Estimation of diabetic retinopathy using deep learning

Diabetes patients may develop the eye condition known as diabetic retinopathy (DR). Diabetic retinopathy is a rapidly expanding medical condition worldwide. Among many two of the DR might cause the diabetic patient’s vision to completely disappear. Early identification of DR is especially more important in this case to help recover vision and for rapid treatment. This condition arises from damage to the retina’s blood vessels brought on by elevated blood glucose levels. Making use of a computer, DR 2 Deep learning’s success has led to the emergence of diagnosis as a promising early detection technique. The proposed study presents an algorithm that takes advantage of deep learning to improve the performance of a computer-assisted diagnostic system for diabetic retinopathy. This system is made for portable diagnostic devices using CNN and ResNet architecture. Using the Messidor dataset, this method was successfully evaluated and yielded results of 96.5% specificity, 91.85% sensitivity, and 96.5% accuracy.