A new model of crypt edge detection using PSO and Bi-cubic interpolation for Iris recognition

Abstract

Several attempts have been made to improve the iris recognition system from first into second generation which is proficient to recognize unique iris features such as crypts. However, the always changing iris features create difficulties in comparison phase to determine the genuineness. Therefore, to determine genuineness, this study proposes a new model of iris recognition using combinational approach of particle swarm optimization (PSO) and Bi-cubic interpolation techniques in selecting the best crypt among unique iris features template. The particles in PSO searches the most optimal crypt features in iris texture meanwhile, Bi-cubic interpolation technique create sharp and refined crypt images. The results indicate an improvement of PSNR rates which are 22.6836 dB for CASIA and 22.3312 dB for UBIRIS database. Moreover, this method has indicates that the output from two databases are within tolerate noise rate. The implication of this study contributes to a new method of feature extraction using bio-inspired, which enhanced the ability of detection in iris recognition.