The human need for system security will continue to increase and it can be obtained from the system that has a different key for each person. To answer the request, recently developed technologies related to the human recognition system with one of the physical body / unique behavior, known as biometrics. Biometric systems are automatically identified based on the characteristics of the individual or the different characteristics of every individual. Current iris recognition technology is a relatively new biometry and predicted needs are increasing. It is therefore important to conduct research aimed at designing iris recognition system, to know the accuracy of the designed system and to know the factors that affect the level of accuracy. Iris recognition system in this research consisted of the localization process of the iris from the other part of eye, feature extraction using wavelet haar transformation is stored as a template and recognition using the Hamming distance in the iris template. Gray eye image data used in this research was taken from the database MMU (Multi Media University). The results showed the accuracy of the test of practice data obtained 100% accuracy rate and of intraclass data obtained the best accuracy rate was 75%. To test the foreign data (interclass), smallest Equal Error rate (EER) is 0.05 with hamming distance threshold value of 0.35. The existence of eyelashes, eyelids, reflecting light and image-making positions affect the iris recognition process.
Keywords: biometrics, iris recognition, haar wavelet transformation, Hamming distance.
Nur Maulidiyah
No comments:
Post a Comment