Certain investigation on an integrated approach for classification and grading of diabetic macular edema using deep learning techniques

Abstract

The Macular Edema (ME) is a condition where the macula swells that leads to vision problems if left undetected and untreated early. It is often connected with the Diabetic Retinopathy (DR) and Diabetic Macular Edema (DME), both of which result from changes in blood vessels due to diabetes. Early diagnosis and timely treatment are critical in preventing vision impairment, especially in working-age individuals. The severity of these conditions depends on the extent of lesions in the retina. According to the International Diabetic Federation (IDF), around 463 million people had diabetes in 2019 and this number is expected to reach 700 million by 2045. The World Health Organization (WHO) recommends a healthy diet, regular exercise, proper medication and routine screenings to prevent the severe complications. Detecting DME early allows for the better treatment and lowers the risk of vision loss. However, many people with DME experience no pain in the early stages, making it difficult to identify without medical tests. Clinically, DME is classified into three grades: 0 (normal), 1 (non-CSME) and 2 (CSME). The key indicators of the condition are exudates that appear as yellow lipid deposits due to the fluid leakage from blood vessels. The severity of the disease increases as the amount of the exudates rises. newline