基于改进的LPP_SIFT人脸识别算法  

Face Recognition Based on Improved Locality Preserving Projection-SIFT

作　　者：任成娟[1] 

机构地区：[1]宝鸡文理学院学院计算机科学系,宝鸡721000

出　　处：《计算机与数字工程》2013年第3期469-471,501,共4页Computer & Digital Engineering

基　　金：院级重点项目"基于视觉感知的人脸识别身份认证新方法与技术"(编号:ZK10168)资助

摘　　要：局部保持映射—尺度不变特征变换(LPP-SIFT)算法是一种有效的特征识别方法,但是基于LPP的算法在人脸识别中容易遇到奇异值问题。为此,论文提出采用奇异值分解(SVD)的LPP-SIFT算法(SVD_LPP_SIFT)。首先用尺度不变特征变换对样本选择特征;然后再根据LPP算法求出新样本空间的低维投影子空间。在算法中,样本数据将被映射到一个非奇异正交矩阵中,以此解决了奇异值问题。在Yale、ORL上实验,实验结果验证SVD_LPP_SIFT算法在人脸识别中的有效性。Locality Preserving Projection-SIFT is an effective method which can extract the feature,but based on LPP is known to suffer from singular value problem.Therefore,using LPP-SIFT of Singular Value Decomposition（SVD） was proposed（SVD_LPP_SIFT）.In the algorithm,SIFT was used to extract feature.Then the data of the low dimensional sample space projection subspace were obtained according to the LPP method.In the algorithm,the sample data were projected on a non-singular orthogonal matrix to solve the problem of singular value.The experimental results demonstrate the efficacy of the SVD_LPP_SIFT approach for face recognition on Yale,ORL.

关 键 词：局部保持映射 尺度不变特征变换 奇异值分解 人脸识别 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]