Discriminative Common Vectors for Face Recognition Using Iterative Approach

Face recognition is the process of identifying individuals from images of their faces by using a stored database of faces labeled with people's identities. Since face images are similar, they are correlated; therefore can be represented in a lower dimensional subspace called feature space without loosing a significant amount of information. LDA method cannot be directly applied because of "small sample size problem". To overcome this problem the discriminative common vectors (DCV) approach is proposed. DCV approach is based on a variation of Fisher's Linear Discriminant Analysis. In this the common vectors are extracted by eliminating the differences of the image samples in each class of images. Then the DCV which will be used for classification are obtained from the common vectors. In this paper, Iterative hierarchical classification by using Discriminative Common Vectors approach is suggested. In this method the Common Vectors obtained from the earlier iteration are grouped together according to their classes and then used as input for the next iteration to obtain the DCV of these classes. The number of iterations required is equal to the number of hierarchy levels in the training set.