Unlock instant, AI-driven research and patent intelligence for your innovation.

Magnetic Resonance Image Reconstruction Method Based on Tensor Dictionary Learning

A technology of magnetic resonance image and dictionary learning, which is used in image enhancement, image analysis, image data processing, etc.

Active Publication Date: 2019-06-04
SOUTHERN MEDICAL UNIVERSITY
View PDF4 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to provide a kind of magnetic resonance image reconstruction method based on tensor dictionary learning to improve the reconstructed image quality

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 Image Reconstruction Method Based on Tensor Dictionary Learning
  • Magnetic Resonance Image Reconstruction Method Based on Tensor Dictionary Learning
  • Magnetic Resonance Image Reconstruction Method Based on Tensor Dictionary Learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0040] A method for reconstructing magnetic resonance images based on tensor dictionary learning, comprising the steps of:

[0041] (1) The original k-space data is obtained by random undersampling with variable density, and the inverse Fourier transform is performed on the sampled data to obtain the initial reconstructed image;

[0042] (2) Establish a compressed sensing reconstruction model based on tensor dictionary learning;

[0043] (3) performing tensor dictionary learning on the randomly extracted part of the three-dimensional sub-image blocks of the reconstructed image to obtain a tensor dictionary for sparse representation;

[0044] (4) carry out the sparse representation under the tensor dictionary to all sub-image blocks with the hard domain value method;

[0045] (5) update the reconstructed image with the least squares method;

[0046] (6) Repeat steps (3)-(5) until convergence to obtain the final reconstructed image.

[0047] In the above step (2), the reconstru...

Embodiment 2

[0069] Take heart computer data as an example, such as figure 1 , figure 2 As shown, under different under-sampling factors for heart computer data, the magnetic resonance image reconstruction method based on tensor dictionary learning of the present invention is used to carry out, specifically comprising the following steps:

[0070] (1) Obtain fully sampled original k-space data by magnetic resonance scanning, and retrospectively undersample the k-space data according to given different undersampling factors to obtain undersampled k-space data Y; for the k-space data Y performs zero-padded Fourier reconstruction to obtain the initial value of the reconstructed image X, and at the same time, let the initial value of G be a zero matrix.

[0071] (2) Establish a compressed sensing reconstruction model based on tensor dictionary learning:

[0072]

[0073] Among them, ||·|| 0 Represents the zero norm, defined by counting the number of non-zero elements, ||·|| F Represent...

Embodiment 3

[0092] Such as image 3 As shown, in Embodiment 3, for and perfusion imaging data under a given undersampling factor, a magnetic resonance image reconstruction method based on tensor dictionary learning is provided, the method includes the following steps:

[0093] (1) For the fully sampled k-space data of the simulation, according to a given undersampling factor, the k-space data is retrospectively undersampled to obtain the undersampled k-space data Y; the k-space data Y is zero-filled Fourier reconstruction, the initial value of the reconstructed image X is obtained, and the initial value of Γ is a zero matrix.

[0094] (2) set up a compressed sensing reconstruction model based on tensor dictionary learning, such as (1) formula;

[0095] (3) Considering X and G as known constants, formula (I) is changed into the following optimization problem:

[0096] Consider X and G as known constants, and change the formula (I) into the following formula (II):

[0097]

[0098] in...

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 relates to a magnetic resonance image reconstruction method based on tensor dictionary learning. The magnetic resonance image reconstruction method is characterized in that (1) original k space data is acquired by adopting a variable density random undersampling way, and inverse Fourier transform of sampling data is carried out to acquire an initial reconstruction image; (2) a compressed sensing reconstruction model is established based on the tensor dictionary learning; (3) a part of three-dimensional sub-image blocks are extracted from a reconstructed image for the tensor dictionary learning, and then a tensor dictionary used for sparse expression is acquired; (4) the sparse expression of the tensor dictionary is used for all of the sub-image blocks by adopting a hard domain method; (5) the reconstructed image is updated by adopting a least square method; (6) the step (3) to the step (5) are repeated until convergence is realized, and the final reconstructed image is acquired. The magnetic resonance image reconstruction method based on the tensor dictionary learning is advantageous in that the reconstructed image quality is improved, and the calculation is simple.

Description

technical field [0001] The invention relates to the technical field of magnetic resonance imaging, in particular to a magnetic resonance image reconstruction method based on tensor dictionary learning under the compressed sensing theory. Background technique [0002] Since Olshausen et al. published a pioneering paper on sparse coding of natural images in top international journals such as Nature in 1996, people have paid more and more attention to dictionary learning. Olshausen et al. derived the l 1 The norm is used as a measure of coefficient sparsity. What is surprising is that the morphology of each atom in the dictionary obtained by learning the sparsity is similar to the feeling of simple cells in the V1 area of ​​the visual cortex. Their research results have laid the foundation for The neurophysiological basis of sparse coding. [0003] In recent years, solving the image inverse problem with signal sparsity prior has attracted extensive attention from scholars, es...

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
Patent Type & Authority Patents(China)
IPC IPC(8): G06T5/00
CPCG06T5/003G06T2207/10088G06T2207/20056G06T2207/20081
Inventor 冯衍秋黄进红冯前进陈武凡
Owner SOUTHERN MEDICAL UNIVERSITY