人脸的民族特征抽取及其识别

人脸的民族特征是人脸信息描述的重要特征之一.首先构建了中国多民族人脸数据库,利用人脸识别技术提取民族面部特征和民族识别.在特征抽取方法中,采集人脸中的代数特征和几何特征,采用LDA(线性判别分析)算法提取人脸图像的代数特征.还构建了能够抽取人脸几何特征的弹性模板,并利用Gabor小波进行特征点定位.实验采用KNN和C5.0分别学习训练集中的代数特征和几何特征,并对测试集进行预测分类.实验结果表明,利用代数特征方法和几何特征方法对藏族、维吾尔族、壮族3个民族的平均识别准确率分别达到79%和90.95%.

Minorities Features Extraction and Recognition of Human Faces

Minorities feature of face is one of the most important features of face features. We created a face database of ethnic minorities and extracted facial features using face recognition technology. In the feature extraction method,we adapted the algebra and geometry from face database, used I_DA algorithm to extract the algebraic features of human face images,this paper also constructed a new face templates to extract the geometric features and used gabor wavelet to locute the points of face templates. KNN and C5. 0 Classifiers were used to learn the train dataset. I}hc result indicates that the average recognition accuracy rates of I}ibetan, Uygur and Zhuang reach 79 0 0 by algebraic features and 90. 95% by geometry features.