A New Method for the Localization of Hard Exudates Based on Analyzing Intensity Incremental-Decremental Trends

Abstract

Diabetic retinopathy (DR) is a serious retinal disease infecting the patients with a long history of diabetes. It damages the vision and may lead to partially sightedness if remains untreated. Therefore, early diagnosis of DR is of significant importance. Hard exudates (HEs) are from important manifestations of DR. They appear as bright lesions in the retinal fundus images. The identification of HEs with automatic approaches helps the ophthalmologists in the process of evaluating disease and its progression. In this article, a new method is proposed for the automatic localization of HEs from retinal fundus images. The primary idea is the trend of intensity changes in different directions inside a sample HE. In fact, as one moves from outside to the center of an HE, the intensity values increase and then decrease gradually. This idea is mathematically modeled in the proposed method. The simulation results indicate that the proposed method has an outstanding performance in terms of simplicity and accuracy in comparison to the state of the art.