Automatic fovea detection in retinal fundus images

The fovea (centre of the retinal macula region) is responsible for central vision. Conditions which can affect the macula region in particular include AMD and macular oedema, which lead to a direct loss in visual acuity due to the effect on central vision. The detection of the fovea is an important step in the assessment of retinal images for pathologies relating to these visual conditions. In this paper, we propose a method to automatically detect the fovea in retinal fundus images. The method makes use of optic disc detection as a starting point. Using information from an extensive database, typical locations of the fovea are used to define search regions for subsequent analysis. We make use of a combination of the image characteristics of the macular region with noise and vessel removal for fovea detection. The proposed method is tested on a large database of 750 retinal fundus images, achieving an accuracy of 86.53%.