Improved Algorithms For Face Recognition Based On Two-Dimensional PCA

With the development of science and technology, information security becomes more and more important.Storing information by ID card and password cannot meet our requirements. So, in the background of globalization, face recognition is produced.Face recognition solves the problems that documents carry has a lot of trouble and easily lost, and password is easy to forget and break. So Face recognition has become a research hotspot of experts and scholars in short several dozen years.This article aims to improve the face recognition algorithm based on principal component analysis.This paper proposes a new algorithm after we finished the analysis about PCA algorithm, and the new algorithm excepts to get a higher recognition rate at a relatively low feature dimensions, and can improve the operating efficiency of the algorithm. Finally, experiments are done in the gray face image database and the color face image database, which have get good results. The main contents of the paper are as follows:(1) This paper introduces some common methods off ace recognition of domestic and foreign as well as their advantages and disadvantages, and generalizes the research of face recognition.(2) This paper reviews the ideas, principles and basic steps of PCA algorithm, and the advantages, disadvantages and the basic steps of the 2DPCA algorithm, C2DPCA algorithm, CS2DPCA algorithm, about PCA to give the reader a clear explanation.(3) CSM2DPCA algorithm is proposed. This new algorithm retains the mirror image of the original algorithm, and joined the odd-evenmean face image for face data matrix, then central processing the face data, the covariance matrix of CS2DPCA algorithm has been improved.(4) Because the color image hasa data dimensionthan the gray image, so we modified the CSM2DPCA algorithm, so thatthe algorithm can be used in color face database.