Method for measuring vascular diameter in coronary angiography image based on depth learning

Abstract

The invention discloses a method for measuring the vascular diameter in a coronary angiography image based on depth learning. The method for measuring the vascular diameter in a coronary angiography image based on depth learning comprises the following steps: a DSA image processing module converts real-time acquired DSA image data into a data stream which can be processed by a subsequent module, and stores the data stream in a memory and sends the data stream to a depth network segmentation module. The depth network segmentation module segments the acquired DSA image data to distinguish the blood vessel pixel from the background pixel. The centerline extracting module extracts the blood vessel centerline from the blood vessel pixel image. The diameter calculation module calculates the measured diameter, the reference diameter and the stenosis rate of the blood vessel based on the segmented image and the blood vessel centerline. Compared with the PVA method, the invention has more objective results and avoids the differences between different doctors and different hospitals due to different interpretation methods. This method does not need manual intervention and is easy to operate.It can be used by doctors in coronary angiography and can provide objective reference for doctors to determine the degree of coronary stenosis.

Description

technical field [0001] The invention relates to a method for measuring blood vessel diameters in coronary angiography images based on deep learning, which belongs to the technical field of computing. Background technique [0002] Coronary angiography is an important method for the diagnosis of heart diseases. Observing and analyzing the shape and movement of blood vessels and estimating the diameter of blood vessels from angiography images can assist doctors in diagnosing cardiovascular diseases and determining appropriate treatment options, which is of great clinical significance. [0003] At present, the standard method for clinical doctors to judge the degree of coronary stenosis in coronary angiography is still PVA (Physician Visual Assessment, physician visual assessment). The criteria were severe stenosis of coronary arteries with a visual lumen diameter of ≥70% or a fractional flow reserve of 0.80 or less. In the past two decades, the decision of whether to perform c...

AI Technical Summary

Problems solved by technology

[0004] Computer-aided QCA can better analyze the lesion location, stenosis degree, and involvement range, but manual participation is required in the operation process, such as manual selection of appropriate frames and pixel-level calibration of DSA images. The degree of influence, and the operation is more complicated, not suitable for coronary angiography

Images