%0 Journal Article
%T Curvelet变换在人脸识别中的应用研究<br>Application Study on Curvelet Transform in Face Recognition
%A 赵庆敏
%A 欧阳欢
%A 辜道平
%J -
%D 2013
%X 针对小波只能反映信号的点奇异性,无法实现人脸图像面部轮廓和五官曲线信息的最优稀疏表示,提出了一种基于第2代Curvelet的人脸识别算法.通过对人脸图像进行第2代Curvelet变换,分解得到表征人脸基本信息的低频系数,再利用双向2维主成分分析(《2D)2PCA)进行降维,并结合最近邻算法进行人脸识别.以ORL人脸数据库进行试验,结果表明:与基于小波变换的算法相比,该算法具有更高识别率和更短的识别时间.<br>As wavelet can only reflect the zero-dimensional singularities of signal and cannot provide an optimally sparse representation of facial contour and five sense organs' curve information of face image,a face recognition algorithm based on second-generation curvelet transform is proposed.The proposed method firstly transforms face image with second-generation curvelet to get low frequency coefficients with facial basic information,and then reduce dimensions with the two-directional two-dimensional principal component analysis((2D)PCA),at last the test face is recognized by using the nearest distance algorithm.Experimental results on ORL and Yale face database show that the proposed method has higher recognition rate and less recognition time than the method based on wavelet
%K 小波
%K 人脸识别
%K 第2代Curvelet
%K 双向2维主成分分析<br>小波 人脸识别 第2代Curvelet 双向2维主成分分析
%K 小波 人脸识别 第2代Curvelet 双向2维主成分分析
%K 小波 人脸识别 第2代Curvelet 双向2维主成分分析
%K 小波 人脸识别 第2代Curvelet 双向2维主成分分析
%U http://lkxb.jxnu.edu.cn//oa/darticle.aspx?type=view&id=20130414