Implementation of Discrete Wavelet Transformation Method with K-Nearest Neighbor for Eye Disease ClassificationWord

Abstract

The eye is an organ in humans that functions as seeing objects around with the reflection of light received by the retina. This sense of vision can be affected by diseases that often occur, including cataracts, as well as other diseases such as glaucoma and retinal disease. The eye disease will interfere with the activities of the sufferer and can also attack his psyche. In examining and ensuring that eye diseases can be done by utilizing technology, with the development of technology to identify eye diseases can be done. Through an image of the patient's retinal fundus, the image can be processed using the image processing method. By combining image processing with classification methods from machine learning, images can be processed until they are identified in the class. This study was conducted with the aim of classifying eye diseases using the discrete wavelet transformation method with K-nearest neighbor, obtaining an accuracy level of 61% in the classification of a class. These results indicate that the classification can be done quite well, but in the results obtained, not all classes can classify well. Using datasets from Kaggle 300 normal eye datasets, 100 cataract eye datasets, 101 glaucoma eye datasets, and 100 retinal disease eye datasets, there are 4 classes of retinal fundus images. The retinal fundus is an image obtained as a result of capturing using a tool called the Ophthalmoscope where this tool helps illuminate and magnify the image in the eye to produce a capture of the retinal fundus.