基于核正交局部判别嵌入的人脸识别  

Face recognition based on kernel orthogonal local discriminant embedding

作　　者：王庆军[1] 张汝波 潘海为[1] 

机构地区：[1]哈尔滨工程大学计算机科学与技术学院,黑龙江哈尔滨150001

出　　处：《光电子．激光》2010年第9期1386-1389,共4页Journal of Optoelectronics·Laser

基　　金：国家"863"计划资助项目(2009AA04Z215);国家自然基金资助项目(60803036)

摘　　要：针对人脸识别中的非线性特征提取问题,提出一种基于核正交局部判别嵌入(KOLDE,kernel orthogonal local discriminant embedding)的人脸识别算法。首先通过引入基向量正交约束,得到OLDE算法,并给出算法的推导过程。然后为了更好地处理高度复杂非线性结构数据,将OLDE向高维空间扩展,在核空间提取图像的高阶非线性信息,得到核空间OLDE算法。在ORL和PIE库上的人脸识别实验验证了算法的有效性。In view of the problems of nonlinear feature extraction in face recognition,an algorithm based on kernel orthogonal local discriminant embedding（KOLDE） for face recognition is proposed in this paper.Firstly,the proposed algorithm gets orthogonal local discriminant embedding（OLDE） algorithm by introducing the orthogonal constraint of basis vectors.Then the OLDE algorithm is expanded to high dimensional feature space by nonlinear mapping in order to deal with dataset of highly nonlinear structure.Finally,the kernel orthogonal local discriminant embedding algorithm is obtained by introducing the orthogonal constraint of basis vectors and traditional LDE method in kernel space.The experimental results on ORL and PIE face database illustrate the effectiveness of the proposed method.

关 键 词：子空间 核正交局部判别嵌入(KOLDE) 核空间 流形 人脸识别 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]