Face Recognition based on Wavelet Transform and Image Comparison

As a time-frequency scale transformation tool for data, function and operator, wavelet transform is a very good method about image compression, by which redundancies of the image are removed and original features of the image are reserved. The pixels of facial images are usually larger, so the wavelet transform is used before image comparison. The low-frequency images can be decomposed and extracted by wavelet transform. The image comparison between the sample images and the testing images, which is based on low-frequency components in different decomposed layer, can reduce effectively the computational complexity and make face recognition very fast. For face recognition, the problem of image comparison can be transformed into sequence comparison. The image comparison offers a similarity degree for human faces after the sequence similarity is defined, using a score function and a comparison function. Finally, a threshold is setting to guarantee the authenticity of correct recognition for human facial images to a certain extent. For the sake of improving computing efficiency, it is necessary to normalizing different values of similarity degree. The experiments show that the proposed method has the characteristics of simple realization, rapid recognition speed and high recognition rate.