Magnetic resonance multi-channel reconstruction method based on deep learning and data self-consistence

A technology of deep learning and magnetic resonance, applied in magnetic resonance measurement, measurement using nuclear magnetic resonance image system, measurement of magnetic variables, etc., can solve the problems of low acceleration factor, low processing capacity of high-magnification undersampling data, etc., to improve quality , improve the generalization ability and shorten the time

Active Publication Date: 2018-09-14
朱高杰
View PDF5 Cites 27 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0006] The object of the present invention is: the present invention provides a magnetic resonance multi-channel reconstruction method based on deep learning and data self-consistency, which solves the problem that the existing magnetic reson

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
  • Magnetic resonance multi-channel reconstruction method based on deep learning and data self-consistence
  • Magnetic resonance multi-channel reconstruction method based on deep learning and data self-consistence
  • Magnetic resonance multi-channel reconstruction method based on deep learning and data self-consistence

Examples

Experimental program
Comparison scheme
Effect test

Example Embodiment

[0043] Example 1

[0044] A magnetic resonance multi-channel reconstruction method based on deep learning and data self-consistent, including the following steps:

[0045] Step 1: Collect multi-channel full-sampled training set K-space data, convert the under-sampled K-space data into curled images as input to the reconstruction network, and use the full-sampled K-space data as the training label data of the reconstruction network based on Fully sampled K-space data to generate convolution kernel;

[0046] Step 2: After inputting the rolled image and fusing the under-sampled K-space data and the convolution kernel, the reconstruction network constructed by repeatedly superimposing the SC layer, CNN network and DC layer, using the training label data as the target, and training the network through back propagation The parameter gets the mapping relationship between the input and output of the reconstructed network;

[0047] Step 3: Input the under-sampled test set data into the recons...

Example Embodiment

[0049] Example 2

[0050] Step 1 includes the following steps:

[0051] Step 1.1: Use the multi-channel receiving coil to collect the K-space data of the full-sample training set, and generate the under-sampled K-space data S through artificial under-sampling u , Its size is N x *N y *N c , Where N x Represents the number of rows of collected data, N y Represents the number of columns of data, N c Represents the number of receiving channels; manual under-sampling does not require manual operation by humans, and can be automated with a variety of algorithms; in the training phase, the under-sampling data is generated by discarding some of the full-sampled data; in the testing phase, the magnetic resonance scanner Under-sampled data can be obtained directly; a full-sampled data can only have one corresponding under-sampled data, so in actual operation, after a full-sampled data is obtained, the under-sampled data is generated artificially on this basis.

[0052] Step 1.2: Undersample K...

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 discloses a magnetic resonance multi-channel reconstruction method based on deep learning and data self-consistence and belongs to the magnetic resonance reconstruction method field. Themethod includes the following steps that: multi-channel full-sampling training set data are collected, under-sampled data are converted into wrap-around images which are used as the input of a reconstruction network, and full-sampled data are used as training mark data, and a convolution kernel is generated based the full-sampled data; 2, the data in the step 1 are inputted, on the basis of the reconstruction network constructed by means of the repeated superposition of SC layers, CNNs, and DC layers, and with the training mark data adopted as an objective, network parameters are trained withback propagation, so that mapping relationships between the input and output of the reconstruction network are obtained; and 3, test set data are inputted into the reconstruction network so as to besubjected to forward propagation, so that unknown mapping data are obtained, and the reconstruction of magnetic resonance is completed. With the method adopted, a problem that an existing resonance reconstruction method can only process single-channel data can be solved; more stable and accurate end-to-end mapping relationships can be obtained; the quality of magnetic resonance reconstruction canbe fundamentally improved; and magnetic resonance scanning time can be significantly shortened.

Description

technical field [0001] The invention relates to the field of magnetic resonance reconstruction methods, in particular to a magnetic resonance multi-channel reconstruction method based on deep learning and data self-consistency. Background technique [0002] Magnetic resonance imaging technology is a technology that uses the nuclear magnetic resonance phenomenon of hydrogen protons for imaging; atomic nuclei that contain a single number of protons in the human body, such as hydrogen nuclei that exist widely, have spin motion; the spin motion of charged atomic nuclei, in physics It is similar to a single small magnet in appearance, and the directional distribution of the small magnet is random without the influence of external conditions; when the human body is placed in an external magnetic field, the small magnet will be rearranged according to the magnetic force lines of the external magnetic field, specifically parallel to or Arranged in two directions antiparallel to the ...

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): G01R33/56
CPCG01R33/56
Inventor 朱高杰
Owner 朱高杰
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap