Face Recognition Based on Wavelet Transform and Improved 2DPCA

Two-dimensional principal component analysis (2DPCA) is a kind of image feature extraction method. The algorithm based on image matrix as the analysis object, before the image feature extraction, it can deal with image data directly and does not need dimension reduction. It process image data without step of vectorization. Although 2DPCA algorithm reduces the computational complexity, it takes up more storage space. In this paper, based on wavelet transform and improved 2DPCA, an approach of face recognition was proposed, Using the improved 2DPCA algorithm can effectively recognize faces. This improved method can eliminate the correlation of image rows and columns at the same time. And it could overcome the drawback mentioned above. The method is the higher efficient recognition rate than 2DPCA algorithm.