It has been demonstrated that shape, area and depth of the optic disc are relevant indices of diabetic retinopathy. In this paper, we present a new fundus optic disc localization and segmentation method based on phase congruency (PC). Firstly, in order to highlight the optic disc, channel images with the highest contrast between optic disc and background are selected in LAB, YUV, YIQ and HSV spaces respectively. Secondly, with the use of PC, features of four selected channel images can be extracted. Multiplication operation is then used to enhance PC detection results. Thirdly, window scanning and gray accumulating are utilized to locate the optic disc. Finally, iterative OTSU automatic threshold segmentation and Hough transform are performed on location images, before the final optic disc segmentation result can be obtained. The experimental results showed that the proposed method can effectively and accurately perform optic disc location and segmentation.
Keywords: Fundus image; diabetic retinopathy; optic disc localization; optic disc segmentation; phase congruency.