM. Correlation coefficients for the angular ROM ranged from moderate to very large (r = 0.41–0.80). The limits of agreement between the two systems for linear movements was ≤ 9.9 mm at all velocities of gait and ≤ 4.6 degrees at all velocities of gait. The single camera system using depth-sensing technology is capable of capturing linear pelvic and trunk ROM during treadmill locomotion with reasonable precision when compared to the criterion method. Further improvements to the measurement of angles and validation across a wider population are recommended.