Free breathing dynamic magnetic resonance reconstruction method and system, medium and electronic equipment

Abstract

The invention provides a free breathing dynamic magnetic resonance reconstruction method and system, a medium and electronic equipment. The method comprises the following steps: constructing a first feature vector based on a two-dimensional space coordinate of a dynamic magnetic resonance image; constructing a second feature vector based on the re-parameterized time latent variable; and inputting the first feature vector and the second feature vector into a first multi-layer perceptron to obtain the reconstructed dynamic magnetic resonance image intensity. According to the free respiration dynamic magnetic resonance reconstruction method and system, the medium and the electronic equipment, free respiration dynamic magnetic resonance reconstruction is achieved based on time latent variable coding and space implicit neural representation.

Description

Technical Field

[0001] The present invention relates to the technical field of magnetic resonance reconstruction, and in particular to a free-breathing dynamic magnetic resonance reconstruction method, system, medium and electronic equipment. Background Art

[0002] Dynamic Magnetic Resonance Imaging (DMRI) is one of the most widely used magnetic resonance technologies. It not only retains the safety, radiation-free advantages and excellent soft tissue contrast of magnetic resonance imaging, but also can capture the dynamic changes of tissues and is used to study organs with intense physiological movements such as the heart. However, the long acquisition time of magnetic resonance makes it difficult to improve the acquisition time and space resolution of DMRI, limiting its application scenarios.

[0003] To improve the spatiotemporal resolution of DMRI, methods based on compressed sensing and deep learning are currently commonly used to reconstruct high-quality dynamic images from downsampled data. These methods usually use the spatiotemporal prior information of the image as regularization constraints to effectively suppress artifacts and improve reconstruction quality. However, these regularization methods are mainly performed in the image space and cannot fully utilize the redundant information contained in periodic motion.

[0004] Manifold learning based on latent variables provides a new idea for DMRI. As a compact low-dimensional representation of high-dimensional data (such as images, etc.), manifold can more effectively describe the characteristic changes contained in the data (such as motion information). The generative model can effectively map the low-dimensional manifold back to high-dimensional data. By optimizing the latent variable as a low-dimensional manifold and combining it with the above-mentioned generative model, complex time series changes can be effectively modeled. Traditional generative models often use convolutional neural networks, but their discrete characteristics limit their ability to deal with continuously changing objects.

[0005] Implicit Neural Representation (INR) is a new unsupervised deep learning framework. Different from traditional deep learning methods, it uses neural networks to map the input spatiotemporal coordinates into high-dimensional continuous functions for implicit expression. The internal continuity of INR enables it to better maintain temporal coherence in DMRI and effectively capture fine structural details. However, since DMRI has different continuity characteristics in time and space dimensions, it is difficult to accurately distinguish spatiotemporal coordinates when directly using INR to process them. Summary of the invention

[0006] In view of the above problems, the purpose of the present invention is to provide a free-breathing dynamic magnetic resonance reconstruction method, system, medium, and electronic device to achieve free-breathing dynamic magnetic resonance reconstruction based on temporal latent variable encoding and spatial implicit neural representation.

[0007] In a first aspect, the present invention provides a free-breathing dynamic magnetic resonance reconstruction method, the method comprising the following steps: constructing a first eigenvector based on the two-dimensional spatial coordinates of the dynamic magnetic resonance image; constructing a second eigenvector based on a reparameterized time latent variable; inputting the first eigenvector and the second eigenvector into a first multilayer perceptron to obtain the reconstructed dynamic magnetic resonance image intensity.

[0008] In an implementation of the first aspect, constructing a first feature vector based on the two-dimensional spatial coordinates of the dynamic magnetic resonance image includes:

[0009] The two-dimensional space coordinates are hash-encoded through a hash table to obtain the first feature vector.

[0010] In an implementation of the first aspect, constructing a second feature vector based on the reparameterized time latent variable includes the following steps:

[0011] Obtain an initialized time latent variable, where the initialized time latent variable is a standard Gaussian random variable with a mean of 0 and a variance of 1;

[0012] Reparameterizing the initialized time latent variable;

[0013] The re-parameterized time latent variable is input into a second multi-layer perceptron to obtain the second feature vector.

[0014] In an implementation of the first aspect, the initialized time latent variable is reparameterized according to c(t)=μ(t)+∈σ(t), wherein c(t) represents the reparameterized time latent variable, ∈ represents a Gaussian random variable with a mean of 0 and a variance of 1, and μ(t) and σ(t) represent the mean and standard deviation of the initialized time latent variable.

[0015] In an implementation of the first aspect, the method further includes: Smooth Regularization Loss and KL loss Constructing a total loss function; optimizing the first multilayer perceptron, the hash table, and the time latent variable based on the minimized total loss function.

[0016] In an implementation of the first aspect, the total loss function where λ S and λ K Represents a hyperparameter.

[0017] In an implementation of the first aspect, the data consistency loss where d θ represents the intensity of reconstructed dynamic magnetic resonance images, F u S c d θ Indicates d θ After the forward model F u S c The predicted k-space data obtained, D c represents the actual k-space data, c represents the cth coil;

[0018] The smooth regularization loss in Represents the gradient of the mean of the latent variable along the time dimension;

[0019] The KL loss Where n represents the dimension of the time latent variable, μ i (t) and σ i (t) represents the mean and standard deviation of the initialized temporal latent variable of the i-th dimension.

[0020] In a second aspect, the present invention provides a free-breathing dynamic magnetic resonance reconstruction system, the system comprising a first building module, a second building module and a reconstruction module;

[0021] The first construction module is used to construct a first feature vector based on the two-dimensional spatial coordinates of the dynamic magnetic resonance image;

[0022] The second construction module is used to construct a second feature vector based on the re-parameterized time latent variable;

[0023] The reconstruction module is used to input the first eigenvector and the second eigenvector into a first multi-layer perceptron to obtain the reconstructed dynamic magnetic resonance image intensity.

[0024] In a third aspect, the present invention provides a storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-mentioned free-breathing dynamic magnetic resonance reconstruction method.

[0025] In a fourth aspect, the present invention provides an electronic device, comprising: a processor and a memory;

[0026] The memory is used to store computer programs;

[0027] The processor is used to execute the computer program stored in the memory, so that the electronic device performs the above-mentioned free breathing dynamic magnetic resonance reconstruction method.

[0028] As described above, the free breathing dynamic magnetic resonance reconstruction method, system, medium, and electronic device of the present invention have the following features:

[0029] Beneficial effects:

[0030] (1) Innovatively, the temporal and spatial dimensions of dynamic MRI images are independently encoded based on spatial implicit neural representation. Through latent variable learning in the temporal dimension, accurate modeling and efficient reconstruction of physiological movements such as heartbeat and breathing are achieved, effectively solving the problems of low spatiotemporal resolution and motion-induced image quality problems in free-breathing dynamic MRI.

[0031] (2) An unsupervised framework based on deep learning that does not require additional training data and can perform well under different data sets, acceleration factors, and acquisition schemes;

[0032] (3) It can generate images of arbitrary temporal resolution using latent variable interpolation and can choose to reconstruct under specific breathing or heartbeat patterns, thereby reducing computational complexity, reducing motion artifacts, and enhancing reconstruction effects under undersampled data, providing a more robust solution and broader application prospects for free-breathing dynamic MRI;

[0033] (4) By combining the latent variable encoding of the time dimension and the implicit neural representation mapping of the spatial dimension, while achieving zero-shot training, it can improve the modeling capabilities of temporal dynamics and spatial details, more accurately capture changes in physiological movements (such as heartbeat and breathing), and ensure temporal continuity and spatial consistency in the reconstruction process. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 Shown is a flow chart of a free-breathing dynamic MRI reconstruction method in one embodiment of the present invention;

[0035] Figure 2 It is a schematic diagram showing a framework of a free-breathing dynamic magnetic resonance reconstruction method according to an embodiment of the present invention;

[0036] Figure 3 A schematic diagram showing latent variables representing heartbeat and breathing changes learned by the free-breathing dynamic MRI reconstruction method of the present invention in one embodiment;

[0037] Figure 4 Shown is a schematic diagram comparing the reconstruction results of the free-breathing dynamic MRI reconstruction method of the present invention and the GRASP reconstruction method in one embodiment;

[0038] Figure 5 Shown is a schematic structural diagram of a free-breathing dynamic magnetic resonance reconstruction system according to an embodiment of the present invention;

[0039] Figure 6 It is a schematic structural diagram of an electronic device of the present invention in one embodiment. DETAILED DESCRIPTION

[0040] The following describes the embodiments of the present invention by specific examples, and those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the following embodiments and features in the embodiments can be combined with each other without conflict.

[0041] It should be noted that the illustrations provided in the following embodiments are only schematic illustrations of the basic concept of the present invention, and thus the drawings only show components related to the present invention rather than being drawn according to the number, shape and size of components in actual implementation. In actual implementation, the type, quantity and proportion of each component may be changed arbitrarily, and the component layout may also be more complicated.

[0042] The free-breathing dynamic magnetic resonance reconstruction method, system, medium, and electronic device of the present invention encode time and space separately, combine the encoded spatial coordinates with the time latent variable, input into a multi-layer perceptron to obtain an image sequence that changes with time, perform data consistency term loss on the actual collected k-space data, and combine the regularization term loss and KL loss of the time dimension to jointly optimize the INR network parameters and latent variables.

[0043] Among them, the (k, t) spatial data D of the cth coil actually collected in DMRI c The relationship between the reconstructed dynamic magnetic resonance image matrix d can be expressed by a linear model as D c =F u S c d. Among them N represents the size of the dynamic magnetic resonance image matrix, M (M<N) represents the number of readout lines acquired per frame in radial sampling, T represents the total number of time frames of the dynamic magnetic resonance image, and C represents the total number of receiving coils. u S c is the forward model, where represents the non-uniform Fourier operator to simulate the undersampling acquisition process of DMRI. is a diagonal matrix representing the sensitivity map of the cth coil. From the undersampled (k, t) spatial data D c The solution to reconstruct the image d is not unique and unstable. It is actually solving an ill-posed inverse problem. Therefore, the optimization process is as follows:

[0044]

[0045] in is a priori regular term, which constrains the target d to achieve the optimal result under ill-posed conditions.

[0046] The technical solutions in the embodiments of the present invention will be described in detail below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0047] like Figure 1 As shown, in one embodiment, the free breathing dynamic MRI reconstruction method of the present invention includes steps S1 to S3.

[0048] Step S1: construct a first feature vector based on the two-dimensional spatial coordinates of a dynamic magnetic resonance image.

[0049] Specifically, in the spatial dimension, the two-dimensional spatial coordinates (x, y) of the dynamic magnetic resonance image are hash-encoded through a hash table to obtain a first eigenvector:

[0050] Step S2: construct a second eigenvector based on the re-parameterized time latent variable.

[0051] Specifically, in the time dimension, first obtain the initialized time latent variable t = {t 1 ,t 2}, the element t of the initialized time latent variable 1 and t 2 is a standard Gaussian random variable with a mean of 0 and a variance of 1. Then, the initialized time latent variable is reparameterized according to c(t)=μ(t)+∈σ(t), where c(t) represents the reparameterized time latent variable, ∈ represents a Gaussian random variable with a mean of 0 and a variance of 1, and μ(t) and σ(t) represent the mean and standard deviation of the initialized time latent variable. Finally, the reparameterized time latent variable c(t) is input into the second multilayer perceptron MLP1, and the second multilayer perceptron MLP1 outputs the second feature vector MLP 1 (c(t)).

[0052] Step S3: input the first eigenvector and the second eigenvector into a first multi-layer perceptron to obtain the reconstructed dynamic magnetic resonance image intensity.

[0053] Specifically, the first feature vector and the second eigenvector MLP 1 (c(t)) is combined as the input of the first multilayer perceptron MLP2, and the output of the first multilayer perceptron MLP2 is the reconstructed dynamic magnetic resonance image intensity. Based on the reconstructed dynamic magnetic resonance image intensity, a reconstructed dynamic magnetic resonance image can be obtained. Wherein, the reconstructed dynamic magnetic resonance image intensity

[0054] It should be noted that when training the first multi-layer perceptron MLP2 and the second multi-layer perceptron MLP1, the data consistency loss Smooth Regularization Loss and KL loss Constructing the total loss function The first multilayer perceptron MLP2, the hash table, the time latent variable and the second multilayer perceptron MLP1 are optimized based on the minimized total loss function. S and λ K represents a hyperparameter, which is used to balance the contribution of the smoothing regularization term and the KL constraint term. Meanwhile, the present invention adopts an algorithm based on gradient descent for iterative optimization.

[0055] In one embodiment, the data consistency loss where d θ represents the intensity of reconstructed dynamic magnetic resonance images, F u S c d θ Indicates d θ After the forward model F u S c The predicted k-space data obtained, D c represents the actual k-space data, and c represents the cth coil.

[0056] In order to constrain the latent variables to capture the smoothness of motion patterns in dynamic images, a smooth regularization loss is calculated on the distribution of the latent variables. in Represents the gradient of the mean of the latent variable along the time dimension. This loss is used to constrain the continuity of the time series.

[0057] In order to constrain the distribution of the reparameterized latent variable to be close to the standard normal distribution, the KL loss is calculated between the distribution of the reparameterized latent variable and its prior distribution. Where n represents the dimension of the time latent variable, μ i (t) and σ i (t) represents the mean and standard deviation of the initialized temporal latent variable of the i-th dimension.

[0058] In the inference process, the arbitrary spatial coordinates (x, y) of the input dynamic magnetic resonance image are combined with Figure 3 The latent variable at a certain moment can be used to generate the image intensity of the position at that moment (i.e., the corresponding respiratory and heartbeat state) through INR. Figure 4As shown, when 13 readout lines are collected per frame (i.e., 20 times acceleration), the reconstruction result of the present invention is significantly improved compared with the compressed sensing method GRASP widely used in DMRI, demonstrating its application potential in free-breathing DMRI scenarios.

[0059] The protection scope of the free-breathing dynamic magnetic resonance reconstruction method described in the embodiment of the present invention is not limited to the execution order of the steps listed in this embodiment. All solutions implemented by adding, reducing or replacing steps in the prior art based on the principles of the present invention are included in the protection scope of the present invention.

[0060] An embodiment of the present invention further provides a free-breathing dynamic MRI reconstruction system, which can implement the free-breathing dynamic MRI reconstruction method described in the present invention. However, the implementation device of the free-breathing dynamic MRI reconstruction system described in the present invention includes but is not limited to the structure of the free-breathing dynamic MRI reconstruction system listed in this embodiment. All structural deformations and replacements of the prior art made according to the principles of the present invention are included in the protection scope of the present invention.

[0061] like Figure 5 As shown, in one embodiment, the free breathing dynamic magnetic resonance reconstruction system of the present invention includes a first construction module 51 , a second construction module 52 and a reconstruction module 53 .

[0062] The first construction module 51 is used to construct a first feature vector based on the two-dimensional spatial coordinates of the dynamic magnetic resonance image.

[0063] The second construction module 52 is used to construct a second feature vector based on the re-parameterized time latent variable.

[0064] The reconstruction module 53 is connected to the first construction module 51 and the second construction module 52, and is used to input the first feature vector and the second feature vector into a first multi-layer perceptron to obtain the reconstructed dynamic magnetic resonance image intensity.

[0065] The structures and principles of the first construction module 51, the second construction module 52 and the reconstruction module 53 correspond one to one with the above-mentioned free breathing dynamic magnetic resonance reconstruction method, so they will not be described in detail here.

[0066] In the several embodiments provided by the present invention, it should be understood that the disclosed system, device or method can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of modules / units is only a logical function division, and there may be other division methods in actual implementation, such as multiple modules or units can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or modules or units, which can be electrical, mechanical or other forms.

[0067] The modules / units described as separate components may or may not be physically separated, and the components displayed as modules / units may or may not be physical modules, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules / units may be selected according to actual needs to achieve the purpose of the embodiments of the present invention. For example, the functional modules / units in the various embodiments of the present invention may be integrated into one processing module, or each module / unit may exist physically separately, or two or more modules / units may be integrated into one module / unit.

[0068] Those of ordinary skill in the art should further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in terms of function in the above description. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0069] The embodiment of the present invention also provides a computer-readable storage medium. A person of ordinary skill in the art can understand that all or part of the steps in the free breathing dynamic magnetic resonance reconstruction method of the above embodiment can be completed by instructing a processor through a program, and the program can be stored in a computer-readable storage medium, and the storage medium is a non-transitory medium, such as a random access memory, a read-only memory, a flash memory, a hard disk, a solid-state hard disk, a magnetic tape, a floppy disk, an optical disc, and any combination thereof. The above storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or a data center that includes one or more available media. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a digital video disc (DVD)), or a semiconductor medium (e.g., a solid-state disk (SSD)), etc.

[0070] An embodiment of the present invention further provides an electronic device, which includes a processor and a memory.

[0071] The memory is used to store computer programs.

[0072] The memory includes: ROM, RAM, disk, USB flash drive, memory card or CD and other media that can store program codes.

[0073] The processor is connected to the memory and is used to execute the computer program stored in the memory so that the electronic device executes the above-mentioned free breathing dynamic magnetic resonance reconstruction method.

[0074] Preferably, the processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components.

[0075] like Figure 6As shown, the electronic device of the present invention is in the form of a general computing device. The components of the electronic device may include but are not limited to: one or more processors or processing units 61, a memory 62, and a bus 63 connecting different system components (including the memory 62 and the processing unit 61).

[0076] Bus 63 represents one or more of several types of bus structures, including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, a processor or a local bus using any of a variety of bus architectures. These architectures include, by way of example, but are not limited to, an Industry Standard Architecture (ISA) bus, a Micro Channel Architecture (MAC) bus, an Enhanced ISA bus, a Video Electronics Standards Association (VESA) local bus, and a Peripheral Component Interconnect (PCI) bus.

[0077] Electronic devices typically include a variety of computer system readable media. These media can be any available media that can be accessed by the electronic device, including volatile and non-volatile media, removable and non-removable media.

[0078] The memory 62 may include computer system readable media in the form of volatile memory, such as random access memory (RAM) 621 and / or cache memory 622. The electronic device may further include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, the storage system 623 may be used to read and write non-removable, non-volatile magnetic media ( Figure 6 not shown, usually called a "hard drive"). Although Figure 6 Not shown, a disk drive for reading and writing to a removable non-volatile disk (e.g., a "floppy disk"), and an optical disk drive for reading and writing to a removable non-volatile optical disk (e.g., a CD-ROM, a DVD-ROM, or other optical media) may be provided. In these cases, each drive may be connected to the bus 63 via one or more data medium interfaces. The memory 62 may include at least one program product having a set (e.g., at least one) of program modules that are configured to perform the functions of the various embodiments of the present invention.

[0079] A program / utility 624 having a set (at least one) of program modules 6241 may be stored, for example, in the memory 62, such program modules 6241 including but not limited to an operating system, one or more application programs, other program modules, and program data, each of which or some combination may include an implementation of a network environment. The program modules 6241 generally perform the functions and / or methods of the embodiments described herein.

[0080] The electronic device may also communicate with one or more external devices (e.g., keyboards, pointing devices, displays, etc.), one or more devices that enable a user to interact with the electronic device, and / or any device that enables the electronic device to communicate with one or more other computing devices (e.g., network cards, modems, etc.). Such communication may be performed via input / output (I / O) interface 64. Furthermore, the electronic device may also communicate with one or more networks (e.g., local area networks (LANs), wide area networks (WANs), and / or public networks, such as the Internet) via network adapter 65. Figure 6 As shown, the network adapter 65 communicates with other modules of the electronic device via the bus 63. It should be understood that, although not shown in the figure, other hardware and / or software modules can be used in conjunction with the electronic device, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0081] The above embodiments are merely illustrative of the principles and effects of the present invention, and are not intended to limit the present invention. Anyone familiar with the art may modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by a person of ordinary skill in the art without departing from the spirit and technical concept disclosed by the present invention shall still be covered by the claims of the present invention.

Claims

1. A free breathing dynamic magnetic resonance reconstruction method, characterized in that: The method comprises the following steps: constructing a first eigenvector based on the two-dimensional spatial coordinates of the dynamic magnetic resonance image; constructing a second eigenvector based on the reparameterized time latent variable; The first eigenvector and the second eigenvector are input into a first multi-layer perceptron to obtain the reconstructed dynamic magnetic resonance image intensity.

2. The free breathing dynamic magnetic resonance reconstruction method according to claim 1, characterized in that: Constructing the first eigenvector based on the two-dimensional spatial coordinates of the dynamic magnetic resonance image includes: The two-dimensional space coordinates are hash-encoded through a hash table to obtain the first feature vector.

3. The free breathing dynamic magnetic resonance reconstruction method according to claim 1, characterized in that: Constructing the second eigenvector based on the reparameterized time latent variable includes the following steps: Obtain an initialized time latent variable, where the initialized time latent variable is a standard Gaussian random variable with a mean of 0 and a variance of 1; Reparameterizing the initialized time latent variable; The re-parameterized time latent variable is input into a second multi-layer perceptron to obtain the second feature vector.

4. The free breathing dynamic magnetic resonance reconstruction method according to claim 3, characterized in that: The initialized time latent variable is reparameterized according to c(t)=μ(t)+∈σ(t), where c(t) represents the reparameterized time latent variable, ∈ represents a Gaussian random variable with a mean of 0 and a variance of 1, and μ(t) and σ(t) represent the mean and standard deviation of the initialized time latent variable.

5. The free breathing dynamic magnetic resonance reconstruction method according to claim 2, characterized in that: Also includes data consistency loss Smooth Regularization Loss and KL loss Constructing a total loss function; optimizing the first multilayer perceptron, the hash table, and the time latent variable based on the minimized total loss function.

6. The free breathing dynamic magnetic resonance reconstruction method according to claim 5, characterized in that: The total loss function where λ S and λ K Represents a hyperparameter.

7. The free breathing dynamic magnetic resonance reconstruction method according to claim 5, characterized in that: The data consistency loss where d θ represents the intensity of reconstructed dynamic magnetic resonance images, F u S c d θ Indicates d θ After the forward model F u S c The predicted k-space data obtained, D c represents the actual k-space data, c represents the cth coil; The smooth regularization loss in Represents the gradient of the mean of the latent variable along the time dimension; The KL loss Where n represents the dimension of the time latent variable, μ i (t) and σ i (t) represents the mean and standard deviation of the initialized temporal latent variable of the i-th dimension.

8. A free breathing dynamic magnetic resonance reconstruction system, characterized in that: The system comprises a first building module, a second building module and a reconstruction module; The first construction module is used to construct a first feature vector based on the two-dimensional spatial coordinates of the dynamic magnetic resonance image; The second construction module is used to construct a second feature vector based on the re-parameterized time latent variable; The reconstruction module is used to input the first eigenvector and the second eigenvector into a first multi-layer perceptron to obtain the reconstructed dynamic magnetic resonance image intensity.

9. A storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the free-breathing dynamic magnetic resonance reconstruction method according to any one of claims 1 to 7 is implemented.

10. An electronic device, characterized in that: include: Processor and memory; The memory is used to store computer programs; The processor is used to execute the computer program stored in the memory, so that the electronic device performs the free breathing dynamic magnetic resonance reconstruction method according to any one of claims 1 to 7.

Images