Lbp And Mixed-2dpca + Pca Based Face Recognition Research

The technology of face recognition is a technology that uses thecomputer to analyze the image and discriminate identity or recognizestatus from the worked image. It is a research area spanning severaldisciplines such as image processing, pattern recognition, computervision, physiology and psychology, it is one of the key issues.Texture feature abstraction refer to abstract characteristic fromimage by some way that definite image processing techniques, therebyobtain veined ration or qualitative descriptive treating process. Thispaper usr LBP vein operator to analysis image texture feature, lt usestructured idea analyses window character, then use statistics integralfeature extraction. In face recognition, first usr face plot somenonredundant region, count every regional ALBP histogram, the face veinwas descripte by each regional ALBP histogram, then use count distance ofeach eigenvector. By use FERET demonstration, ALBP adopt Tamura algorithmanalyse widow size, procure know clearly favorable effect up.In the period of feature extraction we make use of many techniquesincluding PCA, 2DPCA, (2D)~2PCA, A combined 2DPCA plus PCA framework wasproposed to make full use of two kinds of discriminative informationderived from horizontal 2DPCA and vertical 2DPCA. In this framework, thehorizontal feature matrix and the vertical feature matrix are,respectively, processed by a whitened PCA for further dimensionalityreduction. Then, two kinds of features are normalized and fused by a summedweighted-distance based fusion strategy. Our experimental results onFERET database and AR database demonstrate that fusion of the two kindsof features, horizontal and vertical, can achieve better recognitionperformance and, the proposed method, Combined 2DPCA plus PCA, is morepowerful than PCA and 2d-2DPCA plus PCA.