Expression recognition from video using a coupled hidden Markov model
Mingli Song, Jiajun Bu, Chun Chen · 2004
To classify the face expressions in the video, in this paper, a system including both facial feature extraction and expression recognition is presented. The facial feature points are tracked first, then a sequence of FAP based face motion vector are deduced. These vectors are divided into two classifications, one for face expression, the other is for visual speech. Finally, a coupled hidden Markov model is trained to perform the recognition. The experimental results show that our approach is better than conventional face expression recognition system.