A Multiresolution Image Recognition Method Using Fisher Transform and Its Application to Car Type Recognition

This paper introduces a new multiresolution image recognition method. At first, it makes multi-wavelet division over the image, organizes the corresponding components of different resolution into a vector. Then, makes a Fisher transform over the vectors. Recognition is based on the sum of the absolute distance between the vector of the unknown object and the vector of the template or based on the relative distance which is defined as the absolute distance between the vector of the unknown object and that of the template divided by the distance between the vectors of the corresponding templates in the Fisher transformed domain. Because the recognition is mainly dependent on the low frequency component, the method is insensitive noise pixels in the images. Because the recognition procedure has included multiresolution recognizing computation, the right rate of recognition is in sensitive to the dimension size of the image in some degree. The practice of the method proves that the right rate of recognition is very high and the robustness of the method is strong.