A method and system for multispectral fundus image analysis based on adversarial learning

Abstract

A multispectral fundus image analysis method and system based on adversarial learning disclosed in the present invention includes: acquiring multispectral fundus images, including images with vascular labels and images without blood vessels; inputting the multispectral fundus images into the trained fundus images In the analysis model, the registration results of multispectral fundus images are obtained; among them, the fundus image analysis model includes a fundus image registration model and a retinal blood vessel segmentation model. During training, the fundus image registration model and the retinal blood vessel segmentation model respectively adopt The images are trained separately; the fundus image registration model and the retinal vessel segmentation model are adversarially learned from images without vessel labels. The fundus image registration model and the retinal blood vessel segmentation model can be trained independently from images with blood vessel labels, and can also conduct confrontational learning from images without blood vessel labels, which improves the accuracy of image registration and blood vessel segmentation.

Description

technical field [0001] The present disclosure relates to a multispectral fundus image analysis method and system based on adversarial learning. Background technique [0002] The statements in this section merely provide background information related to the present disclosure and do not necessarily constitute prior art. [0003] Effective imaging techniques are key to the diagnosis and successful treatment of ophthalmic diseases. In recent years, a variety of imaging techniques have been developed for fundus imaging, including color fundus photograph (CFP), multispectral imaging (MSI) enhanced depth imaging optical coherence tomography (EDI-OCT), fundus autofluorescence (FAF) , Indocyanine green angiography Fundus fluorescein angiography (ICGA / FFA) was performed. Different ophthalmic imaging techniques show unique advantages in visualizing specific ophthalmic diseases. [0004] Multispectral fundus imaging (MSI) technology is based on light-emitting diode (LED) illuminati...

AI Technical Summary

Problems solved by technology

The second challenge is the misalignment of pixels due to the constant movement of the eyes

Complex eye movements during imaging create great challenges in analyzing MSI images

[0007] Multispectral fundus image registration is usually based on the alignment of images based on the retinal vascular structure, and many ophthalmic diseases can be reflected on retinal blood vessels. The extraction of retinal blood vessels is a key technology for analyzing fundus images. Therefore, the segmentation of retinal blood vessels is a multispectral A key step in fundus image analysis, but currently there is no method for joint analysis of multispectral fundus image alignment and blood vessel segmentation

Images