Enhancement of exudates for the diagnosis of diabetic retinopathy using Fuzzy Morphology

摘要

This paper presents a novel algorithm based on Fuzzy Morphology for the computer-assisted enhancement of exudates in fundus images of human retina for the diagnosis of diabetic retinopathy. Diabetic retinopathy is a common disease in diabetic persons. The disease is diagnosed by the presence of exudates in the macular region. Here, we use Fuzzy Morphology for the enhancement of exudates. The fundus image is first converted to grayscale followed by a series of fuzzy erosion and fuzzy dilation (morphological closing operation) with a diamond shaped structuring element. Finally, the resulting image is added to the original image to transform into enhanced one. Experiments were performed on a database of variety of fundus images with and without the disease. The experiments led to convincing results and the images were enhanced for easier clinical examination.