High-temporal-spatial-resolution magnetic resonance metabolism imaging method and device

Abstract

The invention discloses a high-temporal-spatial-resolution magnetic resonance metabolism imaging method. The method comprises the following steps: obtaining high-frequency information complementary dynamic k-space metabolism imaging data kspt; and performing reconstruction by using the dynamic k-space metabolism imaging data with high-frequency information complementation and a chemical shift imaging data reconstruction model with a multi-level and multi-dimensional similarity problem to obtain magnetic resonance metabolism imaging data of final target solution. The invention further discloses a high-temporal-spatial-resolution magnetic resonance metabolism imaging device, a storage medium and a program product. According to the method, the under-sampling magnification of magnetic resonance metabolism imaging can be improved, the prior information of spatial dimension and spectral dimension multilevel correlation of dynamic metabolism imaging data is fully utilized, and the temporal-spatial resolution of magnetic resonance metabolism imaging is greatly improved.

Description

Technical Field

[0001] The present invention relates to the field of magnetic resonance metabolic imaging, and in particular to a method and device for magnetic resonance metabolic imaging with high temporal and spatial resolution. Background Art

[0002] Magnetic resonance metabolic imaging is an important medical imaging technology that can simultaneously provide structural and metabolic information of metabolites (such as glucose and lactate) in tissues of interest, and is of great value in clinical research. Since metabolic duration is limited, a metabolic imaging experiment often requires the collection of multiple metabolic imaging data in the time dimension to obtain dynamic metabolic information. Therefore, it is of great significance to improve the spatiotemporal resolution of metabolic imaging data.

[0003] The current gold standard for magnetic resonance metabolic imaging is chemical shift imaging (CSI). Since CSI performs phase encoding in the spatial direction and the spectral dimension is obtained through the evolution of chemical shift over time, CSI acquires three-dimensional data containing two-dimensional spatial distribution and one-dimensional spectral information. This three-dimensional characteristic enables CSI to provide both the structural information of tissues and the distribution characteristics of metabolites. However, since CSI requires phase encoding in two directions, the acquisition process is time-consuming, which limits its application in dynamic metabolic studies.

[0004] Compressed sensing is a technology that uses signal sparsity and irregular sampling to reconstruct high-quality data. Using the principle of compressed sensing, undersampling can be performed in the phase encoding dimension of CSI, and the information lost in the undersampling process can be restored through algorithms such as sparse reconstruction, thereby improving the acquisition efficiency while ensuring the quality of metabolic imaging and obtaining metabolic information with high temporal and spatial resolution.

[0005] However, the existing technology of introducing compressed sensing methods into CSI has the following problems: 1. A single under-sampling phase encoding matrix is ​​used for each acquisition of metabolic imaging data during the metabolic process. This practice limits the acquisition of high-frequency information, resulting in the loss of details in dynamic metabolic images, and also limits the under-sampling ratio of the under-sampling phase encoding matrix. 2. Metabolic imaging data are reconstructed as independent units, and the multi-level similar prior information of adjacent dynamic metabolic imaging data in the image dimension (spatial dimension) and spectral dimension is not fully utilized. Therefore, the existing technology has limitations in the under-sampling ratio, and at the same time, the prior information of dynamic metabolic imaging data is insufficiently utilized, which ultimately makes it difficult to significantly improve the spatiotemporal resolution.

[0006] Other high-temporal-resolution MRI metabolic imaging methods mainly include the following two categories, but they all have limitations in obtaining full-spectrum metabolic information, and the technical challenges they face restrict their practical applications: 1. Model-based rapid metabolic imaging. This method uses prior knowledge such as chemical shift to reduce the number of echoes that need to be collected and speed up the acquisition. However, due to the constraints of multiple factors such as T2 relaxation and noise, the selection of the optimal echo spacing is extremely challenging.

[0007] 2. Metabolite-specific imaging. This method uses simultaneous spatial spectrum excitation pulses to selectively excite only the desired compounds, but this method faces two technical challenges: its resolution is limited for metabolites with similar chemical shifts; at the same time, the technical implementation relies on a high-performance radio frequency excitation system, which places high demands on hardware equipment. Summary of the invention

[0008] In view of the problems existing in the prior art such as low spatiotemporal resolution, limited undersampling ratio, and insufficient utilization of prior information of dynamic metabolic imaging data, the present invention proposes a high spatiotemporal resolution magnetic resonance metabolic imaging method and device for obtaining high spatiotemporal resolution magnetic resonance metabolic imaging data.

[0009] The above-mentioned purpose of the present invention is achieved by the following methods: A high temporal and spatial resolution magnetic resonance metabolic imaging method, comprising: Step 1: Acquire dynamic k-space metabolic imaging data ksp t , t=1, 2, ..., T, T is the number of dynamic k-space metabolic imaging numbers, Step 2: Use dynamic k-space metabolic imaging data and chemical shift imaging data to reconstruct the model and obtain the magnetic resonance metabolic imaging data for the final target solution.

[0010] As mentioned above, k-space metabolic imaging data ksp t Based on the under-sampling phase encoding matrix, under-sampling is performed in the phase encoding dimension, data is collected in the central part of the under-sampling phase encoding matrix, and data is randomly collected in the peripheral part of the under-sampling phase encoding matrix. The positions of the peripheral random under-sampling of under-sampling phase encoding matrices with different dynamics are not exactly the same.

[0011] In step 2 above, k-space metabolic imaging data ksp t The data is input into the chemical shift imaging data reconstruction model, and when the preset iteration stop condition is reached after cyclic iteration, the magnetic resonance metabolic imaging data of the final target solution is obtained.

[0012] As mentioned above, the chemical shift imaging data reconstruction model is based on the following formula: , Among them, x t The target is to solve the t-th dynamic magnetic resonance metabolic imaging data, t is the dynamic sequence number, E is the undersampling Fourier transform operator, ksp t is the t-th dynamic k-space metabolic imaging data, ‖‖2 represents the L2 norm, is the MRI metabolic imaging data x in the i-th level t The regularization parameter when performing traversal windows and dynamic dimension merging. S is the total number of levels. is the feature matrix B in The nuclear norm of i is the layer number, n is the image block number, N i is the MRI metabolic imaging data x at the i-th level t The total number of image blocks to be partitioned.

[0013] As described above, the MRI metabolic imaging data x t Performing traversal windows and dynamic dimension merging includes: From MRI metabolic imaging data x t Starting from the vertex of t Traverse and obtain N i Image block b n , each image block b n The dimension size is [j i ,k i ,l i ], J, K are the magnetic resonance metabolic imaging data x t The image dimension, j i , k i is the image dimension of the image block corresponding to the i-th level, L is the magnetic resonance metabolic imaging data x t The spectral dimension, l i is the spectral dimension of the image block corresponding to the i-th level, and the image block b n The rows of the image dimension plane where the two image dimensions are located are spliced ​​in sequence to obtain a one-dimensional vector of the image dimension plane, and then the one-dimensional vectors of the image dimension plane corresponding to each spectrum dimension of the image block are spliced ​​in sequence to obtain a one-dimensional vector of the image block. In a dynamic order, the one-dimensional vectors of each dynamic image block corresponding to the image block at the same position are spliced ​​to obtain the feature matrix B in .

[0014] A high temporal and spatial resolution magnetic resonance metabolic imaging method, further comprising: for each magnetic resonance metabolic imaging data x tThe corresponding feature matrix B in Perform SVD decomposition, each B in After the matrix SVD decomposition, a series of singular values ​​are obtained. , where each B in The largest singular value after matrix SVD decomposition is recorded as the corresponding maximum singular value , keep each feature matrix B in After SVD decomposition, the value is greater than × The singular values ​​of the matrix P are reconstructed by SVD in , each magnetic resonance metabolic imaging data x t The corresponding matrix P in According to the inverse operation of traversal window and dynamic dimension merging, the low-rank constrained MRI matrix L is obtained. t , calculate the updated MRI data x t 1 , , , , in, is the k-space residual Return to the MRI domain to obtain the image domain residual, is the undersampled inverse Fourier transform operator, is the k-space domain residual, and E is the undersampling Fourier transform operator.

[0015] As mentioned above, the preset iteration stop condition is to reach the number of iterations, or The value is less than the set threshold .

[0016] A high spatiotemporal resolution magnetic resonance metabolic imaging device comprises a memory and a processor, wherein the memory stores a computer program, and the processor implements the above steps 1 and 2 when executing the computer program.

[0017] A computer-readable storage medium stores a computer program, which implements the above steps 1 and 2 when executed by a processor.

[0018] A computer program product includes a computer program, and the computer program implements the above steps 1 and 2 when executed by a processor.

[0019] Compared with the prior art, the present invention has the following advantages: 1. In the present invention, the central part of the under-sampling phase encoding matrix is ​​used for data collection, and the peripheral part of the under-sampling phase encoding matrix is ​​randomly used for data collection. The positions of the random under-sampling of the peripheral under-sampling of the under-sampling phase encoding matrices of different dynamics are not exactly the same, which can further improve the under-sampling ratio, thereby improving the spatiotemporal resolution of metabolic imaging. When the above sampling method is not adopted, all the under-sampling phase encoding matrices are the same, so the high-frequency information collected at adjacent time points cannot complement each other. In the present invention, each dynamic uses a different under-sampling phase encoding matrix, and the high-frequency information collected by each under-sampling phase encoding matrix is ​​complementary and dynamically correlated, so the under-sampling ratio can be further improved, and then the chemical shift imaging data reconstruction model of multi-level and multi-dimensional similar problems can be used to reconstruct metabolic images with high spatiotemporal resolution.

[0020] 2. Reconstructing data using the chemical shift imaging data reconstruction model of multi-level and multi-dimensional similarity problems can make full use of the multi-level similarity prior information of adjacent dynamic metabolic imaging data image dimensions (spatial dimensions) and spectral dimensions to further improve the spatiotemporal resolution.

[0021] 3. This method is based on the traditional chemical shift imaging sequence CSI, which has low hardware requirements and is not easily affected by factors such as magnetic field inhomogeneity. It has strong stability and is easier to develop for practical application. It can obtain full-spectrum metabolic information and is suitable for exploratory metabolic research. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 It is a schematic diagram of the process of the present invention; Figure 2 Dynamic k-space metabolic imaging data ksp that complements high-frequency information t The collection map; Figure 3 This is the tissue of interest diagram of the mouse tissue of interest metabolic model in Example 1; Figure 4 This is a comparison diagram of the target magnetic resonance metabolic imaging data diagram of the present invention in Example 1 and the CSI magnetic resonance metabolic image. DETAILED DESCRIPTION

[0023] In order to facilitate those skilled in the art to understand and implement the present invention, the present invention is further described in detail below in conjunction with embodiments. It should be understood that the implementation examples described herein are only used to illustrate and explain the present invention and are not used to limit the present invention.

[0024] Example 1

[0025] This embodiment performs high temporal and spatial resolution magnetic resonance metabolic imaging on a mouse tissue metabolism model. A high temporal and spatial resolution magnetic resonance metabolic imaging method comprises: Step 1: Acquire dynamic k-space metabolic imaging data ksp with high-frequency information complementation at high undersampling ratio t , t=1, 2,…, T, T is the number of dynamic k-space metabolic imaging numbers.

[0026] The high-frequency information of dynamic k-space metabolic imaging data kspt, t=1, 2, …, T, can be complementary, which is mainly based on the k-space characteristics of chemical shift imaging CSI. The data in the peripheral part of the k-space of chemical shift imaging CSI contains high-frequency information, and high-frequency information determines the details and resolution of magnetic resonance metabolic imaging data. By using the same undersampling ratio in different dynamics and the undersampling phase encoding matrix with different positions of peripheral random undersampling, each dynamically acquired k-space metabolic imaging data ksp can be t Contains different high-frequency information. According to the characteristics of dynamic MRI metabolic imaging, adjacent dynamically acquired k-space metabolic imaging data ksp t There is a certain similarity in high-frequency information. This similarity allows the high-frequency information of k-space metabolic imaging data between different dynamics to complement each other after reconstruction using the chemical shift imaging data reconstruction model of multi-level and multi-dimensional similar problems, which not only effectively improves the reconstruction quality under high undersampling ratios, but also maximizes the retention of spatial details and resolution of magnetic resonance metabolic imaging data.

[0027] In the mouse tissue metabolism model, T=30 dynamic k-space metabolic imaging data are required, where T is the number of dynamic k-space metabolic imaging data. Then, 30 different undersampling phase encoding matrices need to be generated (the undersampling matrices have the same undersampling ratio, but the positions of the peripheral random undersampling are not exactly the same) as masks. t (t=1,2,……,T),mask t represents the undersampled phase encoding matrix under the t-th dynamic condition.

[0028] The k-space of chemical shift imaging (CSI) has three dimensions, two of which are phase encoding and the other is chemical shift frequency encoding. The under-sampling phase encoding matrix of the present invention performs under-sampling in the dimension of phase encoding.

[0029] The under-sampled phase encoding matrix consists of 0 and 1, where 0 means that the k-space data of the chemical shift imaging CSI at the phase encoding position corresponding to the image dimension (spatial dimension) is not collected, and 1 means that the k-space data of the chemical shift imaging CSI at the phase encoding position corresponding to the image dimension (spatial dimension) is collected. The values ​​of the central part of the under-sampled phase encoding matrix (approximately one-tenth of the entire under-sampled phase encoding matrix) are all 1, that is, the central part of the under-sampled phase encoding matrix is ​​all data collected. The values ​​of the peripheral part of the under-sampled phase encoding matrix (except the central part) are randomly 0, that is, the peripheral part of the under-sampled phase encoding matrix is ​​randomly data collected.

[0030] According to the under-sampling phase encoding matrix, the data in the center of the k-space of the chemical shift imaging CSI is fully collected, while the data in the peripheral part of the k-space of the chemical shift imaging CSI is randomly collected. In addition, the positions of the peripheral random under-sampling of the under-sampling phase encoding matrix used in different dynamic experiments are not exactly the same.

[0031] Calculation of undersampling ratio: The number of 1s in the undersampling phase encoding matrix is ​​m1, and the size of the undersampling phase encoding matrix is ​​M×N. M and N are the sizes of the undersampling phase encoding matrix in two phase encoding dimensions, respectively. Then, the undersampling ratio = M×N / m1. In this embodiment, the undersampling ratio is 9.

[0032] According to the order of the under-sampling phase encoding matrix, the k-space data of chemical shift imaging CSI at the phase encoding position with a value of 1 in the under-sampling phase encoding matrix is ​​collected to obtain dynamic k-space metabolic imaging data ksp with an under-sampling magnification of 9 t (t=1,2,……,T),ksp t Represents the k-space metabolic imaging data under the t-th dynamic state.

[0033] The order of the undersampling phase encoding matrix is ​​from mask t (t=1,2,……,T) is performed sequentially, and each mask t The phase encoding acquisition order is first from left to right, and then from top to bottom.

[0034] The undersampled phase encoding matrix mask under the t-th dynamic t The data collected in the k-space are placed according to the corresponding phase encoding position to obtain the k-space metabolic imaging data ksp under the t-th dynamic state. t .

[0035] Step 2: Reconstruct the magnetic resonance metabolic imaging data for the final goal by using the dynamic k-space metabolic imaging data with complementary high-frequency information and the chemical shift imaging data reconstruction model for similar problems at multiple levels and dimensions.

[0036] The chemical shift imaging data reconstruction model for multi-level and multi-dimensional similarity problems performs S-level divisions on the image dimension (spatial dimension) and spectrum dimension of the magnetic resonance metabolic imaging data, traversing the window and dynamically merging the dimensions to obtain the feature matrix B in (i=1,2,…,S; n=1,2,…,N i ), and constrain its low rank to maximize the use of multi-level similar prior information of adjacent dynamic magnetic resonance metabolic imaging data in the image dimension (spatial dimension) and spectral dimension, further improving the spatiotemporal resolution.

[0037] (1) The chemical shift imaging data reconstruction model for multi-level and multi-dimensional similarity problems (Formula 1) can be used to find the magnetic resonance metabolic imaging data x by minimizing the equation t (t=1, 2, ..., T): (1) in, x t The target is to solve the t-th dynamic magnetic resonance metabolic imaging data, where t is the dynamic sequence number.

[0038] E is the undersampling Fourier transform operator.

[0039] ksp t is the t-th dynamic k-space metabolic imaging data, t=1,2,…,T.

[0040] ‖‖2 represents the L2 norm.

[0041] For the i-th level S i The method of using MRI data x t The regularization parameter when traversing windows and merging dynamic dimensions is generally an empirical value, where S is the total number of levels.

[0042] S is the total number of levels. For the t-th dynamic MRI metabolic imaging data x t Performing traversal windows and dynamic dimension merging (S-level division) includes the following steps: The tth dynamic MRI metabolic imaging data x t It is a three-dimensional data, the sizes of the three dimensions are J, K and L respectively. For a magnetic resonance metabolic imaging data x of size [J, K, L] t Perform the i-th level S i =[j i ,k i ,l i ], where J and K are the magnetic resonance metabolic imaging data x t The image dimension, j i , ki is the image dimension of the image block corresponding to the i-th level, L is the magnetic resonance metabolic imaging data x t The spectral dimension, l i is the spectral dimension of the image block corresponding to the i-th level: that is, from the magnetic resonance metabolic imaging data x t Starting from the vertex of i ,k i ,l i As the three-dimensional size of the traversal window, the traversal window is used to traverse the magnetic resonance metabolic imaging data x t Traverse and obtain N i image blocks, n is the image block sequence number, N i is the MRI metabolic imaging data x at the i-th level t The total number of image blocks to be divided, each image block b n The dimensions (two image dimensions and one spectrum dimension) are [j i ,k i ,l i ], the image block b n The rows of the image dimension plane where the two image dimensions are located are spliced ​​in sequence to obtain a one-dimensional vector of the image dimension plane, and then the one-dimensional vectors of the image dimension plane corresponding to each spectral dimension of the image block are spliced ​​in sequence to obtain a one-dimensional vector of the image block. The size of the one-dimensional vector of the image block is j i ×k i × i , according to the dynamic order, the one-dimensional vectors of each dynamic image block corresponding to the image block at the same position are spliced ​​to obtain the feature matrix B in , then we get the feature matrix B in The matrix size is [j i ×k i × i ,T]. is the feature matrix B in The nuclear norm is equal to the feature matrix B in The sum of all singular values. That is, the multi-level and multi-dimensional similar prior information of dynamic k-space metabolic imaging data is used to complement the dynamic high-frequency information through similarity constraints.

[0043] Magnetic resonance metabolic imaging datax t It is a three-dimensional data, in which the first two dimensions represent the image dimension and the third dimension is the spectrum dimension. According to the characteristics of chemical shift imaging: chemical shift images have similarity within a range near the spectral peak of a compound. Therefore, according to the half-height width of the spectral peak and the spectral resolution, the hierarchical division size in the spectral dimension is adjusted during hierarchical division.

[0044] For example, if the half-width at half maximum of the spectrum peak is 90 Hz and the spectrum resolution is 15 Hz, the block size of the spectrum dimension is: half-width at half maximum of the spectrum peak / spectrum resolution = 6.

[0045] (2) Substitute the known data into equation (1) and solve it iteratively.

[0046] The target solution of magnetic resonance metabolic imaging data is to convert k-space metabolic imaging data ksp t (t=1,2,...,T) is input into the chemical shift imaging data reconstruction model formula (1) of the multi-level and multi-dimensional similar problem, and after cyclic iteration, the preset iteration stop condition (the number of iterations is reached, or Less than the set threshold ), the target magnetic resonance metabolic imaging data x is obtained.

[0047] The reconstruction iteration process is as follows: Initial MRI data x t (t=1,2,...,T) can be directly obtained from the k-space metabolic imaging data ksp t (t=1,2,…,T) is filled with zeros and then directly Fourier transform is performed to obtain it.

[0048] Magnetic resonance metabolic imaging data x t (t=1,2,...,T) performs the above traversal window and dynamic dimension merging operations to obtain the feature matrix B in , for each magnetic resonance metabolic imaging data x t The corresponding feature matrix B in Perform singular value soft threshold processing, that is, first perform each magnetic resonance metabolic imaging data x t The corresponding feature matrix B in Perform SVD decomposition, each B in After the matrix SVD decomposition, a series of singular values ​​are obtained. , where each B in The largest singular value after matrix SVD decomposition is recorded as the corresponding maximum singular value , keep each feature matrix B in After SVD decomposition, the value is greater than × The singular values ​​of the matrix P are reconstructed by SVD in , for each magnetic resonance metabolic imaging data x t The corresponding feature matrix B in After the above operations are performed, each magnetic resonance metabolic imaging data x t The corresponding P in The matrix is ​​traversed by the inverse operation of window traversal and dynamic dimension merging to obtain the low-rank constrained magnetic resonance metabolic imaging matrix L t(t=1,2,……,T).

[0049] Calculate the k-space residual for each dynamic (t=1,2,……,T): , E is the undersampling Fourier transform operator.

[0050] k-space residual (t=1,2,...,T) returns to the MRI domain to obtain the image domain residual (t=1,2,……,T): , is the undersampled inverse Fourier transform operator.

[0051] Then the updated MRI data x t 1 (t=1,2,……,T) is: , Calculation Judgment Is the value less than the set threshold? ,like If the value is less than , the iteration is stopped, and the updated magnetic resonance metabolic imaging data (t=1,2,...,T) is the magnetic resonance metabolic imaging data to be solved. The value is greater than or equal to the threshold The updated MRI metabolic imaging data As new MRI metabolic imaging data t , continue iterating and updating according to the above process until Less than , or the number of iterations is greater than the set number of iterations q, the last updated magnetic resonance metabolic imaging data As the final target magnetic resonance metabolic imaging data, the set number of iterations q can be set according to actual needs. The value can be set according to actual needs.

[0052] Figure 3 This is a magnetic resonance structural image of a mouse model of tissue metabolism.

[0053] Figure 4This is a comparison of the magnetic resonance metabolic imaging of the present invention and chemical shift imaging CSI on the same mouse tissue model. It can be clearly seen that the present invention improves the temporal and spatial resolution of magnetic resonance metabolic imaging by about 5 times compared with CSI (the temporal and spatial resolution of CSI is 3.5mm×3.5mm×6mm, one dynamic every 10 minutes; the temporal and spatial resolution of the present invention is 2.1mm×2.1mm×6mm, one dynamic every 6 minutes), and the magnetic resonance metabolic image obtained by this method is more consistent with the structural image of the tissue of interest.

[0054] A person skilled in the art can understand that the implementation of step 1 and step 2 in the above embodiment can be completed by instructing related hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processing flow of the above steps.

[0055] Example 2

[0056] In one embodiment, a high spatiotemporal resolution magnetic resonance metabolic imaging device is provided, comprising a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the processing flow of step 1 and step 2 in the above embodiment 1 is implemented.

[0057] Example 3

[0058] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the processing flow of step 1 and step 2 in the above-mentioned embodiment 1 is implemented.

[0059] Example 4

[0060] In one embodiment, a computer program product is provided, including a computer program, which implements the processing flow of step 1 and step 2 in the above embodiment 1 when executed by a processor.

[0061] It should be noted that the embodiments described in the present invention are merely examples of the spirit of the present invention. Those skilled in the art may make various modifications or additions to the described embodiments or replace them in similar ways, but they will not deviate from the spirit of the present invention or exceed the scope defined by the attached claims.

Claims

1. A high temporal and spatial resolution magnetic resonance metabolic imaging method, characterized in that: include: Step 1: Acquire dynamic k-space metabolic imaging data ksp t , t=1, 2, ..., T, T is the number of dynamic k-space metabolic imaging numbers, Step 2: Use dynamic k-space metabolic imaging data and chemical shift imaging data to reconstruct the model and obtain the magnetic resonance metabolic imaging data for the final target solution.

2. A high spatiotemporal resolution magnetic resonance metabolic imaging method according to claim 1, characterized in that: The k-space metabolic imaging data ksp t Based on the under-sampling phase encoding matrix, under-sampling is performed in the phase encoding dimension, data is collected in the central part of the under-sampling phase encoding matrix, and data is randomly collected in the peripheral part of the under-sampling phase encoding matrix. The positions of the peripheral random under-sampling of under-sampling phase encoding matrices with different dynamics are not exactly the same.

3. A high temporal and spatial resolution magnetic resonance metabolic imaging method according to claim 2, characterized in that: In step 2, k-space metabolic imaging data ksp t The data is input into the chemical shift imaging data reconstruction model, and when the preset iteration stop condition is reached after cyclic iteration, the magnetic resonance metabolic imaging data of the final target solution is obtained.

4. The high spatiotemporal resolution magnetic resonance metabolic imaging method according to claim 2, characterized in that: The chemical shift imaging data reconstruction model is based on the following formula: ， Among them, x t The target is to solve the t-th dynamic magnetic resonance metabolic imaging data, t is the dynamic sequence number, E is the undersampling Fourier transform operator, ksp t is the t-th dynamic k-space metabolic imaging data, ‖ ‖2 represents the L2 norm, is the MRI metabolic imaging data x in the i-th level t The regularization parameter when performing traversal windows and dynamic dimension merging. S is the total number of levels. is the feature matrix B in The nuclear norm of i is the layer number, n is the image block number, N i is the MRI metabolic imaging data x at the i-th level t The total number of image blocks to be partitioned.

5. The high spatiotemporal resolution magnetic resonance metabolic imaging method according to claim 4, characterized in that: The magnetic resonance metabolic imaging data x t Performing traversal windows and dynamic dimension merging includes: From MRI metabolic imaging data x t Starting from the vertex of t Traverse and obtain N i Image block b n , each image block b n The dimension size is [j i ,k i ,l i ], J, K are the magnetic resonance metabolic imaging data x t The image dimension, j i , k i is the image dimension of the image block corresponding to the i-th level, L is the magnetic resonance metabolic imaging data x t The spectral dimension, l i is the spectral dimension of the image block corresponding to the i-th level, and the image block b n The rows of the image dimension plane where the two image dimensions are located are spliced ​​in sequence to obtain a one-dimensional vector of the image dimension plane, and then the one-dimensional vectors of the image dimension plane corresponding to each spectrum dimension of the image block are spliced ​​in sequence to obtain a one-dimensional vector of the image block. In a dynamic order, the one-dimensional vectors of each dynamic image block corresponding to the image block at the same position are spliced ​​to obtain the feature matrix B in .

6. The high spatiotemporal resolution magnetic resonance metabolic imaging method according to claim 5, characterized in that: Also includes: For each MRI data x t The corresponding feature matrix B in Perform SVD decomposition, each B in After the matrix SVD decomposition, a series of singular values ​​are obtained. , where each B in The largest singular value after matrix SVD decomposition is recorded as the corresponding maximum singular value , keep each feature matrix B in After SVD decomposition, the value is greater than × The singular values ​​of the matrix P are reconstructed by SVD in , each magnetic resonance metabolic imaging data x t The corresponding matrix P in According to the inverse operation of traversal window and dynamic dimension merging, the low-rank constrained MRI matrix L is obtained. t , calculate the updated MRI data x t 1 , ， ， ， in, is the k-space residual Return to the MRI domain to obtain the image domain residual, is the undersampled inverse Fourier transform operator, is the k-space domain residual, and E is the undersampling Fourier transform operator.

7. The high spatiotemporal resolution magnetic resonance metabolic imaging method according to claim 6, characterized in that: The preset iteration stop condition is reaching the number of iterations, or The value is less than the set threshold .

8. A high temporal and spatial resolution magnetic resonance metabolic imaging device, comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, step 1 and step 2 according to any one of claims 1 to 7 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, step 1 and step 2 according to any one of claims 1 to 7 are implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, step 1 and step 2 according to any one of claims 1 to 7 are implemented.

Images