CN114518555A - Nuclear Overhauser enhanced imaging method and system - Google Patents
Nuclear Overhauser enhanced imaging method and system Download PDFInfo
- Publication number
- CN114518555A CN114518555A CN202210127212.0A CN202210127212A CN114518555A CN 114518555 A CN114518555 A CN 114518555A CN 202210127212 A CN202210127212 A CN 202210127212A CN 114518555 A CN114518555 A CN 114518555A
- Authority
- CN
- China
- Prior art keywords
- noe
- simulated
- graph
- exchange rate
- concentration
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000003384 imaging method Methods 0.000 title claims abstract description 32
- 238000012549 training Methods 0.000 claims abstract description 82
- 238000013528 artificial neural network Methods 0.000 claims abstract description 39
- 238000000034 method Methods 0.000 claims abstract description 34
- 238000010586 diagram Methods 0.000 claims abstract description 25
- 238000007781 pre-processing Methods 0.000 claims abstract description 17
- 238000001914 filtration Methods 0.000 claims description 24
- 238000005070 sampling Methods 0.000 claims description 16
- 238000004088 simulation Methods 0.000 claims description 12
- 238000009499 grossing Methods 0.000 claims description 11
- 238000012545 processing Methods 0.000 claims description 10
- 229920006395 saturated elastomer Polymers 0.000 claims description 10
- 238000010606 normalization Methods 0.000 claims description 4
- 230000000694 effects Effects 0.000 description 9
- 238000002595 magnetic resonance imaging Methods 0.000 description 9
- 238000001208 nuclear magnetic resonance pulse sequence Methods 0.000 description 6
- 238000004364 calculation method Methods 0.000 description 5
- 238000001228 spectrum Methods 0.000 description 5
- 239000011159 matrix material Substances 0.000 description 4
- 239000000126 substance Substances 0.000 description 4
- 238000012546 transfer Methods 0.000 description 4
- 239000013598 vector Substances 0.000 description 3
- 235000013601 eggs Nutrition 0.000 description 2
- 230000005284 excitation Effects 0.000 description 2
- 238000012935 Averaging Methods 0.000 description 1
- 206010028980 Neoplasm Diseases 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 238000007796 conventional method Methods 0.000 description 1
- 238000013527 convolutional neural network Methods 0.000 description 1
- 238000000354 decomposition reaction Methods 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 230000005415 magnetization Effects 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 230000001575 pathological effect Effects 0.000 description 1
- 238000000513 principal component analysis Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 238000011002 quantification Methods 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R33/00—Arrangements or instruments for measuring magnetic variables
- G01R33/20—Arrangements or instruments for measuring magnetic variables involving magnetic resonance
- G01R33/44—Arrangements or instruments for measuring magnetic variables involving magnetic resonance using nuclear magnetic resonance [NMR]
- G01R33/48—NMR imaging systems
- G01R33/54—Signal processing systems, e.g. using pulse sequences ; Generation or control of pulse sequences; Operator console
- G01R33/56—Image enhancement or correction, e.g. subtraction or averaging techniques, e.g. improvement of signal-to-noise ratio and resolution
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Biophysics (AREA)
- Molecular Biology (AREA)
- Biomedical Technology (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Computational Linguistics (AREA)
- Software Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Radiology & Medical Imaging (AREA)
- Signal Processing (AREA)
- High Energy & Nuclear Physics (AREA)
- Condensed Matter Physics & Semiconductors (AREA)
- Magnetic Resonance Imaging Apparatus (AREA)
Abstract
The invention relates to a nuclear Overhauser enhanced imaging method and a nuclear Overhauser enhanced imaging system. The method comprises the following steps: acquiring a CEST image to be processed, and preprocessing the CEST image to obtain a preprocessed CEST image; obtaining the value range of each sample parameter according to the preprocessed CEST image and the Bloch-McConnell equation model; generating a training sample with a set amount according to the value range of the sample parameter; training the deep neural network by adopting a training sample to obtain a trained deep neural network; and inputting the preprocessed CEST image into the trained deep neural network to obtain an NOE comparison graph, wherein the NOE comparison graph comprises an NOE concentration graph and an NOE exchange rate graph. The invention can quickly reconstruct a high-quality NOE contrast diagram.
Description
Technical Field
The invention relates to the technical field of magnetic resonance imaging, in particular to a nuclear Overhauser enhanced imaging method and a nuclear Overhauser enhanced imaging system.
Background
The Nuclear Overhauser Enhancement (NOE) effect is a derivative of the Magnetization Transfer (MT) effect produced by cross-relaxation, and NOE provides a new contrast mechanism for Magnetic Resonance Imaging (MRI). Although the mechanism of the direct chemical exchange action of the NOE effect and the Chemical Exchange Saturation Transfer (CEST) effect is different, the NOE effect and the CEST effect can reduce signals in a specific frequency range in a Z spectrum, the NOE imaging result has strong specificity, can indirectly reflect physiological and pathological information of tissues, and has important application value in the aspects of disease diagnosis, tumor grading and the like. The conventional NOE comparison graph calculation methods include a three-point offset method, a Lorentz difference method, a Bloch-McConnell equation fitting method, a multi-pool Lorentz fitting method and the like. These methods have respective limitations, for example, the three-point offset method and the lorentz difference method cannot avoid the influence of the MT effect on the NOE quantification, and have larger errors in tissues with more fat; the Bloch-McConnell equation fitting method is accurate in calculation, but strict requirements on the range and initial values of fitting parameters and sensitivity to noise are achieved, so that the calculated image is not smooth; the multi-cell lorentz fitting method is sensitive to noise and needs to acquire more frequency points, and particularly when the number of cells is increased, the calculation precision is reduced and the calculation time is increased. Therefore, more efficient methods need to be investigated to reconstruct higher quality NOE contrast maps.
Disclosure of Invention
The invention aims to provide a nuclear Overhauser enhanced imaging method and a nuclear Overhauser enhanced imaging system, which can quickly reconstruct a high-quality NOE contrast map.
In order to achieve the purpose, the invention provides the following scheme:
a method of nuclear Overhauser enhanced imaging comprising:
acquiring a CEST image to be processed;
preprocessing the CEST image to be processed to obtain a preprocessed CEST image;
selecting a plurality of regions of interest in the preprocessed CEST image;
fitting and solving the interested regions respectively by adopting a Bloch-McConnell equation model to obtain the values of the sample parameters in the interested regions; the sample parameters include NOE concentration, NOE exchange rate, magnetic field non-uniformity, and radio frequency field non-uniformity;
determining the value range of each sample parameter according to the maximum value of the values of each sample parameter in all the regions of interest;
generating a training sample with a set amount according to the value range of each sample parameter; the training sample comprises a modulated simulated NOE concentration graph, a modulated simulated NOE exchange rate graph and a simulated CEST image added with noise;
according to each training sample, taking the modulated analog NOE concentration graph and the modulated analog NOE exchange rate graph as labels, and taking the simulated CEST image added with noise as input to train a deep neural network to obtain a trained deep neural network;
and inputting the preprocessed CEST image into the trained deep neural network to obtain an NOE comparison graph.
Optionally, the preprocessing the CEST image to be processed to obtain a preprocessed CEST image specifically includes:
normalizing the CEST image to be processed by taking the saturated image under the maximum frequency offset in the CEST image to be processed as a reference to obtain a normalized CEST image;
filtering the normalized CEST image to obtain a filtered CEST image;
and smoothing the filtered CEST image to obtain a preprocessed CEST image.
Optionally, the generating a set amount of training samples according to the value range of each sample parameter specifically includes:
performing a training sample generation process a plurality of times;
the training sample generation process comprises the following steps:
obtaining a simulated NOE concentration graph and a simulated NOE exchange rate graph according to the value range of the NOE concentration and the value range of the NOE exchange rate;
respectively modulating the analog NOE concentration graph and the analog NOE exchange rate graph by adopting a natural image with textures subjected to filtering processing so as to enable the analog NOE concentration graph and the analog NOE exchange rate graph to generate textures, and obtaining a modulated analog NOE concentration graph and a modulated analog NOE exchange rate graph;
generating a simulated magnetic field map and a simulated radio frequency field map according to the value range of the magnetic field unevenness and the value range of the radio frequency field unevenness;
inputting the modulated simulated NOE concentration graph, the modulated simulated NOE exchange rate graph, the simulated magnetic field graph and the simulated radio frequency field graph into a Bloch-McConnell equation model to obtain a simulated CEST image;
adding random noise into the simulated CEST image to obtain a simulated CEST image added with noise;
and determining the modulated simulated NOE concentration graph, the modulated simulated NOE exchange rate graph and the simulated CEST image added with noise as training samples.
Optionally, the fitting and solving are respectively performed on each region of interest by using a Bloch-McConnell equation model to obtain a value of a sample parameter in each region of interest, and the method specifically includes:
acquiring an equation solution type and pool parameters of a Bloch-McConnell equation model and sampling parameters when the CEST image to be processed is generated;
inputting the sampling parameters, the pool parameters and the equation solution type into the Bloch-McConnell equation model to obtain a determined Bloch-McConnell equation model;
and respectively fitting and solving the interested regions by adopting the determined Bloch-McConnell equation model to obtain the values of the sample parameters in the interested regions.
Optionally, the obtaining of the simulated NOE concentration map and the simulated NOE exchange rate map according to the value range of the NOE concentration and the value range of the NOE exchange rate specifically includes:
creating a blank template;
and respectively generating NOE concentration and NOE exchange rate randomly pixel by pixel in the blank template according to the value range of the NOE concentration and the value range of the NOE exchange rate to obtain a simulated NOE concentration graph and a simulated NOE exchange rate graph.
A nuclear Overhauser enhanced imaging system, comprising:
the image acquisition module is used for acquiring a CEST image to be processed;
the image preprocessing module is used for preprocessing the CEST image to be processed to obtain a preprocessed CEST image;
a region-of-interest determination module for selecting a plurality of regions-of-interest in the preprocessed CEST image;
the sample parameter value determining module is used for respectively fitting and solving the interested regions by adopting a Bloch-McConnell equation model to obtain the values of the sample parameters in the interested regions; the sample parameters include NOE concentration, NOE exchange rate, magnetic field non-uniformity, and radio frequency field non-uniformity;
the sample parameter value range determining module is used for determining the value range of each sample parameter according to the maximum value of the values of each sample parameter in all the regions of interest;
the training sample generating module is used for generating a set amount of training samples according to the value range of each sample parameter; the training sample comprises a modulated simulated NOE concentration graph, a modulated simulated NOE exchange rate graph and a simulated CEST image added with noise;
the network training module is used for training a deep neural network by taking the modulated simulated NOE concentration graph and the modulated simulated NOE exchange rate graph as labels and taking the simulated CEST image added with noise as input according to each training sample to obtain the trained deep neural network;
and the NOE contrast map reconstruction module is used for inputting the preprocessed CEST image into the trained deep neural network to obtain the NOE contrast map.
Optionally, the image preprocessing module specifically includes:
the normalization submodule is used for normalizing the CEST image to be processed by taking the saturated image under the maximum frequency offset in the CEST image to be processed as a reference to obtain a normalized CEST image;
the filtering submodule is used for filtering the normalized CEST image to obtain a filtered CEST image;
and the smoothing submodule is used for smoothing the filtered CEST image to obtain a preprocessed CEST image.
Optionally, the training sample generating module specifically includes:
the training sample generation submodule is used for executing the training sample generation process for multiple times;
the training sample generation submodule includes:
the NOE concentration graph and NOE exchange rate graph determining unit is used for obtaining a NOE concentration graph and a NOE exchange rate graph according to the value range of the NOE concentration and the value range of the NOE exchange rate;
the modulation unit is used for respectively modulating the analog NOE concentration diagram and the analog NOE exchange rate diagram by adopting a natural image with textures subjected to filtering processing so as to enable the analog NOE concentration diagram and the analog NOE exchange rate diagram to generate textures, and obtaining a modulated analog NOE concentration diagram and a modulated analog NOE exchange rate diagram;
the simulated field map determining unit is used for generating a simulated magnetic field map and a simulated radio frequency field map according to the value range of the magnetic field unevenness and the value range of the radio frequency field unevenness;
the simulated CEST image determining unit is used for inputting the modulated simulated NOE concentration graph, the modulated simulated NOE exchange rate graph, the simulated magnetic field graph and the simulated radio frequency field graph into a Bloch-McConnell equation model to obtain a simulated CEST image;
the noise unit is used for adding random noise into the simulated CEST image to obtain a simulated CEST image added with noise;
and the training sample determining unit is used for determining the modulated simulated NOE concentration graph, the modulated simulated NOE exchange rate graph and the simulated CEST image added with noise as training samples.
Optionally, the sample parameter value determining module specifically includes:
the acquisition submodule is used for acquiring the equation solution type and the pool parameter of the Bloch-McConnell equation model and the sampling parameter when the CEST image to be processed is generated;
the Bloch-McConnell equation model determining submodule is used for inputting the sampling parameters, the pool parameters and the equation solution types into the Bloch-McConnell equation model to obtain a determined Bloch-McConnell equation model;
and the sample parameter value determining submodule is used for respectively fitting and solving the interested regions by adopting the determined Bloch-McConnell equation model to obtain the values of the sample parameters in the interested regions.
Optionally, the unit for determining the simulated NOE concentration map and the simulated NOE exchange rate map specifically includes:
the template creating subunit is used for creating a blank template;
and the NOE concentration simulation graph and NOE exchange rate simulation graph determining subunit is used for respectively randomly generating the NOE concentration and the NOE exchange rate by pixel points in the blank template according to the NOE concentration value range and the NOE exchange rate value range to obtain a NOE concentration simulation graph and a NOE exchange rate simulation graph.
According to the specific embodiment provided by the invention, the invention discloses the following technical effects: selecting a plurality of interested areas in a preprocessed CEST image; respectively fitting and solving each region of interest by adopting a Bloch-McConnell equation model to obtain the value of the sample parameter in each region of interest; determining the value range of each sample parameter according to the maximum value of the values of each sample parameter in all the regions of interest; generating training samples with set amount according to the value range of each sample parameter; according to each training sample, taking the modulated analog NOE concentration graph and the modulated analog NOE exchange rate graph as labels, and taking the analog CEST image added with noise as input to train the deep neural network to obtain a trained deep neural network; and inputting the preprocessed CEST image into a trained deep neural network to obtain an NOE comparison graph. Because the Bloch-McConnell equation model has the advantage of high precision and the deep neural network has the advantage of high processing speed, the quality and speed of the reconstructed NOE contrast diagram are improved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings needed to be used in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings without inventive exercise.
Fig. 1 is a flowchart of a nuclear Overhauser enhanced imaging method according to an embodiment of the present invention;
fig. 2 is a structural diagram of a nuclear Overhauser enhanced imaging system according to an embodiment of the present invention;
FIG. 3 is a comparison of NOE of eggs calculated by a conventional method;
fig. 4 is a comparison of NOE of eggs obtained by the imaging method provided by the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
In order to make the aforementioned objects, features and advantages of the present invention comprehensible, embodiments accompanied with figures are described in further detail below.
The embodiment of the invention provides a nuclear Overhauser enhanced imaging method, as shown in fig. 1, the method comprises the following steps:
step 101: acquiring a CEST image to be processed; and the CEST image to be processed consists of saturated images under a plurality of frequency deviations.
Step 102: and preprocessing the CEST image to be processed to obtain a preprocessed CEST image.
Step 103: selecting a plurality of regions of interest in the preprocessed CEST image.
Step 104: fitting and solving the interested regions respectively by adopting a Bloch-McConnell equation model to obtain the values of the sample parameters in the interested regions; the sample parameters include NOE concentration, NOE exchange rate, magnetic field inhomogeneity, and radio frequency field inhomogeneity.
Step 105: and determining the value range of each sample parameter according to the maximum value of the values of each sample parameter in all the regions of interest.
Step 106: generating a training sample with a set amount according to the value range of each sample parameter; the training sample comprises a modulated simulated NOE concentration graph, a modulated simulated NOE exchange rate graph and a simulated CEST image added with noise.
Step 107: and according to each training sample, training a deep neural network by taking the modulated simulated NOE concentration graph and the modulated simulated NOE exchange rate graph as labels and taking the simulated CEST image added with noise as input to obtain the trained deep neural network.
Step 108: and inputting the preprocessed CEST image into the trained deep neural network to obtain an NOE comparison graph, wherein the NOE comparison graph comprises an NOE concentration graph and an NOE exchange rate graph.
In practical application, the preprocessing the CEST image to be processed to obtain a preprocessed CEST image specifically includes:
and normalizing the CEST image to be processed by taking the saturated image under the maximum frequency offset in the CEST image to be processed as a reference to obtain a normalized CEST image.
And filtering the normalized CEST image to obtain a filtered CEST image.
And smoothing the filtered CEST image to obtain a preprocessed CEST image.
In practical application, the generating of the training samples with the set amount according to the value ranges of the sample parameters specifically includes:
the training sample generation process is performed multiple times.
The training sample generation process comprises the following steps:
and obtaining a simulated NOE concentration graph and a simulated NOE exchange rate graph according to the value range of the NOE concentration and the value range of the NOE exchange rate.
And respectively modulating the analog NOE concentration graph and the analog NOE exchange rate graph by adopting a natural image with textures subjected to filtering processing so as to enable the analog NOE concentration graph and the analog NOE exchange rate graph to generate textures, and obtaining a modulated analog NOE concentration graph and a modulated analog NOE exchange rate graph.
And generating a simulated magnetic field map and a simulated radio frequency field map according to the value range of the magnetic field unevenness and the value range of the radio frequency field unevenness.
And inputting the modulated simulated NOE concentration graph, the modulated simulated NOE exchange rate graph, the simulated magnetic field graph and the simulated radio frequency field graph into a Bloch-McConnell equation model to obtain a simulated CEST image.
And adding random noise into the simulated CEST image to obtain the simulated CEST image added with the noise.
And determining the modulated simulated NOE concentration graph, the modulated simulated NOE exchange rate graph and the simulated CEST image added with noise as training samples.
In practical application, the fitting solution is respectively performed on each region of interest by using a Bloch-McConnell equation model to obtain a value of a sample parameter in each region of interest, and the method specifically includes:
and acquiring the equation solution type and the pool parameter of the Bloch-McConnell equation model and the sampling parameter when the CEST image to be processed is generated.
And inputting the sampling parameters, the pool parameters and the equation solution type into the Bloch-McConnell equation model to obtain the determined Bloch-McConnell equation model.
And respectively fitting and solving the interested regions by adopting the determined Bloch-McConnell equation model to obtain the values of the sample parameters in the interested regions.
In practical application, the obtaining of the simulated NOE concentration map and the simulated NOE exchange rate map according to the value range of the NOE concentration and the value range of the NOE exchange rate specifically includes:
a blank template is created.
And respectively generating NOE concentration and NOE exchange rate randomly pixel by pixel in the blank template according to the value range of the NOE concentration and the value range of the NOE exchange rate to obtain a simulated NOE concentration graph and a simulated NOE exchange rate graph.
In practical application, the step of modulating the simulated NOE concentration map and the simulated NOE exchange rate map by using the filtered textured natural image to generate textures, so as to obtain the modulated simulated NOE concentration map and the modulated simulated NOE exchange rate map, which specifically includes:
and filtering the natural image with the texture to obtain a natural image with the texture after filtering.
And respectively modulating the analog NOE concentration graph and the analog NOE exchange rate graph by adopting the natural image with the texture after the filtering treatment to obtain a modulated analog NOE concentration graph and a modulated analog NOE exchange rate graph.
The invention also provides a nuclear Overhauser enhanced imaging system corresponding to the method, as shown in fig. 2, the system comprises:
an image acquisition module a1 for acquiring a CEST image to be processed.
And the image preprocessing module A2 is configured to preprocess the CEST image to be processed to obtain a preprocessed CEST image.
A region-of-interest determining module a3 for selecting a plurality of regions of interest in the preprocessed CEST image.
The sample parameter value determining module A4 is used for respectively fitting and solving the interested regions by adopting a Bloch-McConnell equation model to obtain the values of the sample parameters in the interested regions; the sample parameters include NOE concentration, NOE exchange rate, magnetic field inhomogeneity, and radio frequency field inhomogeneity.
And the sample parameter value range determining module a5 is configured to determine the value range of each sample parameter according to the maximum value of the values of each sample parameter in all the regions of interest.
A training sample generation module A6, configured to generate a set amount of training samples according to the value range of each sample parameter; the training sample comprises a modulated simulated NOE concentration graph, a modulated simulated NOE exchange rate graph and a simulated CEST image added with noise.
And the network training module A7 is configured to train the deep neural network to obtain a trained deep neural network by using the modulated simulated NOE concentration map and the modulated simulated NOE exchange rate map as labels and using the simulated CEST image after noise addition as an input according to each training sample.
And an NOE contrast map reconstruction module A8, configured to input the preprocessed CEST image into the trained deep neural network to obtain an NOE contrast map.
As an optional implementation manner, the image preprocessing module specifically includes:
and the normalization submodule is used for normalizing the CEST image to be processed by taking the saturated image under the maximum frequency offset in the CEST image to be processed as a reference to obtain the normalized CEST image.
And the filtering submodule is used for filtering the normalized CEST image to obtain a filtered CEST image.
And the smoothing submodule is used for smoothing the filtered CEST image to obtain a preprocessed CEST image.
As an optional implementation manner, the training sample generation module specifically includes:
and the training sample generation submodule is used for executing the training sample generation process for multiple times.
The training sample generation submodule includes:
and the NOE concentration graph and NOE exchange rate graph determining unit is used for obtaining a NOE concentration graph and a NOE exchange rate graph according to the value range of the NOE concentration and the value range of the NOE exchange rate.
And the modulation unit is used for respectively modulating the analog NOE concentration graph and the analog NOE exchange rate graph by adopting the natural image with the texture after filtering processing so as to enable the analog NOE concentration graph and the analog NOE exchange rate graph to generate the texture, and obtaining the modulated analog NOE concentration graph and the modulated analog NOE exchange rate graph.
And the simulated field map determining unit is used for generating a simulated magnetic field map and a simulated radio frequency field map according to the value range of the magnetic field unevenness and the value range of the radio frequency field unevenness.
And the simulated CEST image determining unit is used for inputting the modulated simulated NOE concentration graph, the modulated simulated NOE exchange rate graph, the simulated magnetic field graph and the simulated radio frequency field graph into a Bloch-McConnell equation model to obtain a simulated CEST image.
And the noise unit is used for adding random noise to the simulated CEST image to obtain the simulated CEST image added with the noise.
And the training sample determining unit is used for determining the modulated simulated NOE concentration graph, the modulated simulated NOE exchange rate graph and the simulated CEST image added with noise as training samples.
As an optional implementation manner, the sample parameter value determining module specifically includes:
and the acquisition submodule is used for acquiring the equation solution type and the pool parameter of the Bloch-McConnell equation model and the sampling parameter when the CEST image to be processed is generated.
And the Bloch-McConnell equation model determining submodule is used for inputting the sampling parameters, the pool parameters and the equation solution types into the Bloch-McConnell equation model to obtain the determined Bloch-McConnell equation model.
And the sample parameter value determining submodule is used for respectively fitting and solving the interested regions by adopting the determined Bloch-McConnell equation model to obtain the values of the sample parameters in the interested regions.
As an optional implementation manner, the determining unit for the simulated NOE concentration map and the simulated NOE exchange rate map specifically includes:
and the template creating subunit is used for creating a blank template.
And the NOE concentration simulation graph and NOE exchange rate simulation graph determining subunit is used for respectively randomly generating the NOE concentration and the NOE exchange rate by pixel points in the blank template according to the NOE concentration value range and the NOE exchange rate value range to obtain a NOE concentration simulation graph and a NOE exchange rate simulation graph.
As an optional implementation manner, the modulation unit specifically includes:
and the filtering processing subunit is used for carrying out filtering processing on the natural image with the texture to obtain a natural image with the texture after the filtering processing.
And the modulation subunit is used for respectively modulating the analog NOE concentration diagram and the analog NOE exchange rate diagram by adopting the filtered and textured natural image to obtain a modulated analog NOE concentration diagram and a modulated analog NOE exchange rate diagram.
The embodiment of the invention also provides a more detailed nuclear Overhauser enhanced imaging method (the nuclear Overhauser enhanced imaging method based on a Bloch-McConnell equation and a deep neural network), which mainly comprises the following steps: acquiring Chemical Exchange Saturation Transfer (CEST) magnetic resonance imaging data to obtain an original CEST image; preprocessing the original CEST image to obtain a preprocessed CEST image; obtaining NOE concentration, NOE exchange rate, magnetic field unevenness and radio frequency field unevenness parameter ranges of a training sample according to the preprocessed CEST image; generating a training sample of the deep neural network according to the NOE concentration, the NOE exchange rate, the magnetic field unevenness and the radio frequency field unevenness parameter range of the training sample; constructing a deep neural network for reconstructing the NOE contrast map; training the deep neural network by adopting the training sample to obtain a trained deep neural network; and inputting the preprocessed CEST image into the trained deep neural network for reconstruction to obtain an NOE comparison graph comprising an NOE concentration graph and an NOE exchange rate graph. The method comprises the following specific steps:
s1: chemical Exchange Saturation Transfer (CEST) magnetic resonance imaging data are acquired, resulting in an original CEST image (CEST image to be processed).
S1 specifically includes:
s11: a pulse sequence for acquiring CEST magnetic resonance imaging data is determined.
S12: and determining sampling parameters including the shape of the radio frequency saturation pulse, the number of the radio frequency saturation pulses, the intensity of the radio frequency saturation pulses, the duration of the radio frequency saturation pulses, a frequency deviation list of the radio frequency saturation pulses, the flip angle of the radio frequency excitation pulses, the pulse sequence repetition time and the pulse sequence echo time.
S13: and performing magnetic resonance imaging scanning on the imaging object under the sampling parameter in S12 by adopting the pulse sequence in S11 to obtain an original CEST image, wherein the original CEST image consists of saturated images under a plurality of frequency deviations.
S2: and preprocessing the original CEST image in the S1 to obtain a preprocessed CEST image.
S2 specifically includes:
s21: and normalizing the original CEST image by taking the saturated image under the maximum frequency offset in the CEST image to be processed as a reference to obtain a normalized CEST image.
S22: filtering the CEST image normalized in S21.
S23: and smoothing the filtered image in the step S22 to obtain a preprocessed CEST image.
S3: and acquiring NOE concentration, NOE exchange rate, magnetic field unevenness and radio frequency field unevenness parameter ranges of the training sample according to the CEST image preprocessed in the S2.
S3 specifically includes:
s31: a plurality of representative regions of interest (ROIs) are selected within the CEST image preprocessed in S2.
S32: and inputting the sampling parameters in the S11 into a Bloch-McConnell equation model.
S33: pool parameters are determined and input into the Bloch-McConnell equation model obtained at S32.
S34: the type of equation solution is determined and input into the Bloch-McConnell equation model obtained in S33.
S35: and fitting and solving the CEST image data in each ROI by adopting the Bloch-McConnell equation model obtained in S34 to obtain solutions of NOE concentration, NOE exchange rate, magnetic field unevenness and radio frequency field unevenness in each ROI.
S36: and determining the upper limit and the lower limit of the parameter ranges according to the maximum value of the solutions of the NOE concentration, the NOE exchange rate, the magnetic field unevenness and the radio frequency field unevenness in each ROI in S35 to obtain the parameter range of the training sample.
S4: and generating the training sample of the deep neural network according to the NOE concentration, the NOE exchange rate, the magnetic field unevenness and the radio frequency field unevenness parameter range of the training sample in the S3.
S4 specifically includes:
s41: a blank rectangular template is created.
S42: and according to the NOE concentration and NOE exchange rate range of the training sample in the S3, randomly generating the NOE concentration and the NOE exchange rate pixel by pixel in the template to obtain a simulated NOE concentration graph and a simulated NOE exchange rate graph.
S43: and modulating the simulated NOE concentration map and the simulated NOE exchange rate map in the S42 by using the filtered textured natural image to generate textures, and obtaining the modulated simulated NOE concentration map and the modulated simulated NOE exchange rate map, wherein the textures are used for simulating the texture of a real imaging object.
S44: generating a simulated magnetic field diagram and a simulated radio frequency field diagram according to the parameter ranges of the magnetic field unevenness and the radio frequency field unevenness of the training sample in the S3; the simulated magnetic field map and the simulated radio frequency field map are used for simulating the magnetic field and radio frequency field conditions of an actual magnetic resonance imaging scan.
S45: and substituting the modulated simulated NOE concentration graph and the modulated simulated NOE exchange rate graph in the S43, and the simulated magnetic field graph and the simulated radio frequency field graph in the S44 into a Bloch-McConnell equation model to obtain a simulated CEST image.
S46: adding random noise to the simulated CEST image in S45, the random noise being used to simulate the noise of an actual magnetic resonance imaging scan.
S47: and taking the modulated simulated NOE concentration graph and the simulated NOE exchange rate graph in the S43 as labels of the neural network, and taking the simulated CEST image added with noise in the S46 as the input of the neural network to obtain a training sample.
S48: repeating S41-S47 generates a set of training samples, the set of training samples comprising a set amount of training samples.
S5: and constructing a deep neural network for reconstructing the NOE comparison graph.
S6: and training the deep neural network in the S5 by adopting the training samples generated in the S4 to obtain the trained deep neural network.
S7: and inputting the CEST image processed in the S2 process into a trained deep neural network in the S6 for reconstruction, and obtaining an NOE contrast diagram which comprises an NOE concentration diagram and an NOE exchange rate diagram.
In practical applications, in S11, the pulse sequence for acquiring the CEST magnetic resonance imaging data is CEST-FSE.
In practical applications, in S12, the sampling parameters are set as follows: the intensity of the radio frequency saturation pulse is 1.2 muT, the duration is 4s, the shape is rectangular, the number is 1, the frequency deviation list is [ -5, -4.75, -4.5, -4.25, -4, -3.75, -3.5, -3.25, -3, -2.75, -2.5, -2.25, -2, -1.75, -1.5, -1.25, -1, -0.75, -0.5, -0.25,0,0.25,0.5,0.75,1,100] (unit: ppm), the flip angle of the radio frequency excitation pulse is 90 degrees, the pulse sequence repetition time is 6s, the echo time is 40ms, the imaging field of view is 70mm x 70mm, and the matrix is 64 x 96.
In practical applications, in S21, the CEST image at 100ppm is a saturated image at the maximum frequency offset, and is used for signal normalization.
In practical application, in S22, filtering is performed by using a principal component analysis algorithm to filter out main noise, reconstructing Z spectrum signal values of pixel-by-pixel of the normalized CEST image by row to obtain a castorati matrix, averaging the matrix by row to obtain an average Z spectrum signal, performing feature decomposition on the matrix to obtain a feature value and a feature vector, determining the number of optimal feature vectors according to the feature value, mapping the determined optimal feature vectors to the difference between the Z spectrum signal of each row and the average Z spectrum signal, and performing inverse reconstruction to obtain the filtered CEST image.
In practical application, in S23, the specific method of the smoothing process is as follows: filling zero into k-space data of the filtered CEST image to 128 x 128, and performing two-dimensional fast Fourier transform to obtain a processed CEST image;
in practical applications, in S31, the number of representative ROIs is 4, and a single ROI includes 10 pixel points.
In practical applications, the pool parameters in S33 are as follows: the cell number is 3, the longitudinal relaxation time T1 corresponding to direct saturation is 1.2s, the transverse relaxation time T2 is uniformly distributed between [15,35] (unit: ms), the frequency offset is 0ppm, T1 corresponding to NOE cell is 2.2s, T2 is 0.4ms, the frequency offset is-3.6 ppm, the concentration of MT cell is 8mol/L, the exchange rate is 30Hz, T1 is 1s, T2 is 9.1 μ s, the frequency offset is-2.4 ppm, the line type of MT cell is super lorentz line type.
In practical applications, in S34, the equation solution type is a numerical solution.
In practical applications, in S36, the parameter ranges are determined as follows: and respectively setting 50% of the minimum value in numerical solutions of NOE concentration, NOE exchange rate, magnetic field unevenness and radio frequency field unevenness in all the ROIs as the lower limit of the parameter range, and setting 150% of the maximum value as the upper limit of the parameter range to cover the real situation, so as to obtain the parameter range of the training sample.
In practical applications, in S43, the textured natural image is subjected to gaussian filtering to obtain a filtered textured natural image.
In practical applications, the number of training samples in S48 is 900.
In practical applications, the random noise in S46 follows a normal distribution with a mathematical expectation of 0 and a standard deviation of 0.0035.
In practical application, the deep neural network in S5 is a U-shaped convolutional neural network (U-Net), the size of an input single training sample is 128 × 128, the reconstruction result is constrained by using an L1 norm, the block size is 64 × 64, the batch number is 8, and the iteration number is 20 ten thousand.
As shown in fig. 3 and 4, in which fig. 3(a) and 3(b) are used as a comparison, and are respectively a NOE concentration graph and a NOE exchange rate graph obtained by fitting a conventional Bloch-McConnell equation, and fig. 4(a) and 4(b) are respectively a NOE concentration graph and a NOE exchange rate graph obtained by reconstruction according to the present invention. As can be seen from fig. 3 and 4, the present invention can better restore the details of the imaged object and solve the problems of image unevenness and non-uniform magnetic field interference caused by noise. Meanwhile, the calculation time of the traditional Bloch-McConnell equation fitting method is 69 minutes, and the result can be reconstructed by the method only needing a few seconds for the same type of imaging objects under the same acquisition condition.
According to the specific embodiment provided by the invention, the invention discloses the following technical effects:
1. the invention realizes the rapid NOE contrast map reconstruction.
2. The NOE obtained by the invention has high quality compared with the image.
3. The invention effectively corrects the influence of the inhomogeneous magnetic field on the NOE contrast diagram.
The embodiments in the present description are described in a progressive manner, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other. For the system disclosed by the embodiment, the description is relatively simple because the system corresponds to the method disclosed by the embodiment, and the relevant points can be referred to the method part for description.
The principles and embodiments of the present invention have been described herein using specific examples, which are provided only to help understand the method and the core concept of the present invention; meanwhile, for a person skilled in the art, according to the idea of the present invention, the specific embodiments and the application range may be changed. In view of the above, the present disclosure should not be construed as limiting the invention.
Claims (10)
1. A nuclear Overhauser enhanced imaging method is characterized by comprising the following steps:
acquiring a CEST image to be processed;
preprocessing the CEST image to be processed to obtain a preprocessed CEST image;
selecting a plurality of regions of interest in the preprocessed CEST image;
fitting and solving the interested regions respectively by adopting a Bloch-McConnell equation model to obtain the values of the sample parameters in the interested regions; the sample parameters include NOE concentration, NOE exchange rate, magnetic field non-uniformity, and radio frequency field non-uniformity;
determining the value range of each sample parameter according to the maximum value of the values of each sample parameter in all the regions of interest;
generating a training sample with a set amount according to the value range of each sample parameter; the training sample comprises a modulated simulated NOE concentration graph, a modulated simulated NOE exchange rate graph and a simulated CEST image added with noise;
according to each training sample, taking the modulated analog NOE concentration graph and the modulated analog NOE exchange rate graph as labels, and taking the simulated CEST image added with noise as input to train a deep neural network to obtain a trained deep neural network;
and inputting the preprocessed CEST image into the trained deep neural network to obtain an NOE comparison graph.
2. The nuclear Overhauser enhanced imaging method according to claim 1, wherein the preprocessing the CEST image to be processed to obtain a preprocessed CEST image specifically includes:
normalizing the CEST image to be processed by taking a saturated image under the maximum frequency offset in the CEST image to be processed as a reference to obtain a normalized CEST image;
filtering the normalized CEST image to obtain a filtered CEST image;
and smoothing the filtered CEST image to obtain a preprocessed CEST image.
3. The nuclear Overhauser enhanced imaging method according to claim 1, wherein the generating of the training samples of the set amount according to the value ranges of the sample parameters specifically includes:
performing a training sample generation process a plurality of times;
the training sample generation process comprises the following steps:
obtaining a simulated NOE concentration graph and a simulated NOE exchange rate graph according to the value range of the NOE concentration and the value range of the NOE exchange rate;
respectively modulating the analog NOE concentration graph and the analog NOE exchange rate graph by adopting a natural image with textures subjected to filtering processing so as to enable the analog NOE concentration graph and the analog NOE exchange rate graph to generate textures, and obtaining a modulated analog NOE concentration graph and a modulated analog NOE exchange rate graph;
generating a simulated magnetic field map and a simulated radio frequency field map according to the value range of the magnetic field unevenness and the value range of the radio frequency field unevenness;
inputting the modulated simulated NOE concentration graph, the modulated simulated NOE exchange rate graph, the simulated magnetic field graph and the simulated radio frequency field graph into a Bloch-McConnell equation model to obtain a simulated CEST image;
adding random noise into the simulated CEST image to obtain a simulated CEST image added with noise;
and determining the modulated simulated NOE concentration graph, the modulated simulated NOE exchange rate graph and the simulated CEST image added with noise as training samples.
4. The nuclear Overhauser enhanced imaging method according to claim 1, wherein the fitting solution of each region of interest by using a Bloch-McConnell equation model to obtain the value of the sample parameter in each region of interest specifically comprises:
acquiring an equation solution type and pool parameters of a Bloch-McConnell equation model and sampling parameters when the CEST image to be processed is generated;
inputting the sampling parameters, the pool parameters and the equation solution type into the Bloch-McConnell equation model to obtain a determined Bloch-McConnell equation model;
and respectively fitting and solving the interested regions by adopting the determined Bloch-McConnell equation model to obtain the values of the sample parameters in the interested regions.
5. The nuclear Overhauser enhanced imaging method according to claim 3, wherein the obtaining of the simulated NOE concentration map and the simulated NOE exchange rate map according to the value range of the NOE concentration and the value range of the NOE exchange rate specifically comprises:
creating a blank template;
and respectively generating NOE concentration and NOE exchange rate randomly pixel by pixel in the blank template according to the value range of the NOE concentration and the value range of the NOE exchange rate to obtain a simulated NOE concentration graph and a simulated NOE exchange rate graph.
6. A nuclear Overhauser enhanced imaging system, comprising:
the image acquisition module is used for acquiring a CEST image to be processed;
the image preprocessing module is used for preprocessing the CEST image to be processed to obtain a preprocessed CEST image;
a region-of-interest determination module for selecting a plurality of regions-of-interest in the preprocessed CEST image;
the sample parameter value determining module is used for respectively fitting and solving the interested regions by adopting a Bloch-McConnell equation model to obtain the values of the sample parameters in the interested regions; the sample parameters include NOE concentration, NOE exchange rate, magnetic field non-uniformity, and radio frequency field non-uniformity;
the sample parameter value range determining module is used for determining the value range of each sample parameter according to the maximum value of the values of each sample parameter in all the regions of interest;
the training sample generating module is used for generating training samples with set quantity according to the value range of each sample parameter; the training sample comprises a modulated simulated NOE concentration graph, a modulated simulated NOE exchange rate graph and a simulated CEST image added with noise;
the network training module is used for training a deep neural network by taking the modulated simulated NOE concentration graph and the modulated simulated NOE exchange rate graph as labels and taking the simulated CEST image added with noise as input according to each training sample to obtain a trained deep neural network;
and the NOE contrast map reconstruction module is used for inputting the preprocessed CEST image into the trained deep neural network to obtain an NOE contrast map.
7. The nuclear Overhauser enhanced imaging system according to claim 6, wherein the image preprocessing module specifically comprises:
the normalization submodule is used for normalizing the CEST image to be processed by taking a saturated image under the maximum frequency offset in the CEST image to be processed as a reference to obtain a normalized CEST image;
the filtering submodule is used for filtering the normalized CEST image to obtain a filtered CEST image;
and the smoothing submodule is used for smoothing the filtered CEST image to obtain a preprocessed CEST image.
8. The nuclear Overhauser enhanced imaging system according to claim 6, wherein the training sample generating module specifically comprises:
the training sample generation submodule is used for executing the training sample generation process for multiple times;
the training sample generation submodule includes:
the NOE concentration graph and NOE exchange rate graph determining unit is used for obtaining a NOE concentration graph and a NOE exchange rate graph according to the value range of the NOE concentration and the value range of the NOE exchange rate;
the modulation unit is used for respectively modulating the analog NOE concentration diagram and the analog NOE exchange rate diagram by adopting a natural image with textures subjected to filtering processing so as to enable the analog NOE concentration diagram and the analog NOE exchange rate diagram to generate textures, and obtaining a modulated analog NOE concentration diagram and a modulated analog NOE exchange rate diagram;
the simulated field map determining unit is used for generating a simulated magnetic field map and a simulated radio frequency field map according to the value range of the magnetic field unevenness and the value range of the radio frequency field unevenness;
the simulated CEST image determining unit is used for inputting the modulated simulated NOE concentration graph, the modulated simulated NOE exchange rate graph, the simulated magnetic field graph and the simulated radio frequency field graph into a Bloch-McConnell equation model to obtain a simulated CEST image;
the noise unit is used for adding random noise into the simulated CEST image to obtain a simulated CEST image added with noise;
and the training sample determining unit is used for determining the modulated simulated NOE concentration graph, the modulated simulated NOE exchange rate graph and the simulated CEST image added with noise as training samples.
9. The nuclear Overhauser enhanced imaging system of claim 6, wherein the sample parameter value determining module specifically comprises:
the acquisition submodule is used for acquiring the equation solution type and the pool parameter of the Bloch-McConnell equation model and the sampling parameter when the CEST image to be processed is generated;
the Bloch-McConnell equation model determining submodule is used for inputting the sampling parameters, the pool parameters and the equation solution types into the Bloch-McConnell equation model to obtain a determined Bloch-McConnell equation model;
and the sample parameter value determining submodule is used for respectively fitting and solving the interested regions by adopting the determined Bloch-McConnell equation model to obtain the values of the sample parameters in the interested regions.
10. The nuclear Overhauser enhanced imaging system according to claim 8, wherein the analog NOE concentration map and analog NOE exchange rate map determining unit specifically includes:
the template creating subunit is used for creating a blank template;
and the NOE concentration simulation graph and NOE exchange rate simulation graph determining subunit is used for respectively randomly generating the NOE concentration and the NOE exchange rate by pixel points in the blank template according to the NOE concentration value range and the NOE exchange rate value range to obtain a NOE concentration simulation graph and a NOE exchange rate simulation graph.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210127212.0A CN114518555A (en) | 2022-02-11 | 2022-02-11 | Nuclear Overhauser enhanced imaging method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210127212.0A CN114518555A (en) | 2022-02-11 | 2022-02-11 | Nuclear Overhauser enhanced imaging method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114518555A true CN114518555A (en) | 2022-05-20 |
Family
ID=81596319
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210127212.0A Pending CN114518555A (en) | 2022-02-11 | 2022-02-11 | Nuclear Overhauser enhanced imaging method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114518555A (en) |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20160018496A1 (en) * | 2013-03-12 | 2016-01-21 | The Johns Hopkins University | Magnetic Resonance System and Method for Detecting the Buildup of the Transfer of Changes in Magnetization from Nuclei in Mobile Solute Molecules in Tissue |
| US20180268546A1 (en) * | 2017-03-20 | 2018-09-20 | The General Hospital Corporation | Quantification of magnetic resonance data by adaptive fitting of downsampled images |
| US20190011516A1 (en) * | 2017-07-05 | 2019-01-10 | The General Hospital Corporation | System and method for direct saturation-corrected chemical exchange saturation transfer (disc-cest) |
| CN110807492A (en) * | 2019-11-06 | 2020-02-18 | 厦门大学 | Magnetic resonance multi-parameter simultaneous quantitative imaging method and system |
| US20200072931A1 (en) * | 2018-08-30 | 2020-03-05 | Max-Planck-Gesellschaft Zur Foerderung Der Wissenschaften E. V. | Method and apparatus for processing magnetic resonance data |
| CN111062994A (en) * | 2019-11-26 | 2020-04-24 | 厦门大学 | Reconstruction method and system of chemical exchange saturation transfer contrast image |
| CN112924913A (en) * | 2021-02-02 | 2021-06-08 | 厦门大学 | Space-time coding magnetic resonance imaging super-resolution reconstruction method and system |
| CN113052937A (en) * | 2021-04-19 | 2021-06-29 | 厦门大学 | Reconstruction method and reconstruction system of dynamic contrast enhanced nuclear magnetic resonance image |
| CN113777543A (en) * | 2021-07-27 | 2021-12-10 | 华东师范大学 | Chemical exchange saturation transfer imaging post-processing method based on neural network fitting |
-
2022
- 2022-02-11 CN CN202210127212.0A patent/CN114518555A/en active Pending
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20160018496A1 (en) * | 2013-03-12 | 2016-01-21 | The Johns Hopkins University | Magnetic Resonance System and Method for Detecting the Buildup of the Transfer of Changes in Magnetization from Nuclei in Mobile Solute Molecules in Tissue |
| US20180268546A1 (en) * | 2017-03-20 | 2018-09-20 | The General Hospital Corporation | Quantification of magnetic resonance data by adaptive fitting of downsampled images |
| US20190011516A1 (en) * | 2017-07-05 | 2019-01-10 | The General Hospital Corporation | System and method for direct saturation-corrected chemical exchange saturation transfer (disc-cest) |
| US20200072931A1 (en) * | 2018-08-30 | 2020-03-05 | Max-Planck-Gesellschaft Zur Foerderung Der Wissenschaften E. V. | Method and apparatus for processing magnetic resonance data |
| CN110807492A (en) * | 2019-11-06 | 2020-02-18 | 厦门大学 | Magnetic resonance multi-parameter simultaneous quantitative imaging method and system |
| US20210134028A1 (en) * | 2019-11-06 | 2021-05-06 | Xiamen University | Method and system for simultaneous quantitative multiparametric magnetic resonance imaging (mri) |
| CN111062994A (en) * | 2019-11-26 | 2020-04-24 | 厦门大学 | Reconstruction method and system of chemical exchange saturation transfer contrast image |
| CN112924913A (en) * | 2021-02-02 | 2021-06-08 | 厦门大学 | Space-time coding magnetic resonance imaging super-resolution reconstruction method and system |
| CN113052937A (en) * | 2021-04-19 | 2021-06-29 | 厦门大学 | Reconstruction method and reconstruction system of dynamic contrast enhanced nuclear magnetic resonance image |
| CN113777543A (en) * | 2021-07-27 | 2021-12-10 | 华东师范大学 | Chemical exchange saturation transfer imaging post-processing method based on neural network fitting |
Non-Patent Citations (1)
| Title |
|---|
| 向鹏程等: "基于深度神经网络的时空编码磁共振成像超分辨率重建方法", 《物理学报》, vol. 71, no. 5, 17 November 2021 (2021-11-17), pages 363 - 371 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110807492B (en) | Magnetic resonance multi-parameter simultaneous quantitative imaging method and system | |
| Senden et al. | Reconstructing imagined letters from early visual cortex reveals tight topographic correspondence between visual mental imagery and perception | |
| CN109581253A (en) | Method and system for magnetic resonance imaging | |
| CN109100669B (en) | Single-scan synchronous magnetic resonance diffusion and T based on overlapped echoes2Image forming method | |
| US11170543B2 (en) | MRI image reconstruction from undersampled data using adversarially trained generative neural network | |
| CN103218795B (en) | Based on the part K spatial sequence image reconstructing method of self-adaptation doubledictionary study | |
| Deora et al. | Structure preserving compressive sensing MRI reconstruction using generative adversarial networks | |
| CN108132274A (en) | Echo-planar imaging is without reference scan pattern distortion antidote under non-uniform magnetic field | |
| CN110889897B (en) | Method and system for reconstructing incoherent motion magnetic resonance imaging parameters in voxel | |
| Favila et al. | Perception and memory have distinct spatial tuning properties in human visual cortex | |
| Liu et al. | High-performance rapid MR parameter mapping using model-based deep adversarial learning | |
| CN101530324B (en) | Method for generating a fat-reduced, spatially resolved magnetic resonance spectrum | |
| Heunis et al. | The effects of multi-echo fMRI combination and rapid T2*-mapping on offline and real-time BOLD sensitivity | |
| CN113947642A (en) | X-space magnetic particle imaging deconvolution method | |
| Tian et al. | Aliasing artifact reduction in spiral real‐time MRI | |
| DE102012205811A1 (en) | Method for determining sample points of a random subsampling scheme and magnetic resonance system | |
| CN111062994B (en) | Reconstruction method and system of chemical exchange saturation transfer contrast image | |
| CN114518555A (en) | Nuclear Overhauser enhanced imaging method and system | |
| CN115137347B (en) | Myelin quantitative imaging method for three-dimensional ultrashort echo time magnetic resonance fingerprint imaging | |
| CN110992435A (en) | Image reconstruction method and device, and imaging data processing method and device | |
| Thomaz et al. | Undersampled magnetic resonance image reconstructions based on a combination of u-nets and l1, l2, and tv optimizations | |
| Gan et al. | Self-supervised deep equilibrium models with theoretical guarantees and applications to mri reconstruction | |
| He et al. | Compressed sensing-based simultaneous recovery of magnitude and phase MR images via dual trigonometric sparsity | |
| US11435424B1 (en) | Systems and methods of constrained reconstruction of images with white noise | |
| CN117233676B (en) | Echo time-dependent magnetic resonance diffusion imaging signal generation method and device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |