CN112834971A - MRI iterative self-calibration parallel imaging algorithm based on singular value decomposition - Google Patents
MRI iterative self-calibration parallel imaging algorithm based on singular value decomposition Download PDFInfo
- Publication number
- CN112834971A CN112834971A CN202011641459.1A CN202011641459A CN112834971A CN 112834971 A CN112834971 A CN 112834971A CN 202011641459 A CN202011641459 A CN 202011641459A CN 112834971 A CN112834971 A CN 112834971A
- Authority
- CN
- China
- Prior art keywords
- matrix
- calibration
- singular value
- value decomposition
- data
- 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
- 238000000354 decomposition reaction Methods 0.000 title claims abstract description 35
- 238000003384 imaging method Methods 0.000 title claims abstract description 15
- 238000001228 spectrum Methods 0.000 claims abstract description 28
- 238000000034 method Methods 0.000 claims abstract description 13
- 238000005516 engineering process Methods 0.000 claims abstract description 9
- 239000011159 matrix material Substances 0.000 claims description 56
- 238000002595 magnetic resonance imaging Methods 0.000 claims description 17
- 230000035945 sensitivity Effects 0.000 claims description 17
- 238000004364 calculation method Methods 0.000 claims description 16
- 238000005070 sampling Methods 0.000 claims description 15
- 230000008707 rearrangement Effects 0.000 claims description 5
- 230000001133 acceleration Effects 0.000 claims description 4
- 238000001914 filtration Methods 0.000 claims description 3
- 230000014509 gene expression Effects 0.000 claims description 3
- 238000010586 diagram Methods 0.000 description 4
- 230000008569 process Effects 0.000 description 3
- 230000006835 compression Effects 0.000 description 2
- 238000007906 compression Methods 0.000 description 2
- HPTJABJPZMULFH-UHFFFAOYSA-N 12-[(Cyclohexylcarbamoyl)amino]dodecanoic acid Chemical compound OC(=O)CCCCCCCCCCCNC(=O)NC1CCCCC1 HPTJABJPZMULFH-UHFFFAOYSA-N 0.000 description 1
- 238000003759 clinical diagnosis Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification 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
-
- 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
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 MRI iterative self-calibration parallel imaging algorithm (SSCPIT) based on singular value decomposition innovatively adds sensitive spectrum information and combines the parallel acquisition technologies of SENSE and GRAPPA, thereby inheriting the advantages of parallel accelerated acquisition of SENSE and GRAPPA and having the advantages of the GRAPPA self-calibration parallel acquisition technology. By combining the advantages of both SENSE and GRAPPA techniques, the sscipit parallel acquisition technique has a short acquisition time and a higher signal-to-noise ratio compared to SENSE and GRAPPA. The SSCPIT method effectively solves the problems of the existence of artifacts in SENSE and the low signal-to-noise ratio of GRAPPA. Therefore, SSCPIT has high clinical application value.
Description
Technical Field
The invention belongs to the field of magnetic resonance imaging, and particularly relates to an MRI iterative self-calibration parallel imaging algorithm based on singular value decomposition.
Background
In parallel MRI, data is acquired from multiple receiver coils simultaneously, allowing the reconstruction of images from undersampled multi-coil data. Each coil exhibits a different spatial sensitivity as an additional spatial encoding function. This may speed up the acquisition of images by sub-sampling k-space and reconstructing the images using the sensitivity information. The reconstruction algorithms currently used have two different routes: coil sensitivity based explicit reconstruction algorithms (SENSE) and local kernel based reconstruction algorithms in k-space, which exploit correlations between channels of adjacent points in k-space (GRAPPA and SPIRiT). However, the SENSE algorithm has artifacts, the GRAPPA algorithm has a low signal-to-noise ratio, and the obtained image effect is poor.
Disclosure of Invention
The invention aims to solve the technical problem of providing an MRI iterative self-calibration parallel imaging algorithm which is smoother in sensitive spectrum, stronger in robustness and capable of obviously reducing image noise and is based on singular value decomposition.
In order to solve the technical problems, the technical scheme adopted by the invention is An MRI Iterative Self-calibration Parallel Imaging Algorithm (An MRI Iterative Self-calibration Parallel Imaging Algorithm SSCPIT for short) Based on Singular Value Decomposition, which is characterized by comprising the following steps:
(1) the magnetic resonance imaging scanning is realized by applying a parallel acquisition technology in the SPGR and DESPGR sequences;
(2) zero filling, data rearrangement and filtering processing are carried out on the collected original K space data to obtain optimized K space data;
(3) extracting full-sampling data of a center of the K space to serve as calculation data of a calibration matrix, and performing mobile acquisition in the full-sampling data by adopting an n x n correction block to construct the calibration matrix;
(4) performing singular value decomposition on the calibration matrix to obtain a characteristic value and a characteristic vector of the matrix;
(5) performing data rearrangement on the eigenvector to obtain a sensitive spectrum distribution matrix, and taking a row with the largest eigenvalue in the matrix as a sensitive spectrum matrix;
(6) the sensitive spectrum matrix is used as a parameter to enter nonlinear conjugate gradient iteration (CG iteration) to obtain missing K space data;
(7) and performing IFFT on the complete K space data, performing dot product on the complete K space data and the sensitive spectrum matrix, and summing to obtain a complete image.
Because the calculation amount is large, the stability of the program is seriously influenced by the allocation and release of the memory in the calculation process, and the calculation time is increased by repeatedly allocating and releasing the memory, so the memory pre-allocation is carried out before the calculation. The steps (4) - (7) have huge requirements on the CPU memory, and are complex in calculation and long in time consumption; therefore, the steps are calculated in the GPU, so that the calculation time can be greatly reduced, the memory of the CPU can be optimized, and the stability of the program during running is greatly improved.
Furthermore, in the step (1), magnetic resonance scanning acceleration is realized by changing the number of the data lines which are fully acquired in the center, interlaced sampling is performed from the center to two sides, and calibration data is fully acquired in the center. Parallel acquisition technology is adopted in the SPGR and DESPGR sequences, and SSCPIT magnetic resonance scanning acceleration is realized by changing the number of the central full acquisition data lines; and (3) carrying out interlaced sampling from the center to two sides, wherein the interlaced sampling is the same as 2-time speed SENSE, only the center needs to be fully sampled and used for calculating the sensitive spectrum matrix, and the fully sampled part of the center is similar to GRAPPA and needs to be fully sampled and calibrated.
Further, the correction block is 6 x 6.
Further, the method also comprises the following steps: before singular value decomposition of the calibration matrix, the calibration matrix is cut by using a threshold value, and only part of the calibration matrix is reserved. And selecting a proper threshold value to cut the calibration matrix, and only reserving part of the calibration matrix, so that the main purposes of conveniently calculating the sensitive spectrum, saving the memory and ensuring that the threshold value is not too large easily are achieved.
Further, the specific method of singular value decomposition is as follows: the result of weighting the k-space and coil sensitivity, x representing the original image, x is FSm, S is [ S ]1,S2,...,Sn]Is a sensitive spectrum; rewrite the first constraint in a matrix form and merge all the same equations of the second constraint:
Wx＝x
Px＝Py
further expressions can be derived:
WFSm＝FSm
IFFT is carried out on two sides of equal sign at the same time, and the obtained sensitive vector of the coil is F-1Characteristic vector F corresponding to characteristic value of WF of "1-1WFSM＝Sm。
Further, the method for calculating the sensitivity spectrum matrix in the step (5) comprises the following steps: carrying out direct eigenvalue decomposition on W, and realizing decoupling by point-by-point operation of a semi-positive definite matrix in an image threshold:
simplifying the eigenvalue decomposition of operator W to solve for the eigenvalue decomposition of each position in image space
At position m, m (q) is non-zero, the conditions for producing the sensitivity value are obtained:
for all
And only selecting the eigenvector with the eigenvalue being 1 to carry out eigenvalue decomposition to obtain the sensitive spectrum matrix.
In clinical diagnosis, it is necessary to reduce the calculation time to a range acceptable to doctors and patients. The parallel acceleration of the algorithm is realized by adopting a universal parallel computing architecture CUDA parallel computing technology promoted by NVIDIA and an OpenMP-based parallel programming technology. The invention combines the parallel acquisition technology of SENSE and GRAPPA, thereby inheriting the advantages of parallel accelerated acquisition of SENSE and GRAPPA and having the advantages of the GRAPPA self-calibration parallel acquisition technology. By combining the advantages of both SENSE and GRAPPA techniques, the sscipit parallel acquisition technique has a short acquisition time and a higher signal-to-noise ratio compared to SENSE and GRAPPA. The SSCPIT method effectively solves the problems of the existence of artifacts in SENSE and the low signal-to-noise ratio of GRAPPA. Therefore, SSCPIT has high clinical application value.
Drawings
The invention is further described with reference to the following figures and examples:
FIG. 1 is a schematic flow diagram of the process of the present invention;
FIG. 2 is a K-space diagram of the present invention after channel compression;
FIG. 3 is a diagram of all eigenvalues after eigenvalue decomposition of the matrix w;
FIG. 4 is a diagram of the eigenvectors (i.e., sensitivity spectra) after eigenvalue decomposition of the matrix w;
fig. 5 shows the comparison of the sscipit reconstructed picture with the GRAPPA reconstructed picture according to the present invention.
Detailed Description
As shown in fig. 1, an MRI iterative self-calibration parallel imaging algorithm (ssciit for short) based on singular value decomposition according to the present invention includes the following steps: (1) memory pre-allocation: (2) reading K space data; (3) constructing a calibration matrix; (4) the calibration matrix and the K space data enter a GPU; (5) calculating a sensitivity spectrum; (6) CG iteration is carried out to obtain a complete K space; (7) reconstructing an image; (8) the reconstructed result flows back to the CPU. The specific calculation method is as follows.
(1) Realizing magnetic resonance scanning by adopting Cartesian acquisition in the SPGR and DESPGR sequences; in this embodiment, a central calibration line is set in the center of the K space, where the central calibration line adopts a full sampling acquisition mode, performs interlaced sampling from the center to two sides, and is the same as 2-time speed SENSE, but the center needs to be partially full sampling lines for calculating the sensitive spectrum matrix, and the fully-sampled central part is similar to GRAPPA and needs to be fully full sampling calibration data. Fig. 2 shows K-space data after channel compression.
(2) The memory is pre-allocated, because the calculated amount is large, the stability of the program is seriously influenced by the allocation and release of the memory in the calculating process, and the calculation time is increased by repeatedly allocating and releasing the memory;
(3) carrying out zero filling, rearrangement and filtering processing on the acquired K space signal data to obtain optimized K space data;
(4) extracting full-sampling data of a K space center as calculation data of a calibration matrix; adopting n x n correction blocks to perform mobile acquisition in full acquisition data, and finally constructing a calibration matrix; in this embodiment, a 6 × 6 calibration block is used. Selecting a proper threshold value to cut the calibration matrix, and only reserving part of the calibration matrix, so as to conveniently calculate the sensitive spectrum, save the memory, and ensure that the threshold value is not too large easily
(5) And (4) carrying out singular value decomposition on the calibration matrix obtained in the step (4) to obtain the eigenvalue and the eigenvector of the matrix, and using the eigenvalue and the eigenvector to calculate the sensitive spectrum matrix of the SSCPIT. An effective way to analyze the calibration matrix is to perform singular value decomposition:
A＝UΣVH
a is the correction data block constructed from the fully-sampled K-centered matrix, and the columns of the V matrix in SVD are the row basis for A, and thus are the basis for all the overlapping blocks in the calibration data. We can divide V into V⊥And VIIIn which V is⊥Null space, V, representing AIIRepresenting the row space of a. Given k-space data, each k-space block reconstruction must satisfy two conditions:
is expressed as
Wherein
Further, in step 2, x must satisfy Wx ═ x, and the operator W is a semi-positive definite matrix convolution kernel. Thus, by definition, x belongs to the subspace to which the eigenvector of W corresponding to the eigenvalue "1" belongs. The concrete implementation is as follows: weighting the k-space and coil sensitivity of the x-representation as the original image
x＝FSm
Where S is [ S ]1,S2,...,Sn]Is the coil sensitivity vector.
Rewrite the first constraint in a matrix form and merge all the same equations of the second constraint:
Wx＝x
Px＝Py
further expressions can be derived:
WFSm＝FSm
IFFT is carried out on both sides of equal sign, and the sensitivity vector of the coil is F-1And when the characteristic value of the WF is 1, the corresponding characteristic vector is obtained.
F-1WFSM＝Sm
By directly decomposing W by the eigenvalue, the sensitivity can be effectively obtained. Because the operator W is a convolution kernel of a semi-positive definite matrix, decoupling is realized by point-by-point operation of the semi-positive definite matrix in an image threshold:
F-1WFSM＝Sm
simplifying the eigenvalue decomposition of operator W to solve for the eigenvalue decomposition of each position in image space
At position m, m (q) is non-zero, resulting in the conditions that yield sensitivity:
therefore, by applying to all
Only the eigenvector with eigenvalue ═ 1 is selected for eigenvalue decomposition, and a clear sensitive spectrum matrix can be obtained.
In an ideal case, there is only one eigenvector for each position with eigenvalue "1", and the other eigenvalues for that position are much smaller than 1.
After the calculation of the sensitivity spectrum matrix, a standard SENSE reconstruction may be performed. In some cases, errors in acquisition lead to the appearance of multiple eigenvectors with eigenvalues "1" or eigenvalues less than 1, which cannot be solved with the SENSE model in the strict SENSE. Here, the first and second liquid crystal display panels are,
having the form:
here, MqOften the value of (A) is 1 or 2, lambdaj(q)＝1，
By combining the above formulas, the final Soft SENSE reconstruction formula can be obtained:
the step 5 has huge demand on the CPU memory, and simultaneously has complex calculation and longer time consumption; therefore, in this embodiment, step 5 is calculated in the GPU, so that the calculation time can be greatly reduced, the CPU memory can be optimized, and the stability of the program during running is greatly improved.
As shown in fig. 5, the left sscipit right GRAPPA clearly shows that the algorithm of the present embodiment provides significantly better results than the existing algorithm. The SSCPIT innovatively adds sensitive spectrum information, wherein the sensitive spectrum information is obtained by characteristic values, is similar to the traditional mask, but the weight (mask) obtained by the characteristic values is more matched with the edge of a scanned object. After sensitive spectrum information is included, image noise can be obviously reduced.
The above embodiments are merely illustrative of the technical ideas and features of the present invention, and the purpose thereof is to enable those skilled in the art to understand the contents of the present invention and implement the present invention, and not to limit the protection scope of the present invention. All equivalent changes and modifications made according to the spirit of the present invention should be covered within the protection scope of the present invention.
Claims (7)
1. An MRI iterative self-calibration parallel imaging algorithm based on singular value decomposition is characterized by comprising the following steps:
(1) the magnetic resonance imaging scanning is realized by applying a parallel acquisition technology in the SPGR and DESPGR sequences;
(2) zero filling, data rearrangement and filtering processing are carried out on the collected original K space data to obtain optimized K space data;
(3) extracting full-sampling data of a center of the K space to serve as calculation data of a calibration matrix, and performing mobile acquisition in the full-sampling data by adopting an n x n correction block to construct the calibration matrix;
(4) performing singular value decomposition on the calibration matrix to obtain a characteristic value and a characteristic vector of the matrix;
(5) performing data rearrangement on the eigenvector to obtain a sensitive spectrum distribution matrix, and taking a row with the largest eigenvalue in the matrix as a sensitive spectrum matrix;
(6) the sensitive spectrum matrix is used as a parameter to enter nonlinear conjugate gradient iteration to obtain missing K space data;
(7) and performing IFFT on the complete K space data, performing dot product on the complete K space data and the sensitive spectrum matrix, and summing to obtain a complete image.
2. The singular value decomposition-based MRI iterative self-calibration parallel imaging algorithm according to claim 1, wherein in step (1) the magnetic resonance scanning acceleration is realized by changing the number of rows of the center full-sampling data, the center is sampled to two sides in an interlaced manner, and the center is full-sampling the calibration data.
3. The singular value decomposition based MRI iterative self-calibrating parallel imaging algorithm of claim 1, wherein said correction block is 6 x 6.
4. The singular value decomposition-based MRI iterative self-calibrating parallel imaging algorithm of claim 1, further comprising the steps of: before singular value decomposition of the calibration matrix, the calibration matrix is cut by using a threshold value, and only part of the calibration matrix is reserved.
5. The singular value decomposition-based MRI iterative self-calibrating parallel imaging algorithm of claim 4, wherein said threshold value is 60.
6. The singular value decomposition-based MRI iterative self-calibration parallel imaging algorithm according to claim 1, wherein the singular value decomposition is performed by the following specific method: the result of weighting the k-space and coil sensitivity, x representing the original image, x is FSm, S is [ S ]1,S2,...,Sn]Is a sensitive spectrum; rewrite the first constraint in a matrix form and merge all the same equations of the second constraint:
Wx＝x
Px＝Py
further expressions can be derived:
WFSm＝FSm
IFFT is carried out on two sides of equal sign at the same time, and the obtained sensitive vector of the coil is F-1Characteristic vector corresponding to characteristic value of WF of 1
F-1WFSM＝Sm。
7. The singular value decomposition-based MRI iterative self-calibration parallel imaging algorithm according to claim 6, wherein the calculation method of the sensitivity spectrum matrix in the step (5) is as follows: carrying out direct eigenvalue decomposition on W, and realizing decoupling by point-by-point operation of a semi-positive definite matrix in an image threshold:
simplifying the eigenvalue decomposition of operator W to solve for the eigenvalue decomposition of each position in image space
At position m, m (q) is non-zero, the conditions for producing the sensitivity value are obtained:
for all
And only selecting the eigenvector with the eigenvalue being 1 to carry out eigenvalue decomposition to obtain the sensitive spectrum matrix.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011641459.1A CN112834971A (en) | 2020-12-31 | 2020-12-31 | MRI iterative self-calibration parallel imaging algorithm based on singular value decomposition |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011641459.1A CN112834971A (en) | 2020-12-31 | 2020-12-31 | MRI iterative self-calibration parallel imaging algorithm based on singular value decomposition |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112834971A true CN112834971A (en) | 2021-05-25 |
Family
ID=75926878
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011641459.1A Pending CN112834971A (en) | 2020-12-31 | 2020-12-31 | MRI iterative self-calibration parallel imaging algorithm based on singular value decomposition |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112834971A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115097366A (en) * | 2022-06-22 | 2022-09-23 | 浙江朗润医疗科技有限公司 | CG iteration-based abdomen TORNADO image quality improving method |
| CN115952391A (en) * | 2022-12-12 | 2023-04-11 | 海光信息技术股份有限公司 | Data processing method and device, electronic equipment and storage medium |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1799498A (en) * | 2004-12-31 | 2006-07-12 | 西门子（中国）有限公司 | Quick generalized partially parallel acquisition(GRAPA) image reconstruction algorithm for magnetic resonance imaging(MRI) |
| CN1827039A (en) * | 2005-02-28 | 2006-09-06 | 西门子（中国）有限公司 | MRI parallel imaging method utilizing sensitivity encoding in frequency domain |
| US20090091322A1 (en) * | 2007-04-25 | 2009-04-09 | Stefan Posse | Single-shot magnetic resonance spectroscopic imaging with partial parallel imaging |
| US20130044960A1 (en) * | 2011-08-17 | 2013-02-21 | The Regents Of The University Of California | Coil compression for three dimensional autocalibrating parallel imaging with cartesian sampling |
| CN103384836A (en) * | 2010-12-22 | 2013-11-06 | 皇家飞利浦电子股份有限公司 | Rapid parallel reconstruction for arbitrary k-space trajectories containing GRAPPA operator |
| CN106108903A (en) * | 2016-08-08 | 2016-11-16 | 江苏大学 | A kind of parallel MR image rebuilding method of improvement |
| US20190086501A1 (en) * | 2017-09-21 | 2019-03-21 | Centre National De La Recherche Scientifique (Cnrs) | Method and magnetic resonance apparatus for image reconstruction with trimmed autocalibrating k-space estimation based on structured matrix completion |
| CN109934884A (en) * | 2019-01-22 | 2019-06-25 | 昆明理工大学 | A kind of iteration based on changing learning and joint sparse is from consistency parallel imaging reconstruction method |
| CN110148193A (en) * | 2019-04-30 | 2019-08-20 | 台州学院 | Dynamic magnetic resonance method for parallel reconstruction based on adaptive quadrature dictionary learning |
| US20200333416A1 (en) * | 2019-04-19 | 2020-10-22 | Regents Of The University Of Minnesota | Scalable self-calibrated interpolation of undersampled magnetic resonance imaging data |
-
2020
- 2020-12-31 CN CN202011641459.1A patent/CN112834971A/en active Pending
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1799498A (en) * | 2004-12-31 | 2006-07-12 | 西门子（中国）有限公司 | Quick generalized partially parallel acquisition(GRAPA) image reconstruction algorithm for magnetic resonance imaging(MRI) |
| CN1827039A (en) * | 2005-02-28 | 2006-09-06 | 西门子（中国）有限公司 | MRI parallel imaging method utilizing sensitivity encoding in frequency domain |
| US20090091322A1 (en) * | 2007-04-25 | 2009-04-09 | Stefan Posse | Single-shot magnetic resonance spectroscopic imaging with partial parallel imaging |
| CN103384836A (en) * | 2010-12-22 | 2013-11-06 | 皇家飞利浦电子股份有限公司 | Rapid parallel reconstruction for arbitrary k-space trajectories containing GRAPPA operator |
| US20130044960A1 (en) * | 2011-08-17 | 2013-02-21 | The Regents Of The University Of California | Coil compression for three dimensional autocalibrating parallel imaging with cartesian sampling |
| CN106108903A (en) * | 2016-08-08 | 2016-11-16 | 江苏大学 | A kind of parallel MR image rebuilding method of improvement |
| US20190086501A1 (en) * | 2017-09-21 | 2019-03-21 | Centre National De La Recherche Scientifique (Cnrs) | Method and magnetic resonance apparatus for image reconstruction with trimmed autocalibrating k-space estimation based on structured matrix completion |
| CN109934884A (en) * | 2019-01-22 | 2019-06-25 | 昆明理工大学 | A kind of iteration based on changing learning and joint sparse is from consistency parallel imaging reconstruction method |
| US20200333416A1 (en) * | 2019-04-19 | 2020-10-22 | Regents Of The University Of Minnesota | Scalable self-calibrated interpolation of undersampled magnetic resonance imaging data |
| CN110148193A (en) * | 2019-04-30 | 2019-08-20 | 台州学院 | Dynamic magnetic resonance method for parallel reconstruction based on adaptive quadrature dictionary learning |
Non-Patent Citations (5)
| Title |
|---|
| MARTIN UECKER1等: ""ESPIRiT — An Eigenvalue Approach to Autocalibrating Parallel MRI: Where SENSE meets GRAPPA"", 《PUBMED CENTRAL》 * |
| MARTIN UECKER1等: ""ESPIRiT — An Eigenvalue Approach to Autocalibrating Parallel MRI: Where SENSE meets GRAPPA"", 《PUBMED CENTRAL》, vol. 71, no. 3, 6 November 2014 (2014-11-06), pages 3 - 15 * |
| 张久明: ""磁共振并行成像重建算法研究"", 《中国知网》 * |
| 张久明: ""磁共振并行成像重建算法研究"", 《中国知网》, 15 October 2014 (2014-10-15), pages 3 - 4 * |
| 张久明: "磁共振并行成像重建算法研究", 中国知网, pages 1 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115097366A (en) * | 2022-06-22 | 2022-09-23 | 浙江朗润医疗科技有限公司 | CG iteration-based abdomen TORNADO image quality improving method |
| CN115097366B (en) * | 2022-06-22 | 2024-09-27 | 浙江朗润医疗科技有限公司 | Abdomen TORNADO image quality improvement method based on CG iteration |
| CN115952391A (en) * | 2022-12-12 | 2023-04-11 | 海光信息技术股份有限公司 | Data processing method and device, electronic equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Zhao et al. | Applications of a deep learning method for anti-aliasing and super-resolution in MRI | |
| CN108734659B (en) | Sub-pixel convolution image super-resolution reconstruction method based on multi-scale label | |
| Poddar et al. | Dynamic MRI using smoothness regularization on manifolds (SToRM) | |
| US8879811B2 (en) | Alternating direction of multipliers method for parallel MRI reconstruction | |
| CN105046672B (en) | A kind of image super-resolution rebuilding method | |
| CN109360152A (en) | 3 d medical images super resolution ratio reconstruction method based on dense convolutional neural networks | |
| CN111353947A (en) | Magnetic resonance parallel imaging method and related equipment | |
| CN110148215B (en) | Four-dimensional magnetic resonance image reconstruction method based on smooth constraint and local low-rank constraint model | |
| CN104166110B (en) | Magnetic resonance parallel collected image reconstruction method and apparatus | |
| CN109410195B (en) | Magnetic resonance imaging brain partition method and system | |
| CA2403211A1 (en) | Parallel magnetic resonance imaging techniques using radiofrequency coil arrays | |
| CN108776318B (en) | Magnetic resonance multi-contrast image reconstruction | |
| CN111754598B (en) | Local space neighborhood parallel magnetic resonance imaging reconstruction method based on transformation learning | |
| CN112834971A (en) | MRI iterative self-calibration parallel imaging algorithm based on singular value decomposition | |
| CN104635188B (en) | K space reconstructions method and MR imaging method | |
| CN109741416A (en) | Nuclear magnetic resonance image method for reconstructing, device, computer equipment and its storage medium | |
| CN112991483B (en) | Non-local low-rank constraint self-calibration parallel magnetic resonance imaging reconstruction method | |
| CN112819949B (en) | Magnetic resonance fingerprint image reconstruction method based on structured low-rank matrix | |
| CN108287324A (en) | The method for reconstructing and device of the more contrast images of magnetic resonance | |
| CN105184755B (en) | Parallel MR imaging high quality reconstructing method based on the full variation containing joint from uniformity | |
| CN112734869A (en) | Rapid magnetic resonance imaging method based on sparse complex U-shaped network | |
| CN109920017B (en) | Parallel magnetic resonance imaging reconstruction method of joint total variation Lp pseudo norm based on self-consistency of feature vector | |
| CN109934884B (en) | Iterative self-consistency parallel imaging reconstruction method based on transform learning and joint sparsity | |
| CN114187374B (en) | Multi-excitation plane echo diffusion weighted imaging method | |
| US20240290011A1 (en) | Dual-domain self-supervised learning for accelerated non-cartesian magnetic resonance imaging reconstruction |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210525 |
|
| RJ01 | Rejection of invention patent application after publication |