Evaluation of the Reliability of Iris Recognition Biometric Authentication Systems

This study examines the quality of the image in different irises during iris recognition. The main purpose of this study is to provide reliable evidence pertaining to the superiority of the iris to other biometrics particularly in the presence of noise coming with images captured by the iris. The study mainly used the iris database CASIA-IrisV3-interval as input. It also examined and applied iris recognition system for the images of the iris database. During the study, noise was added to the CASIA-IrisV3-interval database in order to prove that images of the iris give the valid and very reliable form of an individuals identity. The images in the databases were compared by use of Hamming Distant (HD) method before and after noises were added to the images. The results obtained from this comparison were the find TPR values and FPR values that were then used to build ROC curves. The resulting ROC curves were then used to evaluate the quality of the image as presented by the CASIA images.