Application of Mean Shift Algorithm in Real-Time Facial Expression Recognition

In a dynamic real-time facial expression recognition, accurate and fast face tracking is a very important preparatory work that is in order to obtain the image sequence of facial expressions. For this problem, we proposed a mean shift algorithm for real-time tracking human faces, and using this method we can obtain the facial expressions image sequence. In order to obtain the initial target of the face image, we used an adaptive skin-color face detection method. Then we used the geometric model based on human face to locate the region of facial expression features, and can estimate the optical flow to calculate the eigen-flow vectors. At last, hidden semi-Markov model is used for facial expression recognition. The experimental results show that the application of mean shift algorithm in realtime facial expression recognition is very effective for obtaining the facial expression image sequence quickly and accurately.