Elliptical Sector Based DCT Feature Extraction for Iris Recognition

Abstract

Iris Recognition (IR) is one of the most reliable Biometric used. This Paper proposes a novel Feature Extraction technique called Elliptical Sector Based DCT Feature Extraction. Also proposed use of pre-processing technique Adaptive Histogram Equalization, Image Adjustment and Image sharpening technique for enhancement of feature considered for selection. The Proposed Extraction technique is used to extract features from transformed segmented Iris image resulting in reduction of features considered for recognition. An attempt is made to improve IR system by analysing each stage of the IR system. A Binary Particle Swarm Optimization (BPSO) method based feature selection algorithm is used to search the feature vector space for the optimal feature subset. Experiments show promising performance of Elliptical Sector Based DCT Feature Extraction and Pre-Processing with top Recognition Rate (RR) being 100% and Average being 97.64%.