Ultra-low field magnetic resonance image enhancement method, system, medium, and electronic device

Abstract

The present invention provides an ultra-low field magnetic resonance image enhancement method, system, medium, and electronic device, including: generating a priori image based on a diffusion model; inputting the priori image into an implicit neural expression, training the implicit neural expression to learn the priori image to generate a high-field magnetic resonance image and inputting the degradation model to generate an ultra-low field magnetic resonance image; training the implicit neural expression to make the ultra-low field magnetic resonance image consistent with the measured ultra-low field magnetic resonance image; adding noise to the high-field magnetic resonance image output by the implicit neural expression and inputting it into the diffusion model for re-sampling and re-training the implicit neural expression until the trained implicit neural expression is obtained after a preset number of samplings; and realizing ultra-low field magnetic resonance image enhancement based on the diffusion model and the trained implicit neural expression. The ultra-low field magnetic resonance image enhancement method, system, medium, and electronic device of the present invention realize zero-sample ultra-low field magnetic resonance image enhancement, which is conducive to the application of ultra-low field magnetic resonance images.

Description

Technical Field

[0001] The present invention relates to the technical field of magnetic resonance imaging (MRI), and in particular to an ultra-low field magnetic resonance image enhancement method, system, medium and electronic equipment. Background Art

[0002] Ultra-low field MRI is a technology that uses ultra-low magnetic fields for MRI. Since the magnetic field strength of ultra-low field MRI is very low, the ultra-low field MRI system can provide a low-cost, electromagnetic shielding-free and convenient MRI method. Ultra-low field MRI is widely used in imaging research of various disease models, such as tumors, spinal diseases, arthritis, muscle trauma and other fields.

[0003] However, when the magnetic field strength is very low, the signal-to-noise ratio of the magnetic resonance image will decrease, and the image quality will deteriorate, thus affecting the diagnosis. Therefore, enhancing ultra-low-field magnetic resonance images to their corresponding high-field magnetic resonance images has always been a hot topic of research. However, as an ill-posed inverse problem, this task is very challenging. Summary of the invention

[0004] In view of the above-mentioned problems, the purpose of the present invention is to provide an ultra-low field magnetic resonance image enhancement method, system, medium, and electronic device, which realize zero-sample ultra-low field magnetic resonance image enhancement based on implicit neural expression driven by a diffusion model, and are conducive to the application of ultra-low field magnetic resonance images.

[0005] In a first aspect, the present invention provides an ultra-low field magnetic resonance image enhancement method, the method comprising the following steps: generating a priori image based on a diffusion model; inputting the priori image into an implicit neural expression, training the implicit neural expression to learn the priori image to generate a high-field magnetic resonance image, and inputting the high-field magnetic resonance image into a degradation model to generate an ultra-low field magnetic resonance image; training the implicit neural expression to make the ultra-low field magnetic resonance image consistent with the measured ultra-low field magnetic resonance image; adding noise to the high-field magnetic resonance image output by the implicit neural expression and inputting it into the diffusion model for re-sampling and re-training the implicit neural expression until a trained implicit neural expression is obtained after a preset number of samplings; and realizing ultra-low field magnetic resonance image enhancement based on the diffusion model and the trained implicit neural expression.

[0006] In an implementation of the first aspect, the implicit neural expression uses a multi-layer perceptron.

[0007] In an implementation of the first aspect, the forward model adopts x LF =PSF(L k (x HF ))+n, where xLF represents the target ultra-low field magnetic resonance image, x HF represents a high-field magnetic resonance image, PSF represents the Gaussian point spread function blur operator, L k represents k-fold linear downsampling, and n represents Gaussian additive noise.

[0008] In an implementation of the first aspect, when the distance between the target ultra-low field magnetic resonance image and the measured ultra-low field magnetic resonance image is When y is the smallest, it is determined that the target ultra-low field magnetic resonance image is consistent with the measured ultra-low field magnetic resonance image, wherein y represents the measured ultra-low field magnetic resonance image, represents the target ultra-low field magnetic resonance image, SSIM represents the structural similarity index, and μ represents the hyperparameter controlling the noise level.

[0009] In a second aspect, the present invention provides an ultra-low field magnetic resonance image enhancement system, the system comprising a generation module, a first training module, a second training module, a circulation module and an enhancement module;

[0010] The generation module is used to generate a priori images based on a diffusion model;

[0011] The first training module is used to input the prior image into an implicit neural expression, train the implicit neural expression to learn the prior image to generate a high-field magnetic resonance image, and input the high-field magnetic resonance image into a degradation model to generate an ultra-low-field magnetic resonance image;

[0012] The second training module is used to train the implicit neural expression so that the ultra-low field magnetic resonance image is consistent with the measured ultra-low field magnetic resonance image;

[0013] The circulation module is used to add noise to the high-field magnetic resonance image output by the implicit neural expression and input it into the diffusion model for re-sampling and re-training the implicit neural expression until the trained implicit neural expression is obtained after a preset number of sampling times;

[0014] The enhancement module is used to achieve ultra-low field magnetic resonance image enhancement based on the diffusion model and the trained implicit neural expression.

[0015] In an implementation of the second aspect, the implicit neural expression uses a multi-layer perceptron.

[0016] In an implementation of the second aspect, the forward model uses x Lf =PSF(L k (x Hf ))+n, where x Lf represents the target ultra-low field magnetic resonance image, x HFrepresents a high-field magnetic resonance image, PSF represents the Gaussian point spread function blur operator, L k represents k-fold linear downsampling, and n represents Gaussian additive noise.

[0017] In an implementation of the second aspect, when the distance between the target ultra-low field magnetic resonance image and the measured ultra-low field magnetic resonance image is When y is the smallest, it is determined that the target ultra-low field magnetic resonance image is consistent with the measured ultra-low field magnetic resonance image, wherein y represents the measured ultra-low field magnetic resonance image, represents the target magnetic resonance image, SSIM represents the structural similarity index, and μ represents the hyperparameter controlling the noise level.

[0018] 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 ultra-low field magnetic resonance image enhancement method.

[0019] In a fourth aspect, the present invention provides an electronic device, comprising: a processor and a memory;

[0020] The memory is used to store computer programs;

[0021] The processor is used to execute the computer program stored in the memory, so that the electronic device executes the above-mentioned ultra-low field magnetic resonance image enhancement method.

[0022] As described above, the ultra-low field magnetic resonance image enhancement method, system, medium, and electronic device of the present invention have the following beneficial effects:

[0023] (1) Zero-sample ultra-low field MRI enhancement is achieved based on implicit neural expression driven by a denoising diffusion model, which is beneficial to the application of ultra-low field MRI;

[0024] (2) Using zero-shot training to overcome the limitations of the diffusion model;

[0025] (3) The diffusion model is used to provide high-quality image priors for implicit neural expressions, thereby overcoming the problem of poor generalization of implicit neural expressions. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 Shown is a flow chart of an ultra-low field magnetic resonance image enhancement method in one embodiment of the present invention;

[0027] Figure 2 Shown is a schematic diagram of the framework of an ultra-low field magnetic resonance image enhancement method of the present invention in one embodiment;

[0028] Figure 3 Shown is a schematic diagram of the effect of the ultra-low field magnetic resonance image enhancement method of one embodiment of the present invention;

[0029] Figure 4 Shown is a schematic structural diagram of an ultra-low field magnetic resonance image enhancement system of the present invention in one embodiment;

[0030] Figure 5 It is a schematic structural diagram of an electronic device of the present invention in one embodiment. DETAILED DESCRIPTION

[0031] 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.

[0032] 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.

[0033] 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.

[0034] like Figure 1 and Figure 2 As shown, in one embodiment, the ultra-low field magnetic resonance image enhancement method of the present invention includes steps S1 to S5.

[0035] Step S1: Generate a priori image based on a diffusion model.

[0036] Specifically, diffusion models are a state-of-the-art generative model that can effectively generate priors for unsupervised inverse problems. Diffusion models generate high-quality image priors by adding a forward process of noise disturbance and a reverse process of training the logarithmic density gradient. In the prior art, most image enhancement methods based on diffusion models are limited by the lack of paired ultra-low field-high field data sets, the fixed degradation model and other conditions, which in turn limits the scope of application of diffusion models. The present invention overcomes the above-mentioned defects, uses a diffusion model to generate a prior image of an ultra-low field magnetic resonance image, and thus achieves zero-sample ultra-low field magnetic resonance image enhancement.

[0037] The generation of the prior image is essentially a diffusion denoising process using the Tweedie formula for the output of the diffusion model. The diffusion model has a sampling process with a sampling number. In the first sampling process, noise is input into the diffusion model. The Tweedie denoising process for each sampling number t in the sampling process can be expressed as:

[0038]

[0039] Among them, σ t is the Gaussian noise level during the t-th sampling process, is the score of the fraction-based diffusion model in the tth sampling process. t is the output of the diffusion model during the t-th sampling process, x 0|t It is the denoised output, i.e., the high-quality prior image, which will also serve as the input of the implicit neural expression.

[0040] Step S2: input the prior image into an implicit neural expression, train the implicit neural expression to learn the prior image to generate a high-field magnetic resonance image, and input the high-field magnetic resonance image into a degradation model to generate an ultra-low-field magnetic resonance image.

[0041] Specifically, the process of degenerating a high-field magnetic resonance image to an ultra-low-field magnetic resonance image can be expressed as Among them, x LF represents an ultra-low field magnetic resonance image, represents the degradation model from high-field MRI to ultra-low field, x HF represents the high-field MRI image that we wish to restore. There is no exact expression, so the present invention uses downsampling, blurring and noise addition to simulate The degradation model represented by is represented by x LF =PSF(L k (x HF ))+n. Where PSF represents the Gaussian point spread function blur operator, L k represents k-fold linear downsampling, and n represents Gaussian additive noise. Therefore, the ultra-low field image enhancement problem can be transformed into the following optimization problem:

[0042]

[0043] Among them, x is the optimization target in this optimization problem; is the optimal solution to the optimization problem, which represents the enhanced target image obtained after optimization. Data consistency term ||x LF -PSF(L k (x))|| 2 Minimize x using the binorm ||·|| LF-PSF(L k (x HF )) to ensure that the target ultra-low field magnetic resonance image obtained by degradation is consistent with the measured ultra-low field magnetic resonance image x LF The consistency of the data is determined by the regularization term R(x), and the regularization term R(x) uses prior knowledge to appropriately constrain the solution space. The present invention uses the semi-quadratic splitting method to decouple the data consistency term and the regularization term, and iterates the two sub-problems to solve the original problem, namely:

[0044]

[0045] Among them, z t is the auxiliary variable in the t-th (1≤t≤T) sampling process, μ is a hyperparameter that controls the noise level, and σ t is the Gaussian noise level of the t-th sampling, and λ is the coefficient of the regularization term. Formula 1 is the prior image generation subproblem, that is, how to generate a high-quality prior image; Formula 2 is the data consistency subproblem. In the present invention, the prior generation subproblem is solved by the diffusion model, and the data consistency subproblem is solved by implicit neural expression.

[0046] Implicit Neural Representation (INR) is a new type of unsupervised zero-shot deep learning framework. Due to the continuity of implicit neural representation, a single well-trained implicit neural representation model can handle super-resolution tasks of any scale. However, in the absence of supplementary information for reconstructing the image, this continuity will cause the generalization of the model to be poor. In the present invention, a diffusion model is used to provide high-quality image priors for implicit neural representation, thereby overcoming the problem of poor generalization of implicit neural representation.

[0047] Among them, implicit neural expression outputs high-field magnetic resonance images Expressed as a continuous function of image coordinates, recorded as:

[0048]

[0049] Where p is the spatial coordinate, and Respectively represent two-dimensional real number space (corresponding to spatial coordinates) and one-dimensional real number space (corresponding to image intensity). The present invention uses a multi-layer perceptron M Φ To approximate this function f.

[0050] Specifically, the implicit neural expression needs to satisfy the following two consistencies:

[0051] (1) Prior consistency: Using the high-quality prior image x provided by the diffusion model 0|t , let the output of the implicit neural expression be a high-field magnetic resonance image With the prior image x0|t Consistency;

[0052] (2) Measurement consistency: Using the actual ultra-low field magnetic resonance data y, the target ultra-low field magnetic resonance image output by the implicit neural expression is made consistent with the measured ultra-low field magnetic resonance image y after being output by the degradation model.

[0053] Therefore, the implicit neural representation is trained with the above two consistencies as the goal, which is achieved by optimizing the following objective function:

[0054]

[0055]

[0056] Among them, A(·) is the degradation model, that is, PSF(L k (·)). The first term of the formula For prior consistency, the difference between the prior image and the high-field MRI image is minimized using the L2 loss, where That is, the implicit neural expression network. In the training of step S2, the implicit neural expression learning diffusion model provides a high-quality prior image, output This output is input into the degradation model, namely where x LF represents the target ultra-low field magnetic resonance image, x HF represents a high-field magnetic resonance image, PSF represents the Gaussian point spread function blur operator, L k represents k-fold linear downsampling, and n represents Gaussian additive noise.

[0057] Step S3: training the implicit neural expression to make the ultra-low field magnetic resonance image consistent with the measured ultra-low field magnetic resonance image.

[0058] Specifically, the second term in the above objective function To measure consistency, the distance between the target ULF MRI image and the measured ULF MRI image is minimized. The implicit neural representation is further trained so that the output of the degraded implicit neural representation is consistent with the measured low-field image. The distance is expressed as: in is a hyperparameter term that uses the noise at the corresponding step level in the diffusion model sampling process to adjust the weight. SSIM (StructureSimilarity Index Measure) represents the structural similarity index, which is used to quantify the similarity between the output of the degraded implicit neural expression and the measured ultra-low field image.

[0059] Step S4, adding noise to the high-field magnetic resonance image output by the implicit neural expression and inputting it into the diffusion model for re-sampling and re-training the implicit neural expression until the trained implicit neural expression is obtained after a preset number of sampling times.

[0060] Specifically, the high-field magnetic resonance image output by the implicit neural expression is added with noise and then input into the diffusion model for further sampling, and the sampling result is input into the implicit neural expression, and the above steps are repeated until the preset sampling times.

[0061] Among them, for each implicit neural expression, the high-field magnetic resonance image output during the tth sampling process is Adding the noise level during the t-1th sampling process, this step can be expressed as:

[0062]

[0063] Among them, σ t-1 is the noise level in the t-1th sampling process, ξ is the parameter that controls the noise level, and ∈ is Gaussian noise. t-1 Return the input to the diffusion model in step S1 to perform the next diffusion model sampling. Execute steps S1 to S4 in a loop until the diffusion model sampling of the preset sampling times is completed. After the sampling process of the diffusion model of the preset sampling times is completed, the high-field magnetic resonance image output by the implicit neural expression is the image after the ultra-low-field magnetic resonance image enhancement.

[0064] Step S5: realizing ultra-low field magnetic resonance image enhancement based on the diffusion model and the trained implicit neural expression.

[0065] Specifically, in actual use, the ultra-low field magnetic resonance image to be enhanced is input into the diffusion model and the trained implicit neural expression in sequence to obtain an enhanced high field magnetic resonance image.

[0066] Figure 3 The figure shows the effect of a high-field magnetic resonance image obtained by using the ultra-low-field magnetic resonance image enhancement method of the present invention.

[0067] The first column is the measured ultra-low field magnetic resonance image input into the present invention, the second column is the enhanced magnetic resonance image outputted by the ultra-low field magnetic resonance image enhancement method of the present invention, and the third column is the measured high field magnetic resonance image corresponding to the input ultra-low field magnetic resonance image, which serves as a high field reference image. As can be seen from the figure, the ultra-low field magnetic resonance image enhancement method of the present invention has an enhancement effect on the ultra-low field magnetic resonance image that exceeds other existing algorithms, realizes high-quality enhancement of the ultra-low field magnetic resonance image, and is conducive to diagnosis using the ultra-low field magnetic resonance image.

[0068] The protection scope of the ultra-low field magnetic resonance image enhancement 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.

[0069] An embodiment of the present invention further provides an ultra-low field magnetic resonance image enhancement system, which can implement the ultra-low field magnetic resonance image enhancement method described in the present invention. However, the implementation device of the ultra-low field magnetic resonance image enhancement system described in the present invention includes but is not limited to the structure of the ultra-low field magnetic resonance image enhancement 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.

[0070] like Figure 4 As shown, in one embodiment, the ultra-low field magnetic resonance image enhancement system provided by the present invention includes a generation module 41, a first training module 42, a second training module 43, a circulation module 44 and an enhancement module 45.

[0071] The generating module 41 is used to generate a priori images based on a diffusion model.

[0072] The first training module 42 is connected to the generation module 41, and is used to input the prior image into implicit neural expression, train the implicit neural expression to learn the prior image to generate a high-field magnetic resonance image, and input the high-field magnetic resonance image into a degradation model to generate an ultra-low-field magnetic resonance image.

[0073] The second training module 43 is connected to the first training module 42 and is used to train the implicit neural expression so that the ultra-low field magnetic resonance image is consistent with the measured ultra-low field magnetic resonance image.

[0074] The circulation module 44 is connected to the generation module 41 and the second training module 43, and is used to add noise to the high-field magnetic resonance image output by the implicit neural expression and input it into the diffusion model for re-sampling and re-training the implicit neural expression until the trained implicit neural expression is obtained after a preset number of sampling times.

[0075] The enhancement module 45 is connected to the circulation module 44 and is used to realize ultra-low field magnetic resonance image enhancement based on the diffusion model and the trained implicit neural expression.

[0076] The structures and principles of the generation module 41, the first training module 42, the second training module 43, the circulation module 44 and the enhancement module 45 correspond to the above-mentioned ultra-low field magnetic resonance image enhancement method one by one, so they will not be described in detail here.

[0077] 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.

[0078] 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.

[0079] 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.

[0080] The embodiment of the present invention also provides a computer-readable storage medium. A person skilled in the art can understand that all or part of the steps in the ultra-low field magnetic resonance image enhancement 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 (for example, a floppy disk, a hard disk, a tape), an optical medium (for example, a digital video disc (DVD)), or a semiconductor medium (for example, a solid state disk (SSD)), etc.

[0081] An embodiment of the present invention further provides an electronic device, which includes a processor and a memory.

[0082] The memory is used to store computer programs.

[0083] The memory includes: ROM, RAM, disk, USB flash drive, memory card or CD and other media that can store program codes.

[0084] 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 ultra-low field magnetic resonance image enhancement method.

[0085] 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.

[0086] like Figure 5As 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 51, a memory 52, and a bus 53 connecting different system components (including the memory 52 and the processing unit 51).

[0087] Bus 53 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. For example, these architectures include but are not limited to Industry Standard Architecture (ISA) bus, Micro Channel Architecture (MAC) bus, Enhanced ISA bus, Video Electronics Standards Association (VESA) local bus and Peripheral Component Interconnect (PCI) bus.

[0088] 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.

[0089] The memory 52 may include computer system readable media in the form of volatile memory, such as random access memory (RAM) 521 and / or cache memory 522. 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 523 may be used to read and write non-removable, non-volatile magnetic media ( Figure 5 not shown, usually called a "hard drive"). Although Figure 5 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 53 via one or more data medium interfaces. The memory 52 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.

[0090] A program / utility 524 having a set (at least one) of program modules 5241 may be stored, for example, in the memory 52, such program modules 5241 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 thereof may include an implementation of a network environment. The program modules 5241 generally perform the functions and / or methods of the embodiments described herein.

[0091] 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 54. 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 55. Figure 5 As shown, the network adapter 55 communicates with other modules of the electronic device via the bus 53. 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.

[0092] 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 method for ultra-low field magnetic resonance image enhancement, characterized in that: The method comprises the following steps: Generate a priori images based on diffusion models; Inputting the prior image into an implicit neural expression, training the implicit neural expression to learn the prior image to generate a high-field magnetic resonance image, and inputting the high-field magnetic resonance image into a degradation model to generate an ultra-low-field magnetic resonance image; Training the implicit neural expression to make the ultra-low field magnetic resonance image consistent with the measured ultra-low field magnetic resonance image; The high-field magnetic resonance image output by the implicit neural expression is added with noise and then input into the diffusion model for re-sampling and re-training the implicit neural expression until the trained implicit neural expression is obtained after a preset number of sampling times; Implementing ultra-low field magnetic resonance image enhancement based on the diffusion model and the trained implicit neural expression; The degradation model uses x LF =PSF(L k (x HF ))+n, where x LF represents the target ultra-low field magnetic resonance image, x HF represents a high-field magnetic resonance image, PSF represents the Gaussian point spread function blur operator, L k represents k-fold linear downsampling, and n represents Gaussian additive noise.

2. The ultra-low field magnetic resonance image enhancement method according to claim 1, characterized in that: The implicit neural expression adopts a multi-layer perceptron.

3. The ultra-low field magnetic resonance image enhancement method according to claim 1, characterized in that: When the distance between the target ultra-low field magnetic resonance image and the measured ultra-low field magnetic resonance image is When y is the smallest, it is determined that the target ultra-low field magnetic resonance image is consistent with the measured ultra-low field magnetic resonance image, wherein y represents the measured ultra-low field magnetic resonance image, represents the target ultra-low field magnetic resonance image, SSIM represents the structural similarity index, and μ represents the hyperparameter controlling the noise level.

4. An ultra-low field magnetic resonance image enhancement system, characterized in that: The system comprises a generation module, a first training module, a second training module, a circulation module and an enhancement module; The generation module is used to generate a priori images based on a diffusion model; The first training module is used to input the prior image into an implicit neural expression, train the implicit neural expression to learn the prior image to generate a high-field magnetic resonance image, and input the high-field magnetic resonance image into a degradation model to generate an ultra-low-field magnetic resonance image; The second training module is used to train the implicit neural expression so that the ultra-low field magnetic resonance image is consistent with the measured ultra-low field magnetic resonance image; The circulation module is used to add noise to the high-field magnetic resonance image output by the implicit neural expression and input it into the diffusion model for re-sampling and re-training the implicit neural expression until the trained implicit neural expression is obtained after a preset number of sampling times; The enhancement module is used to realize ultra-low field magnetic resonance image enhancement based on the diffusion model and the trained implicit neural expression; The degradation model uses x LF =PSF(L k (x HF ))+n, where x LF represents the target ultra-low field magnetic resonance image, x HF represents a high-field magnetic resonance image, PSF represents the Gaussian point spread function blur operator, L k represents k-fold linear downsampling, and n represents Gaussian additive noise.

5. The ultra-low field magnetic resonance image enhancement system according to claim 4, characterized in that: The implicit neural expression adopts a multi-layer perceptron.

6. The ultra-low field magnetic resonance image enhancement system according to claim 4, characterized in that: When the distance between the target ultra-low field magnetic resonance image and the measured ultra-low field magnetic resonance image is When y is the smallest, it is determined that the target ultra-low field magnetic resonance image is consistent with the measured ultra-low field magnetic resonance image, wherein y represents the measured ultra-low field magnetic resonance image, represents the target magnetic resonance image, SSIM represents the structural similarity index, and μ represents the hyperparameter controlling the noise level.

7. A storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the ultra-low field magnetic resonance image enhancement method according to any one of claims 1 to 3 is implemented.

8. 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 ultra-low field magnetic resonance image enhancement method according to any one of claims 1 to 3.

Images