Optic disc positioning and segmentation method, device and storage medium in fundus image

Abstract

The positioning and segmentation method, device and storage medium of the optic disc in the fundus image. In the initial positioning stage of the optic disc, the method first performs improved weighted grayscale conversion on the color fundus image; according to the ratio between the fundus radius and the optic disc radius, the minimum square Multiplicative ellipse fitting is used to estimate the radius of the optic disc; then multiple morphological operations are used to remove the interference of arteries and veins, and the histogram statistics are used to determine the adaptive threshold. The geometric center of the optic disc is used as the estimated center of the optic disc; finally, the rectangular frame of the optic disc is determined. In the stage of accurate positioning and segmentation of the optic disc, the GrabCut method is used to segment the optic disc, and ellipse fitting is used to determine the position and size parameters of the segmented area contour to complete the positioning of the optic disc. For the case where the radius of the optic disc in the GrabCut segmentation result is too large, the second segmentation is performed, a new threshold is determined, and the contour is fitted to obtain the final segmentation result.

Description

technical field [0001] The present application relates to the field of image analysis, in particular to a method, device and storage medium for accurately locating and segmenting the optic disc in a color fundus image. Background technique [0002] Obtaining the optic disc from the fundus image, locating and segmenting the optic disc are the most urgent needs of fundus image measurement. For the location and segmentation of the optic disc, the conventional method is to segment the optic disc by using image processing methods such as image preprocessing, filtering, ellipse fitting and morphology. The above methods have certain effects, but still have certain defects. [0003] For example, Lalonde (Lalonde M, Beaulieu M, Gagnon L. Fast and robust optic disc detection using pyramidal decomposition and Hausdorff-based template matching [J]. Medical Imaging, IEEE Transactions on, 2001, 20(11): 1193-1200) etc. Using Hausdorff-based template matching technology combined with pyram...

AI Technical Summary

Problems solved by technology

[0006] The current method is good for its own test data set, but it may not achieve good segmentation results for other data sets, which may be due to the fact that the training parameters are not necessarily suitable for color fundus images of various types and sizes

The segmentation ideas and methods of filtering, morphology, ellipse fitting and graph theory used in these methods have been tested to have obvious effects on segmentation, but they need to be implemented step by step to adapt to various fundus maps

Images