Supercharge Your Innovation With Domain-Expert AI Agents!

Blood vessel identification method based on SWI image and recurrent neural network

A technology of cyclic neural network and recognition method, applied in the field of nuclear magnetic resonance imaging, can solve problems such as inability to automatically distinguish, and achieve detailed and accurate results

Active Publication Date: 2020-05-08
THE FIRST AFFILIATED HOSPITAL OF ARMY MEDICAL UNIV
View PDF8 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] Aiming at the defects in the prior art, the present invention provides a blood vessel recognition method based on SWI images and cyclic neural network to solve the problem in the prior art that when performing blood vessel recognition on SWI images, other blood vessels that are also displayed as low signal cannot be automatically distinguished. organizations, technical issues that require human judgment

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Blood vessel identification method based on SWI image and recurrent neural network
  • Blood vessel identification method based on SWI image and recurrent neural network
  • Blood vessel identification method based on SWI image and recurrent neural network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0029] The invention provides a blood vessel recognition method based on SWI image and cyclic neural network, comprising the following steps:

[0030] S1. Use the nuclear magnetic resonance instrument to scan with the SWI sequence to obtain the original magnetic moment image and the original phase image generated by multiple pairs. The original magnetic moment image and the original phase image are fused after image post-processing to obtain the SWI image;

[0031] S2. Segmenting the SWI image through an edge detection algorithm to obtain an image to be recognized;

[0032] S3. Selecting part of the images to be recognized and the original phase images corresponding to the part of the images to be recognized as a training set; using the remaining images to be recognized by manual marking, and the original phase images corresponding to the remaining images to be recognized as a test set;

[0033] S4. Input the training set into the cyclic neural network, associate the image to ...

Embodiment 2

[0064] Judging from the physiological and anatomical images of the human body, blood vessels have their structural characteristics. The overall structure of blood vessels is tree-like with multiple branches; the local structure is tubular with a certain curvature. In the SWI image, if the tissue shows low signal, and at the same time it is tree-like as a whole and has multiple branches, it means that the tissue is a venous vessel.

[0065] On the technical scheme of embodiment 1, change. The SWI image is obtained by scanning with a nuclear magnetic resonance apparatus, and the image segmentation is performed on the SWI image to obtain the image to be recognized. These two steps refer to the technical solution of Embodiment 1. From the point of view of human anatomy section, the SWI image scanned by MRI is a group of layered, continuous, multiple two-dimensional images; on the space coordinate axis of the section, the distance between layers is 0.5-1mm . When carrying out th...

Embodiment 3

[0068] An MRI machine is used to scan with a SWI sequence, and usually images of veins are shown on the images. However, in the actual work of medical personnel, it is often necessary to observe both venous blood vessels and arterial blood vessels at the same time.

[0069] In order to solve the above-mentioned technical problems, on the basis of Embodiment 1 and Embodiment 2, the technical solution adopted is: when using the nuclear magnetic resonance instrument to scan with the SWI sequence, apply an inclination angle to the second echo of the SWI, Three-dimensional fully flow-compensated arterial angiography was performed. In this way, arterial vessels and venous vessels can be clearly imaged at the same time, and a SWI image that clearly displays both arterial vessels and venous vessels on the same image can be obtained. Preferably, the applied inclination angle is 15 degrees to 25 degrees.

[0070] In this way, for the SWI image that clearly shows arterial blood vessels...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention provides a blood vessel identification method based on an SWI image and a recurrent neural network. The blood vessel identification method comprises the following steps: S1, scanning toobtain an SWI image synthesized by multiple groups of original magnetic moment images and original phase images generated by pairing; S2, carrying out the image segmentation of the SWI image through an edge detection algorithm, and obtaining a to-be-recognized image; S3, selecting a part of the to-be-identified image and an original phase image generated by pairing the part of the to-be-identifiedimage as a training set; the remaining to-be-identified images and original phase images generated by pairing the remaining to-be-identified images are used as a test set in an artificial identification mode; S4, inputting the training set into a recurrent neural network, and associating the to-be-identified image with the original phase image for image identification to obtain a blood vessel identification model; and S5, performing image segmentation on the SWI image needing to be subjected to blood vessel identification according to the step S2, and inputting the SWI image into the blood vessel identification model to obtain a blood vessel structure image; the technical problem of automatic identification of the blood vessel in the SWI image can be solved.

Description

technical field [0001] The invention relates to the technical field of nuclear magnetic resonance imaging, in particular to a blood vessel recognition method based on SWI images and a circulatory neural network. Background technique [0002] In nuclear magnetic resonance imaging technology, SWI susceptibility weighted imaging is a newly developed magnetic resonance contrast-enhanced imaging technology in recent years. SWI provides image contrast enhancement based on differences in susceptibility between different tissues, and it can be applied to all sequences that are sensitive to magnetization effects between different tissues or between subvoxels. [0003] Prior art CN107248155A discloses a method for segmenting cerebral veins and vessels based on SWI images. The method reads each two-dimensional SWI cerebral veins and vessels image, and uses image processing algorithms to eliminate noises such as low contrast of veins and vessels and brain artifacts. And so on the impac...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
IPC IPC(8): G06T7/00G06T7/13G06T7/136G06T11/00
CPCG06T7/0012G06T7/13G06T7/136G06T11/003G06T2207/10088G06T2207/20081G06T2207/20084G06T2207/30101
Inventor 陈家飞王洁甄志铭何敏
Owner THE FIRST AFFILIATED HOSPITAL OF ARMY MEDICAL UNIV
Features
  • R&D
  • Intellectual Property
  • Life Sciences
  • Materials
  • Tech Scout
Why Patsnap Eureka
  • Unparalleled Data Quality
  • Higher Quality Content
  • 60% Fewer Hallucinations
Social media
Patsnap Eureka Blog
Learn More

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2025 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com