Iris recognition using contrast enhancement and spectrum-based feature extraction

The statistical description of the iris varies drastically with changes in contrast. These variations make Iris Recognition (IR) even more challenging. In this paper, two novel techniques are proposed, viz., Contrast Enhancement using Top Hat and Bottom Hat filters for enhancing the gradience between brighter and darker pixels and DWT+DCT feature extractor to select the salient features. A Binary Particle Swarm Optimization based feature selection algorithm is used to search the feature space for the optimal feature subset. A complete IR system for enhanced recognition performance is presented. Experimental results on two benchmark iris databases, namely, IITD and MMU, illustrate the promising performance of the proposed techniques for Iris Recognition.