Utilizing Discrete Wavelet Transform and Discrete Cosine Transform for Iris Recognition

Abstract

The analysis of the iris images of individuals has proven that iris is very distinctiveness and permanence for biometric uses. The task of iris analysis requires precise steps to yield accurate decisions. Considerable studies have shown that the extraction of the most informative features is one of the important keys for resulting in high level of accuracy. Discrete wavelet transform (DWT) and discrete cosine transform (DCT) have been intensively utilized to extract the features of iris images. This paper provides a technique for analyzing the combination of the features extracted by DWT and DCT all at once. The proposed technique is applied on CASIA interval-v4 image database. For the classification task, the extracted features are fed into the multiclass SVM. The accuracy rates yielded by the proposed technique reached 100%. This is quite promised comparing with those results of processing DWT and DCT separately.