ABSTRACT
Diabetes is considered to be the most common disease among all the age group people from older to young children. This disease becomes more chronic and deadly if it remains untreated and unidentified. It majorly affects the human visibility and in later stages causes microvascular complications such as diabetic retinopathy in the retina. Non-invasive detection of ophthalmic disorder due to blood sugar is quite challenging, but the elegant use of machine learning approaches can simplify the process of diagnosis. It is not only minimizing bare-bones of disease detection process but it is also inherently fast and robust. In this chapter, a new convolution neural network has been proposed for ocular disease detection and it develops a framework, which gives medical attention and treatment for severe disorder such as diabetic retinopathy. The proposed model is a structural modification of existing VGG-16 for analysing the complex artefacts in the retinal region. This intelligent platform is capable of handling a large set of image samples and detects the appropriate disease very profoundly. The efficiency of the proposed deep neural network has obtained an accuracy of 89% for binary class classification.
