Localization of the optic disc center in retinal images based on the Harris corner detector

摘要

Purpose: Localizing the optic disc and its center is the first step of most identification, segmentation algorithms and diagnosing some diseases on fundus photographs such as diabetic retinopathy. Despite the importance of optic disc localization, there is not very accurate and fast method for localizing the center of optic disc in retinal images. Therefore, we propose a robust and fast algorithm for localizing the center of optic disc. Methods: Based on the property of optic disc, vessels originate from the center of optic disc and the number of the vessels in the vicinity of optic disc is more than others regions in the retinal images. Therefore, we can see the largest number of corners and bifurcations around the optic disc in the retinal images. In this paper, a robust method based on Harris corner detector is proposed. Using the Harris corner detector, corners and bifurcations are found in the retinal images. Then, we use a moving window near the size of optic disc to count the number of corners. Finally, the center of windows in which the high number of corners are located, is obtained and the mean of these centers is considered as the center of optic disc. The DRIVE, STARE and a local dataset including 273 retinal images are used to evaluate the proposed algorithm. Results: The success rate is 97.5%, 87.65% and 97.8% for DRIVE, STARE and a local dataset. The average distance between the estimated and the manually identified optic disc centers is 4.61, 11 and 9 pixels for the DRIVE, STARE and local dataset respectively. Comparing the results of our proposed method and counterpart methods verifies the effectiveness of the proposed method. Conclusions: In this paper, we proposed a new method for localizing the center of optic disc based on corners and bifurcations obtained using Harris corner detector. Comparing the results of our proposed method and counterpart methods verifies the effectiveness of the proposed method.