Novel DCT based feature extraction for enhanced Iris Recognition

Iris recognition (IR) under varying live-tissues is challenging, and extracting tissue edge features is an effective approach to solve this problem. In this paper, we propose a unique combination of edge detection plus DCT based feature extraction for enhanced Iris Recognition. Two novel methods, namely circular sector and triangular shaped DCT feature extraction techniques are proposed. Individual stages of the IR system are examined and an attempt is made to improve each stage. A Binary Particle Swarm Optimization (BPSO)-based feature selection algorithm is used to search the feature vector space for obtaining the optimal feature subset. Experimental results show the promising performance of circular sector DCT extraction for iris recognition on Phoenix, MMU, IITD databases.