Estimation of the Accuracy of a Shape-from-silhouette Markerless Motion Capture System

INTRODUCTION The most common methods for accurate capture of three-dimensional human motion require a laboratory environment and the attachment of markers or fixtures to the body segments. These laboratory conditions can cause adaptive changes to normal patterns of locomotion that introduce experimental artifact. For example, it has been shown that attaching straps to the thigh or shank alters joint kinematics and kinetics (Fisher 2003). Thus our understanding of normal and pathological human movement would be enhanced by a method that allows capture of human movement without the constraint of fixtures or markers placed on the body. Markerless methods are not widely available because the accurate capture of human motion without markers is technically challenging. A recently reported method in the computer vision literature using shapefrom-silhouette (SFS) reconstruction offers an attractive approach (Cheung 2003). However, the factors influencing the accuracy of this method have not been addressed. The purpose of this study was to evaluate the accuracy of SFS reconstruction of a body for different camera placements, number of cameras, camera resolution and object contours.