Calibration and Error Model Analysis of 3D Monocular Vision Model Based Hand Posture Estimation

In this paper we investigate the accuracy of vision based hand posture estimation using a single camera. We use a two step process to estimate the hand posture: we first generate 3D posture hypothesis using geometric and kinematics inverse transformations, and then validate the hypothesis by projecting the postures on the image plane and comparing the projected model with the ground truth using extracted hand color and silhouette features. We also calibrate the estimated posture joint angles with the corresponding angles acquired using a CyberGlove, and we use the interval of confidence of the estimated values as a measure of the vision based posture estimation accuracy.