Detection of red lesions in retinal images using image processing and machine learning techniques

Abstract

Diabetic Retinopathy (DR) is a diabetes complication that causes damage to the blood vessels of the light sensitive tissue at the back of the eye. All the people who are suffering from diabetes have a high risk of subjecting to DR which may lead to total blindness. Red lesions, cotton-wool spots and exudates are symptoms of non proliferative diabetic retinopathy which is the early stage of diabetic retinopathy. When the disease develops to proliferative diabetic retinopathy fluid leaking from retinal capillaries and the formation of new vessels on the surface of the retina happens. At this stage there is a very low possibility of preventing total blindness. Therefore, early detection of DR is important to prevent vision loss. So, if there is an easy way of detecting early signs of DR accurately that will be beneficial. Red lesion detection is more important for the early identification of DR. In this paper, we are proposing a method for the automated detection of red lesions in retinal images using image processing techniques and machine learning. The developed algorithm has sensitivity and specificity of 92.05% and 88.68% respectively.