IDENTIFIKASI RETINA MENGGUNAKAN ALIHRAGAM GELOMBANG SINGKAT

  • Aris Wijayanti

Abstract

Identification of a retinal biometric identification method with a low error rate due to the unique patterns in the retina of blood vessels behind the retina. These patterns can be used as training data for the recognition system is then used for comparison when the identification is done. This study aims to identify the image of the human eye retina, either the left or right side, using image processing techniques and measuring the normalized Euclidean distance. So far, research on biometric systems, particularly with the object of the eye's retina, the eye is done at the owner from the owner's eyes with diverse backgrounds, such as the Messidor database. In this study created a system that can recognize the retinal image using the transformation Haar short waves by measuring the normalized Euclidean distance. The retinal image will be the initial pretreatment process of changing the original image into a grey image, which is then performed using the Haar wavelet feature extraction to obtain the energy that will be used for the normalization of the Euclidean distance so that the process of recognition by Euclidean values ​​are compared. Testing is done using eye retinal image database taken from Messidor many as 100 of the 300 images taken at random were then stored in a database, the database is one of 100 images stored, the database of 100 images stored 2, and 3 as many as 100 images database stored. Of the best database testing should be done as much as 6 levels of decomposition levels. From the test results have identified the recognition accuracy rate of up to 98%. The greatest degree of familiarity is level 1 that is equal to 98%. Followed usage by 80% level 2, level 3 is 59%, level 4 is 47%, level 5 is 45% and the lowest is the last level 6 that is equal to 37%.

Downloads

Download data is not yet available.
Published
2020-02-03
How to Cite
Wijayanti, A. (2020). IDENTIFIKASI RETINA MENGGUNAKAN ALIHRAGAM GELOMBANG SINGKAT. SAINTEKBU, 12(1), 10-17. https://doi.org/https://doi.org/10.32764/saintekbu.v12i1.852