Perception of hard exudates using Fuzzy Optimization and Discrete Wavelet Transformation in Fundus Image

Exudates are the clinical symptoms of Diabetic retinopathy, a retinal disorder caused due to the existence of diabetes for a long period, which leads to vision problem if not diagnosed in the earlier stage. A clinical screening system has been designed to facilitate the process of identifying and classifying exudates from the retinal fundus image. Initially the exudates are segmented by employing fuzzy c-means algorithm and subjected to discrete wavelet transformation. The Haralick texture features extracted by the gray level co-occurrence matrix are fed to the classifier. The performance of the proposed system was appraised by comparing with the ground truth of the publically available dataset IDRID. A graphical user interface has been designed and implemented using MATLAB.