Study Of Facial Expression Recognition Based Hidden Markov Model

Facial expression recognition is one of the most challenging problems in the fields of pattern recognition, machine vision, affective computing and psychology. It has turned into an active research topic in the recent decades. Although facial expression automatic recognition technique develops quickly along with various applications, there are still many problems unsettled yet. Therefore the automatic recognition technology of facial expression has received the researchers' extensive concern.Facial expression recognition consists of localization of human eyes, feature extraction of facial expression and facial expression identified through HMM. Thus, these main three problems are researched in this paper.In the part of the pretreatment of facial expression image, firstly vertical integral projection curve is used to acquire face's left and right boundaries. Secondly, eye-brow area is approximately located through vertical integral projection and horizontal integral projection. Thirdly, eye location can be realized through integral projection and automatic threshold segmentation. Finally, by rotating ,cropping and zooming the image, thus a standard image can be achieved.In the second part, expression feature vectors of the expression sub-regions are extracted by log Gabor wavelet transformation to form expression feature vector. Experiment shows that expression features can be extracted effectively based on log Gabor wavelet transformation, which is insensitive to illumination variety and individual difference.As a statistical method, Hidden Markov Model has become popular in the field of speech recognition and thesis in recent years, So we select it as facial expression recognition.Experimental results based on Jaffe facial expression database demonstrate that facial expression can be recognized greatly through the statistical method-HMM.