A human face recognition technique called modular 2DPCA is presented in this paper.First

the original images are divided into modular images in proposed approach.Then

an image covariance matrix is constructed directly using the sub-images

and its eigenvectors are derived for image feature extraction.Compared with previous techniques based on image vectors such as PCA

there are two advantages for this way: 1)the sub-image matrices don't need to be transformed into vectors prior to feature extraction

and dimension reduction of discriminant features can be effected conveniently;2)singular value decomposition of matrix is absolutely avoided in the process of feature extraction so the features for recognition can be gained easily.Moreover

2DPCA is the special case of modular 2DPCA.To test modular 2DPCA and evaluate its performance

a series of experiments were performed on two human face image databases: ORL and NJUST603 human face databases.The experimental results indicated that the recognition performance of modular 2DPCA is superior to that of PCA and is more robust than that of 2DPCA as well.