Magnetic resonance imaging method, device, storage medium and electronic equipment

Abstract

The present specification discloses a magnetic resonance imaging method, device, storage medium and electronic device. When the sampling method is partial Fourier sampling, partial Fourier sampling is performed on the magnetic resonance signal generated by the scanning part to obtain the first sampling data, and a partial Fourier reconstruction algorithm and / or a generation network reconstruction algorithm are applied to perform image reconstruction. When the sampling method selected by the user is non-partial Fourier sampling, non-partial Fourier sampling is performed on the magnetic resonance signal generated by the scanning part to determine the second sampling data, and a regression network reconstruction algorithm is applied to determine the second reconstruction data for image reconstruction. When performing magnetic resonance imaging through this method, the user can flexibly select the sampling method according to actual needs, and perform magnetic resonance reconstruction through the sampling method selected by the user. The flexibility of this method is high, and the method can flexibly adapt to different imaging needs to obtain a target reconstructed image that matches the user's imaging needs.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a magnetic resonance imaging method, device, storage medium and electronic device. Background Art

[0002] In existing magnetic resonance imaging devices, a device is usually configured with a specific magnetic resonance imaging algorithm. When using the magnetic resonance imaging device for magnetic resonance imaging, only the magnetic resonance imaging algorithm configured in the device can be used. Users cannot select the magnetic resonance imaging algorithm according to imaging requirements, and the application of the magnetic resonance imaging device is not flexible enough.

[0003] Therefore, the present specification provides a method for magnetic resonance imaging. Summary of the invention

[0004] The present specification provides a magnetic resonance imaging method, device, storage medium and electronic device to at least partially solve the above-mentioned problems existing in the prior art.

[0005] This manual adopts the following technical solutions:

[0006] This specification provides a magnetic resonance imaging method, comprising:

[0007] Determine the sampling method selected by the user;

[0008] When the sampling mode is partial Fourier sampling, partial Fourier sampling is performed on the magnetic resonance signal generated by the scanning part to obtain first sampling data; a partial Fourier reconstruction algorithm and / or a generated network reconstruction algorithm are applied to determine first reconstruction data; and a target reconstructed image is determined based on the first reconstruction data;

[0009] When the sampling method selected by the user is non-partial Fourier sampling, non-partial Fourier sampling is performed on the magnetic resonance signal generated by the scanning part to determine the second sampling data; a regression network reconstruction algorithm is applied to determine the second reconstruction data, and a target reconstructed image is determined based on the second reconstruction data.

[0010] Optionally, applying a partial Fourier reconstruction algorithm and / or a generation network reconstruction algorithm to determine the first reconstruction data specifically includes:

[0011] When the sampling method is partial Fourier sampling, an algorithm selection interface is displayed;

[0012] Determining the reconstruction algorithm selected by the user in the algorithm selection interface, and judging whether the reconstruction algorithm is a combined reconstruction algorithm;

[0013] If so, applying a partial Fourier reconstruction algorithm and a generative network reconstruction algorithm to determine first reconstruction data;

[0014] If not, a partial Fourier reconstruction algorithm or a generation network reconstruction algorithm is applied to determine the first reconstruction data.

[0015] Optionally, applying a partial Fourier reconstruction algorithm and a generation network reconstruction algorithm to determine the first reconstruction data specifically includes:

[0016] Applying a partial Fourier reconstruction algorithm to determine first reconstructed sub-data corresponding to the first sampled data; applying a generative network reconstruction algorithm to determine second reconstructed sub-data corresponding to the first sampled data;

[0017] First reconstruction data are determined according to the first reconstruction sub-data and the second reconstruction sub-data.

[0018] Optionally, applying a partial Fourier reconstruction algorithm to determine first reconstructed sub-data corresponding to the first sampled data specifically includes:

[0019] Inputting the first sampling data into a first reconstruction network to obtain first initial reconstruction data;

[0020] determining an initial reconstructed image according to the first initial reconstruction data;

[0021] Phase correction is performed on the initial reconstructed image to obtain first reconstructed sub-data.

[0022] Optionally, applying a generated network reconstruction algorithm to determine second reconstructed sub-data corresponding to the first sampled data specifically includes:

[0023] Inputting the first sampling data into a second reconstruction network to obtain second initial reconstruction data;

[0024] The second initial reconstruction data is input into the generation network to obtain second reconstruction sub-data.

[0025] Optionally, determining a target reconstructed image according to the first reconstruction data specifically includes:

[0026] Performing a data fidelity operation on the first reconstruction data to obtain target reconstruction data;

[0027] The target reconstruction data is subjected to inverse Fourier transform to obtain a target reconstructed image.

[0028] Optionally, applying a regression network reconstruction algorithm to determine second reconstruction data, and determining a target reconstructed image according to the second reconstruction data specifically includes:

[0029] Inputting the second sampling data into a third reconstruction network to obtain third initial reconstruction data;

[0030] Inputting the third initial reconstruction data into a regression network to obtain target reconstruction data;

[0031] The target reconstruction data is subjected to inverse Fourier transform to obtain a target reconstructed image.

[0032] This specification provides a magnetic resonance imaging device, the device comprising:

[0033] A sampling method determination module determines the sampling method selected by the user;

[0034] A partial Fourier sampling module, when the sampling mode is partial Fourier sampling, determines a sampling range, performs partial Fourier sampling on the magnetic resonance signal generated by the scanning part within the sampling range, and obtains first sampling data; applies a partial Fourier reconstruction algorithm and / or generates a network reconstruction algorithm to determine first reconstruction data; and determines a target reconstructed image according to the first reconstruction data;

[0035] The non-partial Fourier sampling module, when the sampling method selected by the user is non-partial Fourier sampling, determines the sampling range, performs non-partial Fourier sampling on the magnetic resonance signal generated by the scanning part, and determines the second sampling data; applies the regression network reconstruction algorithm to determine the second reconstruction data, and determines the target reconstructed image according to the second reconstruction data. This specification provides a computer-readable storage medium, the storage medium stores a computer program, and the computer program implements the above-mentioned magnetic resonance imaging method when executed by a processor.

[0036] The present specification provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned magnetic resonance imaging method when executing the program.

[0037] At least one of the above technical solutions adopted in this specification can achieve the following beneficial effects:

[0038] In the magnetic resonance imaging method provided in this specification, the sampling method selected by the user is determined. When the sampling method is partial Fourier sampling, partial Fourier sampling is performed on the magnetic resonance signal generated by the scanning part to obtain the first sampling data, and a partial Fourier reconstruction algorithm and / or a generation network reconstruction algorithm are applied to perform image reconstruction. When the sampling method selected by the user is non-partial Fourier sampling, non-partial Fourier sampling is performed on the magnetic resonance signal generated by the scanning part to determine the second sampling data, and a regression network reconstruction algorithm is applied to determine the second reconstruction data for image reconstruction. When performing magnetic resonance imaging through this method, the user can flexibly select the sampling method according to actual needs, and perform magnetic resonance reconstruction through the sampling method selected by the user. The flexibility of this method is relatively high. Through this method, different imaging requirements can be flexibly adapted to obtain a target reconstructed image that matches the user's imaging requirements. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] The drawings described herein are used to provide a further understanding of this specification and constitute a part of this specification. The illustrative embodiments and descriptions of this specification are used to explain this specification and do not constitute an improper limitation on this specification. In the drawings:

[0040] Figure 1 A schematic diagram of a magnetic resonance imaging method in this specification;

[0041] Figure 2 A schematic diagram of a partial Fourier sampling range provided in this specification;

[0042] Figure 3 A schematic diagram of a non-partial Fourier sampling range provided for this specification;

[0043] Figure 4 A schematic diagram of algorithm selection for a magnetic resonance imaging process provided in an embodiment of this specification;

[0044] Figure 5 A network configuration structure diagram of a magnetic resonance reconstruction algorithm provided in an embodiment of this specification;

[0045] Figure 6 A schematic diagram of a magnetic resonance imaging device provided in this specification;

[0046] Figure 7 The corresponding Figure 1 Schematic diagram of electronic equipment. DETAILED DESCRIPTION

[0047] In order to make the purpose, technical solutions and advantages of this specification more clear, the technical solutions of this specification will be clearly and completely described below in combination with the specific embodiments of this specification and the corresponding drawings. Obviously, the described embodiments are only part of the embodiments of this specification, not all of the embodiments. Based on the embodiments in this specification, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0048] The technical solutions provided by the embodiments of this specification are described in detail below in conjunction with the accompanying drawings.

[0049] Figure 1 The following is a flow chart of a magnetic resonance imaging method in this specification, which specifically includes the following steps:

[0050] S100: Determine the sampling method selected by the user.

[0051] All steps in the magnetic resonance imaging method provided in this specification can be implemented by any electronic device with computing functions, such as a terminal, a server, etc. For ease of description, the magnetic resonance imaging method provided in this specification is described below with the server as the execution subject.

[0052] The magnetic resonance imaging method provided in this specification can simultaneously support a magnetic resonance imaging algorithm based on partial Fourier sampling and a magnetic resonance imaging algorithm based on non-partial Fourier sampling.

[0053] Because different sampling methods require different imaging algorithms, the server needs to determine the sampling method selected by the user before performing magnetic image reconstruction.

[0054] Specifically, the server may display selection information to the user, and determine the sampling method used in magnetic resonance imaging in response to the user's selection. The selection information may be used to prompt the user to select a sampling method, and this specification does not limit the specific content of the selection information.

[0055] Because different sampling methods require different imaging algorithms, existing imaging devices usually only support specific sampling methods and configure sampling algorithms suitable for the specific sampling methods. Imaging algorithms cannot be selected according to user needs, and the flexibility is poor, and the diverse user needs cannot be met. Different imaging algorithms have different advantages and can meet different user needs. The method of this manual can meet more diverse user needs.

[0056] S102: When the sampling method is partial Fourier sampling, partial Fourier sampling is performed on the magnetic resonance signal generated by the scanning part to obtain first sampling data; a partial Fourier reconstruction algorithm and / or a network reconstruction algorithm is applied to determine first reconstruction data; and a target reconstructed image is determined based on the first reconstruction data.

[0057] In magnetic resonance imaging, the spatial frequency information of the magnetic resonance signal of the scanned part is stored in K space. K space is a frequency domain space, and each data point in K space contains the spatial frequency and phase information of all pixels in the image domain space.

[0058] In order to obtain a magnetic resonance image in the image domain space, it is necessary to sample in the K space to obtain a K space data matrix, and then perform an inverse Fourier transform on the sampled K space data matrix to convert the K space data from the frequency domain space to the image domain space. In the inverse Fourier transform, the data dimension of the K space data matrix is ​​consistent with the pixel data dimension contained in the converted magnetic resonance image.

[0059] In magnetic resonance scanning, the more K-space data collected, the longer the scanning time. In order to speed up the scanning process, only part of the K-space data is collected in this manual to obtain an undersampled K-space data matrix, and a reconstruction algorithm is used to obtain a nearly fully sampled K-space data matrix based on the undersampled K-space data matrix. In a fully sampled K-space data matrix, the element value of each element position is obtained by sampling in the K-space. In an undersampled K-space data matrix, only the element values ​​of some element positions are obtained by sampling in the K-space.

[0060] When the sampling method selected by the user is partial Fourier sampling, the server performs partial Fourier sampling on the magnetic resonance signal generated by the scanning part in K space to obtain first sampling data. The first sampling data is the under-sampled K space data obtained by using the partial Fourier sampling method.

[0061] In the partial Fourier sampling method, the conjugate symmetry of K space is applied, and the server samples in K space within a preset sampling range according to a preset sampling range, wherein the preset sampling range is at least half of the data range in K space.

[0062] Figure 2 This is a partial Fourier sampling range diagram provided in this specification. Figure 2 As shown, the rectangular part represents the K space, the part filled with horizontal lines represents the sampling range, and the sampling range including the horizontal lines is a part of the data range in the K space.

[0063] In order to obtain nearly fully sampled K-space data based on the first sampling data, the server applies a partial Fourier reconstruction algorithm and / or generates a network reconstruction algorithm to determine the first reconstruction data.

[0064] Then, the server determines a target reconstructed image according to the first reconstruction data.

[0065] The first reconstruction data is nearly fully sampled K-space data generated by the server based on the first sampling data. The server can obtain a magnetic resonance image in a corresponding image domain according to the first reconstruction data.

[0066] Specifically, the server may use the first reconstruction data as target reconstruction data and perform inverse Fourier transform on the target reconstruction data to obtain a target reconstructed image.

[0067] In the magnetic resonance imaging method of the present specification, there are three imaging algorithms corresponding to partial Fourier sampling that can be selected by the user, namely, a combined reconstruction algorithm using a partial Fourier reconstruction algorithm and a generative network reconstruction algorithm, a generative network reconstruction algorithm, and a partial Fourier reconstruction algorithm.

[0068] In one or more embodiments of the present specification, in order to facilitate user operation, the server may display an algorithm selection interface to the user after determining that the sampling method is partial Fourier sampling.

[0069] On the algorithm selection page, users can choose to use a combined reconstruction algorithm or a non-combined reconstruction algorithm. The combined reconstruction algorithm is to use both the partial Fourier reconstruction algorithm and the generative network reconstruction algorithm for image reconstruction. The non-combined reconstruction algorithm is to use one of the partial Fourier reconstruction algorithm or the generative network reconstruction algorithm for image reconstruction.

[0070] The server determines the reconstruction algorithm selected by the user in the algorithm selection interface, and determines whether the reconstruction algorithm selected by the user is a combined reconstruction algorithm.

[0071] When the reconstruction algorithm selected by the user is a combined reconstruction algorithm, the server simultaneously applies a partial Fourier reconstruction algorithm and a generative network reconstruction algorithm to perform image reconstruction and determine first reconstruction data.

[0072] When the reconstruction algorithm selected by the user is not a combined reconstruction algorithm, the server applies a partial Fourier reconstruction algorithm or a generated network reconstruction algorithm to perform image reconstruction, and applies the partial Fourier reconstruction algorithm or the generated network reconstruction algorithm to determine the first reconstruction data.

[0073] The non-combined reconstruction algorithm in this specification can be configured as either a partial Fourier reconstruction algorithm or a generative network reconstruction algorithm, or can be configured to include both a partial Fourier reconstruction algorithm and a generative network reconstruction algorithm.

[0074] If the non-combined reconstruction algorithm is configured as a generated network reconstruction algorithm, when the user selects the non-combined reconstruction algorithm, the server uses the generated network reconstruction algorithm to determine the first reconstruction data. If the non-combined reconstruction algorithm is configured as a partial Fourier reconstruction algorithm, when the user selects the non-combined reconstruction algorithm, the server uses the partial Fourier reconstruction algorithm to determine the first reconstruction data.

[0075] If the non-combined reconstruction algorithm is configured to include both the partial Fourier reconstruction algorithm and the generative network reconstruction algorithm, the user can further choose to use the partial Fourier reconstruction algorithm or the generative network reconstruction algorithm for magnetic resonance imaging according to the reconstruction requirements. In this way, when the user selects the partial Fourier sampling method, the server can provide three reconstruction algorithms for the user to choose from, which can meet more user needs.

[0076] When partial Fourier sampling is used, the number of samples is relatively small and the imaging speed is fast. When the user has higher requirements for imaging speed, partial Fourier sampling can be selected.

[0077] S104: When the sampling method selected by the user is non-partial Fourier sampling, non-partial Fourier sampling is performed on the magnetic resonance signal generated by the scanning part to determine the second sampling data; a regression network reconstruction algorithm is applied to determine the second reconstruction data, and a target reconstructed image is determined based on the second reconstruction data.

[0078] In the non-partial Fourier sampling mode, the server performs sampling in the entire K-space range according to a preset sampling interval to obtain second sampling data, which is also under-sampled K-space data.

[0079] Figure 3 This is a schematic diagram of a non-partial Fourier sampling range provided in this specification. Figure 3 As shown in the figure, the rectangular part represents the K space, and the horizontal line filled part represents the sampling range. Figure 3 It can be seen that the sampling range of the non-partial Fourier sampling method is the entire K space.

[0080] Then, the server reconstructs the image based on the second sampling data through a regression network reconstruction algorithm to determine a target reconstructed image.

[0081] Specifically, the server inputs the second sampling data into the third reconstruction network to obtain third initial reconstruction data, and inputs the third initial reconstruction data into the regression network to reconstruct the third initial reconstruction data into K-space data closer to full sampling through the regression network to obtain target reconstruction data.

[0082] The third reconstruction network is used to perform feature extraction based on the second sampling data, and perform preliminary data reconstruction based on the feature extraction result to obtain third initial reconstruction data, where the third initial reconstruction data is K-space data that has undergone preliminary reconstruction.

[0083] The regression network reconstructs the K-space data based on the third initial reconstruction data to obtain K-space data close to full sampling, namely, target reconstruction data. The target reconstruction data is subjected to inverse Fourier transform to obtain a target reconstructed image.

[0084] When non-partial Fourier sampling is used, the number of samples is relatively large. Based on the large number of samples, the target reconstruction data obtained is closer to the fully sampled K space and contains more detailed features. The reconstructed image obtained after inverse Fourier transform of such target reconstruction data has richer details and higher image quality. However, the increase in the number of samples also leads to a slower imaging speed in this embodiment. When the user has higher requirements for the image quality of the target reconstructed image, the non-partial Fourier sampling method can be selected.

[0085] In the magnetic resonance imaging method provided in this specification, the sampling method selected by the user is determined. When the sampling method is partial Fourier sampling, partial Fourier sampling is performed on the magnetic resonance signal generated by the scanning part to obtain the first sampling data, and the partial Fourier reconstruction algorithm and / or the generation network reconstruction algorithm are applied to determine the first reconstruction data, and the target reconstructed image is determined based on the first reconstruction data. When the sampling method selected by the user is non-partial Fourier sampling, non-partial Fourier sampling is performed on the magnetic resonance signal generated by the scanning part to determine the second sampling data, and the regression network reconstruction algorithm is applied to determine the second reconstruction data, and the target reconstructed image is determined based on the second reconstruction data. When performing magnetic resonance imaging through this method, the user can flexibly select the sampling method according to actual needs, and perform magnetic resonance reconstruction through the sampling method selected by the user. The flexibility of this method is high, and the method can flexibly adapt to different imaging needs to obtain a target reconstructed image that matches the user's imaging needs.

[0086] Figure 4 A schematic diagram of algorithm selection for a magnetic resonance imaging process provided in an embodiment of this specification. Figure 4 In the illustrated embodiment, before magnetic resonance reconstruction begins, the user can choose whether to use partial Fourier sampling according to the requirements for imaging speed and imaging quality. If the user requires faster imaging speed, partial Fourier sampling can be selected, and then the desired reconstruction algorithm can be selected. A combined reconstruction algorithm combining the partial Fourier reconstruction algorithm and the generative network reconstruction algorithm, or a generative network reconstruction algorithm, or a partial Fourier reconstruction algorithm is selected to reconstruct the target reconstruction image. If the user requires higher imaging quality, non-partial Fourier sampling can be selected, and a regression network reconstruction algorithm can be applied to reconstruct the target reconstruction image.

[0087] In the above step S102, when partial Fourier sampling is applied, there are multiple imaging algorithms that the user can select. In the magnetic resonance imaging process, the user needs to select the imaging algorithm again after selecting the sampling method, and then the server can determine the magnetic resonance imaging algorithm required by the user.

[0088] Because the user needs to perform two selection operations in this process, the operation process is relatively cumbersome. In one or more embodiments of this specification, in order to simplify the user's operation steps in magnetic resonance imaging, the server can set a default imaging algorithm so that when the user does not select an imaging algorithm, the default imaging algorithm can be directly applied to perform magnetic resonance imaging through partial Fourier sampling.

[0089] The default imaging algorithm can be set to any one of a combined reconstruction algorithm of a partial Fourier reconstruction algorithm and a generative network reconstruction algorithm, a generative network reconstruction algorithm, and a partial Fourier reconstruction algorithm.

[0090] In this embodiment, after determining that the sampling method is partial Fourier sampling, the server may not display the algorithm selection interface to the user. If the user does not need to change the default imaging algorithm, the user does not need to select the algorithm, and the server reconstructs the magnetic resonance image according to the default imaging algorithm.

[0091] If the user needs to change the default imaging algorithm, the server responds to the user's modification requirements and displays an algorithm selection interface so that the user can change the imaging algorithm used in the current imaging process, or replace other default imaging algorithms so that in subsequent imaging processes, the changed magnetic resonance imaging algorithm can be applied to perform magnetic resonance imaging.

[0092] In the above step S102, when the reconstruction algorithm selected by the user is a combined reconstruction algorithm, the server simultaneously applies a partial Fourier reconstruction algorithm and a generative network reconstruction algorithm to perform image reconstruction.

[0093] The server applies a partial Fourier reconstruction algorithm to determine first reconstructed sub-data corresponding to the first sampled data, and applies a generated network reconstruction algorithm to determine second reconstructed sub-data corresponding to the first sampled data.

[0094] This specification does not limit the order of determining the first reconstruction sub-data and the second reconstruction sub-data. The server may choose to determine the first reconstruction sub-data first, or determine the second reconstruction sub-data first, or determine the first reconstruction sub-data and the second reconstruction sub-data at the same time.

[0095] The specific process of the server determining the first reconstruction sub-data is as follows:

[0096] First, the server inputs the first sampling data into the first reconstruction network to obtain the first initial reconstruction data.

[0097] Because the first sampling data is under-sampled K-space data, there are certain element positions in the first sampling data without corresponding sampling values. Therefore, an accurate magnetic resonance image cannot be determined directly based on the first sampling data.

[0098] The first reconstruction network in this step is used to perform feature extraction based on the first sampling data, and perform preliminary K-space data reconstruction based on the feature extraction result to obtain first initial reconstruction data, which is the K-space data that has undergone preliminary reconstruction.

[0099] Then, the server determines an initial reconstructed image according to the first initial reconstructed data, and performs phase correction on the initial reconstructed image to obtain first reconstructed sub-data.

[0100] Specifically, the server performs an inverse Fourier transform on the first initial reconstruction data to obtain an initial reconstructed image, and performs a low-pass filter on the first initial reconstruction data to obtain low-frequency data.

[0101] Specifically, the server may determine a low-pass filter and a preset cutoff frequency, and the low-pass filter may be any existing low-pass filter, such as a Gaussian filter, a Butterworth filter, etc. Then, the server filters the data values ​​in the first initial reconstruction data whose frequency is greater than the cutoff frequency to obtain low-frequency data.

[0102] Then, the server performs an inverse Fourier transform on the low-frequency data to determine the phase of the low-frequency data as a correction phase, and performs phase correction on the initial reconstructed image using the correction phase to obtain the first reconstructed sub-data.

[0103] Specifically, the server performs an inverse Fourier transform on the low-frequency data to obtain a low-resolution image. For each pixel in the low-resolution image, according to the coordinates of the pixel, a pixel with the same coordinates as the pixel is determined in the initial reconstructed image as the original pixel of the pixel. The low-resolution image is an image obtained by performing low-frequency feature extraction on the initial reconstructed image, and the pixel is the pixel after the low-frequency feature extraction is performed on the original pixel of the pixel.

[0104] The server uses the phase of the pixel as the correction phase of the original pixel of the pixel. The server replaces the original pixel phase of the pixel with the correction phase of the original pixel of the pixel to obtain a phase-corrected pixel corresponding to the pixel.

[0105] The following is a further explanation of this process through the complex representation of pixels:

[0106] For each pixel I in the low-resolution image low-res (x,y), the complex form of the pixel is:

[0107]

[0108] Among them, |I low-res (x, y)| represents the amplitude of the pixel, Φ1(x, y) represents the correction phase of the pixel, and e is the base of the natural logarithm.

[0109] For the original pixel point I of the pixel point in the initial reconstructed image initial (x,y), the complex form of the original pixel is:

[0110]

[0111] Among them, |I initial (x, y)| represents the amplitude of the original pixel, and Φ2(x, y) represents the correction phase of the original pixel.

[0112] According to the amplitude of the original pixel point and the phase of the pixel point, a pixel point after phase correction is constructed, and the pixel point I corresponding to the phase correction of the pixel point is constructed. corrected (x,y) is represented as:

[0113]

[0114] The server determines a phase-corrected image according to the phase-corrected pixel points corresponding to each pixel point, and performs Fourier transform on the phase-corrected image to obtain first reconstructed sub-data.

[0115] In this embodiment, because the low-resolution image is an image obtained after the initial reconstructed image is low-pass filtered, high-frequency noise is removed from the low-resolution image compared to the initial reconstructed image, and the phase information contained in the low-resolution image is more accurate, the first reconstructed sub-data obtained through the phase correction step is also more accurate. Since the first reconstructed sub-data is phase-corrected based on the first initial reconstructed data, the first reconstructed sub-data is more accurate full-sampled K-space data, and a more accurate magnetic resonance image can be obtained through the first reconstructed sub-data.

[0116] The above is only an embodiment of a partial Fourier reconstruction algorithm. In addition to the partial Fourier reconstruction algorithm provided above, the server can also use other partial Fourier reconstruction algorithms such as conjugate synthesis algorithm, zero difference detection algorithm, Projections onto ConvexSets (POCS) algorithm, etc. to determine the first reconstructed sub-data.

[0117] The specific process of the server determining the second reconstruction sub-data is as follows:

[0118] First, the server inputs the first sampling data into the second reconstruction network to obtain second initial reconstruction data.

[0119] The second reconstruction network is used to perform feature extraction based on the first sampling data, and perform preliminary data reconstruction based on the feature extraction result to obtain second initial reconstruction data, where the second initial reconstruction data is K-space data that has undergone preliminary reconstruction.

[0120] Then, the server inputs the initial reconstruction data into the generation network to obtain the second reconstruction sub-data.

[0121] After determining the first reconstruction sub-data and the second reconstruction sub-data, the server determines the first reconstruction data according to the first reconstruction sub-data and the second reconstruction sub-data.

[0122] Specifically, the server fuses the first reconstructed sub-data with the second reconstructed sub-data to obtain the first reconstructed data. This specification does not limit the specific fusion method, and the first reconstructed sub-data and the second reconstructed sub-data may be fused by weighted summation, matrix splicing, etc.

[0123] The partial Fourier reconstruction algorithm provided in the embodiments of this specification reduces the influence of noise data and noise sensitivity through phase correction, and can obtain K-space data that is closer to full sampling. In the generative network reconstruction algorithm, by training the generative network, the generative network can learn the subtle structural features and texture information of the magnetic resonance image, and obtain a target reconstructed image with rich details.

[0124] Therefore, in this embodiment, the target reconstruction data is determined based on the first reconstruction sub-data determined by the partial Fourier reconstruction algorithm and the second reconstruction sub-data determined by the generative network reconstruction algorithm. The reconstruction method of this embodiment combines the advantages of the partial Fourier reconstruction algorithm and the generative network reconstruction algorithm, and can obtain a target reconstructed image with low noise sensitivity and rich details.

[0125] At the same time, because the present embodiment adopts a partial Fourier sampling method, the amount of sampling data is small, the scanning time is short, and the magnetic resonance imaging efficiency is high.

[0126] In the above step S102, when the reconstruction algorithm selected by the user is a non-combination reconstruction algorithm, the server may adopt a partial Fourier reconstruction algorithm to determine the first reconstruction data.

[0127] Specifically, the server inputs the first sampling data into the first reconstruction network to obtain the first initial reconstruction data, determines the initial reconstruction image according to the first initial reconstruction data, performs phase correction on the initial reconstruction image to obtain the first reconstruction sub-data, and uses the first reconstruction sub-data as the first reconstruction data.

[0128] In this embodiment, the specific process of the server determining the first reconstruction sub-data is the same as the process of determining the first reconstruction sub-data involved in determining the first reconstruction data using the combined reconstruction algorithm in the above embodiment. For details, please refer to the description of the above corresponding content.

[0129] By performing magnetic resonance reconstruction in this embodiment, a partial Fourier sampling method is adopted, and a partial Fourier reconstruction algorithm is used to reduce noise sensitivity, so that a target reconstructed image with low noise sensitivity can be obtained while accelerating the imaging speed.

[0130] In the above step S102, when the reconstruction algorithm selected by the user is a non-combined reconstruction algorithm, the server may adopt a generated network reconstruction algorithm to determine the first reconstruction data.

[0131] Specifically, the server inputs the first sampling data into the second reconstruction network to obtain the second initial reconstruction data. Then, the server inputs the second initial reconstruction data into the generation network to obtain the second reconstruction sub-data, and uses the second reconstruction sub-data as the first reconstruction data.

[0132] In this embodiment, the specific process of the server determining the second reconstruction sub-data is the same as the process of determining the second reconstruction sub-data involved in determining the first reconstruction data using the combined reconstruction algorithm in the above embodiment. For details, please refer to the description of the above corresponding content.

[0133] The magnetic resonance reconstruction method of this embodiment uses a partial Fourier sampling method and a generative network reconstruction algorithm. The number of samples in the partial Fourier sampling method is small, and the conjugate symmetry of the K space is used to fill the under-sampled part of the K space data, which can improve the imaging speed. The generative network can learn the subtle structural features and texture information of the magnetic resonance image and obtain a target reconstructed image with rich details. Therefore, this embodiment can obtain a target reconstructed image with rich details while speeding up the imaging speed.

[0134] Figure 5 is a network configuration structure diagram of a magnetic resonance reconstruction algorithm provided in an embodiment of this specification, Figure 5 In the corresponding magnetic resonance imaging algorithm, when the user selects partial Fourier sampling, a combined reconstruction algorithm combining the partial Fourier reconstruction algorithm and the generative network reconstruction algorithm, as well as two optional reconstruction algorithms of the generative network reconstruction algorithm are provided. When the user selects non-partial Fourier sampling, the regression network reconstruction algorithm is applied to determine the target reconstructed image.

[0135] like Figure 5 As shown, when the user selects partial Fourier sampling and selects to use a combined reconstruction algorithm combining partial Fourier sampling and a generative network reconstruction algorithm for image reconstruction, the server adopts a partial Fourier sampling method to determine the first sampling data. Then the first sampling data is input into the first reconstruction network to obtain the first initial reconstruction data, and the input of the partial Fourier reconstruction algorithm is determined based on the first initial reconstruction data. Through the partial Fourier reconstruction algorithm, the reconstructed full-sampled K-space data, i.e., the first reconstruction sub-data, is obtained. And the first sampling data is input into the second reconstruction network to obtain the second initial reconstruction data, and the input of the generative network is determined based on the second initial reconstruction data. Through the generative network reconstruction algorithm, the reconstructed full-sampled K-space data, i.e., the second reconstruction sub-data, is obtained. Finally, the server fuses the second reconstruction sub-data with the first reconstruction sub-data to obtain the first reconstruction data, and performs an inverse Fourier transform on the first reconstruction data to obtain the target reconstructed image.

[0136] When the user selects partial Fourier sampling and chooses to apply the generative network reconstruction algorithm for image reconstruction, the server adopts partial Fourier sampling to determine the first sampling data. Then the first sampling data is input into the first reconstruction network to obtain the first initial reconstruction data, and the input of the partial Fourier reconstruction algorithm is determined based on the first initial reconstruction data. Through the generative network reconstruction algorithm, the reconstructed full-sampled K-space data is obtained, which is the first reconstruction data.

[0137] When the user selects the non-partial Fourier sampling method, the server adopts the non-partial Fourier sampling method to determine the second sampling data, and inputs the second sampling data into the third reconstruction network to obtain the third initial reconstruction data. The input of the regression network is determined based on the third initial reconstruction data, and the K-space data close to full sampling is obtained by generating a network reconstruction algorithm, which is the second reconstruction data.

[0138] In the above step S102, the server may also perform a data fidelity operation on the first reconstruction data to obtain target reconstruction data.

[0139] Specifically, the server may use the position of each sampled data point in the K-space data matrix corresponding to the first sampled data as a marked position, and use the position of each unsampled data point in the K-space corresponding to the first sampled data as a non-marked position.

[0140] The server determines the target reconstructed data according to the element value of each marked position in the first sampled data and the element value of each unmarked position in the first reconstructed data.

[0141] Specifically, the server may determine the data dimension of the mask matrix according to the data dimension of the K-space data matrix, where the mask matrix is ​​used to mark the positions of the data points collected in the K-space corresponding to the first sampling data.

[0142] In the mask matrix, the element values ​​of the marked positions are set to the first preset value, and the element values ​​of the non-marked positions are set to the second value. Usually, the mask matrix is ​​a binary matrix, the first preset value can be 1, and the second preset value can be 0.

[0143] The data fidelity process can be expressed by the following formula:

[0144] k dc =k0·mask+k1·(1-mask)

[0145] Among them, k dc represents the target reconstructed data obtained after data fidelity, k0 represents the first sampling data, mask represents the mask matrix, and k1 represents the first reconstructed data.

[0146] Although the first reconstructed data is full-sampled K-space data, the first reconstructed data is based on the first sampling data and is obtained through the reconstruction algorithm. During the operation of the reconstruction algorithm, the element value of the marked position in the first sampling data may change. However, because the element value of the marked position in the first sampling data is obtained by real sampling in the K-space, the element value of the first sampling data is more accurate for the marked position.

[0147] Therefore, through the data fidelity operation, the consistency of the element value of the marked position in the first sampling data and the first reconstruction data is maintained. While the reconstructed K-space data is obtained through the reconstruction algorithm, the real sampling data is retained. The target reconstruction data used to convert the target reconstructed image represents a more accurate magnetic resonance signal strength at a certain level of the scanning part, and can determine a clearer and more accurate target reconstructed image.

[0148] Similarly, in one or more embodiments of the present specification, in order to obtain a more accurate target reconstructed image, when the server uses the partial Fourier reconstruction algorithm to determine the target reconstructed image, the server may also perform a data fidelity operation on the first initial reconstruction data to obtain first input data. The first input data is used as an input of the partial Fourier reconstruction algorithm.

[0149] The server may use, for each sampled data point corresponding to the first sampled data in the K space, the position of the data point in the K space data matrix as a marked position, and for each unsampled data point corresponding to the first sampled data in the K space, the position of the data point in the K space as a non-marked position.

[0150] The server determines the first input data according to the element value of each marked position in the first sampled data and the element value of each unmarked position in the first initial reconstructed data.

[0151] The specific implementation process of the data fidelity operation performed by the server on the first initial reconstruction data is similar to the implementation process of the data fidelity operation performed on the first reconstruction data. The specific implementation process of the data fidelity operation in this embodiment can refer to the above corresponding description.

[0152] After the data fidelity operation is performed on the first initial reconstruction data, the consistency of the element value of the marked position in the first sampling data and the first input data can be enhanced, so that a clearer and more accurate target reconstructed image can be determined based on the first input data for which the data fidelity is determined.

[0153] Similarly, in one or more embodiments of the present specification, when the server uses a partial Fourier reconstruction algorithm to determine the target reconstructed image, the server may perform a data fidelity operation on the second initial reconstruction data to obtain second input data. The second input data is used as an input for generating a network.

[0154] The server may use, for each sampled data point corresponding to the first sampled data in the K space, the position of the data point in the K space data matrix as a marked position, and for each unsampled data point corresponding to the first sampled data in the K space, the position of the data point in the K space as a non-marked position.

[0155] The server determines the second input data according to the element value of each marked position in the first sampled data and the element value of each unmarked position in the first initial reconstructed data.

[0156] In this embodiment, the specific implementation process of the server's data fidelity operation on the second initial reconstructed data is similar to the above-mentioned implementation process of the data fidelity operation on the first reconstructed data. The specific implementation process of the data fidelity operation in this embodiment can refer to the above-mentioned corresponding description.

[0157] In this embodiment, after the data fidelity operation is performed on the first initial reconstructed data, the consistency of the element value of the marked position in the first sampling data and the second input data can be enhanced, so that a clearer and more accurate target reconstructed image can be determined based on the second input data for which the data fidelity is determined.

[0158] In the above step S104, the server may perform a data fidelity operation on the third initial reconstruction data to obtain third input data. The third input data is used as an input of the regression network.

[0159] Specifically, the server may use the position of each sampled data point in the K-space data matrix corresponding to the first sampled data as a marked position, and use the position of each unsampled data point in the K-space corresponding to the first sampled data as a non-marked position.

[0160] The server determines the third input data according to the element value of each marked position in the first sampled data and the element value of each unmarked position in the third initial reconstructed data.

[0161] The specific implementation process of the data fidelity operation performed by the server on the third initial reconstruction data is similar to the implementation process of the data fidelity operation performed on the first reconstruction data. The specific implementation process of the data fidelity operation in this embodiment can refer to the above corresponding description.

[0162] After the data fidelity operation is performed on the third initial reconstruction data, the consistency of the element value of the marked position in the first sampling data and the third input data can be enhanced, so that a clearer and more accurate target reconstructed image can be determined based on the third input data determined by the data fidelity.

[0163] In the above step S104, the server may perform a data fidelity operation on the second reconstructed data to obtain target reconstructed data.

[0164] The server may use, for each sampled data point corresponding to the first sampled data in the K space, the position of the data point in the K space data matrix as a marked position, and for each unsampled data point corresponding to the first sampled data in the K space, the position of the data point in the K space as a non-marked position.

[0165] The server determines the target reconstructed data according to the element value of each marked position in the first sampled data and the element value of each unmarked position in the second reconstructed data.

[0166] The specific implementation process of the data fidelity operation performed by the server on the second reconstructed data is similar to the implementation process of the data fidelity operation performed on the first reconstructed data. The specific implementation process of the data fidelity operation in this embodiment can refer to the above corresponding description.

[0167] After performing data fidelity operation on the second reconstructed data, the consistency of the element value of the marked position in the first sampling data and the target reconstructed data can be enhanced. The target reconstructed data determined based on the data fidelity operation can more accurately reflect the strength of the magnetic resonance signal of the scanning part, and the target reconstructed image obtained by inverse Fourier transform of the target reconstructed data is also more accurate.

[0168] In the above step S104, a method for magnetic resonance reconstruction based on non-partial Fourier sampling is provided. Because partial Fourier sampling is only performed within the partial data range of the K space, while non-partial Fourier sampling needs to be sampled within the entire K space range. Therefore, the sampling speed based on non-partial Fourier sampling is slower than partial Fourier sampling, that is, when non-partial Fourier sampling is used for magnetic resonance reconstruction, the imaging speed is slow. However, because the sampling range of the non-partial Fourier sampling method is relatively large and the sampled K space data points are also large, the second sampling data obtained based on the non-partial Fourier sampling can be used to reconstruct a target reconstructed image with rich details and higher image quality.

[0169] In the above step S104, in order to provide users with more flexibility when applying non-partial Fourier sampling, the server may preset multiple imaging time levels to freely adjust the imaging speed of non-partial Fourier sampling according to the required reconstructed image quality and the acceptable reconstruction time.

[0170] Different imaging time levels correspond to different amounts of data collected in K space. The shorter the imaging time level, the less data needs to be collected in K space, but the imaging quality will be reduced accordingly. The longer the imaging time level, the more data needs to be collected in K space, and the higher the imaging quality.

[0171] When the server determines that the sampling method selected by the user is non-partial Fourier sampling, the server displays an imaging time level selection interface to the user, and determines the imaging time level selected by the user for magnetic resonance reconstruction in response to the user operation.

[0172] Then, the server determines the amount of K-space data that needs to be collected corresponding to the imaging time level selected by the user, and performs sampling in the K-space according to the amount of K-space data that needs to be collected to obtain second sampling data.

[0173] The target reconstructed image obtained by the second sampling data in this embodiment matches the user's requirements in terms of image quality and imaging time. Combined with this embodiment, the flexibility of magnetic resonance reconstruction is further increased, and more imaging requirements can be met.

[0174] The above embodiments discuss a method of adding data fidelity operations in multiple steps of magnetic resonance imaging to enhance the accuracy of the target reconstructed image. The data fidelity operations involved in the above embodiments are: in the partial Fourier sampling mode, a data fidelity operation is performed on the first initial reconstruction data to determine the input of the partial Fourier reconstruction algorithm; in the partial Fourier sampling mode, a data fidelity operation is performed on the second initial reconstruction data to determine the input of the generating network; in the partial Fourier sampling mode, a data fidelity operation is performed on the first reconstruction data to determine the target reconstruction data; in the non-partial Fourier sampling mode, a data fidelity operation is performed on the third initial reconstruction data to determine the input of the regression network; in the non-partial Fourier sampling mode, a data fidelity operation is performed on the second reconstruction data to determine the target reconstruction data.

[0175] In the magnetic resonance imaging method of the present specification, one or more of the above data fidelity operations may be selected according to imaging requirements, or none of the above data fidelity operations may be selected.

[0176] In the above step S102, the generative network used for image reconstruction of the first sampling data of partial Fourier sampling is generatively trained together with the first reconstruction network.

[0177] In the present specification, the first reconstruction network for determining the first initial reconstruction data, the third reconstruction network for determining the second initial reconstruction data, and the third reconstruction network for determining the third initial reconstruction data have the same function, and are all reconstructed based on the under-sampled K-space data to obtain the initially reconstructed fully-sampled K-space data.

[0178] This specification does not limit the specific network structures of the first reconstruction network, the second reconstruction network, and the third reconstruction network. The first reconstruction network, the second reconstruction network, and the third reconstruction network may adopt the same network structure or different network structures. For example, in one embodiment, the first reconstruction network, the second reconstruction network, and the third reconstruction network may all adopt a U-net (Convolutional Networks for Biomedical Image Segmentation, U-Net) structure, and the first reconstruction network, the second reconstruction network, and the third reconstruction network are networks with the same network structure but different network parameters.

[0179] In the above step S102, the generative network used to reconstruct the first sampled data is obtained by generative training. During training, the generative network and the first reconstruction network are used as generators and matched with a discriminator. The generator and the discriminator form a generative adversarial network and perform generative training together.

[0180] This specification does not limit the specific network structure of the generative adversarial network used for generative training. The network structure can be built according to any model structure used to generate adversarial models, such as classic generative adversarial networks (GAN), cycle generative adversarial networks (CycleGAN), super-resolution generative adversarial networks (SRGAN), etc.

[0181] In the above step S104, the regression network used to reconstruct the second sampled data is trained end-to-end together with the third generation network.

[0182] The specific server determines each sample data and determines the sample reconstructed image corresponding to each sample data. The sample data and the sample reconstructed image can be obtained from any existing magnetic resonance signal sampling data set, or from the magnetic resonance detection record of the hospital.

[0183] The sample data in this embodiment is under-sampled K-space data obtained by non-partial Fourier sampling. The sample reconstructed image corresponding to the sample data is the annotation of the sample data during training, and is used to supervise the third generation network and the regression network to learn the mapping between under-sampled K-space data and fully sampled K-space data.

[0184] The above is the magnetic resonance imaging method provided in this specification. Based on the same idea, this specification also provides a corresponding magnetic resonance imaging device, such as Figure 6 shown.

[0185] Figure 6 A schematic diagram of a magnetic resonance imaging device provided in this specification specifically includes:

[0186] The sampling mode determination module 200 is used to determine the sampling mode selected by the user;

[0187] The partial Fourier sampling module 202 is used to determine a sampling range when the sampling mode is partial Fourier sampling, perform partial Fourier sampling on the magnetic resonance signal generated by the scanning part within the sampling range to obtain first sampling data; apply a partial Fourier reconstruction algorithm and / or a generated network reconstruction algorithm to determine first reconstruction data; and determine a target reconstructed image according to the first reconstruction data;

[0188] The non-partial Fourier sampling module 204 is used to determine the sampling range when the sampling method selected by the user is non-partial Fourier sampling, perform non-partial Fourier sampling on the magnetic resonance signal generated by the scanning part, and determine the second sampling data; apply the regression network reconstruction algorithm to determine the second reconstruction data, and determine the target reconstructed image based on the second reconstruction data.

[0189] Optionally, the partial Fourier sampling module 202 is specifically used to display an algorithm selection interface when the sampling method is partial Fourier sampling, determine the reconstruction algorithm selected by the user in the algorithm selection interface, and judge whether the reconstruction algorithm is a combined reconstruction algorithm. If so, apply the partial Fourier reconstruction algorithm and the generated network reconstruction algorithm to determine the first reconstruction data; if not, apply the partial Fourier reconstruction algorithm or the generated network reconstruction algorithm to determine the first reconstruction data.

[0190] Optionally, the partial Fourier sampling module 202 is specifically used to apply a partial Fourier reconstruction algorithm to determine first reconstructed sub-data corresponding to the first sampling data; apply a generated network reconstruction algorithm to determine second reconstructed sub-data corresponding to the first sampling data, and determine first reconstructed data based on the first reconstructed sub-data and the second reconstructed sub-data.

[0191] Optionally, the partial Fourier sampling module 202 is specifically used to input the first sampling data into a first reconstruction network to obtain first initial reconstruction data, determine an initial reconstructed image based on the first initial reconstruction data, perform phase correction on the initial reconstructed image, and obtain first reconstructed sub-data.

[0192] Optionally, the partial Fourier sampling module 202 is specifically configured to input the first sampling data into a second reconstruction network to obtain second initial reconstruction data, and input the second initial reconstruction data into a generation network to obtain second reconstruction sub-data.

[0193] Optionally, the partial Fourier sampling module 202 is specifically used to perform a data fidelity operation on the first reconstructed data to obtain target reconstructed data, and perform an inverse Fourier transform on the target reconstructed data to obtain a target reconstructed image.

[0194] Optionally, the non-partial Fourier sampling module 204 is specifically used to input the second sampling data into a third reconstruction network to obtain third initial reconstruction data, input the third initial reconstruction data into a regression network to obtain target reconstruction data, and perform an inverse Fourier transform on the target reconstruction data to obtain a target reconstructed image.

[0195] This specification also provides a computer-readable storage medium, which stores a computer program, which can be used to execute the above Figure 1 A magnetic resonance imaging method is provided.

[0196] This manual also provides Figure 7 The schematic structure diagram of the electronic device shown in FIG. Figure 7 As mentioned above, at the hardware level, the electronic device includes a processor, an internal bus, a network interface, a memory, and a non-volatile memory, and may also include other hardware required for the business. The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs it to achieve the above Figure 1 Of course, in addition to the software implementation, this specification does not exclude other implementations, such as logic devices or a combination of software and hardware, etc., that is, the execution subject of the following processing flow is not limited to each logic unit, but can also be hardware or logic devices.

[0197] For the improvement of a technology, it can be clearly distinguished whether it is a hardware improvement (for example, improvement of the circuit structure of diodes, transistors, switches, etc.) or a software improvement (improvement of the method flow). However, with the development of technology, many improvements of the method flow today can be regarded as direct improvements of the hardware circuit structure. Designers almost always obtain the corresponding hardware circuit structure by programming the improved method flow into the hardware circuit. Therefore, it cannot be said that the improvement of a method flow cannot be implemented with a hardware entity module. For example, a programmable logic device (PLD) (such as a field programmable gate array (FPGA)) is such an integrated circuit whose logical function is determined by the user's programming of the device. Designers can "integrate" a digital system on a PLD by programming themselves, without having to ask chip manufacturers to design and make dedicated integrated circuit chips. Moreover, nowadays, instead of manually making integrated circuit chips, this kind of programming is mostly implemented by "logic compiler" software, which is similar to the software compiler used when developing and writing programs, and the original code before compilation must also be written in a specific programming language, which is called hardware description language (HDL). There is not only one HDL, but many kinds, such as ABEL (Advanced Boolean Expression Language), AHDL (Altera Hardware Description Language), Confluence, CUPL (Cornell University Programming Language), HDCal, JHDL (Java Hardware Description Language), Lava, Lola, MyHDL, PALASM, RHDL (Ruby Hardware Description Language), etc. The most commonly used ones are VHDL (Very-High-Speed ​​Integrated Circuit Hardware Description Language) and Verilog. Those skilled in the art should also know that it is only necessary to program the method flow slightly in the above-mentioned hardware description languages ​​and program it into the integrated circuit, and then it is easy to obtain the hardware circuit that implements the logic method flow.

[0198] The controller can be implemented in any appropriate manner, for example, the controller can take the form of a microprocessor or processor and a computer-readable medium storing a computer-readable program code (such as software or firmware) that can be executed by the (micro)processor, a logic gate, a switch, an application-specific integrated circuit (ASIC), a programmable logic controller, and an embedded microcontroller. Examples of controllers include, but are not limited to, the following microcontrollers: ARC 625D, Atmel AT91SAM, Microchip PIC18F26K20, and Silicone Labs C8051F320. The memory controller can also be implemented as part of the control logic of the memory. Those skilled in the art also know that in addition to implementing the controller in a purely computer-readable program code manner, the controller can be implemented in the form of a logic gate, a switch, an application-specific integrated circuit, a programmable logic controller, and an embedded microcontroller by logically programming the method steps. Therefore, this controller can be considered as a hardware component, and the devices included therein for implementing various functions can also be regarded as structures within the hardware component. Or even, the devices for implementing various functions can be regarded as both software modules for implementing the method and structures within the hardware component.

[0199] The systems, devices, modules or units described in the above embodiments may be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, the computer may be, for example, a personal computer, a laptop computer, a cellular phone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.

[0200] For the convenience of description, the above device is described in various units according to their functions. Of course, when implementing this specification, the functions of each unit can be implemented in the same or multiple software and / or hardware.

[0201] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0202] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0203] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0204] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0205] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0206] The memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0207] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0208] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0209] Those skilled in the art will appreciate that the embodiments of this specification may be provided as methods, systems or computer program products. Therefore, this specification may take the form of a complete hardware embodiment, a complete software embodiment or an embodiment combining software and hardware. Moreover, this specification may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0210] This specification may be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. This specification may also be practiced in distributed computing environments where tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules may be located in local and remote computer storage media, including storage devices.

[0211] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0212] The above description is only an embodiment of this specification and is not intended to limit this specification. For those skilled in the art, this specification may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of this specification should be included in the scope of the claims of this application.

Claims

1. A magnetic resonance imaging method, characterized in that: include: Determine the sampling method selected by the user; When the sampling mode is partial Fourier sampling, partial Fourier sampling is performed on the magnetic resonance signal generated by the scanning part to obtain first sampling data; a partial Fourier reconstruction algorithm and / or a generated network reconstruction algorithm are applied to determine first reconstruction data; and a target reconstructed image is determined based on the first reconstruction data; When the sampling method selected by the user is non-partial Fourier sampling, non-partial Fourier sampling is performed on the magnetic resonance signal generated by the scanning part to determine the second sampling data; a regression network reconstruction algorithm is applied to determine the second reconstruction data, and a target reconstructed image is determined based on the second reconstruction data.

2. The method according to claim 1, characterized in that Applying a partial Fourier reconstruction algorithm and / or a generation network reconstruction algorithm to determine first reconstruction data specifically includes: When the sampling method is partial Fourier sampling, an algorithm selection interface is displayed; Determining the reconstruction algorithm selected by the user in the algorithm selection interface, and judging whether the reconstruction algorithm is a combined reconstruction algorithm; If so, applying a partial Fourier reconstruction algorithm and a generative network reconstruction algorithm to determine first reconstruction data; If not, a partial Fourier reconstruction algorithm or a generation network reconstruction algorithm is applied to determine the first reconstruction data.

3. The method according to claim 2, characterized in that Applying a portion of the Fourier reconstruction algorithm and the generation network reconstruction algorithm to determine the first reconstruction data specifically includes: Applying a partial Fourier reconstruction algorithm to determine first reconstructed sub-data corresponding to the first sampled data; applying a generative network reconstruction algorithm to determine second reconstructed sub-data corresponding to the first sampled data; First reconstruction data are determined according to the first reconstruction sub-data and the second reconstruction sub-data.

4. The method according to claim 3, characterized in that Applying a partial Fourier reconstruction algorithm to determine first reconstructed sub-data corresponding to the first sampled data specifically includes: Inputting the first sampling data into a first reconstruction network to obtain first initial reconstruction data; determining an initial reconstructed image according to the first initial reconstruction data; Phase correction is performed on the initial reconstructed image to obtain first reconstructed sub-data.

5. The method according to claim 3, characterized in that Applying a generated network reconstruction algorithm to determine second reconstructed sub-data corresponding to the first sampling data specifically includes: Inputting the first sampling data into a second reconstruction network to obtain second initial reconstruction data; The second initial reconstruction data is input into the generation network to obtain second reconstruction sub-data.

6. The method according to claim 1, characterized in that Determining a target reconstructed image according to the first reconstruction data specifically includes: Performing a data fidelity operation on the first reconstruction data to obtain target reconstruction data; The target reconstruction data is subjected to inverse Fourier transform to obtain a target reconstructed image.

7. The method according to claim 1, characterized in that Applying a regression network reconstruction algorithm to determine second reconstruction data, and determining a target reconstructed image according to the second reconstruction data, specifically includes: Inputting the second sampling data into a third reconstruction network to obtain third initial reconstruction data; Inputting the third initial reconstruction data into a regression network to obtain target reconstruction data; The target reconstruction data is subjected to inverse Fourier transform to obtain a target reconstructed image.

8. A magnetic resonance imaging device, characterized in that: include: A sampling method determination module determines the sampling method selected by the user; A partial Fourier sampling module, when the sampling mode is partial Fourier sampling, determines a sampling range, performs partial Fourier sampling on the magnetic resonance signal generated by the scanning part within the sampling range, and obtains first sampling data; applies a partial Fourier reconstruction algorithm and / or generates a network reconstruction algorithm to determine first reconstruction data; and determines a target reconstructed image according to the first reconstruction data; The non-partial Fourier sampling module determines the sampling range when the sampling method selected by the user is non-partial Fourier sampling, performs non-partial Fourier sampling on the magnetic resonance signal generated by the scanning part, and determines the second sampling data; applies the regression network reconstruction algorithm to determine the second reconstruction data, and determines the target reconstructed image based on the second reconstruction data.

9. A computer-readable storage medium, characterized in that: The storage medium stores a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the method according to any one of claims 1 to 7 is implemented.

Images