Iris recognition using savitzky-golay filter energy feature through symbolic data modeling

The iris has remained a preferred biometric trait compared to other biometrics because of its uniqueness, stability properties. However, degraded iris images captured under less constrained acquisition setups and varying lighting conditions will affect the performance of iris based biometrics systems. This paper presents a novel method for iris recognition through symbolic modeling approach and also investigates the applicability of Savitzky-Golay filter for iris feature extraction. The approach also proves that the symbolic representation effectively handles noise and degradations, including low resolution, specular reflection, and occlusion of eyelids present in the eye images and uses minimum number of features to represent iris image. Hence reduction in the number of features for iris representation will decrease testing times during classification. The modified content based symbolic similarity analysis method is devised and used to compare the symbolic object of test image and symbolic knowledge base of trained eye images. The CASIA 1.1 and CASIA-Iris 4.0 Interval databases were selected to evaluate the proposed system and recognition is over 99 % and 95 % respectively. The experimental results show that, the symbolic representation of Savitzky-Golay filter energy features of iris outperforms other state-of-the-art methods.