IRIS RECOGNITION USING GRAY LEVEL RUN LENGTH MATRIX AND KNN CLASSIFIER

Biometric devices are great tools for the security. Iris is a very unique identifying characteristic amongst all human biometric traits. In the proposed system, the biometric authentication system using iris recognition is presented. In this iris image is preprocessed then image localized with the Hough transform, normalized using Daugman’s rubbersheet model finally image sharpening is used with the morphological toggle filter. Feature extraction is done using Gray Level Run Length Matrix (GLRLM) technique with 0 directions and classification is done using multiclass KNN. This system is evaluated on CASIA database and it gives 92.66% accuracy.