CN112557980A - Magnetic resonance image correction method, magnetic resonance image correction device, medium, and electronic apparatus - Google Patents
Magnetic resonance image correction method, magnetic resonance image correction device, medium, and electronic apparatus Download PDFInfo
- Publication number
- CN112557980A CN112557980A CN202011205975.XA CN202011205975A CN112557980A CN 112557980 A CN112557980 A CN 112557980A CN 202011205975 A CN202011205975 A CN 202011205975A CN 112557980 A CN112557980 A CN 112557980A
- Authority
- CN
- China
- Prior art keywords
- data
- space data
- space
- target
- initial
- 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.)
- Granted
Links
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
- 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
- G01R33/4818—MR characterised by data acquisition along a specific k-space trajectory or by the temporal order of k-space coverage, e.g. centric or segmented coverage of k-space
Landscapes
- Physics & Mathematics (AREA)
- High Energy & Nuclear Physics (AREA)
- Condensed Matter Physics & Semiconductors (AREA)
- General Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Radiology & Medical Imaging (AREA)
- Engineering & Computer Science (AREA)
- Signal Processing (AREA)
- Magnetic Resonance Imaging Apparatus (AREA)
Abstract
The present disclosure relates to a magnetic resonance image rectification method, apparatus, medium, and electronic device, the method comprising: acquiring initial K space data corresponding to a magnetic resonance image, wherein the initial K space data are multi-channel data; performing multi-channel image reconstruction processing on the initial K space data according to a down-sampling convolution core to obtain first reconstructed K space data; performing multi-channel image reconstruction processing on the first reconstructed K space data according to a full sampling convolution core to obtain second reconstructed K space data; correcting the second reconstructed K space data according to the initial K space data to obtain target K space data; and carrying out channel combination on the target K space data to obtain a target magnetic resonance image. Therefore, the data acquisition efficiency is not required to be reduced, the motion artifact in the magnetic resonance image can be effectively eliminated, and the effectiveness of the magnetic resonance image correction result is improved.
Description
Technical Field
The present disclosure relates to the field of image processing, and in particular, to a magnetic resonance image correction method, apparatus, medium, and electronic device.
Background
In magnetic resonance imaging in general, image artifacts arise during the magnetic resonance signal acquisition due to any voluntary or involuntary motion of the subject, and severe motion artifacts may lead to misdiagnosis.
In the related art, it is common to avoid the generation of motion artifacts by accelerating the magnetic resonance data acquisition, or a single-shot based Imaging method such as Echo Planar Imaging (EPI) can be used to reduce the motion artifacts, but it is generally required to reduce the data acquisition efficiency and is limited to a lower spatial resolution.
Disclosure of Invention
The purpose of the present disclosure is to provide a magnetic resonance image correction method, apparatus, medium, and electronic device with high accuracy without reducing data acquisition efficiency.
In order to achieve the above object, according to a first aspect of the present disclosure, there is provided a magnetic resonance image rectification method including:
acquiring initial K space data corresponding to a magnetic resonance image, wherein the initial K space data are multi-channel data;
performing multi-channel image reconstruction processing on the initial K space data according to a down-sampling convolution core to obtain first reconstructed K space data;
performing multi-channel image reconstruction processing on the first reconstructed K space data according to a full sampling convolution core to obtain second reconstructed K space data;
correcting the second reconstructed K space data according to the initial K space data to obtain target K space data;
and carrying out channel combination on the target K space data to obtain a target magnetic resonance image.
Optionally, the modifying the second reconstructed K-space data according to the initial K-space data to obtain target K-space data includes:
determining whether data of each target row in the initial K space data is valid data, wherein the target row is any row in the initial K space data;
and under the condition that the data of the target line is determined to be valid data, replacing the data of the target line in the second reconstruction K-space data with the data of the target line of the initial K-space data to obtain the target K-space data.
Optionally, the determining whether the data of each target row in the initial K-space data is valid data includes:
for each target row of the initial K-space data:
determining a difference value of the data of the target row in the initial K-space data and the second reconstructed K-space data;
determining the ratio of the standard deviation corresponding to the module value of the difference value to the mean value as a first parameter of the target row;
determining the ratio of the statistical information of the module value of the difference value to the statistical information of the data of the target row in the initial K space data as a second parameter of the target row;
and determining whether the data of the target line in the initial K space data is valid data according to the first parameter and the second parameter of the target line.
Optionally, the determining whether the data of the target line in the initial K-space data is valid data according to the first parameter and the second parameter of the target line includes:
and determining the data of the target row in the initial K space data as valid data under the condition that the first parameter is smaller than a first threshold value and the second parameter is smaller than a second threshold value.
Optionally, the performing multi-channel reconstruction processing on the magnetic resonance image according to the initial K-space data and the downsampling convolution kernel to obtain first reconstructed K-space data includes:
performing down-sampling on the initial K space data according to preset interval information to obtain a plurality of K space sub-regions, wherein the K space sub-regions are not overlapped with each other, and the sum of the K space sub-regions is the region of the K space data;
and performing multi-channel reconstruction processing on each K space subregion according to the downsampling convolution kernel, and determining data obtained by averaging reconstruction data corresponding to each K space subregion as the first reconstruction K space data.
Optionally, for each channel of the initial K-space data, the number of the downsampling convolution kernels corresponding to the channel is one less than the number of the K-space sub-regions.
Optionally, the determining of the downsampled convolution kernel comprises:
acquiring low-frequency data, wherein the low-frequency data is determined from pre-scanned K-space data or from the initial K-space data;
determining a plurality of groups of data pairs comprising source data and target data according to the size of the downsampling convolution kernel, the preset interval information and the low-frequency data;
determining a source matrix and a target matrix corresponding to each channel according to the multiple groups of data pairs, wherein the source matrix corresponding to each channel is determined according to source data corresponding to all the channels, and the target matrix corresponding to each channel is determined according to target data corresponding to the channel;
and aiming at each channel, determining a downsampling convolution kernel of the channel according to the source matrix and the target matrix corresponding to the channel, wherein the product of the source matrix and a weight matrix formed by arranging the downsampling convolution kernels is the target matrix.
According to a second aspect of the present disclosure, there is provided a magnetic resonance image rectification apparatus, the apparatus including:
the acquisition module is used for acquiring initial K space data corresponding to the magnetic resonance image, wherein the initial K space data are multi-channel data;
the first processing module is used for carrying out multi-channel image reconstruction processing on the initial K space data according to the downsampling convolution kernel to obtain first reconstructed K space data;
the second processing module is used for carrying out multi-channel image reconstruction processing on the first reconstruction K space data according to the full sampling convolution kernel to obtain second reconstruction K space data;
the correction module is used for correcting the second reconstruction K space data according to the initial K space data to obtain target K space data;
and the merging module is used for performing channel merging on the target K space data to obtain a target magnetic resonance image.
Optionally, the second processing module includes:
a first determining submodule, configured to determine whether data of each target row in the initial K-space data is valid data, where the target row is any row in the initial K-space data;
and the replacing submodule is used for replacing the data of the target line in the second reconstruction K-space data with the data of the target line of the initial K-space data to obtain the target K-space data under the condition that the data of the target line is determined to be valid data.
Optionally, the first determining sub-module includes:
a second determining sub-module, configured to determine, for data of each target row of the initial K-space data, a difference between data of the target row in the initial K-space data and the second reconstructed K-space data;
the second determining submodule is used for determining the ratio of the standard deviation corresponding to the modulus of the difference value and the mean value as the first parameter of the target row;
a third determining submodule, configured to determine, as a second parameter of the target line, a ratio of statistical information of a modulus of the difference to statistical information of data of the target line in the initial K-space data;
a fourth determining submodule, configured to determine whether data of the target line in the initial K-space data is valid data according to the first parameter and the second parameter of the target line.
Optionally, the fourth determining submodule is configured to:
and determining the data of the target row in the initial K space data as valid data under the condition that the first parameter is smaller than a first threshold value and the second parameter is smaller than a second threshold value.
Optionally, the first processing module includes:
the down-sampling sub-module is used for down-sampling the initial K space data according to preset interval information to obtain a plurality of K space sub-regions, wherein the K space sub-regions are not overlapped with each other, and the sum of the K space sub-regions is the region of the K space data;
and the processing sub-module is used for performing multi-channel reconstruction processing on each K space sub-region according to the downsampling convolution kernel, and determining data obtained by averaging reconstruction data corresponding to each K space sub-region as the first reconstruction K space data.
Optionally, for each channel of the initial K-space data, the number of the downsampling convolution kernels corresponding to the channel is one less than the number of the K-space sub-regions.
Optionally, the determining module for the downsampling convolution kernel includes:
the acquisition submodule is used for acquiring low-frequency data, wherein the low-frequency data is determined from pre-scanned K space data or the initial K space data;
a fifth determining submodule, configured to determine, according to the size of the downsampling convolution kernel, the preset interval information, and the low-frequency data, a plurality of groups of data pairs including source data and target data;
a sixth determining submodule, configured to determine, according to the multiple groups of data pairs, a source matrix and a target matrix corresponding to each channel, where the source matrix corresponding to each channel is determined according to source data corresponding to all the channels, and the target matrix corresponding to each channel is determined according to target data corresponding to the channel;
and a sixth determining submodule, configured to determine, for each channel, a downsampling convolution kernel of the channel according to the source matrix and the target matrix corresponding to the channel, where a product of the source matrix and a weight matrix formed by arranging the downsampling convolution kernels is the target matrix.
According to a third aspect of the present disclosure, there is provided a computer readable storage medium having stored thereon a computer program which, when executed by a processor, performs the steps of the method of any one of the above-mentioned first aspects.
According to a fourth aspect of the present disclosure, there is provided an electronic device comprising:
a memory having a computer program stored thereon;
a processor for executing the computer program in the memory to implement the steps of the method of any of the first aspect above.
In the technical scheme, initial K space data corresponding to a magnetic resonance image is obtained, multichannel image reconstruction processing is performed on the initial K space data according to a downsampling convolution kernel to obtain first reconstructed K space data, multichannel image reconstruction processing is performed on the first reconstructed K space data according to a full-sampling convolution kernel to obtain second reconstructed K space data, and then the second reconstructed K space data is corrected according to the initial K space data to obtain target K space data, so that channel combination can be performed on the target K space data to obtain a target magnetic resonance image. Therefore, according to the technical scheme, the data acquisition efficiency does not need to be reduced, the image contrast is not affected, and the application range of the magnetic resonance image correction method is widened. And through the sequential combination of multi-channel reconstruction based on the downsampling convolution kernel and the full-sampling convolution kernel, the motion artifact in the magnetic resonance image can be effectively eliminated, meanwhile, the reconstructed K space data is corrected through the initial K space data, the signal-to-noise ratio and the accuracy of the determined target K space data can be effectively guaranteed, and the effectiveness and the stability of the correction result of the magnetic resonance image are improved.
Additional features and advantages of the disclosure will be set forth in the detailed description which follows.
Drawings
The accompanying drawings, which are included to provide a further understanding of the disclosure and are incorporated in and constitute a part of this specification, illustrate embodiments of the disclosure and together with the description serve to explain the disclosure without limiting the disclosure. In the drawings:
fig. 1 is a flow chart of a magnetic resonance image rectification method provided according to an embodiment of the present disclosure;
FIG. 2 is a flow chart of an implementation of performing a multi-channel reconstruction process on the magnetic resonance image from initial K-space data and a downsampled convolution kernel to obtain first reconstructed K-space data;
FIG. 3A is a schematic illustration of initial K-space data;
FIGS. 3B-3F are schematic views of K space sub-regions;
FIG. 4 is a schematic diagram of a process for determining a downsampled convolution kernel;
FIG. 5 is a flow diagram of an exemplary implementation of determining whether data of each target row in the initial K-space data is valid data;
FIG. 6A is a schematic representation of second reconstructed K-space data;
FIG. 6B is a diagram of target K-space data;
figure 7 is a comparison of results from processing magnetic resonance images based on the prior art and embodiments of the present disclosure;
figure 8 is a block diagram of a magnetic resonance image rectification apparatus provided in accordance with an embodiment of the present disclosure;
FIG. 9 is a block diagram illustrating an electronic device in accordance with an exemplary embodiment;
FIG. 10 is a block diagram illustrating an electronic device in accordance with an example embodiment.
Detailed Description
The following detailed description of specific embodiments of the present disclosure is provided in connection with the accompanying drawings. It should be understood that the detailed description and specific examples, while indicating the present disclosure, are given by way of illustration and explanation only, not limitation.
Fig. 1 is a flowchart illustrating a magnetic resonance image rectification method according to an embodiment of the present disclosure. As shown in fig. 1, the method may include:
in step 11, initial K-space data corresponding to the magnetic resonance image is acquired, wherein the initial K-space data is multi-channel data. Wherein the magnetic resonance image is obtained by scanning according to the data acquisition efficiency of the existing magnetic resonance scanning, such as the scanning based on the T2 TSE sequence.
Each signal obtained by magnetic resonance scanning contains global information, and the magnetic resonance signals need to be spatially positioned and numbered. The K-space is fourier space, which is a space filled with raw data of MR (Magnetic Resonance) signals with spatially localized encoded information. The MR signals acquired by the receiving coil performing the magnetic resonance scanning are radio waves with actually spatially encoded information, and can be converted into digital information through analog-to-digital conversion, so as to fill K space, thereby obtaining the initial K space data.
In step 12, the initial K-space data is subjected to multi-channel image reconstruction processing according to the downsampled convolution kernel, so as to obtain first reconstructed K-space data.
For the current channel, according to each channel of the initial K space data, the data corresponding to the central position of the downsampling convolution kernel corresponding to the channel in the current channel can be predicted according to the discontinuous row data corresponding to the downsampling convolution kernel in each channel of the initial K space data and the downsampling convolution kernel corresponding to the downsampling convolution kernel, so as to reconstruct the image. In this embodiment, the first reconstructed K-space data may be obtained by reconstructing based on a magnetic resonance parallel imaging technique by the downsampled convolution kernel in each channel. Among them, the magnetic resonance parallel imaging technology is a conventional technology in the art, and is not described herein again.
In step 13, the first reconstructed K-space data is subjected to multi-channel image reconstruction processing according to the full-sampling convolution kernel to obtain second reconstructed K-space data.
Similarly, each channel has a corresponding full-sampling convolution kernel, and for the current channel, according to adjacent symmetrical continuous line data corresponding to the full-sampling convolution kernel in each channel of the initial K-space data and the corresponding full-sampling convolution kernel, data corresponding to the position of the center of the full-sampling convolution kernel corresponding to the channel in the current channel is predicted to perform image reconstruction, so as to obtain second reconstructed K-space data.
In step 14, the second reconstructed K-space data is corrected according to the initial K-space data to obtain target K-space data.
In step 12 and step 13, image reconstruction is performed by using the downsampling convolution kernel and the full-sampling convolution kernel, so that motion artifacts can be eliminated to a greater extent, the second reconstructed K-space data is corrected by using the initial K-space data, the signal-to-noise ratio in the originally acquired magnetic resonance image can be ensured, and data support is provided for subsequently providing the accuracy of the corrected target magnetic resonance image.
In step 15, channel merging is performed on the target K-space data to obtain a target magnetic resonance image.
Illustratively, methods of performing channel merging include, but are not limited to: the method comprises a multi-channel image square sum root-opening method, a method for calculating a sensitivity distribution diagram by scanning a low-resolution image and then carrying out multi-channel combination, a self-adaptive channel combination method of a phased array line diagram and the like. The above channel merging method is a conventional technique in the art, and is not described herein again.
Each MR image has its corresponding K-space data. The target K space data is subjected to Fourier transform, so that the spatial positioning coding information in the original digital data can be decoded, MR signals with different frequencies, phases and amplitudes are resolved, the different frequencies and phases represent different spatial positions, and the amplitudes represent the MR signal intensity. By distributing the MR signals with different frequencies, phases and signal strengths to the corresponding pixels, MR image data, i.e. a reconstructed MR image, i.e. a magnetic resonance image, can be obtained, so that a target magnetic resonance image can be obtained.
Therefore, in the technical scheme, initial K-space data corresponding to a magnetic resonance image is acquired, multi-channel image reconstruction processing is performed on the initial K-space data according to a downsampling convolution kernel to obtain first reconstructed K-space data, multi-channel image reconstruction processing is performed on the first reconstructed K-space data according to a full-sampling convolution kernel to obtain second reconstructed K-space data, and then the second reconstructed K-space data is corrected according to the initial K-space data to obtain target K-space data, so that channel combination can be performed on the target K-space data to obtain a target magnetic resonance image. Therefore, according to the technical scheme, the data acquisition efficiency does not need to be reduced, the image contrast is not affected, and the application range of the magnetic resonance image correction method is widened. And through the sequential combination of multi-channel reconstruction based on the downsampling convolution kernel and the full-sampling convolution kernel, the motion artifact in the magnetic resonance image can be effectively eliminated, meanwhile, the reconstructed K space data is corrected through the initial K space data, the signal-to-noise ratio and the accuracy of the determined target K space data can be effectively guaranteed, and the effectiveness and the stability of the correction result of the magnetic resonance image are improved.
In order to make those skilled in the art understand the technical solutions provided by the embodiments of the present invention, the following detailed descriptions are provided for the above steps.
In a possible embodiment, in step 12, the magnetic resonance image is subjected to a multi-channel reconstruction process according to the initial K-space data and the downsampled convolution kernel, and the first reconstructed K-space data is obtained as follows, as shown in fig. 2, and this step may include:
in step 21, down-sampling the initial K-space data according to preset interval information to obtain a plurality of K-space sub-regions, where each K-space sub-region is not overlapped with each other, and the sum of each K-space sub-region is the region of the K-space data.
The preset interval information can be set according to an actual use scene, and corresponds to the number of the K space sub-regions. For example, if the number of the K space sub-regions is 2, the preset interval information may be one line at intervals, that is, one line of data is sampled at intervals of one line of data, so that the K space sub-regions are not overlapped with each other, and the sum of the K space sub-regions is the region of the K space data.
For example, as shown in fig. 3A, which is a schematic diagram of the initial K-space data, when the preset interval information is one line away, two K-space sub-regions obtained by down-sampling are shown in fig. 3B and 3C, where solid dots are used to represent data points in the K-space sub-region, as shown by P dots, and open dots are used to represent data points not in the K-space sub-region, as shown by Q dots.
Likewise, when the preset interval information is two lines apart, 3K-space sub-regions can be obtained by down-sampling, and for example, the K-space sub-regions obtained by down-sampling the preset interval information in two lines apart in fig. 3A are shown in fig. 3D-3F. The down-sampling process of other preset interval information is similar to that described above, and is not described herein again.
In step 22, performing multi-channel reconstruction processing on each K space sub-region according to the downsampled convolution kernel, and determining data obtained by averaging reconstruction data corresponding to each K space sub-region as first reconstruction K space data.
In one possible embodiment, for each channel of the initial K-space data, the number of the downsampled convolution kernels for that channel is one less than the number of the K-space sub-regions.
Wherein each downsampled convolution kernel is used to predict missing data in the K-space sub-region. As an example, as described in fig. 3B and fig. 3C above, if the preset interval information is one line at an interval, the number of the determined K space sub-regions is 2, and the number of the corresponding downsampling convolution kernels is 1, and the downsampling convolution kernels are used to predict missing data in the K space sub-regions, that is, to predict data of empty points in the K space sub-regions.
As another example, as shown in fig. 3D-3F above, if the preset interval information is two lines apart, the number of the determined K-space sub-regions is 3, and the number of the corresponding downsampling convolution kernels is 2, then the downsampling convolution kernel 1 may be used to predict data in one line adjacent to the data in the K-space sub-region in the target direction, and then the downsampling convolution kernel 2 is used to predict data in two lines adjacent to the data in the K-space sub-region in the target direction. The target direction may be a positive direction and a negative direction of the phase encoding direction in the K space, which is not limited by the present disclosure, and one of the directions may be selected.
Therefore, each missing data in the K space subregion can be predicted through the downsampling convolution kernel, and data support is provided for multi-channel image reconstruction processing.
And if the initial K space data is multi-channel data, determining that the K space sub-region is also multi-channel data. In this step, for each channel, the parallel imaging algorithm based on K space reconstructs each K space sub-region based on the downsampling convolution kernel, and determines data of a missing part in the K space sub-region based on the downsampling convolution kernel, so as to obtain reconstructed data corresponding to the K space sub-region. Then, the corresponding reconstructed data of the plurality of K-space sub-regions corresponding to each channel may be averaged, so as to obtain the reconstructed data corresponding to each channel, that is, the first reconstructed K-space data of the multiple channels is formed.
Therefore, according to the technical scheme, the initial K space region is subjected to down-sampling, so that a plurality of K space sub-regions can be determined, and missing data in each K space sub-region can be supplemented to effectively inhibit motion artifacts. Meanwhile, through multi-channel consistency constraint inside the K space sub-regions, unified processing in the multiple sub-regions and channels of the K space can be effectively guaranteed, the consistency of data processing is improved, and therefore the accuracy of the first reconstruction K space data is guaranteed.
Optionally, the determining of the downsampled convolution kernel comprises:
acquiring low frequency data, wherein the low frequency data is determined from pre-scanned K-space data or from the initial K-space data.
As an example, a pre-scan is performed based on a magnetic resonance scanning apparatus, so that pre-scanned K-space data can be obtained, where data at a central position portion in the K-space data is low-frequency data, and data at a central position and at positions adjacent to the central position in the K-space data can be selected as the low-frequency data according to actual use requirements, as is well known to those skilled in the art.
As another example, the low-frequency data may be determined from the initial K-space data to be corrected obtained by performing the scan this time, so that the determined down-sampling convolution kernel may be more suitable for the correction process of the magnetic resonance image this time. The manner of determining the low frequency data is similar to the above, and is not described herein again.
And then, determining a plurality of groups of data pairs comprising source data and target data according to the size of the downsampling convolution kernel, the preset interval information and the low-frequency data.
The size of the downsampling convolution kernel can be set according to an actual use scene, which is not limited by the present disclosure.
As shown in fig. 4, the left data is determined low-frequency data, where the size of the determined downsampling convolution kernel is 7 × 5, the preset interval information is one line at intervals, as shown in fig. 4, a plurality of stacked solid frames represent a plurality of channels, data of solid points in the solid frames are source data, data of grid points at the center position are target data, and as shown in fig. 4, a plurality of sets of data pairs of the source data and the target data may be determined by sliding the downsampling convolution kernel by a preset step length.
And determining a source matrix and a target matrix corresponding to each channel according to the multiple groups of data pairs, wherein the source matrix corresponding to each channel is determined according to the source data corresponding to all the channels, and the target matrix corresponding to each channel is determined according to the target data corresponding to the channel.
As shown in fig. 4, when determining the source matrix and the target matrix corresponding to each channel, the source matrix is generated by the source data determined by the stacked coils, and the target matrix corresponding to each channel is determined according to the target data corresponding to the source data in each channel, so that the data in each channel can be predicted based on the determined downsampling convolution kernel.
And aiming at each channel, determining a downsampling convolution kernel of the channel according to the source matrix and the target matrix corresponding to the channel, wherein the product of the source matrix and a weight matrix formed by arranging the downsampling convolution kernels is the target matrix.
After the source matrix and the target matrix corresponding to each channel are determined, the downsampling convolution kernel can be solved according to the principle that the product of the source matrix and a weight matrix formed by arranging the downsampling convolution kernels is the target matrix, wherein the solution can be carried out through a least square method, and the specific calculation mode of the least square method is not repeated here.
In the determination process of the full-sampling convolution kernel, when determining the source data, the determination is performed based on adjacent symmetrical continuous data, and the specific determination mode of the full-sampling convolution kernel is similar to that of the downsampling convolution kernel, and is not described herein again.
Therefore, by the technical scheme, the downsampling convolution kernel for image reconstruction can be determined to reconstruct missing data in each K space subregion, and the whole data can be predicted through partial data in the original image based on the downsampling convolution kernel through the determined downsampling convolution kernel, so that motion artifacts in the original image can be eliminated to a certain extent, and accurate data support is provided for accurate geomagnetic resonance image correction.
Optionally, in step 13, performing multi-channel image reconstruction processing on the first reconstructed K-space data according to the full-sampling convolution kernel to obtain an exemplary implementation manner of the second reconstructed K-space data and an implementation manner of step 12, which are not described herein again. In the technical scheme of the present disclosure, the multichannel consistency constraint inside the whole K space is determined by the full sampling convolution kernel, so that the correction effect of the motion artifact can be further enhanced.
In a possible embodiment, in step 14, the second reconstructed K-space data is modified according to the initial K-space data, and an exemplary implementation manner of obtaining the target K-space data is as follows, and the step may include:
determining whether data of each target row in the initial K-space data is valid data, wherein the target row is any row in the initial K-space data.
For example, as shown in fig. 3A, a schematic diagram of the initial K-space data is shown, where the Kx direction is a frequency encoding direction in the K-space, and the Ky direction is a phase encoding direction in the K-space, then the data of the target row is each row of data parallel to the Kx direction in the K-space data.
Determining whether the data of each target line in the initial K-space data is valid data, that is, determining whether the data of the target line in the initial K-space data is data affected by operation, if the data of the target line is not affected by motion, determining that the data of the target line is valid data, and if the data of the target line is affected by motion, determining that the data of the target line is invalid data.
In one possible embodiment, an exemplary implementation manner of determining whether the data of each target row in the initial K-space data is valid data is as follows, and as shown in fig. 5, the step may include:
for data of each target row of the initial K-space data, performing the following steps:
in step 51, the difference of the data of the target row in the initial K-space data and the second reconstructed K-space data is determined.
Illustratively, the data dimensions in the initial K-space data and the second reconstructed K-space data are the same, as shown in FIG. 6A, which is a schematic diagram of the second reconstructed K-space data, in the target row S { S }1,S2,…,SmFor example, m is used to represent the number of dimensions of the K-space data in the frequency encoding direction, as shown in fig. 6A, there are 12 data points in a row, that is, m is 12. For convenience of explanation, the initial K-space data is shown as solid dots in fig. 3A, and the second reconstructed K-space data is shown as diagonal dots in fig. 6A.
As an example, the difference between the data of the target line in the initial K-space data and the data of the target line in the second reconstructed K-space data, i.e. the difference T between the data of the target line S in the initial K-space (denoted as SS) and the data of the target line S in the second reconstructed K-space (denoted as ST), is determined as follows:
T＝{T1,T 2 ,…,Tm}＝{SS1-ST1,SS2-ST 2 ,…,SSm-STm}。
the calculation method of the difference corresponding to each target row is the same as the above, and is not described herein again.
In step 52, a ratio of a standard deviation corresponding to a module value of the difference to the mean is determined as a first parameter of the target row.
In this step, after the module value of the difference T is determined, a standard deviation and a mean value corresponding to a sequence formed by the module values of the difference T may be determined, where a calculation manner of the standard deviation and the mean value is the prior art, and is not described herein again. The first parameter may then be determined, where the first parameter may be used to characterize similarity between the distribution of the data of the target row in the initial K-space and the distribution of the data of the target row in the second reconstructed K-space, and the smaller the first parameter, the more similar the first parameter and the second parameter, and vice versa, the larger the first parameter, the greater the difference between the first parameter and the second parameter.
In step 53, the ratio of the statistical information of the module value of the difference value to the statistical information of the data of the target row in the initial K-space data is determined as the second parameter of the target row.
The statistical information may be data and information or data mean information, that is, the mean of the model of the difference is determined as the statistical information of the module value of the difference, and the mean of the data of the target line in the initial K space data is used as the statistical information of the data of the target line in the initial K space data, so that the second parameter of the target line can be determined. The second parameter is used to characterize the similarity between the data of the target row after reconstruction and the data in the original K space, wherein the smaller the second parameter is, the more similar the second parameter is, and conversely, the larger the second parameter is, the larger the difference between the second parameter and the original K space is.
In step 54, it is determined whether the data of the target line in the initial K-space data is valid data according to the first parameter and the second parameter of the target line.
In a possible embodiment, the determining whether the data of the target line in the initial K-space data is valid data according to the first parameter and the second parameter of the target line may include:
and determining the data of the target row in the initial K space data as valid data under the condition that the first parameter is smaller than a first threshold value and the second parameter is smaller than a second threshold value.
In this embodiment, the first threshold and the second threshold may be set according to an actual usage scenario. When the first parameter is smaller than the first threshold and the second parameter is smaller than the second threshold, it may be indicated that the data of the target line in the initial K-space data is similar to the original data after reconstruction, that is, the data of the target line is less affected by motion, and at this time, the data of the target line in the initial K-space data may be determined to be valid data.
Therefore, according to the technical scheme, whether the data of the target line in the initial K space data is influenced by motion can be determined by comparing the difference between the data at the same position in the initial K space data and the data at the same position in the second reconstruction K space data, so that the effective data in the initial K space data can be determined, on one hand, the accuracy of the determined effective data can be ensured, and on the other hand, accurate data support can be provided for the follow-up correction of the second reconstruction K space data.
And then, under the condition that the data of the target line is determined to be valid data, replacing the data of the target line in the second reconstruction K-space data with the data of the target line of the initial K-space data to obtain the target K-space data.
If the target line S is determined to be valid data, the data of the target line S in the second reconstructed K space data may be replaced with the data of the target line S in the initial K space, and the obtained target K space data is as shown in fig. 6B.
Therefore, according to the technical scheme, whether the data of the target row in the initial K space data is valid data or not can be determined, so that the corresponding data in the second reconstruction K space can be corrected, on one hand, accurate elimination of motion artifacts in the magnetic resonance image can be guaranteed, on the other hand, the problem of noise amplification caused by convolution reconstruction can be reduced to a certain extent by replacing the original data which are not influenced by motion into the reconstruction data, the signal-to-noise ratio in the target K space data is effectively guaranteed, the stability and robustness of the magnetic resonance image correction result can be improved, and the application range of the magnetic resonance image correction method is widened.
As shown in fig. 7, the left diagram (a) is the original imaging result after performing multi-slice 2D acquisition on a liver region by using a conventional multi-shot based T2 TSE sequence and then performing fourier transform on the acquired K-space data directly, and the right diagram (b) is the reconstruction result obtained based on the magnetic resonance image rectification method provided by the embodiment of the present disclosure. As can be seen from the comparison graph, the method provided based on the present disclosure can greatly reduce the originally serious respiratory motion artifact, and the detail display of the liver is clearer.
The present disclosure also provides a magnetic resonance image rectification apparatus, as shown in fig. 8, the apparatus 10 including:
an obtaining module 100, configured to obtain initial K-space data corresponding to a magnetic resonance image, where the initial K-space data is multi-channel data;
the first processing module 200 is configured to perform multi-channel image reconstruction processing on the initial K-space data according to a downsampling convolution kernel to obtain first reconstructed K-space data;
the second processing module 300 is configured to perform multi-channel image reconstruction processing on the first reconstructed K-space data according to a full-sampling convolution kernel to obtain second reconstructed K-space data;
a modification module 400, configured to modify the second reconstructed K-space data according to the initial K-space data to obtain target K-space data;
and a merging module 500, configured to perform channel merging on the target K-space data to obtain a target magnetic resonance image.
Optionally, the second processing module includes:
a first determining submodule, configured to determine whether data of each target row in the initial K-space data is valid data, where the target row is any row in the initial K-space data;
and the replacing submodule is used for replacing the data of the target line in the second reconstruction K-space data with the data of the target line of the initial K-space data to obtain the target K-space data under the condition that the data of the target line is determined to be valid data.
Optionally, the first determining sub-module includes:
a second determining sub-module, configured to determine, for data of each target row of the initial K-space data, a difference between data of the target row in the initial K-space data and the second reconstructed K-space data;
the second determining submodule is used for determining the ratio of the standard deviation corresponding to the modulus of the difference value and the mean value as the first parameter of the target row;
a third determining submodule, configured to determine, as a second parameter of the target line, a ratio of statistical information of a modulus of the difference to statistical information of data of the target line in the initial K-space data;
a fourth determining submodule, configured to determine whether data of the target line in the initial K-space data is valid data according to the first parameter and the second parameter of the target line.
Optionally, the fourth determining submodule is configured to:
and determining the data of the target row in the initial K space data as valid data under the condition that the first parameter is smaller than a first threshold value and the second parameter is smaller than a second threshold value.
Optionally, the first processing module includes:
the down-sampling sub-module is used for down-sampling the initial K space data according to preset interval information to obtain a plurality of K space sub-regions, wherein the K space sub-regions are not overlapped with each other, and the sum of the K space sub-regions is the region of the K space data;
and the processing sub-module is used for performing multi-channel reconstruction processing on each K space sub-region according to the downsampling convolution kernel, and determining data obtained by averaging reconstruction data corresponding to each K space sub-region as the first reconstruction K space data.
Optionally, for each channel of the initial K-space data, the number of the downsampling convolution kernels corresponding to the channel is one less than the number of the K-space sub-regions.
Optionally, the determining module for the downsampling convolution kernel includes:
the acquisition submodule is used for acquiring low-frequency data, wherein the low-frequency data is determined from pre-scanned K space data or the initial K space data;
a fifth determining submodule, configured to determine, according to the size of the downsampling convolution kernel, the preset interval information, and the low-frequency data, a plurality of groups of data pairs including source data and target data;
a sixth determining submodule, configured to determine, according to the multiple groups of data pairs, a source matrix and a target matrix corresponding to each channel, where the source matrix corresponding to each channel is determined according to source data corresponding to all the channels, and the target matrix corresponding to each channel is determined according to target data corresponding to the channel;
and a sixth determining submodule, configured to determine, for each channel, a downsampling convolution kernel of the channel according to the source matrix and the target matrix corresponding to the channel, where a product of the source matrix and a weight matrix formed by arranging the downsampling convolution kernels is the target matrix.
With regard to the apparatus in the above-described embodiment, the specific manner in which each module performs the operation has been described in detail in the embodiment related to the method, and will not be elaborated here.
Fig. 9 is a block diagram illustrating an electronic device 700 in accordance with an example embodiment. As shown in fig. 9, the electronic device 700 may include: a processor 701 and a memory 702. The electronic device 700 may also include one or more of a multimedia component 703, an input/output (I/O) interface 704, and a communication component 705.
The processor 701 is configured to control the overall operation of the electronic apparatus 700 to complete all or part of the steps in the magnetic resonance image rectification method. The memory 702 is used to store various types of data to support operation at the electronic device 700, such as instructions for any application or method operating on the electronic device 700 and application-related data, such as contact data, transmitted and received messages, pictures, audio, video, and the like. The Memory 702 may be implemented by any type of volatile or non-volatile Memory device or combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read-Only Memory (EPROM), Programmable Read-Only Memory (PROM), Read-Only Memory (ROM), magnetic Memory, flash Memory, magnetic disk, or optical disk. The multimedia components 703 may include screen and audio components. Wherein the screen may be, for example, a touch screen and the audio component is used for outputting and/or inputting audio signals. For example, the audio component may include a microphone for receiving external audio signals. The received audio signal may further be stored in the memory 702 or transmitted through the communication component 705. The audio assembly also includes at least one speaker for outputting audio signals. The I/O interface 704 provides an interface between the processor 701 and other interface modules, such as a keyboard, mouse, buttons, etc. These buttons may be virtual buttons or physical buttons. The communication component 705 is used for wired or wireless communication between the electronic device 700 and other devices. Wireless Communication, such as Wi-Fi, bluetooth, Near Field Communication (NFC), 2G, 3G, 4G, NB-IOT, eMTC, or other 5G, etc., or a combination of one or more of them, which is not limited herein. The corresponding communication component 705 may thus include: Wi-Fi module, Bluetooth module, NFC module, etc.
In an exemplary embodiment, the electronic Device 700 may be implemented by one or more Application Specific Integrated Circuits (ASICs), Digital Signal Processors (DSPs), Digital Signal Processing Devices (DSPDs), Programmable Logic Devices (PLDs), Field Programmable Gate Arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components for performing the magnetic resonance image rectification method.
In another exemplary embodiment, a computer readable storage medium comprising program instructions is also provided, which when executed by a processor, implement the steps of the magnetic resonance image rectification method described above. For example, the computer readable storage medium may be the memory 702 described above comprising program instructions executable by the processor 701 of the electronic device 700 to perform the magnetic resonance image rectification method described above.
Fig. 10 is a block diagram illustrating an electronic device 1900 according to an example embodiment. For example, the electronic device 1900 may be provided as a server. Referring to fig. 10, an electronic device 1900 includes a processor 1922, which may be one or more in number, and a memory 1932 for storing computer programs executable by the processor 1922. The computer program stored in memory 1932 may include one or more modules that each correspond to a set of instructions. Further, the processor 1922 may be configured to execute the computer program to perform the magnetic resonance image rectification method described above.
In addition, the electronic device 1900 may further include a power component 1926 and a communication component 1950, the power component 1926 can be configured to perform power management of the electronic device 1900, and the communication component 1950 can be configured to enable communication, e.g., wired or wireless communication, of the electronic device 1900. In addition, the electronic device 1900 may also include input/output (I/O) interfaces 1958. The electronic device 1900 may operate based on an operating system, such as Windows Server, stored in memory 1932TM，Mac OSXTM，UnixTM，LinuxTMAnd so on.
In another exemplary embodiment, a computer readable storage medium comprising program instructions is also provided, which when executed by a processor, implement the steps of the magnetic resonance image rectification method described above. For example, the computer readable storage medium may be the memory 1932 described above that includes program instructions that are executable by the processor 1922 of the electronic device 1900 to perform the magnetic resonance image rectification method described above.
In another exemplary embodiment, a computer program product is also provided, which comprises a computer program executable by a programmable apparatus, the computer program having code portions for performing the magnetic resonance image rectification method described above when executed by the programmable apparatus.
The preferred embodiments of the present disclosure are described in detail with reference to the accompanying drawings, however, the present disclosure is not limited to the specific details of the above embodiments, and various simple modifications may be made to the technical solution of the present disclosure within the technical idea of the present disclosure, and these simple modifications all belong to the protection scope of the present disclosure.
It should be noted that the various features described in the above embodiments may be combined in any suitable manner without departing from the scope of the invention. In order to avoid unnecessary repetition, various possible combinations will not be separately described in this disclosure.
In addition, any combination of various embodiments of the present disclosure may be made, and the same should be considered as the disclosure of the present disclosure, as long as it does not depart from the spirit of the present disclosure.
Claims (10)
1. A magnetic resonance image rectification method, characterized in that the method comprises:
acquiring initial K space data corresponding to a magnetic resonance image, wherein the initial K space data are multi-channel data;
performing multi-channel image reconstruction processing on the initial K space data according to a down-sampling convolution core to obtain first reconstructed K space data;
performing multi-channel image reconstruction processing on the first reconstructed K space data according to a full sampling convolution core to obtain second reconstructed K space data;
correcting the second reconstructed K space data according to the initial K space data to obtain target K space data;
and carrying out channel combination on the target K space data to obtain a target magnetic resonance image.
2. The method of claim 1, wherein the modifying the second reconstructed K-space data according to the initial K-space data to obtain target K-space data comprises:
determining whether data of each target row in the initial K space data is valid data, wherein the target row is any row in the initial K space data;
and under the condition that the data of the target line is determined to be valid data, replacing the data of the target line in the second reconstruction K-space data with the data of the target line of the initial K-space data to obtain the target K-space data.
3. The method of claim 2, wherein said determining whether data of each target row in the initial K-space data is valid data comprises:
for each target row of the initial K-space data:
determining a difference value of the data of the target row in the initial K-space data and the second reconstructed K-space data;
determining the ratio of the standard deviation corresponding to the module value of the difference value to the mean value as a first parameter of the target row;
determining the ratio of the statistical information of the module value of the difference value to the statistical information of the data of the target row in the initial K space data as a second parameter of the target row;
and determining whether the data of the target line in the initial K space data is valid data according to the first parameter and the second parameter of the target line.
4. The method of claim 3, wherein the determining whether the data of the target line in the initial K-space data is valid data according to the first parameter and the second parameter of the target line comprises:
and determining the data of the target row in the initial K space data as valid data under the condition that the first parameter is smaller than a first threshold value and the second parameter is smaller than a second threshold value.
5. The method of claim 1, wherein performing a multi-channel reconstruction process on the magnetic resonance image from the initial K-space data and the downsampled convolution kernel to obtain first reconstructed K-space data comprises:
performing down-sampling on the initial K space data according to preset interval information to obtain a plurality of K space sub-regions, wherein the K space sub-regions are not overlapped with each other, and the sum of the K space sub-regions is the region of the K space data;
and performing multi-channel reconstruction processing on each K space subregion according to the downsampling convolution kernel, and determining data obtained by averaging reconstruction data corresponding to each K space subregion as the first reconstruction K space data.
6. The method of claim 5, wherein for each channel of the initial K-space data, the number of downsampled convolution kernels for that channel is one less than the number of K-space sub-regions.
7. The method of claim 5, wherein the determining of the downsampled convolution kernel comprises:
acquiring low-frequency data, wherein the low-frequency data is determined from pre-scanned K-space data or from the initial K-space data;
determining a plurality of groups of data pairs comprising source data and target data according to the size of the downsampling convolution kernel, the preset interval information and the low-frequency data;
determining a source matrix and a target matrix corresponding to each channel according to the multiple groups of data pairs, wherein the source matrix corresponding to each channel is determined according to source data corresponding to all the channels, and the target matrix corresponding to each channel is determined according to target data corresponding to the channel;
and aiming at each channel, determining a downsampling convolution kernel of the channel according to the source matrix and the target matrix corresponding to the channel, wherein the product of the source matrix and a weight matrix formed by arranging the downsampling convolution kernels is the target matrix.
8. An apparatus for magnetic resonance image rectification, characterized in that the apparatus comprises:
the acquisition module is used for acquiring initial K space data corresponding to the magnetic resonance image, wherein the initial K space data are multi-channel data;
the first processing module is used for carrying out multi-channel image reconstruction processing on the initial K space data according to the downsampling convolution kernel to obtain first reconstructed K space data;
the second processing module is used for carrying out multi-channel image reconstruction processing on the first reconstruction K space data according to the full sampling convolution kernel to obtain second reconstruction K space data;
the correction module is used for correcting the second reconstruction K space data according to the initial K space data to obtain target K space data;
and the merging module is used for performing channel merging on the target K space data to obtain a target magnetic resonance image.
9. A computer-readable storage medium, on which a computer program is stored which, when being executed by a processor, carries out the steps of the method according to any one of claims 1 to 7.
10. An electronic device, comprising:
a memory having a computer program stored thereon;
a processor for executing the computer program in the memory to carry out the steps of the method of any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011205975.XA CN112557980B (en) | 2020-11-02 | 2020-11-02 | Magnetic resonance image correction method, magnetic resonance image correction device, medium, and electronic apparatus |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011205975.XA CN112557980B (en) | 2020-11-02 | 2020-11-02 | Magnetic resonance image correction method, magnetic resonance image correction device, medium, and electronic apparatus |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112557980A true CN112557980A (en) | 2021-03-26 |
| CN112557980B CN112557980B (en) | 2022-05-03 |
Family
ID=75041761
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011205975.XA Active CN112557980B (en) | 2020-11-02 | 2020-11-02 | Magnetic resonance image correction method, magnetic resonance image correction device, medium, and electronic apparatus |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112557980B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114266728A (en) * | 2021-11-27 | 2022-04-01 | 深圳先进技术研究院 | Magnetic resonance image acquisition method, device, equipment and medium |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080154115A1 (en) * | 2004-02-10 | 2008-06-26 | Koninklijke Philips Electronics N.V. | Magnetic Resonance Imaging Method |
| CN102435966A (en) * | 2011-09-02 | 2012-05-02 | 中国科学院深圳先进技术研究院 | Three-dimensional magnetic resonance imaging method and three-dimensional magnetic resonance imaging system |
| WO2012085810A2 (en) * | 2010-12-22 | 2012-06-28 | Koninklijke Philips Electronics N.V. | Rapid parallel reconstruction for arbitrary k-space trajectories |
| CN104635188A (en) * | 2013-11-12 | 2015-05-20 | 上海联影医疗科技有限公司 | K-space reconstruction method and magnetic resonance imaging method |
| TW201618034A (en) * | 2014-08-04 | 2016-05-16 | 輝達公司 | Deinterleaving interleaved high dynamic range image by using YUV interpolation |
| CN107092805A (en) * | 2014-01-09 | 2017-08-25 | 上海联影医疗科技有限公司 | magnetic resonance parallel imaging device |
| CN109613460A (en) * | 2018-12-28 | 2019-04-12 | 上海联影医疗科技有限公司 | MR imaging method, device, computer equipment and storage medium |
| CN111128222A (en) * | 2018-10-30 | 2020-05-08 | 富士通株式会社 | Speech separation method, speech separation model training method, and computer-readable medium |
| WO2021034960A1 (en) * | 2019-08-21 | 2021-02-25 | The Regents Of The University Of California | Systems and methods for imputing real-time physiological signals |
-
2020
- 2020-11-02 CN CN202011205975.XA patent/CN112557980B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080154115A1 (en) * | 2004-02-10 | 2008-06-26 | Koninklijke Philips Electronics N.V. | Magnetic Resonance Imaging Method |
| WO2012085810A2 (en) * | 2010-12-22 | 2012-06-28 | Koninklijke Philips Electronics N.V. | Rapid parallel reconstruction for arbitrary k-space trajectories |
| CN102435966A (en) * | 2011-09-02 | 2012-05-02 | 中国科学院深圳先进技术研究院 | Three-dimensional magnetic resonance imaging method and three-dimensional magnetic resonance imaging system |
| CN104635188A (en) * | 2013-11-12 | 2015-05-20 | 上海联影医疗科技有限公司 | K-space reconstruction method and magnetic resonance imaging method |
| CN107092805A (en) * | 2014-01-09 | 2017-08-25 | 上海联影医疗科技有限公司 | magnetic resonance parallel imaging device |
| TW201618034A (en) * | 2014-08-04 | 2016-05-16 | 輝達公司 | Deinterleaving interleaved high dynamic range image by using YUV interpolation |
| CN111128222A (en) * | 2018-10-30 | 2020-05-08 | 富士通株式会社 | Speech separation method, speech separation model training method, and computer-readable medium |
| CN109613460A (en) * | 2018-12-28 | 2019-04-12 | 上海联影医疗科技有限公司 | MR imaging method, device, computer equipment and storage medium |
| WO2021034960A1 (en) * | 2019-08-21 | 2021-02-25 | The Regents Of The University Of California | Systems and methods for imputing real-time physiological signals |
Non-Patent Citations (3)
| Title |
|---|
| DIZHI MA 等: "Flexible Spatial and Angular Light Field Super Resolution", 《2020 IEEE INTERNATIONAL CONFERENCE ON IMAGE PROCESSING (ICIP)》 * |
| PETERSEN, TC 等: "Local electron tomography using angular variations of surface tangents: Stomo version 2", 《COMPUTER PHYSICS COMMUNICATIONS》 * |
| 王一达 等: "卷积神经网络重建欠采的磁共振图像", 《磁共振成像》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114266728A (en) * | 2021-11-27 | 2022-04-01 | 深圳先进技术研究院 | Magnetic resonance image acquisition method, device, equipment and medium |
| WO2023092742A1 (en) * | 2021-11-27 | 2023-06-01 | 深圳先进技术研究院 | Method and apparatus for acquiring magnetic resonance image, and device and medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112557980B (en) | 2022-05-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20220036516A1 (en) | Training methods of a denoising model and image denoising methods and apparatuses | |
| KR102144320B1 (en) | Magnetic resonance image processing apparatus and method | |
| US10274569B2 (en) | Magnetic resonance imaging system with ghost artifact reduction and method operation thereof | |
| US10317487B2 (en) | Method and magnetic resonance apparatus for acquisition of MR data of a two-dimensional volume segment | |
| CN111093495B (en) | Magnetic resonance imaging apparatus, nyquist ghost correction method, and storage device | |
| CN112557980B (en) | Magnetic resonance image correction method, magnetic resonance image correction device, medium, and electronic apparatus | |
| US20210049461A1 (en) | Data compression apparatus, data compression method, and learning apparatus | |
| KR20220082302A (en) | MAGNETIC RESONANCE IMAGE PROCESSING APPARATUS AND METHOD USING ARTIFICIAL NEURAL NETWORK AND RESCAlING | |
| JP2001292980A (en) | Method for image pickup by nuclear spin resonance | |
| CN112272535B (en) | Method and system for magnetic resonance diffusion imaging | |
| CN113534031A (en) | Image domain data generating method, computer device and readable storage medium | |
| US10852380B2 (en) | Magnetic resonance image reconstruction | |
| CN113391251A (en) | Magnetic resonance image reconstruction method, device and equipment | |
| CN109975343B (en) | Magnetic resonance imaging method, system and storage medium | |
| JP2019162369A (en) | Magnetic resonance imaging apparatus, sensitivity distribution calculation method, and sensitivity distribution calculation program | |
| CN111537930B (en) | Magnetic resonance parameter imaging method and equipment based on gradient waveform adjustment | |
| US10955507B2 (en) | Magnetic resonance imaging | |
| US11119173B2 (en) | Dynamic imaging based on echo planar imaging sequence | |
| Ponce et al. | Auto‐calibration approach for k–t SENSE | |
| CN114325528B (en) | Magnetic resonance imaging method and related equipment | |
| CN113359076B (en) | Magnetic resonance imaging method, magnetic resonance imaging apparatus, computer device, and storage medium | |
| KR102572311B1 (en) | Magnetic resonance image processing apparatus and method using artificial neural network and substitute map | |
| US20230221388A1 (en) | Magnetic resonance imaging apparatus, image reconstruction apparatus, and image reconstruction method | |
| CN113384260B (en) | Acceleration factor adjusting method, magnetic resonance imaging scanning method, device and equipment | |
| EP4318048A1 (en) | Image processing device, image processing system, and image processing method |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |