Automatic detection and grading of diabetic maculopathy using fundus images

摘要

Diabetic maculopathy is one of the leading complications of diabetes that is measured as the significant reason for vision misfortune among individuals around the globe. In this paper, a two stage approach is proposed for automatic detection and grading of diabetic maculopathy based on mathematical morphology. Mainly two features are needed for the detection of maculopathy. One is the macula and other one is hard exudates, based on macula location and Optic Disc (OD) diameter, macular section has been drawn. We validate the presence or absence of hard exudates in the macula section and severity of the maculopathy was assessed. The experiment results performed on images from public database i.e., MESSIDOR indicate that the proposed method achieved encouraging results.