Diffusion magnetic resonance imaging microstructure parameter estimation method, system, equipment and medium

Abstract

The invention discloses a diffusion magnetic resonance imaging microstructure parameter estimation method, system and device and a medium, and relates to the technical field of image processing, and the method comprises the steps: constructing simulation imaging regions with preset biological tissue characteristics, and generating microstructure parameter diagrams of different microstructure parameters of each simulation imaging region; synthesizing a plurality of simulation diffusion magnetic resonance images according to the microstructure parameter diagram corresponding to each simulation imaging area, different diffusion sensitive factors and diffusion directions; each simulated diffusion magnetic resonance image and the corresponding microstructure parameter diagram form a training set; training a deep learning network by adopting the training set and taking the simulated diffusion magnetic resonance image as input and the microstructure parameter diagram as output to obtain a microstructure parameter estimation model; and inputting a clinically acquired actual diffusion magnetic resonance image into the microstructure parameter estimation model, and outputting a microstructure parameter diagram. According to the method, the micro-structure parameter estimation efficiency and robustness can be improved.

Description

Technical Field

[0001] The present application relates to the field of image processing technologies, and particularly to a method, a system, a device and a medium for estimating microstructure parameters of diffusion magnetic resonance imaging. Background Art

[0002] As one of the best non-invasive ways to detect the microstructure imaging of living tissues, diffusion magnetic resonance directly delves into the scale of the microstructure of biological tissues to the cellular and even molecular level. Diffusion Weighted Imaging (DWI) reflects the microscopic structural characteristics of tissues based on the diffusion motion of water molecules. However, its assumption that the diffusion motion of water molecules is free and Gaussian-distributed deviates from the actual situation to a certain extent. Especially in pathological tissues, the diffusion of water molecules is significantly restricted and hindered by the surrounding tissues. Diffusion Kurtosis Imaging (DKI) is based on a non-Gaussian diffusion model. By introducing a kurtosis tensor to describe the non-Gaussian characteristics of water molecule diffusion and estimating the degree of deviation from the Gaussian distribution through diffusion kurtosis. It reflects the complexity of the tissue microstructure, can more accurately describe the tissue microstructure, provides richer tissue microstructure information than traditional DWI, and provides an important basis for formulating personalized treatment plans and treatment monitoring.

[0003] In the DKI model, common parameters include microstructural indices such as mean diffusivity (MD) and mean kurtosis (MK). MD reflects the overall diffusion ability of water molecules and is similar to the apparent diffusion coefficient (ADC) in related DWI, but it takes into account the influence of non-Gaussian diffusion; MK is specifically used to measure the non-Gaussian degree of diffusion and is more sensitive to changes in tissue microstructure. The estimation of DKI microstructural parameters is the core content of this technology, and its accuracy and reliability directly affect the application effect of DKI in various fields. The method based on model fitting is a common method for estimating DKI microstructural parameters. This method fits the acquired diffusion-weighted signal with the DKI model to obtain the corresponding parameter values. Specifically, magnetic resonance signals need to be acquired at multiple different diffusion-sensitizing gradient directions and intensities (b-values), and then optimization algorithms such as the nonlinear least squares method and the maximum likelihood estimation method are used to fit the signals. The advantage of this method is that it has a sound theoretical basis and can accurately estimate parameter values, but the calculation process is relatively complex, has high requirements for data quality, and is easily disturbed by noise. Machine learning and deep learning methods can also be used to estimate the microstructural parameters of diffusion imaging. They establish a mapping relationship between the input data (such as diffusion signals) and the output parameters by learning a large number of sample data with known parameters. However, machine learning methods require a large amount of high-quality sample data for training, and the generalization ability of the model needs to be further verified. Otherwise, it is prone to overfitting, resulting in poor performance on new data. Summary of the Invention

[0004] The purpose of this application is to provide a method, system, device, and medium for estimating microstructural parameters of diffusion magnetic resonance imaging, which can improve the efficiency and robustness of microstructural parameter estimation.

[0005] To achieve the above purpose, this application provides the following solutions:

[0006] In the first aspect, this application provides a method for estimating microstructural parameters of diffusion magnetic resonance imaging, including:

[0007] Construct a simulated imaging region with preset biological tissue characteristics, and generate microstructural parameter maps with different microstructural parameters for each simulated imaging region;

[0008] According to the microstructural parameter map corresponding to each simulated imaging region, as well as different diffusion-sensitizing factors and diffusion directions, synthesize multiple simulated diffusion magnetic resonance images; each simulated diffusion magnetic resonance image and the corresponding microstructural parameter map form a training set;

[0009] Use the training set to train a deep learning network with the simulated diffusion magnetic resonance image as the input and the microstructural parameter map as the output to obtain a microstructural parameter estimation model;

[0010] Input the actually acquired diffusion magnetic resonance images in clinical into the microstructural parameter estimation model, and output the microstructural parameter map.

[0011] Optionally, construct a simulated imaging region with preset biological tissue characteristics, and generate microstructural parameter maps with different microstructural parameters for each of the simulated imaging regions, specifically including:

[0012] Initialize the imaging region;

[0013] In the imaging region, randomly generate a plurality of geometric figures according to the preset biological tissue characteristics; each geometric figure serves as a simulated imaging region, and the preset biological tissue characteristics include a preset size range, a preset shape range, and a preset position;

[0014] Use the diffusion magnetic resonance imaging model to generate microstructural parameter maps with different microstructural parameters for each simulated imaging region.

[0015] Optionally, the diffusion magnetic resonance imaging model is a diffusion kurtosis imaging model;

[0016] The diffusion kurtosis imaging model is expressed as:

[0017] where S(b) is the magnetic resonance signal intensity, b is the diffusion sensitivity factor, S0 is the initial magnetic resonance signal intensity, D is the diffusion coefficient or apparent diffusion coefficient, and K is the diffusion kurtosis or apparent diffusion kurtosis.

[0018] Optionally, the microstructural parameter map includes an MD parameter map and an MK parameter map.

[0019] Optionally, using the diffusion magnetic resonance imaging model to generate microstructural parameter maps with different microstructural parameters for each simulated imaging region, specifically including:

[0020] Set the D parameter in the diffusion kurtosis imaging model corresponding to each simulated imaging region to obtain an MD parameter map containing average diffusion coefficient information;

[0021] Set the K parameter in the diffusion kurtosis imaging model corresponding to each simulated imaging region to obtain an MK parameter map containing average kurtosis information.

[0022] Optionally, input the actually acquired diffusion magnetic resonance images in clinical into the microstructural parameter estimation model, and output the microstructural parameter map, specifically including:

[0023] Compile the written diffusion magnetic resonance pulse sequence to generate an executable file on the magnetic resonance imaging instrument;

[0024] Acquire actual clinical data by executing the executable file, and generate k-space data according to the actual clinical data; the k-space data is frequency-domain data;

[0025] Perform a Fourier transform on the k-space data to obtain the actual diffusion magnetic resonance image;

[0026] Input the actual diffusion magnetic resonance image into the microstructural parameter estimation model to output a microstructural parameter map.

[0027] Optionally, the diffusion magnetic resonance pulse sequence includes information on experimental parameters, radio frequency pulses, and gradient fields;

[0028] The experimental parameters include magnetic resonance hardware parameters, imaging parameters, and diffusion-related parameters; the radio frequency pulses include excitation pulses and refocusing pulses; the gradient fields include slice selection gradients, frequency encoding gradients, phase encoding gradients, and diffusion gradients.

[0029] In a second aspect, the present application provides a diffusion magnetic resonance imaging microstructural parameter estimation system, which includes:

[0030] A microstructural parameter map simulation generation module for constructing a simulated imaging region with preset biological tissue characteristics and generating microstructural parameter maps with different microstructural parameters for each of the simulated imaging regions;

[0031] A training set construction module for synthesizing a plurality of simulated diffusion magnetic resonance images according to the microstructural parameter maps corresponding to each of the simulated imaging regions, different diffusion sensitivity factors, and diffusion directions; each simulated diffusion magnetic resonance image and the corresponding microstructural parameter map form a training set;

[0032] A deep learning network training module for using the training set to train a deep learning network with the simulated diffusion magnetic resonance image as the input and the microstructural parameter map as the output to obtain a microstructural parameter estimation model;

[0033] A microstructural parameter map output module for inputting the actual diffusion magnetic resonance image collected clinically into the microstructural parameter estimation model to output a microstructural parameter map.

[0034] In a third aspect, the present application provides a computer device, including: a memory, a processor, and a computer program stored on the memory and executable on the processor, where the processor executes the computer program to implement the steps of the diffusion magnetic resonance imaging microstructural parameter estimation method described in any one of the above.

[0035] In a fourth aspect, the present application provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the diffusion magnetic resonance imaging microstructural parameter estimation method described in any one of the above are implemented.

[0036] According to the specific embodiments provided in this application, the following technical effects are disclosed in this application:

[0037] This application provides a method, system, device and medium for estimating microstructural parameters in diffusion magnetic resonance imaging. A simulated imaging region with preset biological tissue characteristics is constructed, and a microstructural parameter map with different microstructural parameters for each simulated imaging region is generated. According to the microstructural parameter map corresponding to each simulated imaging region, as well as different diffusion sensitivity factors and diffusion directions, a plurality of simulated diffusion magnetic resonance images are synthesized; each simulated diffusion magnetic resonance image and the corresponding microstructural parameter map form a training set, and a microstructural parameter estimation model is determined according to the training set, realizing that directly obtaining a microstructural parameter map by inputting an actual diffusion magnetic resonance image into the microstructural parameter estimation model, improving the construction efficiency of the microstructural parameter map, and moreover, the model prediction based on the actual diffusion magnetic resonance image can make full use of the correlation between pixels, improving the robustness of microstructural parameter estimation. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the following will briefly introduce the drawings required in the embodiments. Obviously, the drawings in the following description are only some embodiments of this application. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0039] Figure 1 It is a schematic flow chart of a method for estimating microstructural parameters in diffusion magnetic resonance imaging provided in an embodiment of this application;

[0040] Figure 2 It is a schematic comparison diagram of the fitting method and the method of this application for estimating microstructural parameters in diffusion kurtosis imaging;

[0041] Figure 3 It is a schematic diagram of the functional modules of a system for estimating microstructural parameters in diffusion magnetic resonance imaging provided in an embodiment of this application;

[0042] Figure 4 It is a schematic diagram of the structure of a computer device provided in an embodiment of this application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0043] The following will clearly and completely describe the technical solutions in the embodiments of this application with reference to the drawings in the embodiments of this application. Obviously, the described embodiments are only some embodiments of this application, rather than all embodiments. Based on the embodiments in this application, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of this application.

[0044] To make the above objects, features, and advantages of the present application more obvious and understandable, the present application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0045] The present application provides a method for estimating diffusion magnetic resonance imaging microstructural parameters. As Figure 1 shown, the method for estimating diffusion magnetic resonance imaging microstructural parameters includes:

[0046] Step 101: Construct a simulated imaging region with preset biological tissue characteristics, and generate a microstructural parameter map with different microstructural parameters for each simulated imaging region.

[0047] Step 102: According to the microstructural parameter map corresponding to each simulated imaging region, as well as different diffusion sensitivity factors and diffusion directions, synthesize a plurality of simulated diffusion magnetic resonance images; each simulated diffusion magnetic resonance image and the corresponding microstructural parameter map form a training set.

[0048] Step 103: Use the training set, with the simulated diffusion magnetic resonance image as the input and the microstructural parameter map as the output, to train a deep learning network to obtain a microstructural parameter estimation model.

[0049] Step 104: Input the actual diffusion magnetic resonance image collected clinically into the microstructural parameter estimation model, and output a microstructural parameter map.

[0050] The present application trains a deep learning network by synthesizing data with biological tissue characteristics, and uses the trained deep learning network, that is, the microstructural parameter estimation model, to achieve reliable estimation of magnetic resonance diffusion microstructural parameters. The present application overcomes the limitation of related methods that focus on estimating the parameter characteristics of individual pixels, fully utilizes the correlation between pixels, can achieve an estimation effect close to that of traditional least squares, improves the reconstruction quality of the microstructural parameter map, and enhances the efficiency and robustness of microstructural parameter estimation. The non-invasive microstructural parameter estimation method involved in the present application has certain clinical application prospects in tumors, cardiovascular and cerebrovascular diseases, etc. By analyzing the changes in the microstructural parameters of diseased tissues, the disease grade, malignancy, and lesion invasion range can be more accurately evaluated, providing more valuable information for early disease diagnosis.

[0051] In an exemplary embodiment, step 101 constructs a simulated imaging region with biological tissue characteristics and obtains a parameter map corresponding to different microstructural parameters. Specifically, it includes:

[0052] Initialize the imaging region.

[0053] In the imaging region, randomly generate a plurality of geometric figures according to the preset biological tissue characteristics; each geometric figure serves as a simulated imaging region, and the preset biological tissue characteristics include a preset size range, a preset shape range, and a preset position.

[0054] A diffusion magnetic resonance imaging model is used to generate microstructure parameter maps with different microstructure parameters for each simulated imaging region. According to the diffusion magnetic resonance imaging model, imaging regions with corresponding parameters are set so that similar magnetic resonance microstructure parameter information exists within the structure of the simulated imaging object.

[0055] The diffusion magnetic resonance imaging model is a diffusion kurtosis imaging model.

[0056] The diffusion kurtosis imaging model is expressed as:

[0057] where S(b) is the magnetic resonance signal intensity, b is the diffusion sensitivity factor, S0 is the initial magnetic resonance signal intensity, D is the diffusion coefficient or apparent diffusion coefficient, D is used to describe the average rate of water molecule diffusion, K is the diffusion kurtosis or apparent diffusion kurtosis, and K is used to describe the degree of deviation of diffusion from the Gaussian distribution, reflecting the complexity and heterogeneity of tissue microstructure.

[0058] Based on the diffusion kurtosis imaging model, by changing the value of the diffusion sensitivity factor, different magnetic resonance signal intensities are obtained, and a large number of simulated diffusion magnetic resonance images are synthesized.

[0059] The microstructure parameter maps include an MD parameter map and an MK parameter map.

[0060] In an exemplary embodiment, using the diffusion magnetic resonance imaging model to generate microstructure parameter maps with different microstructure parameters for each simulated imaging region specifically includes:

[0061] Adding texture information within the local region and neighborhood of the simulated imaging region to make the simulated imaging region closer to actual biological tissue.

[0062] Adding Rice noise within the local region and neighborhood of the simulated imaging region to make the simulated imaging region closer to the real signal environment.

[0063] Setting the D parameter in the diffusion kurtosis imaging model corresponding to each simulated imaging region to obtain an MD parameter map containing average diffusion coefficient information.

[0064] Setting the K parameter in the diffusion kurtosis imaging model corresponding to each simulated imaging region to obtain an MK parameter map containing average kurtosis information.

[0065] In an exemplary embodiment, step 102 specifically includes: synthesizing simulated diffusion magnetic resonance images corresponding to different diffusion sensitivity factors and diffusion directions according to the diffusion model, and a large number of simulated diffusion magnetic resonance images are generated in this process.

[0066] In an exemplary embodiment, the training of the deep learning network in step 103 specifically includes: The training is in the form of supervised learning. The diffusion kurtosis images with different b-values or diffusion directions synthesized by the diffusion imaging model are used as the input of the deep learning model, and the synthesized microstructure parameter map is used as the label. During the training process, the potential non-linear mapping relationship between the diffusion magnetic resonance images and their corresponding microstructure parameter maps is learned.

[0067] In an exemplary embodiment, step 104 specifically includes:

[0068] The prepared diffusion magnetic resonance pulse sequence is compiled to generate an executable file on the magnetic resonance imaging device.

[0069] Actual clinical data is acquired by executing the executable file, and k-space data corresponding to a specific imaging region under different b-values or diffusion directions is generated according to the actual clinical data; the k-space data is frequency domain data.

[0070] The k-space data is subjected to Fourier transform to obtain image domain data, that is, the actual diffusion magnetic resonance image.

[0071] The actual diffusion magnetic resonance image is input into the microstructure parameter estimation model, and a microstructure parameter map is output to complete the estimation of the microstructure parameters.

[0072] In an exemplary embodiment, the diffusion magnetic resonance pulse sequence includes information on experimental parameters, radiofrequency pulses, and gradient fields.

[0073] For the diffusion magnetic resonance pulse sequence, experimental parameters, radiofrequency pulses, and gradient fields are set.

[0074] The experimental parameters include magnetic resonance hardware parameters, imaging parameters, and diffusion-related parameters; the magnetic resonance hardware parameters include the main magnetic field strength; the imaging parameters include the imaging field of view, matrix size, and slice thickness; the diffusion-related parameters include the diffusion sensitivity factor (b-value) and the diffusion gradient direction.

[0075] The radiofrequency pulses include excitation pulses and refocusing pulses; the design of the excitation pulses includes the selection of the type of excitation pulse and the optimization of pulse parameters. According to actual requirements, parameters such as the bandwidth and phase of the excitation pulse may need to be adjusted to ensure a uniform excitation effect and good image quality. In the spin echo sequence, a 180° refocusing pulse is required to compensate for the inhomogeneity of the main magnetic field, thereby generating a spin echo signal.

[0076] The gradient field includes a slice selection gradient, a frequency encoding gradient, a phase encoding gradient, and a diffusion gradient. The slice selection gradient is applied during the excitation pulse and the refocusing pulse to select a specific imaging slice. The amplitude and duration of the slice selection gradient are calculated and set according to the slice thickness and the position of the imaging region; the frequency encoding gradient is applied during the readout signal to perform spatial encoding of the signal in the frequency dimension and determine its amplitude and duration; before signal acquisition, the signal is spatially encoded in the phase dimension by applying phase encoding gradients with different amplitudes. The number of phase encoding steps is determined according to the matrix size, and the corresponding phase encoding gradient amplitude and duration are set for each step; according to the set b value and the diffusion gradient direction, a diffusion-sensitized gradient module with a specific amplitude, duration, and direction is applied on both sides of the 180° refocusing pulse in the spin echo sequence.

[0077] The diffusion magnetic resonance pulse sequence is compiled to generate an executable file on a magnetic resonance imaging device.

[0078] In an exemplary embodiment, actual clinical data is acquired by executing the executable file, and k-space data is generated based on the actual clinical data, which specifically includes: actual clinical data is acquired by executing the executable file, and an excitation pulse, a slice selection gradient, a 180° refocusing pulse, a diffusion gradient, a phase encoding gradient, and a frequency encoding gradient are applied in sequence. According to the echo time, the signal generated by the transverse magnetization vector is detected by a magnetic resonance receiving coil. For each b value, diffusion direction, and each phase encoding step, a signal acquisition is required, and thus signals with different magnetic resonance parameter weighting information are obtained. The acquired signals are filled into the k-space in sequence according to the phase encoding and frequency encoding order. For each b value and diffusion direction, there is a corresponding k-space matrix.

[0079] In an exemplary embodiment, the k-space data is subjected to Fourier transform to obtain the actual diffusion magnetic resonance image, which specifically includes: zero-padding is performed on the edge of the k-space data, and the filled k-space data is subjected to Fourier transform to convert it from the spatial frequency domain to the image space, obtaining diffusion kurtosis magnetic resonance images under different b values and diffusion directions, that is, the actual diffusion magnetic resonance image.

[0080] In an exemplary embodiment, using the clinically acquired diffusion kurtosis magnetic resonance imaging data as a verification, the method of the present application is compared with the results of the least squares estimation method, and the comparison results are as Figure 2 shown Figure 2In part (a), it is the MD parameter map obtained by the least squares estimation method. In part (b), it is the MD parameter map obtained by the method of this application. In part (c), it is the MK parameter map obtained by the least squares estimation method. In part (d), it is the MK parameter map obtained by the method of this application. It can be seen that the method of this application can achieve microstructural parameter estimation results very close to those of the least squares estimation method. However, the method of this application is robust to noise and its execution efficiency is significantly higher than that of the least squares estimation method. The in-vivo diffusion kurtosis imaging includes 30 diffusion directions and 4 diffusion sensitivity factors, and the 4 diffusion sensitivity factors are 700, 1000, 1400, and 2000 respectively.

[0081] Based on the same inventive concept, the embodiment of this application also provides a diffusion magnetic resonance imaging microstructural parameter estimation system for implementing the diffusion magnetic resonance imaging microstructural parameter estimation method involved above. The implementation solutions provided by this system to solve problems are similar to the implementation solutions recorded in the above method. Therefore, the specific limitations in one or more embodiments of the diffusion magnetic resonance imaging microstructural parameter estimation system provided below can refer to the limitations on the diffusion magnetic resonance imaging microstructural parameter estimation method in the above text, and will not be elaborated here.

[0082] In an exemplary embodiment, as Figure 3 shown, this application provides a diffusion magnetic resonance imaging microstructural parameter estimation system, including:

[0083] A microstructural parameter map simulation generation module, configured to construct a simulated imaging region with preset biological tissue characteristics, and generate microstructural parameter maps with different microstructural parameters for each of the simulated imaging regions.

[0084] A training set construction module, configured to synthesize a plurality of simulated diffusion magnetic resonance images according to the microstructural parameter maps corresponding to each of the simulated imaging regions, as well as different diffusion sensitivity factors and diffusion directions; each simulated diffusion magnetic resonance image and the corresponding microstructural parameter map form a training set.

[0085] A deep learning network training module, configured to use the training set, take the simulated diffusion magnetic resonance images as inputs, and the microstructural parameter maps as outputs to train a deep learning network, and obtain a microstructural parameter estimation model.

[0086] A microstructural parameter map output module, configured to input the actually acquired clinical diffusion magnetic resonance images into the microstructural parameter estimation model, and output microstructural parameter maps.

[0087] In an exemplary embodiment, a computer device is provided. This computer device can be a server or a terminal, and its internal structure diagram can be as Figure 4As shown in the figure. The computer device includes a processor, a memory, an input / output interface (Input / Output, abbreviated as I / O), and a communication interface. Among them, the processor, the memory, and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store diffusion magnetic resonance imaging microstructure parameter estimation data. The input / output interface of the computer device is used to exchange information between the processor and external devices. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, it implements a method for estimating diffusion magnetic resonance imaging microstructure parameters.

[0088] Those skilled in the art can understand that Figure 4 the structure shown in the figure is only a block diagram of some structures related to the solution of this application, and does not constitute a limitation on the computer device to which the solution of this application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements. In an exemplary embodiment, a computer device is provided, including a memory and a processor. A computer program is stored in the memory, and when the processor executes the computer program, the steps in the above method embodiments are implemented.

[0089] In an exemplary embodiment, a computer-readable storage medium is provided, storing a computer program, and when the computer program is executed by the processor, the steps in the above method embodiments are implemented.

[0090] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use, and processing of relevant data need to comply with relevant regulations.

[0091] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant 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 processes of the embodiments of the above methods. Among them, any reference to a memory, database, or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memories. Non-volatile memories can include read-only memory (ROM), magnetic tapes, floppy disks, flash memories, optical memories, high-density embedded non-volatile memories, resistive random access memories (ReRAM), magnetoresistive random access memories (MRAM), ferroelectric random access memories (FRAM), phase change memories (PCM), graphene memories, etc. Volatile memories can include random access memory (RAM) or external cache memories, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0092] The databases involved in the embodiments provided in the present application can include at least one of relational databases and non-relational databases. Non-relational databases can include distributed databases based on blockchain, etc., without limitation. The processors involved in the embodiments provided in the present application can be general-purpose processors, central processors, graphics processors, digital signal processors, data processing logics of programmable logics, etc., without limitation.

[0093] The technical features of the above embodiments can be combined arbitrarily. For the sake of concise description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered to be within the scope described in this specification.

[0094] In this article, specific examples are used to elaborate on the principles and implementation manners of the present application. The descriptions of the above embodiments are only used to help understand the methods and core ideas of the present application; at the same time, for those of ordinary skill in the art, according to the ideas of the present application, there will be changes in the specific implementation manners and application scopes. In summary, the content of this specification should not be construed as a limitation to the present application.

Claims

1. A diffusion magnetic resonance imaging microstructure parameter estimation method, characterized in that: The diffusion magnetic resonance imaging microstructure parameter estimation method comprises: Constructing a simulated imaging region with preset biological tissue characteristics, and generating a microstructure parameter map of different microstructure parameters of each simulated imaging region; According to the microstructure parameter map corresponding to each simulated imaging area, as well as different diffusion sensitivity factors and diffusion directions, a plurality of simulated diffusion magnetic resonance images are synthesized; each simulated diffusion magnetic resonance image and the corresponding microstructure parameter map constitute a training set; The training set is used to train a deep learning network with a simulated diffusion magnetic resonance image as input and a microstructure parameter map as output to obtain a microstructure parameter estimation model; The actual diffusion magnetic resonance image collected clinically is input into the microstructure parameter estimation model, and a microstructure parameter map is output.

2. The method for estimating microstructure parameters of diffusion magnetic resonance imaging according to claim 1, characterized in that: Constructing a simulated imaging region with preset biological tissue characteristics and generating a microstructure parameter map of different microstructure parameters of each simulated imaging region specifically includes: Initialize the imaging area; In the imaging area, a plurality of geometric figures are randomly generated according to preset biological tissue characteristics; each geometric figure serves as a simulated imaging area, and the preset biological tissue characteristics include a preset size range, a preset shape range, and a preset position; A diffusion MRI model was used to generate microstructural parameter maps with different microstructural parameters for each simulated imaging region.

3. The method for estimating microstructure parameters of diffusion magnetic resonance imaging according to claim 2, characterized in that: The diffusion magnetic resonance imaging model is a diffusion kurtosis imaging model; The diffusion kurtosis imaging model is expressed as: Wherein, S(b) is the magnetic resonance signal intensity, b is the diffusion sensitivity factor, S0 is the initial magnetic resonance signal intensity, D is the diffusion coefficient or the apparent diffusion coefficient, and K is the diffusion kurtosis or the apparent diffusion kurtosis.

4. The method for estimating microstructure parameters of diffusion magnetic resonance imaging according to claim 3, characterized in that: The microstructure parameter map includes an MD parameter map and an MK parameter map.

5. The method for estimating microstructure parameters of diffusion magnetic resonance imaging according to claim 4, characterized in that: The diffusion MRI model was used to generate microstructural parameter maps of different microstructural parameters for each simulated imaging area, including: Setting the D parameter in the diffusion kurtosis imaging model corresponding to each simulated imaging region to obtain an MD parameter map containing average diffusion coefficient information; The K parameter in the diffusion kurtosis imaging model corresponding to each simulated imaging region is set to obtain an MK parameter map of average kurtosis information.

6. The method for estimating microstructure parameters of diffusion magnetic resonance imaging according to claim 1, characterized in that: Inputting the actual diffusion magnetic resonance image collected clinically into the microstructure parameter estimation model and outputting the microstructure parameter map specifically includes: The programmed diffusion magnetic resonance pulse sequence is compiled to generate an executable file on the magnetic resonance imaging device; The actual clinical data is collected by executing the executable file, and k-space data is generated according to the actual clinical data; the k-space data is frequency domain data; Performing Fourier transformation on the k-space data to obtain the actual diffusion magnetic resonance image; The actual diffusion magnetic resonance image is input into the microstructure parameter estimation model, and a microstructure parameter map is output.

7. The method for estimating microstructure parameters of diffusion magnetic resonance imaging according to claim 6, characterized in that: The diffusion magnetic resonance pulse sequence includes information on experimental parameters, radio frequency pulses and gradient fields; The experimental parameters include magnetic resonance hardware parameters, imaging parameters and diffusion-related parameters; the radio frequency pulses include excitation pulses and refocusing pulses; the gradient fields include slice selection gradients, frequency encoding gradients, phase encoding gradients and diffusion gradients.

8. A diffusion magnetic resonance imaging microstructure parameter estimation system, characterized in that: The diffusion magnetic resonance imaging microstructure parameter estimation system comprises: A microstructure parameter map simulation generation module is used to construct a simulated imaging area with preset biological tissue characteristics and generate a microstructure parameter map of different microstructure parameters in each simulated imaging area; A training set construction module is used to synthesize a plurality of simulated diffusion magnetic resonance images according to the microstructure parameter map corresponding to each simulated imaging area, as well as different diffusion sensitivity factors and diffusion directions; each simulated diffusion magnetic resonance image and the corresponding microstructure parameter map constitute a training set; A deep learning network training module is used to use the training set, take the simulated diffusion magnetic resonance image as input, and the microstructure parameter map as output to train the deep learning network to obtain a microstructure parameter estimation model; The microstructure parameter map output module is used to input the actual diffusion magnetic resonance image collected clinically into the microstructure parameter estimation model and output a microstructure parameter map.

9. A computer device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method for estimating microstructure parameters of diffusion magnetic resonance imaging according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for estimating microstructure parameters of diffusion magnetic resonance imaging according to any one of claims 1 to 7 is implemented.

Images