Detection of Diabetic Retinopathy Using Deep Learning

Diabetic retinopathy is a condition that damages the blood vessels in the retina due to poorly controlled blood sugar levels. Various detection techniques such as ophthalmoscope and retinal imaging are expensive and time-consuming. An accurate machine learning model with an image processing technique can improve early detection at a lower cost. In this research, a deep learning model that helps to increase identification accuracy is proposed. It consists of a convolution layer, a max pool layer, a flattening layer, and extra layers including normalization and drop-out layers. The proposed model experiments on diabetic retinopathy experimental results and datasets for retinal color fundus images demonstrate that it provides a 95.23 recognition accuracy.