National conference on Digital Image and Signal Processing |
Foundation of Computer Science USA |
DISP2015 - Number 2 |
April 2015 |
Authors: G. G. Rajput, B M Reshmi |
5c3fbf68-7a6d-49a6-bb57-35dee41ac5b2 |
G. G. Rajput, B M Reshmi . Mathematical Morphology based Optic Disc Detection in Retinal Fundus Images. National conference on Digital Image and Signal Processing. DISP2015, 2 (April 2015), 1-5.
Optic disc segmentation is a prerequisite in automatic analysis of retinal disease in color fundus images. In this paper, an algorithm for detection of optic disc (OD) in fundus images is presented. Mathematical morphology is used for segmentation of candidate optic disc region, followed by determining its center and approximating the circular optic disc boundary using circular model. Proposed method is evaluated on public database, MESSIDOR. The proposed optic disc segmentation algorithm yielded an average overlapping score of 99. 47% between true OD and segmented OD and success rate of 93%.