Facial Expression Recognition Based On Gabor And Neural Network

Facila Expression plays a very important role in information transmission when people communicate with each other. Mehrabian pointed out,if we analyzed the information of transferring between people, language accounted for 7%, tone accounted for 38%, and expression accounted for 55%. These data indicate the important effect of face expression in exchanges between people. Human Facial expression recognition is a process of feature extraction and classification, which makes the computer detect the expression state from given expression images and ascertain the subject's specific emotion in order to achieve smarter and more natural inter-action between human beings and computers. Facial expression recognition becomes a hotspot in scientific research field in recent decades. Facial expression recognition is an important part of affective computing and intelligent human-machine interactive, which has a wide range of applications and potential market value.Generally, facsial recognition system consists of mainly four parts, which are expression image acquisition, expression image preprocessing, expression feature extraction and classification. This paper presents a detailed description of the research on the key issues of expression feature extraction and classification.Firstly, the thesis introduces the significance, application future and the history of research on the Facial Expression Recognition, summarizes its main method. Preprocessors are executed firstly.Secondly, we propose the method of extracting human facial expression features which bases on Discrete Cosine Tranform,Gabor wavelet transform and 2DPCA.In order to reduce the amount of data, DCT compressed images are proceed after the preprocessor. And to make the compressed image Gabor transform to extract the main characteristics of expression. Then, 2DPCA be used on the characteristics of the Gabor matrix dimensionality reduction treatment. In this way, while retaining the main expressions of information, we reduce the dimension of the characteristic and get a low-dimensional matrix of the characteristic. Finally, we study the neural network and the ensemble neural networks. And design the BP network and the ensemble neural networks respectively, to classify the facial characteristics. And compare the result of these two classifications. Then, we experiment with the CUM's Cohn-Kanade facial expression database to validate the validity of the method which proposes in this thesis.