Automatic Detection of Hard Exudate in Color Retinal Fundus Image and Diabetic Maculopathy Grading

Diabetic macular degeneration (DM) is often characterized by the early accumulation of lipid deposits on the retina, referred to as hard exudates (HEs), which originate from damaged retinal blood vessels. Timely detection of HEs is crucial to prevent vision loss and ensure prompt treatment for DM patients. This study introduces a novel approach leveraging mathematical morphology for the automated detection of HEs and grading of diabetic maculopathy using color fundus images. The identification and grading of diabetic maculopathy are based on two primary factors: precise delineation of the macular region, including the fovea and its center, using the optic disc's location and diameter, and accurate detection of hard exudates. The severity of diabetic maculopathy is determined by the presence of HEs within the macula. When tested on the publicly available MESSIDOR database, the proposed method demonstrated an impressive 92% accuracy in detecting diabetic maculopathy and identifying hard exudates within the macular region.