Proposed retinal abnormality detection and classification approach: Computer aided detection for diabetic retinopathy by machine learning approaches

Diabetes occurs when the pancreas fails to secrete enough insulin, slowly affecting the retina of the human eye, leading to diabetic retinopathy. The blood vessels in the retina get altered and have abnormality. Exudates are secreted, micro-aneurysms and haemorrhages occur in the retina. The appearance of these features represents the degree of severity of the disease. Early detection of diabetic retinopathy plays a major role in the success of such disease treatment. The main challenge is to extract exudates which are similar in colour property and size of the optic disk, and then micro-aneurysms are similar in colour and proximity with blood vessels. The main objective of the paper is to develop a computer aided detection system to find the abnormality of retinal imaging and detects the presence of abnormality features from retinal fundus images. There is few existing research works have been undergone by applying machine learning techniques, but existing approaches have not achieved a good accuracy of detection and they have not yielded successful performance in different datasets. The proposed methodology is to enhance the image and filter the noise, detect blood vessel and identify the optic disc, extract the exudates and micro aneurysms, extract the features and classify different stages of diabetic retinopathy into mild, moderate, severe non-proliferative diabetic retinopathy (NPDR) and proliferative Diabetic retinopathy (PDR) by using proposed machine learning methods. The expected output of proposed work in this paper will be a preliminary design and pilot prototype development.