【摘要】 采用基于非线性核空间的主分量分析法(KPCA)和线性主元空间鉴别分析法(LDA)相结合的算法,首先将人脸图像在非线性高维空间中进行主成分分量降维,然后采用基于主元空间的LDA方法对子空间再度降维,同时利用欧式距离分类器(KNN)对样本进行有效的分类识别。采用Matlab和ORL人脸库对该算法进行验证,实验证明,该算法识别性能显著提高,明显优于其他算法。更多还原

【Abstract】 This paper uses a new algorithm,which combines nonlinear kernel space based principal component analysis method (KPCA) and principal component space linear discriminant analysis method (LDA).The first step is to reduce the dimension of the principal components of the face image in non-linear high dimensional space. The second step is to use LDA method based on principal component subspace to reduce further dimension. And the Euclidean distance separator is embedded to LDA to classify the sample effectively. This paper at last uses Matlab and the ORL face database to verify the algorithm,and experimental results show that the algorithm is much better than other algorithms and is significantly improved the performance.更多还原