Lip-tracking algorithm based on elliptical deformable templates

A new lip-tracking algorithm based on deformable template, optical flow and the extended Kalman filter techniques is proposed, which combines the shape, motion and color cues of the object in image to acquire a fast and robust scheme of lip tracking. The algorithm uses several elliptical arcs as deformable template to capture the edge of the lip. With no needs of any special smoothness constraint, the optical flow is estimated by model-based optical flow method in YCrCb color space. By using extended Kalman filter, the measurements of shape and motion are naturally integrated to provide an effective fusion solution without explicit calculation of optical flow. Besides, an optical-flow based measurement error and the estimated covariance given by the EKF filter are used to detect and reject the contour sample points introduced by noise, or spurious edge to improve the robust of the scheme. Experiment results are presented to validate the algorithm.