LDPC and SHA Based Iris Recognition for Smart Card Security

Abstract

We introduce a novel way to use Low Density Parity Check (LDPC) error correction code to reduce the variability and noise in iris code, which is generated from Iris Recognition System (IRS) and Secure Hash Algorithm (SHA-512) to transform codewords into hash string to make them as a Cancellable Biometric. In the enrolment process, by using majority voting scheme a unique iris code is generated from n-iris codes of IRS from n-eye samples of same person in different time interval. Then the codewords from LDPC encoding, parity check matrix of LDPC and hash string of codewords from SHA-512 are stored into smart card for high security access environment. In the verification process, iris code from live person is generated by using IRS then LDPC decoding is applied with the help of stored codewords and parity check matrix in the smart card. For authentication, new hash string, produced by employing SHA-512 on corrected iris code is compared with hash string stored in smart card. The LDPC code reduces the Hamming distance for genuine comparisons by a larger amount than for the impostor comparisons. This results in better separation between genuine and impostor users which improves the verification performance. Security of this scheme is very high due to the security complexity of SHA-512, which is 2256 under birthday attack. Experimental results show that this approach can assure a higher security with a low false rejection or false acceptance rate.