Automated Identification of Diabetic Retinopathy - A Survey

Diabetes strikes when the pancreas stops to produce sufficient insulin, gradually disturbing the retina of the human eye, leading to diabetic retinopathy. The blood vessels in the retina become changed and have abnormality. Exudates are concealed, micro-aneurysms and haemorrhages occur in the retina of eye, which intern leads to blindness. The presence of these structures signifies the harshness of the disease. A systematized Diabetic Retinopathy screening system will enable the detection of lesions accurately, consequently facilitating the ophthalmologists. Micro-aneurysms are the initial clinical signs of diabetic retinopathy. Timely identification of diabetic retinopathy plays a major role in the success of managing the disease. The main task is to extract exudates, which are similar in color property and size of the optic disk; afterwards micro-aneurysms are alike in color and closeness with blood vessels. The primary objective of this review is to survey the methods, techniques potential benefits and limitations of automated detection of micro-aneurysm in order to better manage translation into clinical practice, based on extensive experience with systems used by opthalmologists treating diabetic retinopathy. Keywords-Microaneurysms, Optic Disc, Exudates, Diabetes Mellitus, Machine Learning _________________________________________*****_________________________________________