作　　者： ; ;

机构地区： 电子科技大学中山学院

出　　处： 《计算机应用与软件》 2009年第11期223-226,共4页

摘　　要： 研究光照变化条件下的人脸识别问题。基于最近提出的二阶特征脸方法和(2D)2PCA方法,提出了二阶(2D)2PCA方法。该方法将(2D)2PCA技术分别应用到原始图像矩阵集和剩余图像矩阵集。在extended Yale人脸库上的实验表明,在光照变化条件下,二阶(2D)2PCA方法是一种有效的人脸识别方法。该方法与传统的特征脸、二阶特征脸方法和(2D)2PCA相比,具有更高的识别精度;且比特征脸和二阶特征脸方法节省计算时间。 In this paper,face recognition under various lighting conditions is studied. Based on the second-order eigenface method and the （2D）^2 PCA method come out recently, a novel technique for face image feature extraction and recognition, the second-order （2D）^2 PCA, is proposed. This method applies （2D）^2PCA in original image matrix set the residual image matrix set respectively. Experiment resuhs in the extended Yale face database show that,the proposed method is effective in face image identification under various lighting conditions. Comparing with traditional eigenfaee, second-order eigenfaee and （ 2D ）^2pCA methods, this method achieves higher recognition accuracy, and has much less computational cost than eigenface and second-order eigenface methods.

关 键 词： 人脸识别 特征脸 二阶特征脸 光照变化

领　　域： [自动化与计算机技术] [自动化与计算机技术] [理学] [理学]