Face tracking system using H∞ approach

An automatic face detection and tracking based on H_infin filter is presented in this paper. The proposed system is used to track and predict the location of a moving person´s face in different background. To overcome the lack of a prior knowledge of the system model and the noise statistic, face is tracked using the prediction values generated by H_infin which minimizes the worst case estimation error. The proposed system consists of two parts: face detection and tracking. The face is first segmented out from the complex background using elliptical skin boundary. The position of the segmented face is then passed to the tracking system as the parameters for H_infin to predict the appearance of the face in the succeeding frame. The performance of the proposed system is evaluated and compared with Kalman filter based tracking system. The algorithm is tested on various video sequences and the simulation results reveal good performance of the proposed method.