Search

Diabetes, a widespread disease found in nearly everyone of the persons in working age group. Diabetes is a reason for cluster of disease. One such disease that affects the human vision is the Diabetic Retinopathy (DR). Current advancement in the medical science can only control it and cannot be cured. Screening the retina periodically for DR by ophthalmologists is difficult and costlier. Hence it has to be automatically detected at its earlier stage. Most researchers use deep learning as tool for earlier detection of DR. In this research paper, we used the modified GoogLeNet Convolution Neural Network architecture as a transfer learning technique. With publicly available kaggle dataset for our research, we achieved accuracy of 94%. [ABSTRACT FROM AUTHOR]