No abstract

No abstract

No abstract

Diabetic retinopathy (DR) in patients affects retina function owing to chronic elevated excessive blood glucose rates. The DR is a severe medical disorder. Patients with diabetes are easily prone to this medical complication which when not detected and treated at earlier stages, leads to vision loss. Globally, diabetes mellitus is the fifth leading cause of vision loss. So, active research is being conducted in this area to find novel ways in identifying the stages of DR. Specific image recognition methods and computer simulation algorithms were initially used to classify DR, but their usefulness was inadequate in real-time clinical practice. The evolution of deep learning models like convolution neural network performed better in identifying DR and non-referable DR compared to conventional machine learning models. Different variations of CNN architecture are being developed over the period, but more analysis and experimentation needs to be carried out to choose the appropriate architecture for detecting Diabetic Retinopathy. The aim of this research is to apply and understand how the performance of pre-trained Deep learning model – ResNet, a deep layered neural network performs to identify non-referable DR and different types of referable DR.

No abstract