全局与局部特征融合在人脸识别中的应用研究

摘　要主成分分析（PCA）和线性判别分析（LDA）是基于全局结构的特征提取方法，局部保持投影（LPP）和正交拉普拉斯脸（OLF）是基于局部结构的特征提取方法，全局结构特征的弱点是忽略了局部结构特征，局部结构特征的弱点是忽略了整体结构。基于此，本文提出了一种全局与局部结构特征融合（GLSF）的提取方法，将PCA和LDA的提取结果融合到LPP中，既描述了全局结构，又考虑了局部结构。在ORL及Yale上的实验结果表明，GLSF方法比PCA、LDA、LPP、OLF等方法具有更高的识别率。

Research and Application of Face Recognition with Feature Fusion of Global and Local Structure

AbstractPrincipal component analysis (PCA) and linear descriminant analysis (LDA) are an extraction method based on the global structure features. Locality preserving projection (LPP) and orthogonal laplacianfaces (OLF) methods are based on the local structure features. The local structure features can not be characterized in the global structure features, and the global structure features are ignored in the local structure. For this, it is proposed in this paper a novel method named fusion of global and local structure (GLSF) to fusion the feature extracted from PCA and LDA into LPP, considering both the global and the local structure. Experiments on ORL and Yale show higher recognition accuracy than PCA, LDA, LPP, OLF, and so on.