Abstract

The detection of microaneurysms is crucial, as it is an early indicator of a complication of prolonged diabetes called Diabetic Retinopathy. In this paper, an automated approach is proposed to detect microaneurysms from retinal fundus images. Firstly, the magenta plane of the input image is extracted and a few preprocessing techniques are carried out. This is followed by the localization and the removal of the optic disk. The threshold value is determined and is optimized using Firefly algorithm. Then top hat transform is applied to detect the microaneurysms. The image quality parameters and the performance parameters were calculated and analyzed on the images of the DIARETDB1 database. The experimental results yielded a sensitivity of 99.80% before optimization and 100% after optimization.