Facial Expression Recognition Based On Hidden Semi-Markov Models

In recent years, the reasons for renewed interest in facial expression recognition are multiple, but mainly due to people have more interest about human computer interaction (HCI). Facial expression recognition is to analyze and detect the special expression state from given expression images or video frames and then to ascertain the subject's specific inborn emotion, achieving smarter and more natural interaction between human beings and computers. The study of facial expression recognition has found important applied values.This paper first described the Background, Significance and research of Expression Recognition, then analyzed and summarized the common methods of expression recognition from face detection and feature extraction area combined the domestic and abroad research technology focus on the expression classification, analyzed the Advantages and disadvantages between different technologies briefly.Traditional feature extraction methods mostly extract the features against the whole face, but couldn't properly resolve the problem that part of the face is covered. Two-Dimension Discrete cosine transform, 2D-DCT feature extraction method was proposed in this paper. First, partition the facial expression image, find the two Key areas of eye brows and mouse, then extract 2D-DCT coefficient of the two areas, put the 2D-DCT coefficient as observation sequence. The qualitative analysis showed this method is reasonable and available.The partially covered expression recognition based on Hidden Markov Models is few, and there is no expression recognition method based on Hidden Semi-Markov Models. This paper presents two expression recognition method respectively based on HMM and HSMM, recognized and simulated the two partially covered expression recognition methods. Experiments showed that HSMM got better Recognition rate. Because HSMM has the Characteristics that each state has a number of observations and allowing default values, so it could got better recognition rate. This model improved the recognition results on partially covered images. Meanwhile the uncovered facial expression recognition has also improved.Finally, the facial expression recognition prototype system was designed and implemented, and proved the validity of this method through experiments.