Early detection of diabetic lesions can halt progression of the disease and prevent consequent loss of vision, as exudates are a visible sign of diabetic retinopathy and a marker for the presence of coexistent retinal oedema. If present in the macular area, they are a major cause of treatable visual loss in the non proliferative forms of diabetic retinopathy. It would be useful to have an automated method of detecting exudates in digital retinal images produced from diabetic retinopathy screening programmes In this work we apply digital image processing techniques to detect the main anatomical features of the retina, and to extract diabetic retinopathy signs from retinal images. Various approaches are explored to localize the optic disc. First template matching is implemented. Then we use texture analysis combined with morphological operations and markercontrolled watershed to localize the optic disc and its boundary. Once the optic disc has been located, a Region-of-Interest (ROI) for the macular region is created. Then the macula and fovea are segmented. We have proposed various techniques to detect exudates, hemorrhages and microaneurysms by morphological operations combined with Otsu’s thresholding algorithm and FCM clustering. The methodologies are tested on a set of 130 images. The success rates of disk localization by template matching and morphological operations are 94.6%, and 96.9% respectively, and 92% for the macular region. It is also found that the proposed method for exudates detection by morphological operations detects exudates successfully with sensitivity, specificity, and accuracy of 80.19 %, 98.85%, and 98.72 % respectively.
© 2019 MTPR. All Rights Reserved