一种核正交局部敏感辨别分析算法

为了发掘嵌入在人脸样本的非线性结构信息,把核方法和基向量正交化思想引入局部敏感分析算法中,提出一种新的人脸识别算法-核正交局部敏感辨别分析(Kernel based Orthogonal Locality Sensitive Discriminant Analysis).并给出了算法的推导过程及计算步骤.首先用核方法提取人脸样本的非线性信息,并将其投影至高维非线性空间,然后采用局部敏感辨别分析做线性映射,最后采用施密特正交化方法得到正交的基向量,从而使算法更好地描述人脸非线性流形结构特征.在ORL和YaleB人脸库的人脸识别实验证明了所提算法的有效性.

Kernel Based Orthogonal Locality Sensitive Discriminant Analysis

To discover the nonlinear structure information embedded in face image, a new face recognition algorithm called Kernel based Orthogonal Locality Sensitive Discriminant Analysis is proposed,which considers the kernel method and orthogonal basis functions. The derivation process of the algorithm and the calculation steps are also presented. First the face images are mapped into high nonlinear space by kernel method to obtain the nonlinear information. Then a linear transformation is performed to preserve locality geometric structures. Finally,a set of orthogonal basis vectors are obtained by Schmidt orthogonalization method to present the nonlinear manifold structure of the face image. Experiments on ORL and YaleB demonstrate the effectiveness of the new algorithm.