IntroductionInvestigating human motion with expensive and accurate optical marker based systems has been the state of the art since long ago. However, markerless low-cost systems have always been a desideratum in the field of biomechanics and sports science. Due to increasing computer chip power and the corresponding progress in image processing techniques the realization of such a system has become feasible. With the advent of the Microsoft Kinect sensor in 2011 a flexible low-cost tool has entered the computer game market that enables markerless tracking of human motion. At first sight the Kinect provides an amazing accuracy. The goal of the present work is to quantitatively investigate the tracking accuracy of the Kinect sensor by studying the human gait cycle on a treadmill. The Kinect results are compared with data stemming from a VICON system which is regarded as a kind of gold standard in terms of spatial resolution. Subsequently, a post processing step is applied to the Kinect data using anthropometric data as a priori information in order to enhance Kinect's tracking results.
MethodsThe Kinect sensor provides a skeleton model whose measured joint positions and segment lengths only roughly match those of the VICON system. A first inspection of the measurements reveals that segment lengths determined by the Kinect system are considerably varying in time (up to +/-20% at the most). In order to freeze the length of all segments the Levenberg-Marquardt-algorithm (LMA) is applied to the Kinect data. LMA manipulates the Kinect joint positions such that the segment lengths of the Kinect skeleton to be become identical with pre-defined lengths. These segment length of the original human being can easily be determined by previous manual measurements. In our special case we use the data provided by the VICON system.
ResultsThe comparison between both systems shows that the Kinect sensor is currently not able to provide anthropometrically reliable segment lengths. However, this failure of the Kinect system is healed with our above described optimization procedure. When looking at the angles relevant for the human gait cycle the Kinect sensor provides astonishing accuracy. The curves describing the time dependency of those angles are very similar when comparing the data from both systems (except for a small scaling factor). The shapes of these curves are only marginally affected by our optimization process.
ConclusionOur work shows that the Kinect sensor is already capable of describing the motion of walking and running with acceptable accuracy in the frontal plane. Additionally it is proved that a posteriori optimization using a priori available knowledge can further improve the tracking of the anatomical landmarks. That does not mean the Kinect sensor is now ready for use in highly sophisticated scientific applications, but it can already be applied for basic research in simple clinical test cases. It is expected that regarding both accuracy and measurement frequency the near future will see significant improvemen...
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.