Detection of Hard Exudate from Diabetic Retinopathy Image Using Fuzzy Logic

Abstract

Diabetic retinopathy, otherwise called diabetic eye illness, is a therapeutic condition in which damage occurs to the retina because of diabetes and is a main source of visual deficiency. As exudates (exudates are mass of cells and fluid that has seeped out of blood vessels or an organ) are among early clinical indications of diabetic retinopathy, their location would be a basic resource for the mass screening errand and fill in as an essential. A procedure is proposed which depends on morphological image processing and fuzzy logic to recognize hard exudates from diabetic retinopathy retinal image in this dissertation. At the underlying stage, the exudates are distinguished utilizing mathematical morphology that incorporates image preprocessing utilizing HSV colour model and elimination of optic disc. The hard exudates are separated utilizing an adaptive fuzzy logic algorithm that utilizations values in the RGB colour space of retinal image to form fuzzy sets and membership function. The fuzzy output for all the pixels in every exudate is calculated for a given input set corresponding to red, green and blue channels of a pixel in exudates. Since, digital image is formed from combination of pixels, during image acquisition process, the quality of the image diminishes from the point they are captured. To get a quality image, image quality metrics are applied on the proposed algorithm. Then, fuzzy output is computed for hard exudates according to the proportion of the hard exudates detected. By comparing the results with hand-drawn ground truths, it has been obtained that the sensitivity and specificity of detecting hard exudates are 81.75% and 99.99%, respectively.