Font Size: a A A

Face Recognition Based On Parameter-less Two-dimensional Discriminant Locality Preserving Projections

Posted on:2016-11-28Degree:MasterType:Thesis
Country:ChinaCandidate:K WangFull Text:PDF
GTID:2308330479983548Subject:Computational Mathematics
Abstract/Summary:PDF Full Text Request
Face recognition is a kind of biometric identification technology. In recent years, face recognition has become the most important research topic in the field of pattern recognition and artificial intelligence. With the rapid development of computer technology, face recognition technology has gain large achievement. Generally, a face recognition system consist of four main steps: face detection, image preprocessing, feature extraction and feature matching. Among, feature extraction and matching is the core of face recognition which will directly affect the recognition results. This paper mainly aimed at the feature extraction algorithm and feature matching and doing the following works:① Recently a linear manifold algorithm was proposed: locality preserving projections(LPP) Algorithm, which preserve the local information of faces and detect the intrinsic manifold structure of face data by the adjacency matrix between sample points. We investigated LPP in two-dimensional sense based directly on image matrices, which is called two-dimensional locality preserving projections(2D-LPP). Research shows that 2D-LPP recover the nonlinear manifold structure of human faces, but 2D-LPP is an unsupervised algorithm, which deemphasizes discriminant information.② By introducing between-class scatter constraint and label information into two-dimensional locality preserving projections(2D-LPP) algorithm, two-dimensional discriminant locality preserving projections(2D-DLPP) has more discriminant power than 2D-LPP. However, 2D-DLPP is confronted with the difficulty of parameter selection, which limits its power on solving recognition problem.③ In this paper, to solve this problem we constructing parameter-less matrix on the basses of Locality Preserving Projections Algorithm, an algorithm called parameter-less two-dimensional discriminant locality preserving projections(parameter-less 2D-DLPP) is proposed. This method can reduce the recognition time and more efficient. At last, we do experiments on the ORL face database and the Yale face database to test the performance of 2D-DLPP. The simulation results on Yale and ORL face database show that the method in this paper can get higher recognition rate than 2D-DLPP, 2D-LPP and 2D-LDA.
Keywords/Search Tags:Face recognition, Feature extraction, Manifold learning, Two-dimensional locality preserving projections, Parameter-less
PDF Full Text Request
Related items