Method, system and apparatus for parameter optimization of time-dependent diffusion magnetic resonance imaging

Abstract

This invention provides a method, system, and device for optimizing parameters in time-dependent diffusion magnetic resonance imaging (TD-MRI), belonging to the field of medical image processing technology. The method includes: acquiring the hardware constraints of the target MRI scanner and the physiological constraints of the target anatomical region; iteratively simulating noisy signals within a biophysical model framework using a global optimization algorithm, calculating the estimation error of microstructural parameters and updating the candidate parameter set until convergence; and outputting the combination of acquired parameters with the minimum error. This invention, employing the aforementioned method, system, and device for optimizing parameters in time-dependent diffusion magnetic resonance imaging, can improve the quantitative accuracy, repeatability, and cross-center versatility of td-dMRI, overcoming the shortcomings of traditional empirical parameters such as large errors, poor repeatability, and difficulty in generalization. It is applicable to clinical and research scenarios with different field strengths and different anatomical regions.

Description

Technical Field

[0001] This invention relates to the field of medical image processing technology, and in particular to a method, system and device for parameter optimization in time-dependent diffusion magnetic resonance imaging. Background Technology

[0002] Time-dependent diffusion magnetic resonance imaging (td-dMRI) is a magnetic resonance microstructure imaging method. Implementing td-dMRI typically requires combining two or more different diffusion gradient sequences, such as pulsed gradient spin echo (PGSE) and oscillating gradient spin echo (OGSE). By combining and analyzing signals from different sequences, biophysical models such as IMPULSE (Imaging Microstructural Parameters Using Limited Spectrally Edited Diffusion) can be constructed, allowing for the precise calculation of microstructural parameters. Despite the promising future of td-dMRI, its clinical application faces significant challenges, which constitute the fundamental limitations of current technologies.

[0003] Lack of standardized optimization methods: Currently, the selection of td-dMRI scanning parameters (such as OGSE frequency, b-value, PGSE diffusion time, etc.) mainly relies on researchers' "experience" or directly citing published "literature values." This approach is very crude and has not formed a scientific and systematic method for parameter optimization.

[0004] High parameter sensitivity and unstable results: Quantitative results from td-dMRI are extremely sensitive to the acquired parameters. More importantly, the optimal parameter combination is influenced by a variety of factors:

[0005] Differences in MRI hardware systems: Different models and field strengths (e.g., 3.0T vs 5.0T) of MRI equipment have vastly different hardware performance characteristics (e.g., maximum gradient field strength, gradient switching rate). Directly applying "empirical parameters" from one device to another often yields inaccurate results.

[0006] Anatomical region specificity: Different human tissues (such as the liver, kidneys, and prostate) have unique physiological characteristics, such as different T2 relaxation times and baseline signal-to-noise ratios (SNR). A set of parameters suitable for the liver may become ineffective in the kidneys due to excessively low SNR or excessively rapid T2 relaxation time decay.

[0007] Poor quantitative accuracy and repeatability: For the reasons mentioned above, using unoptimized, universal “literature parameters” for scanning will result in large errors in the final calculated microstructural indicators (such as cell size), and the results will be inconsistent (i.e., low repeatability) when repeated scanning is performed on the same subject.

[0008] In summary, existing technologies lack a systematic and automated solution to find the optimal acquisition parameters for td-dMRI by comprehensively considering factors such as MRI hardware performance and the physiological characteristics of the imaging area. As a result, the accuracy and stability of its quantitative results cannot meet clinical requirements. Summary of the Invention

[0009] The purpose of this invention is to provide a method, system, and device for optimizing parameters in time-dependent diffusion magnetic resonance imaging (TD-MRI), and to construct an automated parameter optimization technology that couples "device and tissue" to systematically replace empirical settings, thereby fundamentally improving the accuracy, repeatability, and cross-center versatility of td-dMRI quantitative results.

[0010] To achieve the above objectives, the present invention provides a parameter optimization method for time-dependent diffusion magnetic resonance imaging, comprising the following steps:

[0011] Step S1: Obtain the hardware constraints of the target MRI scanner and the physiological constraints of the target anatomical region;

[0012] Step S2: Based on the constraints, iteratively execute the global optimization algorithm:

[0013] Step S21: For the candidate acquisition parameter set, generate a noisy magnetic resonance signal based on a biophysical model simulation.

[0014] Step S22: Fit the microstructure parameters of the simulated magnetic resonance signal with noise, and calculate the error between the estimated microstructure parameters obtained after fitting and the preset true values.

[0015] Step S23: Update the candidate parameter set based on the error until convergence;

[0016] Step S3: Output the set of collected parameters with the smallest error in the iteration as the optimization result.

[0017] Preferably, in step S1, the hardware constraints of the target MRI scanner include the maximum gradient field strength, gradient switching rate, and physical minimum echo time.

[0018] Preferably, in step S1, the physiological constraints of the target anatomical region are tissue T2 relaxation time, microstructural information, and baseline signal-to-noise ratio.

[0019] Preferably, the basic signal-to-noise ratio is obtained as follows:

[0020] Acquire at least two b0 images of the target anatomical region;

[0021] The pixel-level signal-to-noise ratio is calculated using the difference method and then averaged. The specific calculation formula is as follows:

[0022] ;

[0023] in, Indicates the fundamental signal-to-noise ratio. This represents the average signal intensity measured within the region of interest of the average image obtained by adding two b0 images acquired in the same manner. This represents the standard deviation measured within the region of interest of the difference image obtained by subtracting two b0 images acquired in the same manner.

[0024] Preferably, in step S2, the global optimization algorithm is one of the following: genetic algorithm, particle swarm optimization algorithm, or simulated annealing algorithm.

[0025] Preferably, in step S21, the generation of the noisy magnetic resonance signal satisfies:

[0026] ;

[0027] in, Indicates in a specific Value and effective diffusion time Magnetic resonance signal intensity under certain conditions express value, Indicates the effective diffusion time. Indicates intracellular water content. This represents the apparent diffusion coefficient under the current effective diffusion time. Indicates the extracellular water diffusion coefficient;

[0028] The noise level is dynamically adjusted according to the echo time and T2 relaxation time of the candidate parameters. The specific calculation formula is as follows:

[0029] ;

[0030] in, This indicates the adjusted signal-to-noise ratio. Represents the echo time of the analog signal and the generation of the fundamental signal-to-noise ratio image. difference, Indicates the number of times data was collected. Indicates the current analog signal strength. This represents the signal strength when the value of b is 0.

[0031] Preferably, in step S21, the biophysical model is the IMPULSED model.

[0032] Preferably, the set of collected parameters includes:

[0033] The frequency and b-value of the oscillating gradient spin echo sequence, and the diffusion time and b-value of the pulsed gradient spin echo sequence.

[0034] This invention also provides a parameter optimization system for time-dependent diffusion magnetic resonance imaging, comprising:

[0035] The constraint acquisition module is used to acquire the hardware constraints of the target MRI scanner and the physiological constraints of the target anatomical region;

[0036] The iterative optimization module is used to search for the optimal set of acquisition parameters by using a global optimization algorithm based on the constraints obtained by the constraint acquisition module and by iteratively simulating and generating noisy magnetic resonance signals and calculating their fitting error.

[0037] The parameter output module is used to output the set of collected parameters that minimizes the fitting error, as searched by the iterative optimization module.

[0038] The present invention also provides a computer device including a memory and a processor, the memory being used to store instructions and the processor being used to execute the instructions to implement the parameter optimization method for time-dependent diffusion magnetic resonance imaging as described above.

[0039] Therefore, the present invention employs the above-described method, system, and device for optimizing parameters in time-dependent diffusion magnetic resonance imaging, and the beneficial technical effects are as follows:

[0040] (1) Improve quantitative accuracy: Through systematic optimization, the parameters determined by this invention can significantly reduce the estimation error of microstructural parameters, making the measurement results closer to the true values ​​of histopathology.

[0041] (2) Enhanced measurement repeatability: Parameters tailored for specific devices and tissues significantly improve the consistency of measurement results when repeated scans are performed on the same subject, which is crucial for longitudinal monitoring of disease.

[0042] (3) Improve clinical reliability: The quantitative maps generated using optimized parameters have fewer artifacts and are more evenly distributed, which improves the reliability of clinical diagnosis and diagnostic confidence.

[0043] (4) Promote the standardization and promotion of technology: This invention is universal and can be easily applied to different models of MRI equipment. By customizing standardized protocols for each center, it effectively solves the problem of the technology being difficult to promote among multiple centers. Attached Figure Description

[0044] Figure 1 This is a system workflow diagram;

[0045] Figure 2 A simplified diagram illustrating the input and output. Detailed Implementation

[0046] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0047] Unless otherwise defined, the technical or scientific terms used in this invention shall have the ordinary meaning as understood by one of ordinary skill in the art to which this invention pertains.

[0048] Example 1

[0049] A method for optimizing parameters in time-dependent diffusion magnetic resonance imaging includes the following steps:

[0050] Step S1: Constraint setting module, obtain the hardware constraints of the target MRI scanner and the physiological constraints of the target anatomical region.

[0051] like Figure 1 As shown, the goal of this embodiment is to optimize td-dMRI parameters for human kidney imaging using a 5.0T MRI scanner.

[0052] The constraint setting module, serving as the input to the optimization process, is responsible for defining the search boundaries and simulation environment. For example... Figure 2 As shown, in this embodiment, it specifically receives:

[0053] Hardware constraints of the target MRI scanner: Hardware parameters of a United Imaging Healthcare 5.0T uMR Jupiter MRI system, including maximum gradient field strength (120mT / m), maximum gradient switching rate (200mT / m / ms), and minimum echo time (TE) limits.

[0054] The physiological constraints of the target anatomical region are tissue T2 relaxation time, microstructural information, and baseline signal-to-noise ratio.

[0055] Evaluation (Baseline Signal-to-Noise Ratio): The actual signal-to-noise ratio value obtained by performing two b0 image scans on the target (e.g., kidney) and calculating using the difference method. In this embodiment... Approximately 20.

[0056] Simulated physiological tissue: A digital model containing preset kidney tissue T2 values ​​(e.g., 80ms) and microstructural information (including cell diameter, intracellular volume fraction, extracellular water diffusivity, and intracellular water diffusivity).

[0057] The basic signal-to-noise ratio is obtained as follows:

[0058] Acquire at least two b0 images of the target anatomical region;

[0059] The pixel-level signal-to-noise ratio is calculated using the difference method and then averaged.

[0060] Step S2: Based on the above constraints, the following operations are performed iteratively using a global optimization algorithm (repeated ten times).

[0061] The iterative optimization module, which serves as the core optimization engine, receives the aforementioned constraints and performs iterative optimization.

[0062] In this embodiment, the iterative optimization module incorporates a genetic algorithm. It randomly initializes a population containing multiple candidate parameter sets and performs iterative signal simulation based on the IMPULSE (Imaging Microstructure Parameters Edited with Finite Spectrum) model and a dynamic noise model. The genetic algorithm's optimization settings are as follows: population size is 500, maximum number of iterations is 240, and the 12 elite individuals with the highest fitness values ​​are retained in each generation to ensure the transmission of superior genes (EliteCount). The crossover ratio is set to 0.8, meaning 80% of the population individuals generate the next generation through crossover. The mutation operation uses an adaptive feasible mutation function (mutationadaptfeasible) to dynamically adjust the mutation strategy. The convergence criterion for the fitness function is 1e-6; the algorithm considers convergence met when the change in the objective function between two consecutive iterations is less than this value.

[0063] The generation of noisy magnetic resonance signals satisfies:

[0064] ;

[0065] in, Indicates in a specific Value and effective diffusion time Magnetic resonance signal intensity under certain conditions express value, Indicates the effective diffusion time. Indicates intracellular water content. This represents the apparent diffusion coefficient under the current effective diffusion time. This represents the extracellular water diffusion coefficient.

[0066] The noise level is dynamically adjusted according to the echo time and T2 relaxation time of the candidate parameters. The specific calculation formula is as follows:

[0067] ;

[0068] in, This indicates the adjusted signal-to-noise ratio. Represents the echo time of the analog signal and the generation of the fundamental signal-to-noise ratio image. difference, Indicates the number of times data was collected. Indicates the current analog signal strength. express b Signal strength when the value is 0.

[0069] The noise level is dynamically adjusted according to the echo time and T2 relaxation time of the candidate parameters.

[0070] In each iteration, the root mean square error between the estimated value and the true value is minimized as the evaluation criterion. A better parameter set is generated through selection, crossover and mutation operations until the preset convergence condition (such as the number of iterations or the error threshold) is met, and the optimal candidate parameter set in the optimization process is output.

[0071] The testing unit performs final verification and selection of the candidate optimal parameter set output by the iterative optimization module. The testing unit performs fitting tests on an independent, broader simulation test set, accumulates the fitting error for each pixel, and finally selects and outputs the set of parameters with the smallest global fitting error from the candidate set as the final optimization result.

[0072] Step S3: Output the set of collected parameters with the smallest error in the iteration as the optimization result.

[0073] Ultimately, an optimal combination of acquisition parameters was obtained for this 5.0T MRI system and kidney tissue. This combination includes specific OGSE (oscillatory gradient spin echo) frequency, PGSE (pulse gradient spin echo) diffusion time, and corresponding b-values, which can be directly loaded onto the MRI scanner for high-quality clinical data acquisition. The specific settings are as follows:

[0074] OGSE 36Hz b-value 0 / 250 / 500 / 750;

[0075] OGSE 54Hz b-value 0 / 120 / 240 / 360;

[0076] The PGSE gradient duration is 10ms, the gradient interval is 54ms, and the b value is 0 / 250 / 500 / 750.

[0077] Example 2

[0078] Hardware constraints of the target MRI scanner: Hardware parameters of a Siemens 3.0T Prisma MRI system, including maximum gradient field strength (80 mT / m), maximum gradient switching rate (200 mT / m / ms), and minimum echo time (TE) limits.

[0079] The physiological constraints of the target anatomical region are tissue T2 relaxation time, microstructural information, and baseline signal-to-noise ratio.

[0080] Evaluation (Baseline Signal-to-Noise Ratio): The actual signal-to-noise ratio value obtained by performing two b0 image scans on the target (e.g., prostate) and calculating it using the difference method. In this embodiment... Approximately 5.

[0081] Simulated physiological tissue: A digital model containing preset prostate tissue T2 values ​​(e.g., 100ms) and microstructural information (including cell diameter, intracellular volume fraction, extracellular water diffusion coefficient, and intracellular water diffusion coefficient).

[0082] The basic signal-to-noise ratio is obtained as follows:

[0083] Acquire at least two b0 images of the target anatomical region;

[0084] The pixel-level signal-to-noise ratio is calculated using the difference method and then averaged.

[0085] Step S2: Based on the above constraints, the following operations are performed iteratively using a global optimization algorithm.

[0086] In this embodiment, the iterative optimization module incorporates a particle swarm optimization algorithm. A particle swarm containing multiple candidate parameter sets (i.e., particles) is randomly initialized, and iterative signal simulation is performed based on the IMPULSE (Imaging Microstructure Parameters Edited with Finite Spectrum) model and a dynamic noise model. The particle swarm optimization algorithm is configured as follows: particle size is 500, and the maximum number of iterations is 240. The particle update strategy employs dynamically adjusted inertia weights to balance the algorithm's global search and local convergence capabilities. Individual learning factors (c1) and social learning factors (c2) guide particles to learn towards their own historical best position and the global best position. The convergence criterion for the fitness function is 1e-6; the algorithm considers convergence met when the change in the global best fitness value between two consecutive iterations is less than this value.

[0087] The generation of noisy magnetic resonance signals satisfies:

[0088] ;

[0089] The noise level is dynamically adjusted according to the echo time and T2 relaxation time of the candidate parameters. The specific calculation formula is as follows:

[0090] ;

[0091] The noise level is dynamically adjusted according to the echo time and T2 relaxation time of the candidate parameters.

[0092] In each iteration, the root mean square error between the estimated value and the true value is minimized as the evaluation criterion. The velocity and position of each particle are updated to explore a better parameter region until the preset convergence condition (such as the number of iterations or the error threshold) is met. The global optimal candidate parameter set recorded during the optimization process is then output.

[0093] The testing unit performs final verification and selection of the candidate optimal parameter set output by the iterative optimization module. The testing unit performs fitting tests on an independent, broader simulation test set, accumulates the fitting error for each pixel, and finally selects and outputs the set of parameters with the smallest global fitting error from the candidate set as the final optimization result.

[0094] Step S3: Output the set of collected parameters with the smallest error in the iteration as the optimization result.

[0095] Ultimately, an optimal combination of acquisition parameters was obtained for this MRI system and prostate tissue. This combination includes specific OGSE (oscillating gradient spin echo) frequencies, PGSE (pulse gradient spin echo) diffusion times, and corresponding b-values, which can be directly loaded onto the MRI scanner for high-quality clinical data acquisition. The specific settings are as follows:

[0096] OGSE 32Hz b-value 0 / 400 / 800 / 1200;

[0097] OGSE 51Hz b-value 0 / 100 / 250 / 350;

[0098] The PGSE gradient duration is 12ms, the gradient interval is 67ms, and the b value is 0 / 300 / 600 / 900.

[0099] Example 3

[0100] Hardware constraints of the target MRI scanner: Hardware parameters of a Philips 3.0T Ingenia Elition MRI system, including maximum gradient field strength (45 mT / m), maximum gradient switching rate (220 mT / m / ms), and minimum echo time (TE) limits.

[0101] The physiological constraints of the target anatomical region are tissue T2 relaxation time, microstructural information, and baseline signal-to-noise ratio.

[0102] Evaluation (Base Signal-to-Noise Ratio): The actual signal-to-noise ratio value obtained by performing two b0 image scans on the target (e.g., liver) and calculating using the difference method. In this embodiment... Approximately 10.

[0103] Simulated physiological tissue: A digital model containing preset liver tissue T2 values ​​(e.g., 40ms) and microstructure information (including cell diameter, intracellular volume fraction, extracellular water diffusion coefficient, and intracellular water diffusion coefficient).

[0104] The basic signal-to-noise ratio is obtained as follows:

[0105] Acquire at least two b0 images of the target anatomical region;

[0106] The pixel-level signal-to-noise ratio is calculated using the difference method and then averaged.

[0107] Step S2: Based on the above constraints, the following operations are performed iteratively using a global optimization algorithm.

[0108] In this embodiment, the iterative optimization module incorporates a simulated annealing algorithm. It randomly initializes a candidate parameter set as the initial solution and performs iterative signal simulation based on the IMPULSE (Imaging Microstructure Parameters Edited with Finite Spectrum) model and a dynamic noise model. The optimized settings for the simulated annealing algorithm are as follows: initial temperature is set to 100, maximum number of iterations is 5000, and the annealing function employs an exponential cooling strategy with a cooling coefficient of 0.95. The generation of new solutions utilizes an adaptive perturbation function, which dynamically adjusts the perturbation range of parameters based on the current iteration count and the quality of the solution to balance global search and local exploitation capabilities. The convergence criterion for the objective function is 1e-6; the algorithm considers convergence met when the change in the optimal objective function value found in 100 consecutive iterations is less than this value.

[0109] The generation of noisy magnetic resonance signals satisfies:

[0110] ;

[0111] The noise level is dynamically adjusted according to the echo time and T2 relaxation time of the candidate parameters. The specific calculation formula is as follows:

[0112] ;

[0113] The noise level is dynamically adjusted according to the echo time and T2 relaxation time of the candidate parameters.

[0114] In each iteration, the objective function is to minimize the root mean square error between the estimated and true values. The algorithm generates a new solution from the neighborhood of the current solution and decides whether to accept it based on the Metropolis criterion: if the new solution has a better objective function value, it is accepted directly; if the new solution is worse, it is accepted with a probability related to the current temperature, which decreases as the temperature decreases, thus allowing the algorithm to escape local optima in the early stages. The iteration continues until the preset convergence conditions are met (such as reaching the maximum number of iterations or the objective function stabilizing), and the optimal candidate parameter set found during the optimization process is output.

[0115] The testing unit performs final verification and selection of the candidate optimal parameter set output by the iterative optimization module. The testing unit performs fitting tests on an independent, broader simulation test set, accumulates the fitting error for each pixel, and finally selects and outputs the set of parameters with the smallest global fitting error from the candidate set as the final optimization result.

[0116] Step S3: Output the set of collected parameters with the smallest error in the iteration as the optimization result.

[0117] Ultimately, an optimal combination of acquisition parameters was obtained for this MRI system and liver tissue. This combination includes specific OGSE (oscillating gradient spin echo) frequencies, PGSE (pulse gradient spin echo) diffusion times, and corresponding b-values, which can be directly loaded onto the MRI scanner for high-quality clinical data acquisition. The specific settings are as follows:

[0118] OGSE 30Hz b-value 0 / 200 / 400 / 600;

[0119] OGSE 60Hz b-value 0 / 75 / 125 / 200;

[0120] The PGSE gradient duration is 12ms, the gradient interval is 56ms, and the b value is 0 / 400 / 800 / 1200.

[0121] Example 4

[0122] A parameter optimization system for time-dependent diffusion magnetic resonance imaging includes:

[0123] The constraint acquisition module is used to acquire the hardware constraints of the target MRI scanner and the physiological constraints of the target anatomical region;

[0124] The iterative optimization module is used to search for the optimal set of acquisition parameters by using a global optimization algorithm based on the constraints obtained by the constraint acquisition module and by iteratively simulating and generating noisy magnetic resonance signals and calculating their fitting error.

[0125] The parameter output module is used to output the set of collected parameters that minimizes the fitting error, as searched by the iterative optimization module.

[0126] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0127] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-including system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0128] More specific examples of computer-readable media (a non-exhaustive list) include: electrical connections (electronic devices) having one or more wires, portable computer disk drives (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which the program can be printed, because the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0129] It is worth noting that all contents not described in detail in this invention are existing technologies and are well known to those skilled in the art.

[0130] Therefore, this invention employs the aforementioned time-dependent diffusion magnetic resonance imaging parameter optimization method, system, and device to construct an automated parameter optimization technology that couples "device and tissue," systematically replacing empirical settings, thereby fundamentally improving the accuracy, repeatability, and cross-center universality of td-dMRI quantitative results.

[0131] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A parameter optimization method for time-dependent diffusion magnetic resonance imaging, characterized in that, Includes the following steps: Step S1: Obtain the hardware constraints of the target MRI scanner and the physiological constraints of the target anatomical region; Step S2: Based on the constraints, iteratively execute the global optimization algorithm: Step S21: For the candidate acquisition parameter set, generate a noisy magnetic resonance signal based on a biophysical model simulation. Step S22: Fit the microstructure parameters of the simulated magnetic resonance signal with noise, and calculate the error between the estimated microstructure parameters obtained after fitting and the preset true values. Step S23: Update the candidate parameter set based on the error until convergence; Step S3: Output the set of collected parameters with the smallest error in the iteration as the optimization result.

2. The parameter optimization method for time-dependent diffusion magnetic resonance imaging according to claim 1, characterized in that, In step S1, the hardware constraints of the target MRI scanner include the maximum gradient field strength, gradient switching rate, and physical minimum echo time.

3. The parameter optimization method for time-dependent diffusion magnetic resonance imaging according to claim 1, characterized in that, In step S1, the physiological constraints of the target anatomical region are tissue T2 relaxation time, microstructural information, and baseline signal-to-noise ratio.

4. The parameter optimization method for time-dependent diffusion magnetic resonance imaging according to claim 3, characterized in that, The basic signal-to-noise ratio is obtained as follows: Acquire at least two b0 images of the target anatomical region; The pixel-level signal-to-noise ratio is calculated using the difference method and then averaged. The specific calculation formula is as follows: ； in, Indicates the fundamental signal-to-noise ratio. This represents the average signal intensity measured within the region of interest of the average image obtained by adding two b0 images acquired in the same manner. This represents the standard deviation measured within the region of interest of the difference image obtained by subtracting two b0 images acquired in the same manner.

5. The parameter optimization method for time-dependent diffusion magnetic resonance imaging according to claim 1, characterized in that, In step S2, the global optimization algorithm is one of the following: genetic algorithm, particle swarm optimization algorithm, or simulated annealing algorithm.

6. The parameter optimization method for time-dependent diffusion magnetic resonance imaging according to claim 4, characterized in that, In step S21, the generation of the noisy magnetic resonance signal satisfies: ； in, Indicates in a specific Value and effective diffusion time Magnetic resonance signal intensity under certain conditions express value, Indicates the effective diffusion time. Indicates intracellular water content. This represents the apparent diffusion coefficient under the current effective diffusion time. Indicates the extracellular water diffusion coefficient; The noise level is dynamically adjusted according to the echo time and T2 relaxation time of the candidate parameters. The specific calculation formula is as follows: ； in, This indicates the adjusted signal-to-noise ratio. Represents the echo time of the analog signal and the generation of the fundamental signal-to-noise ratio image. difference, Indicates the number of times data was collected. Indicates the current analog signal strength. This represents the signal strength when the value of b is 0.

7. The parameter optimization method for time-dependent diffusion magnetic resonance imaging according to claim 1, characterized in that, In step S21, the biophysical model is the IMPULSED model.

8. The parameter optimization method for time-dependent diffusion magnetic resonance imaging according to claim 1, characterized in that, The set of collected parameters includes: The frequency and b-value of the oscillating gradient spin echo sequence, and the diffusion time and b-value of the pulsed gradient spin echo sequence.

9. A parameter optimization system for time-dependent diffusion magnetic resonance imaging, characterized in that, include: The constraint acquisition module is used to acquire the hardware constraints of the target MRI scanner and the physiological constraints of the target anatomical region; The iterative optimization module is used to search for the optimal set of acquisition parameters by using a global optimization algorithm based on the constraints obtained by the constraint acquisition module and by iteratively simulating and generating noisy magnetic resonance signals and calculating their fitting error. The parameter output module is used to output the set of collected parameters that minimizes the fitting error, as searched by the iterative optimization module.

10. A computer device, characterized in that, It includes a memory and a processor, the memory being used to store instructions, and the processor being used to execute the instructions to implement the parameter optimization method for time-dependent diffusion magnetic resonance imaging as described in any one of claims 1-8.

Images