Detection and Grading of Diabetic Retinopathy Using Deep Learning

Retinopathy, a complication affecting the eyes, is associated with diabetes and high blood pressure, and it results in damage to the retinal Blood Vessels. This damage leads to a negative impact on a person’s vision. There is potential for the early detection, diagnosis, and treatment of retinopathy through automated deep learning algorithms utilizing computer vision technology. Diabetic retinopathy has four distinct stages and is a leading cause of blindness in adults of working age. Similarly, hypertensive retinopathy also has four stages and can lead to vision issues or blindness if not addressed promptly. It is characterized as inflammation in retinal blood vessels and if left untreated could heighten the risk of cardiovascular disease and stroke. Therefore, to identify blood vessel damage, we have compared different CNN architectures to find the best solution among them. After our comparisons, we get the best precision of 83.0 from ResNet Architecture which will be further optimized in the future.