Research On Facial Recognition Of China Ethnic Minorities

China is a multi-minority country, which has 56 ethnic minorities. Studying and analyzing the facial characters of each minority is an extremely challenging issue. There are not only theoretical research values but also wide applications. In this paper, the automatic face recognition technology is used to study the classification of ethnic minorities, in which the feature extraction and recognition technology is the keystone. The main contents of this paper are as follows:First, the development of face recognition is summarized and the research background and meaning of ethnic minorities recognition are given. Then several methods and the face datasets used in this paper are introduced briefly, in addition, the difficulties are analyzed. Finally, the main tasks of this paper are presented.Second, relevant mathematic theory of feature extraction and classification are introduced. The feature extraction is composed of three parts:feature produce, transformation and election. Then, the pattern recognition methods are presented briefly.Third, the statistical feature extraction algorithms that are principal component analysis and linear discriminate analysis are introduced in detail, and the k-nearesst neighbor classifier is depicted, and then the experiments are made in multi-minority database.Fourth, the geometrical feature extraction method is studied. Firstly, the face images are pre-processed, and then key feature points are located, and the feature vectors are conformed. Lastly, the decision tree classifier is used to classify the ethnic minorities.Fifth, the elastic bunch graph is presented. Firstly, the theory of Gabor wavelet transformation and elastic bunch graph matching are introduced detailedly. and then the matching process is depicted and the special ethnic face bunch graph is constructed.Finally, the contents of this paper are summarized and the prospect of ethnic minorities recognition is made.