Pattern Analysis Using Zernike Moments

Abstract

Complex Zernike moments derive independent spatial information from an image due to the orthogonality of the Zernike polynomials. The orthogonality, rotation invariance and spatial feature representation of the Zernike moments have been the main motivations to study their characteristics. Statistical variability of iris patterns has encouraged us to apply Zernike moments to a database of iris images in order to evaluate their feature representation performance. The features are in complex form and represented by phase and magnitude information. In this experiment, characteristics of the magnitude information have been studied and the results of 250 comparisons among 50 different eye images are presented as well.