Diabetic Retinopathy Detection Using Statistical Features

Diabetic Retinopathy is a medical condition that occurs due to prolonged diabetes, ensuing in harmful to small blood vessels present in the retina, which in later stage leads to complete loss of vision. Moreover, this disease is asymptomatic at its earlier stages and hence, constant screening and precise diagnosis of Diabetic Retinopathy is required. Fundus image analysis proves to be an indispensable tool for the diagnosis of patients. This Paper focuses on using a neural network algorithm aimed to diagnosis of the disease by extracting the key features in the retinal fundus images are lesions, hemorrhages, exudates then the statistical parameters which henceforth help in the classification. This algorithm proves to achieve better accuracy than the traditional unsupervised classification algorithms.