Fundus image optic cup optic disc segmentation method and system for assisting glaucoma screening

Abstract

The invention discloses a fundus image optic disc segmentation method and system for assisting glaucoma screening, and relates to the technical field of image information processing. The invention includes collecting several fundus images, preprocessing the fundus images, and obtaining training image samples set and validation image sample set; use the training image sample set to train the constructed W‑Net‑Mcon fully convolutional neural network to obtain the best W‑Net‑Mcon fully convolutional neural network; preprocess the fundus images to be segmented Input the best W‑Net‑Mcon fully convolutional neural network to obtain the predicted target result graph; use polar coordinate inverse transformation and ellipse fitting to process the predicted target result graph to obtain the final segmentation result, and then obtain the cup-to-plate ratio, and finally obtain According to the preliminary screening results of glaucoma, the present invention can effectively extract image semantic information in multiple dimensions, fuse features at different levels, fuse global features with detailed features, encourage feature reuse, improve gradient backpropagation, and improve image segmentation accuracy.

Description

technical field [0001] The invention relates to the technical field of image information processing, and more particularly to a method and system for optic cup and optic disc segmentation of fundus images for assisting glaucoma screening. Background technique [0002] Glaucoma is the second leading cause of blindness worldwide (after cataracts) and the leading cause of irreversible blindness. Because the visual impairment caused by glaucoma is irreversible, early screening and diagnosis of glaucoma is crucial. Fundus images have become important reference information for doctors to diagnose glaucoma due to their low cost and easy availability. They play a huge role in recording the process of glaucoma lesions, diagnosing glaucoma diseases, evaluating treatment effects, and preventing some potential glaucoma lesions. . The current mainstream glaucoma screening technique is to use the optic nerve head in the fundus image to evaluate, and use a binary classification to determ...

AI Technical Summary

Problems solved by technology

[0008] The purpose of the present invention is: in order to solve the problem that the existing network structure has insufficient levels of semantic information of fundus images and insufficient fusion of deep and shallow features, resulting in unsatisfactory segmentation of optic cup and disc. The present invention provides an auxiliary glaucoma screening The fundus image cup-optic-disc segmentation method and system can improve the accuracy of fundus image cup-optic-disc segmentation to obtain an accurate cup-to-disc ratio, thereby providing a favorable assistance for automatic glaucoma screening

Images