Fundus Images-Based Detection and Grading of Macular Edema Using Robust Macula Localization

Abstract

The macula is an oval-shaped area near the center of the human retina, and at its center, there is a small pit known as the fovea. The fovea contains large concentrations of cone cells and is responsible for sharp, colored vision. Macular disorders are the group of diseases that damage the macula, resulting in blurred vision or even blindness. Macular edema (ME), one of the most common types of macular disorder, is caused by fluid accumulation beneath the macula. In this paper, we present an automated system for the detection of ME from fundus images. We introduce a new automated system for the detailed grading of the severity of disease using knowledge of exudates and maculae. A new set of features is used along with a minimum distance classifier for accurate localization of the fovea, which is important for the grading of ME. The proposed system uses different hybrid features and support vector machines for segmentation of exudates. The detailed grading of ME—as both clinically significant ME and non-clinically significant ME—is done using localized foveae and segmented exudates. The proposed algorithm is validated using public and local data sets. We have achieved an average accuracy of 96.1% in the detection and grading of ME with our proposed method.