基于Gabor小波和有监督2DNPE的人脸识别方法  

Face Recognition Method Based on Gabor Wavelet and Supervised 2DNPE

作　　者：徐秀秀[1] 梁久祯[1] 

机构地区：[1]江南大学物联网工程学院智能系统与网络计算研究所,江苏无锡214122

出　　处：《小型微型计算机系统》2015年第8期1896-1901,共6页Journal of Chinese Computer Systems

基　　金：国家自然科学基金项目(61170121)资助

摘　　要：针对人脸图像识别易受光照、表情等外部因素影响这一问题,提出一种基于Gabor小波和有监督二维近邻保持嵌入算法(S2DNPE)的人脸识别方法.传统的2DNPE算法是一种无监督的流形学习算法,没有考虑类成员之间的关系,为了提高算法的鉴别能力,在其基础上引入类别信息使其变为有监督的2DNPE算法.首先利用Gabor小波对人脸图像进行特征提取,得到对光照、表情等因素都具有一定鲁棒性的图像特征,然后应用S2DNPE算法对其进行降维、提取映射到低维子空间的特征向量,最后采用最近邻分类器分类识别.在Yale、FERET和AR人脸库上进行实验,结果表明,与2DPCA,2DLPP,2DNPE,B2DLPP及G2DPCA算法相比,该方法具有较好的识别率.To solve the problem that the face recognition is easily affected by external factors such as illumination, expression, a new face recognition algorithm based on Gabor wavelet and supervised 2DNPE is proposed. Two-dimensional neighborhood preserving embedding is an unsupervised manifold learning method and it does not take any class membership relation into account. In order to enhance the performance of 2DNPE, the class information is used in S2DNPE. Gabor wavelet is used to extract the feature of human face images and face image features which are robust to illumination variations and face expression changes can be got. Then its dimension is reduced and eigenvectors mapping into low-dimensional subspace are extracted by the method of S2DNPE. Nearest neighbor classifier is adopted for face classification and recognition. The performance of the proposed method is evaluated and compared with 2DPCA, 2DLPP ,2DNPE, B2DLPP and G2DPCA on the Yale, FERET and AR face databases. The experiment results demonstrate the effectiveness and superior of the proposed method.

关 键 词：人脸识别 二维近邻保持嵌入 有监督学习 GABOR小波 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]