关键词: 人脸识别, 二维主成分分析+二维线性判别分析（2DPCA+2DLDA）, 局部保持投影（LPP）, 改进的局部保持投影（ILPP）, 局部流形信息

Abstract: Aiming at the unsupervised learning problem of Locality Preserving Projection（LPP）, which is only contained the local information, a new face recognition algorithm based on 2DPCA+2DLDA and improved LPP is presented. Each image in the training set is mapped by 2DPCA+2DLDA to retain the spatial information and class label information. Then, the weight matrix of LPP is optimized with the within-class and between-class label information, which makes the LPP become a supervised nonlinear learning method. Local manifolds information of training set is obtained by improved LPP（ILPP） algorithm. The experimental results on the Yale and ORL face databases validate the effectiveness of the proposed algorithm.

Key words: face recognition, Two-Dimensional Principal Component Analysis+Two-Dimensional Linear Discriminant Analysis（2DPCA+2DLDA）, Locality Preserving Projection（LPP）, Improved Locality Preserving Projection（ILPP）, local manifolds information