A method that can be used for spotting segments that display facial expression is proposed. The motion of the face is modeled by HMM in such a way that each state corresponds to the conditions of facial muscles, e.g., relaxed, contracting, apex and relaxing. The probability assigned to each state is updated iteratively as the feature vector is obtained from image processing. A spotted segment is placed into a certain category when the probability of that category exceeds a threshold value. Experiments show that the segments for the six basic expressions can be spotted accurately in near real time.
展开▼