PENGENALAN POLA IRIS MATA MENGGUNAKAN METODE TEMPLATE MATCHING DENGAN LIBRARY OPENCV

Abstract

Biometrics is a technology that used on computer systems in the introduction of a person using a part of the human body. One part of the body that can be used in biometric systems are human iris, due to the nature of the iris of the eye that will not change and is unique among person with another person. One application of biometrics using human iris is on the smartphone. The system was built using OpenCV as the main library. Proposed recognition system was divided into two stages, there was iris recognition stage and iris matching stage. Recognition stage began with taking the image, and then do the pre-processing stage which consists of scaling operations to a smaller size and grayscaling, then proceed to the segmentation stage, using Canny edge detection method, circle Hough transform method for detecting circle of the iris and the pupil, and Daugman's rubber sheet models for normalization, then proceed to the feature extraction stage using Gabor Filter and Average Absolute Deviation. Matching phase was made using euclidean distance method to measure the similarity distance between two iris features. Tests carried out with 10 eye images, each of which 5 left and right eye images are derived from the research subjects. The results of this study concluded that the optimal threshold for this system was 475 with a percentage of 36% False Reject Rate, False Acceptance Rate 40%, 38% system error ratio, and the Genuine Acceptance Rate 64%.