Detection of Eye Disease in Retinal Images Based on Haar Wavelets

It is known that there are many eye diseases in the world nowadays. Diabetic retinopathy is one of the diseases that lead to blindness. This disease is observed especially in premature babies, early detection and correct diagnosis of this disease is an urgent issue today. In order to detect this disease, highly qualified ophthalmologists are needed to observe the process in the retinal or fundus images of the eye. Unfortunately, the number of leading specialists in eye diseases is very small in today’s growing population. The method proposed in this article is used to classify the image data processed on the basis of Haar wavelets in the detection of retinopathy in the eye using inference tree and ResNet. Until now, a lot of research is being conducted on such classification systems, but when it comes to human health, these systems need to be further improved and more accurate. Fundus images were downloaded from open source for this research. In the research conducted, the accuracy was $86 \%$ with ResNet and $78 \%$ with inference tree.