Intracranial artery segmentation method and system

Abstract

The invention relates to an intracranial artery segmentation method and system. The method comprises the following steps: acquiring a three-dimensional time leap magnetic resonance blood vessel image to be segmented; inputting the to-be-segmented three-dimensional time leap magnetic resonance blood vessel image into an intracranial artery segmentation model to obtain an intracranial artery image in the to-be-segmented three-dimensional time leap magnetic resonance blood vessel image; obtaining the intracranial artery segmentation model by training a three-dimensional convolutional neural network model, and the three-dimensional convolutional neural network model comprises a first convolutional block module and a second convolutional block module which are connected in sequence; the first convolution block module comprises a plurality of cavity convolution blocks which are connected in sequence, the second convolution block module comprises a plurality of three-dimensional convolution block units which are connected in sequence, and each three-dimensional convolution block unit comprises a three-dimensional convolution block and a three-dimensional voxel rendering neural network module which are connected in sequence. According to the invention, the accuracy of intracranial artery segmentation can be improved.

Description

technical field [0001] The present invention relates to the technical field of image segmentation, in particular to a method and system for intracranial artery segmentation. Background technique [0002] Intracranial artery stenosis is an important cause of ischemic stroke, if not detected and treated in time, it may lead to serious neurological sequelae and be fatal. The current intracranial artery screening is mainly for doctors to observe the maximum intensity projection (MIP) at all angles of the blood vessel through a Magnetic Resonance Imaging (MRI) monitor, so as to further evaluate the patient's intracranial artery stenosis. screening. Clinically, three-dimensional time-of-flight magnetic resonance angiography (Three Dimensional Time of Flight Magnetic Resonance Angiography, 3D TOF MRA) has the advantage of not requiring the injection of contrast agents, so it is widely used for screening of intracranial arteries. [0003] However, the existing deep learning segmen...

AI Technical Summary

Problems solved by technology

[0003] However, the existing deep learning segmentation model uses layer-by-layer prediction of vascular regions for 3D TOF MRA, without using context information, which limits the effect of intracranial artery segmentation to a certain extent, resulting in inaccurate segmentation results

Images