Retinal vessel segmentation map generation method based on credibility and deep learning

Abstract

The invention discloses a retinal vessel segmentation map generation method based on credibility and deep learning. The retinal vessel segmentation map generation method comprises the steps: 1, obtaining training data, and building a training set through employing a preset credibility model and the training data; 2, selecting data from the training set and inputting the data into a deep learning model based on a convolutional neural network for training to obtain a classifier; 3, obtaining a to-be-detected image, and performing image preprocessing on the to-be-detected image; 4, inputting the to-be-detected image after image preprocessing in the step 3 into the classifier in the step 2 to obtain five prediction probability values of pixel points in the to-be-detected image in the five types of credibility areas; and 5, generating a retinal vessel segmentation map according to the predicted probability value of the pixel point in the to-be-detected image in the step 4 in the five types of credibility areas. By means of the retinal vessel segmentation map generation method, thick blood vessels can be accurately segmented, and tiny blood vessels can also be accurately segmented.

Description

technical field [0001] The invention belongs to the technical field of retinal vessel segmentation in fundus images, and in particular relates to a method for generating retinal vessel segmentation images based on reliability and deep learning. Background technique [0002] The morphological structure (width, branching and tortuosity, etc.) of retinal vessels is often used as an important biomarker for the diagnosis and evaluation of various cardiovascular and ophthalmic diseases such as diabetes, hypertension and choroidal neovascularization. The best way to segment blood vessels today is to let trained experts manually calibrate blood vessels, but this task is extremely tedious and time-consuming. This facilitates the development of automatic vessel segmentation methods. [0003] Many automated vessel segmentation methods have been proposed over the past two decades, but none have proven accurate enough to be used as a standard in the medical community. These segmentatio...

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to provide a method for generating a retinal vessel segmentation map based on reliability and deep learning, which overcomes the problem of inaccurate segmentation of small blood vessels in existing methods, and can accurately segment both thick blood vessels and small blood vessels

Images