Diabetic retinopathy detection using image-processing

Abstract

Diabetic retinopathy is a leading problem throughout the world and many people are losing their vision because of this disease. The disease can get severe if it is not treated properly at its early stages. The damage in the retinal blood vessel eventually blocks the light that passes through the optical nerves which makes the patient with Diabetic Retinopathy blind. Therefore, in our research we wanted to find out a way to overcome this problem and thus using the help of convolutional neural network (ConvNet), we wereable to detect multiple stages of severity for Diabetic Retinopathy.There are other processes present to detect Diabetic Retinopathy and one such process is manual screening, but this requires a skilled ophthalmologist and takes up a huge amount of time. Thus our automatic diabetic retinopathy detection technique can be used to replace such manual processes and theophthalmologist can spend more time taking proper care of the patient or at least decrease the severity of this disease.