Fundus Vessel Segmentation Technology Based on Image Recognition

Abstract

The lower part of the eye is filled with abundant blood vessels, in the normal course of infusion of blood vessels, and the reason for this phenomenon is various eyeground vessels, for example, diabetes retinopathy, retina venous obstruction, hypertension retinopathy, hyperoptic retinopathy, etc. If the bleeding volume is small, people don't need to look for a doctor, but eyeground can be self-relieved. If, however, the hemorrhage in the blood vessel is extensive and hard to heal, it may be necessary to make a new diagnosis and operation. So it is of great importance to treat and recover the entire eyeground. The objective of eyeground image segmentation is to extract the blood vessels from complicated fundusimages. This method can offer non-invasive, high resolution data of retinal blood vessels to assist in the correct and reliable diagnosis of eye diseases. This article gives a brief introduction to some of the technology of segmentation, and discusses the features of them. The purpose of this article is to provide the reader with an insight into how this area develops and to become familiar with the techniques used to segment the vessel.