A modified PCA algorithm for face recognition

In principal component analysis (PCA) algorithm for face recognition, the eigenvectors associated with the large eigenvalues are empirically regarded as representing the changes in the illumination; hence, when we extract the feature vector, the influence of the large eigenvectors should be reduced. In this paper, we propose a modified principal component analysis (MPCA) algorithm for face recognition, and this method is based on the idea of reducing the influence of the eigenvectors associated with the large eigenvalues by normalizing the feature vector element by its corresponding standard deviation. The Yale face database and Yale face database B are used to verify our method and compare it with the commonly used algorithms, namely, PCA and linear discriminant analysis (LDA). The simulation results show that the proposed method results in a better performance than the conventional PCA and LDA approaches, and the computation at cost remains the same as that of the PCA, and much less than that of the LDA.