Abstract

Abstract In this paper, an efficient iris recognition algorithm is presented based on features extracted from Integer Wavelet Transform (IWT). At first, for proficient iris region localization and segmentation, Circular Hough Transform (CHT) and Relative Total Variation (RTV) model is utilized in combination. The segmented iris region is normalized and decomposed by four level IWT. 256 sub-bands are generated by applying four level Integer Wavelet Transform on the input image, out of which only 192 lower sub-bands are taken into consideration. High frequency sub-bands are ignored as it adds noise to the system and degrades the accuracy of the system. Features such as energy are extracted from each of 192 sub-bands, thus generating a 192 bit binary code. Unique iris code is computed by comparing energy of each sub-band with a pre-computed individual personalized threshold value. For image authentication, test iris image is compared with registered iris template iris code by calculating normalized hamming distance between them. IWT is preferred over Discrete Wavelet Transform (DWT), as IWT is computationally efficient since IWT compares the two integer values, while DWT may be complex in nature. The proposed algorithm provides the accuracy of 98.9% with the average execution time of 1.03 sec that is better than DWT and previous reported algorithms. False Accept Rate (FAR), False Reject Rate (FRR) and Equal Error Rate (EER) have also improved.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call