| Face recognition technology is one research topic in pattern recognition and image processing. There are many different descriptions about human face, such as age, gender and others. Ethnic facial characters are one important attribute descriptive set of human face. Different origins, genes, resident zones and dietary habits develop respective facial characters among different ethnics. So the study is focused on the research of different facial characters in different ethnics. This way to the research of ethnic characteristics can digitally retain the facial features of whole nations. And meantime it is of great importance to find new features in the process of national integration. In terms of technology, it can serve as one weak classifier in large-scale face searching to accelerate the speed of face matching.The main contribution of this paper is as following:First, the significance of ethnic facial characters to face recognition is introduced, and the process of face recognition is summarized in this paper.Second, To study facial characters of different ethnics, flexible template is constructed to extract facial characters by using Java language.Third, tree augmented naive Bayesian classifier is selected to analyze the ethnic facial characters extracted by the flexible template. Bayesian classifier is firstly introduced, then three main classification model of Bayesian classifier are elaborated. Finally, experiment is carried out by using characteristic data of Tibetan, the Zhuang and Uygur. and the result is analyzed in detail.Forth, support vector machine is selected as the classifier to analyze the character data of the ethnics. Through the recognition performance in different kernel functions, this paper analyzes the spatial distribution of the ethnic facial characters, finally, the best kernel function to identify the ethnic facial characters is obtained.Finally, the contents of this paper are summarized and the prospect of ethnic minorities recognition is made. |
PDF Full Text Request