Fundus optic disc segmentation method based on U-Net neural network

Abstract

The invention provides a fundus optic disc segmentation method based on a U-Net neural network. The fundus optic disc segmentation method based on the U-Net neural network comprises the following steps: collecting a plurality of original fundus images; preprocessing the plurality of original eye fundus images; marking optic disc areas of the plurality of preprocessed original eye fundus images; constructing a U-Net neural network, and performing training verification on the original fundus image and the fundus image marked with the optic disk area by using the U-Net neural network to obtain a U-Net neural network model; and using the U-Net neural network model to identify and segment the optic disc region of the fundus image. The method can assist a doctor in obtaining an accurate and effective cup-to-disk ratio, assists the doctor in making a treatment method, accelerates the speed of distinguishing the degree of glaucoma of a patient, and reduces misdiagnosis caused by subjective factors of the doctor.

Description

technical field [0001] The invention relates to the technical field of image processing, in particular to a U-Net neural network-based fundus optic disc segmentation method. Background technique [0002] Glaucoma is a common irreversible blinding eye disease. In the early stage of glaucoma disease, there are generally no obvious eye symptoms. With the development of the disease, when there is a large visual field damage, the best treatment period is often missed. Therefore, timely diagnosis and treatment of glaucoma is an important way to delay the development of the disease. Optic nerve damage in most glaucoma patients is due to the increase of intraocular pressure caused by the blockage of aqueous humor circulation in the eyeball, which exceeds the intraocular pressure limit that the optic nerve can tolerate, resulting in slow and progressive damage to the optic nerve. Glaucoma is mainly manifested as the expansion of the central bright area (cup) of the optic disc. The ...

AI Technical Summary

Problems solved by technology

[0005] Based on this, in order to solve the problem that the existing glaucoma diagnosis method has certain subjectivity and relies on the professional knowledge and experience of sonographers, which may easily lead to misdiagnosis, the present invention provides a fundus optic disc segmentation based on U-Net neural network Method, its specific technical scheme is as follows:

Images