Iris based human verification algorithms

Abstract

In this paper three algorithms for iris verification have been presented. Iris detection algorithms include the normalization and iris extraction steps. Three algorithms for verification process are (a) Algorithm using radial and circular features, (b) Algorithm using Fourier transforms and (c) Algorithm using Circular-Mellin transforms. Proposed algorithms have been tested on CASIA database and some non-infrared Iris images. The experimental results show that the algorithm based on Circular - Mellin Transform gives the best result with an accuracy of 95.45%. Some initial experiments on non-infrared iris images shows that this algorithm can work on such images but it still requires some more attention and this is our future work. © Springer-Verlag Berlin Heidelberg 2004.