Coronary artery segmentation method based on CTA image

Abstract

The invention discloses a coronary artery segmentation method based on a CTA image. Firstly, non-coronary tissues are effectively inhibited through image preprocessing, and the contrast ratio of coronary arteries and background is improved; secondly, an irregular ascending aorta layer with coronary artery bifurcations is detected in combination with an optical flow method and heart anatomical structure priori knowledge, and manual initialization of starting points of left and right coronary arteries is avoided; and finally, compared with a traditional region growing method, the provided self-adaptive region growing method combined with endpoint detection has better segmentation capability and accuracy for fine branches with uneven gray levels and complex topological structures.

Description

technical field [0001] The invention belongs to the field of medical image processing, and specifically relates to a non-diagnostic coronary artery segmentation method based on a CTA image, which is a method for identifying the starting points of left and right coronary arteries on a coronary CT angiography (Computed Tomography Angiography, CTA) image and Methods for coronary artery segmentation. Background technique [0002] The heart is an organ that pumps blood to the whole body through the circulatory system. It ensures that the tissues of the whole body can obtain enough oxygen and nutrients through the blood, and remove carbon dioxide and other waste products. The coronary arteries are wrapped around the heart for the heart itself. Blood vessels that supply blood are critical to the proper functioning of the heart. The topology of the entire coronary artery tree is complex, with many and small branches, and its diameter generally ranges from 2mm to 7mm. Starting from...

AI Technical Summary

Problems solved by technology

At present, the manual segmentation results of cardiologists are the most accurate, but manual segmentation is very time-consuming, and the human subjective factors of different experts will also lead to subtle differences in the segmentation results

In recent years, various computer-aided diagnosis techniques (Computer Aided Diagnosis, CAD) have been used for coronary artery segmentation, but most of them are still semi-automatic segmentation methods, which require manual setting of the starting seed points and corresponding thresholds of the branches of the left and right coronary arteries. Complete the subsequent coronary artery segmentation, and it is easy to miss the small branches of the coronary artery

A small number of fully automatic methods are prone to over-segmentation, and non-coronary blood vessels are also mistakenly segmented into coronary arteries, thereby reducing the accuracy of segmentation

Images