Automatic Optic Disk Localization and Diameter Estimation in Retinal Fundus Images

Optic disc localization is a necessary step in most automatic screening systems for ophthalmic pathologies. In this paper, we propose a novel approach to automatically determine the location and the diameter of the optic disc in retinal fundus images. Our proposed method is based on the circular shape of the optic disc and the distribution of blood vessels. We first detect blood vessels by a vessel detector named B-COSFIRE filter. Then we apply a bank of anti-symmetric Gabor filters on the grayscale fundus image to detect edges. Finally, we identify the optic disc by applying the Circular Hough Transform (CHT) to detect candidates for circles and select among them the one that encloses the highest proportion of vessel pixels. We evaluate the proposed approach on two publicly available data sets: CHASEDB1 and ONH Hunter. We correctly localize 96.43% and 96.97% of the optic discs in these two sets, respectively. For each image, we compare the manually determined hight (vertical diameter) of the optic disc with the automatically computed one by the CHT. The relative errors we obtain for the two data sets of the optic disc are 10.69% and 8.78%, respectively