基于聚类算法和层次SVM的姿态人脸识别方法  

Method of pose-varied face recognition based on clustering algorithm and hierarchical support vector machines

作　　者：石良武[1] 陈荣元[1] 陈海军[1] 蒋加伏[2] 

机构地区：[1]湖南商学院现代教育技术中心,湖南长沙410205 [2]长沙理工大学计算机与通信工程学院,湖南长沙410076

出　　处：《计算机工程与设计》2007年第12期2922-2924,共3页Computer Engineering and Design

基　　金：湖南省自然科学基金项目(04JJ40048);湖南省教育厅科研基金项目(04C076)

摘　　要：提出了一种基于神经网络和层次支持向量机的多姿态人脸识别方法。该方法在训练阶段先利用神经网络把姿态人脸图像特征向准标准人脸图像特征映射,再根据聚类结果来训练支持向量机。识别阶段是先利用神经网络变换得到待识别图像所对应的准标准图像的特征,再让层次支持向量机初步判断待识别图像最可能所属的人,最后利用否定算法对待识别的人脸图像进行确认,实验表明该算法效果较佳。A method of pose-varied face recognition based on clustering algorithm and hierarchical support vector machines. At the stage of training, the feature vector of pose-varied image is transformed to the feature vector of standard image using neural network, then the standard image feature vector is clustered, lastly the hierarchical support vector machines is trained using the result of clustering. At the stage of the recognition, firstly it transforms the feature vector of pose-varied image to the feature vector of standard image using the neural network, estimates which person the image mostly belongs to using hierarchical support vector machines, lastly confirms estimation using negative algorithm. The experiment shows that the effect is better.

关 键 词：人脸识别 神经网络 层次支持向量机 离散余弦变换 聚类算法 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]