Ordinal Measure of Discrete Cosine Transform Coefficients and its Application to Fingerprint Matching

Abstract

Recently, the identification system is not limited in using an ID and personal identification number (PIN) but also in using biometriccharacteristics.One of biometric characteristics that has been widely used is fingerprint.This paper proposes a fingerprint matching algorithms using ordinal measure of DCT coefficient. The ordinal measure of DCT coefficient is generated from DCT blocks with size 8x8 pixels. Matching level was determined by computing the Minkowski distance between features of input fingerprint image and fingerprint images in the database. The simulations were accomplished using 128 fingerprints that have been normalized, from which as many as 1024 genuine attempts and 15360 impostor attempts were generated. The proposed algorithms achievedan Equal Error Rate (EER) at threshold 0.3. At the EER, it resulted in FAR value of 0.82%, and FRR value of 78.41% respectively.The low value of FAR showed that the system wasconsiderably secure.

DOI:http://dx.doi.org/10.11591/ijece.v3i6.4373