Special Issue Article
An Effective Analysis of Macular edema Severity for Diabetic Retinopathy
Recently, we have many researches on the fundus image for the detection of abnormality. Diabetic retinopathy (DR) is the damage of retina caused by complication of diabetes which results complete vision loss. Macula is responsible for our pinpoint vision. Diabetic macular edema (DME) is the major problem for the diabetic patients. Several techniques have been reported about an automated solution for the diabetic macular edema detection. An automated system for early detection of macular edema should classify all possible exudates present on the surface of retina. In this paper, two simple single class classifiers are used for the detection of abnormality. The normal retinal images are trained in these classifiers for the classification. The performance of the proposed methodology with the existing systems is evaluated based on classification accuracy. By finding the exudate, the proposed PCA DD classifier yields the highest classification accuracy compare to the Gaussian DD classifier. The overall severity accuracy for Gaussian DD and PCA DD is 84% and 92% respectively. Experimental result shows the superior nature of PCA classifier in terms of performance measures.