Zernike's Feature Descriptors for Iris Recognition with SVM

Abstract

Valuable information of the iris is intrinsically located in its natural texture, therefore preserve and extract the most relevant features for biometric recognition is of paramount importance. The iris pattern is subject to translation, scaling and rotation, consequently the variations produced by these artifacts must be minimized. The main contribution of this work consists on performing a comparison between the descriptive power of the Zernike and pseudo Zernike polynomials for the identification of iris images using a Support Vector Machine (SVM) as a classifier. Experiments with the iris data set obtained from the Bath University repository show that our proposal yields high levels of accuracy.