High signal-to-noise ratio magnetic resonance data acquisition methods, signal processing methods and systems

Abstract

This invention discloses a high signal-to-noise ratio (SNR) magnetic resonance (MRI) data acquisition method, signal processing method, and system, belonging to the field of MRI technology. The invention aims to maximize the image SNR while maintaining spatial resolution by simultaneously optimizing the radio frequency flip angle sequence of the echo train and the k-space filtering weights through a joint optimization framework. This invention can obtain high-quality MRI images, fundamentally improving the image quality of FSE sequences.

Description

Technical Field

[0001] This invention belongs to the field of magnetic resonance technology, specifically relating to a high signal-to-noise ratio magnetic resonance data acquisition method, signal processing method, and system. Background Technology

[0002] Magnetic resonance imaging (MRI) primarily acquires images by repeatedly exciting and encoding the signal of an object in a magnetic field. Each encoding step is equivalent to acquiring a line in the frequency domain of the image, conforming to the Nyquist sampling theorem. After all encodings are completed, an inverse Fourier transform can be performed to reconstruct the image. Alternatively, advanced reconstruction algorithms such as multi-coil or parallel processing can be used to reconstruct the image even with undersampling. However, the object's signal requires a certain amount of time to recover after being excited. Therefore, the repeated excitations required for encoding necessitate waiting for the object's signal to recover, directly leading to problems such as long imaging time and low signal-to-noise ratio in MRI images.

[0003] Fast echo sequences (also known as Fast Spin Echo, Turbo Spin Echo, or Rapid Acquisition with Relaxation Enhancement) are a fundamental signal acquisition method in magnetic resonance imaging (MRI). They are widely used in various body parts, including the head and abdomen, and are also employed in numerous MRI techniques such as diffusion imaging, perfusion imaging, T2-weighted imaging, and parametric mapping. It is one of the most widely used basic sequences. For ease of description, fast echo sequences will be uniformly referred to as FSE in this invention. FSE is an extension of the typical spin echo (SE), consisting of a radio frequency excitation pulse followed by a series of refocusing pulses. After the excitation pulse, a signal is generated that decays rapidly. However, after each refocusing pulse, a signal refocusing occurs; typically, an FSE sequence can contain several to hundreds of refocusing pulses. Therefore, the FSE sequence can achieve multiple echo signals by repeatedly encoding the refocusing signal with a single excitation, thereby reducing the need for repeated excitation of the signal and waiting for object signal recovery, and realizing rapid image acquisition.

[0004] Early FSE sequences used 180° radio frequency pulses for signal refocusing. There was a relatively slow decay between multiple refocused echo signals. According to the magnetic field characteristics of matter, T2 decreases exponentially; typically, T2 for head gray matter is approximately 100 ms. This phenomenon leads to two problems:

[0005] First, the decrease in echo refocusing signal leads to a weaker signal after multiple echoes, resulting in a decrease in image signal-to-noise ratio and affecting quality.

[0006] Secondly, since each echo refocusing signal is encoded as part of the image frequency domain space, the exponential decrease between multiple echoes will cause the frequency domain space to form an exponentially decreasing filter, which will result in image blurring.

[0007] To address the aforementioned issues, existing technologies have proposed a variable flip angle design. This involves not using a constant 180-degree flip angle for the echo refocusing pulse in the FSE sequence, but instead designing a specific response flip angle based on experience, which can reduce image blur to some extent. Specific variable flip angle design methods fall into two categories: the first method directly ensures that the frequency domain signal meets a designed function shape (e.g., a Fermi window function) by designing the flip angle, thereby guaranteeing image resolution; the second method uses k-space filtering, applying a window function (e.g., a Fermi window after correction) to the k-space data during image reconstruction to weight the signal attenuated by T2 decay, correcting its amplitude and thus suppressing blur. However, existing designs still have limitations and cannot maximize the control of the image's signal-to-noise ratio. Summary of the Invention

[0008] This invention addresses the image acquisition problems of existing FSE sequence acquisition technologies by providing a high signal-to-noise ratio (SNR) magnetic resonance (MRI) data acquisition method, signal processing method, and system. Based on a joint framework of the point spread function (PSF), this invention maximizes the image SNR while maintaining spatial resolution by simultaneously optimizing the radio frequency flip angle sequence and k-space filtering weights, thereby obtaining high-quality MRI images.

[0009] The specific technical solution adopted in this invention is as follows:

[0010] In a first aspect, the present invention provides a method for optimizing fast echo imaging parameters for magnetic resonance data acquisition, the specific steps of which are as follows:

[0011] When the fast echo sequence to be optimized is executed by the magnetic resonance imaging (MRI) device, after applying the echo refocusing pulse sequence with a flip angle, the signal needs to be acquired and the k-space lines in the k-space space filled line by line; a weighting factor is set for signal filtering corresponding to the signal filled in each k-space line. This weighting factor The design signal strength for this k-space line Compared with theoretical signal strength The ratio;

[0012] The optimization method includes:

[0013] The flip angle sequence corresponding to the echo refocusing pulse sequence in the fast echo sequence and design signal strength sequence As parameters to be optimized, to maximize the design signal strength at the center of k-space. Sum of squared weighting factors corresponding to all k-space lines The ratio between the two values ​​is the optimization objective. Under image quality and security constraints, a numerical optimization algorithm is used to solve for the parameters to be optimized, thus obtaining the optimal flip angle sequence corresponding to the fast echo sequence. and the optimal design signal strength sequence .

[0014] As a preferred embodiment of the first aspect above, the image quality constraints include one or more of image resolution constraints, image artifact constraints, and image contrast constraints. The image resolution constraint ensures that the spatial resolution of the magnetic resonance image reconstructed based on k-space data meets a preset resolution requirement. The image artifact constraint ensures that the artifact evaluation index of the magnetic resonance image reconstructed based on k-space data meets a preset artifact control requirement. The image contrast constraint ensures that the contrast between tissues of interest in the magnetic resonance image reconstructed based on k-space data meets a preset contrast requirement. The safety constraint ensures that the energy deposition per unit time does not exceed an energy threshold.

[0015] As a preferred embodiment of the first aspect above, the designed signal strength sequence After inverse Fourier transform, the point spread function is used. One or more of the image resolution constraint, image artifact constraint, image contrast constraint, and security constraint satisfy the following definition:

[0016] The image resolution constraint is defined as follows: the full width of the point spread function at any height within the range of 1 / 3 to 2 / 3 of the peak height does not exceed a preset width value; more preferably, the image resolution constraint is defined as the full width of the point spread function at 1 / 2 of the peak height does not exceed a preset width value;

[0017] The image artifact constraint is defined as follows: the absolute size of the side lobe of the point spread function or the relative size of the side lobe to the main lobe does not exceed a preset size value, and the size type includes area or height; more preferably, the image artifact constraint is defined as the ratio of the area of ​​the first side lobe of the point spread function to the area of ​​the main lobe does not exceed a preset ratio.

[0018] The image contrast constraint is defined as follows: the relative difference in theoretical signal intensity of the two tissues of interest at the center of k space is greater than a preset contrast value; more preferably, the image contrast constraint is set to the ratio obtained by dividing the difference in theoretical signal intensity of the two tissues of interest at the center of k space by the sum of the theoretical signal intensity of the two tissues of interest at the center of k space by a value greater than the preset contrast value.

[0019] The safety constraint is defined as follows: the sum of the squares of the flip angles of all echo refocusing pulses does not exceed a preset energy threshold.

[0020] As a preferred embodiment of the first aspect above, when performing the solution using a numerical optimization algorithm, the optimization objective is converted into its equivalent form, which preferably transforms the maximization objective function into the minimization objective function.

[0021] Secondly, this invention provides a high signal-to-noise ratio magnetic resonance data acquisition method, the specific steps of which are as follows:

[0022] After pre-executing the fast echo imaging parameter optimization method for magnetic resonance data acquisition as described in the first aspect above, the optimal flip angle sequence corresponding to the fast echo sequence is obtained. .

[0023] The magnetic resonance imaging (MRI) device is controlled to run a fast echo sequence to scan the imaging target. During the operation, a radio frequency excitation pulse is applied first, followed by an optimized flip angle sequence. Echo refocusing pulses are applied sequentially and signals are acquired to fill the k-space lines in the k-space one by one. After the scan is completed, the k-space data corresponding to the imaging target is obtained.

[0024] Thirdly, the present invention provides a magnetic resonance signal processing method, which, after obtaining the k-space data corresponding to the imaging target according to the high signal-to-noise ratio magnetic resonance data acquisition method described in the second aspect above, continues to process the optimally designed signal intensity sequence obtained through optimization. The signals in all k-space lines are filtered sequentially. The original signal strength of the k-space line is multiplied by the signal strength sequence of the k-space line in the optimal design. The corresponding weighting factors are used to complete the filtering, and the filtered k-space data is saved for subsequent magnetic resonance image reconstruction.

[0025] Fourthly, the present invention provides a computer program product, including a computer program / instructions, which, when executed by a processor, can implement the fast echo imaging parameter optimization method for magnetic resonance data acquisition as described in any of the first aspects above, or can implement the high signal-to-noise ratio magnetic resonance data acquisition method as described in the second aspect above, or can implement the magnetic resonance signal processing method as described in the third aspect above.

[0026] Fifthly, the present invention provides a computer electronic device, which includes a memory and a processor;

[0027] The memory is used to store computer programs;

[0028] The processor is configured to, when executing the computer program, implement the fast echo imaging parameter optimization method for magnetic resonance data acquisition as described in any of the first aspects above, or implement the high signal-to-noise ratio magnetic resonance data acquisition method as described in the second aspect above, or implement the magnetic resonance signal processing method as described in the third aspect above.

[0029] In a sixth aspect, the present invention provides a magnetic resonance imaging system, which includes a magnetic resonance scanner and a control unit, wherein the control unit stores a computer program, and when the computer program is executed, it is used to control the magnetic resonance scanner to implement the high signal-to-noise ratio magnetic resonance data acquisition method as described in the second aspect above.

[0030] As a preferred embodiment of the sixth aspect above, it further includes a signal processing device, wherein the signal processing device processes the optimized design signal strength sequence obtained through optimization. The signals in all k-space lines are filtered sequentially. The original signal strength of the k-space line is multiplied by the signal strength sequence of the k-space line in the optimal design. The corresponding weighting factors are used to complete the filtering, and then the filtered k-space data is reconstructed to obtain a high signal-to-noise ratio magnetic resonance image.

[0031] Compared with the prior art, the present invention has the following advantages:

[0032] 1. This invention achieves a significant improvement in image signal-to-noise ratio (SNR). By simultaneously optimizing the RF flip angle sequence and k-space filtering weights within a unified optimization framework, this invention achieves synergistic design between echo signal evolution and the k-space weighting process. By adjusting the evolution of the echo signal throughout the echo chain, the signal distribution in different spatial frequency regions of k-space before filtering becomes more balanced, effectively reducing the noise amplification effect caused by k-space filtering. Its most significant effect is the improvement in image SNR. Compared to a constant flip angle design, the existing FFA-Fermi method improves SNR by approximately 50%. This invention achieves a 101% to 125% improvement in SNR (i.e., reaching 201% to 225% of the baseline value) in T1-weighted, T2-weighted, and proton density-weighted brain images, demonstrating a clear performance advantage. This result has been consistently verified in both theoretical calculations and volunteer experiments.

[0033] 2. Compared with existing methods, this invention directly constrains the point spread function (PSF), controlling the optimization process directly from the perspective of spatial resolution, thus enabling more flexible and effective control over image quality. By constraining the full width at half maximum (FWHM) of the PSF, the image achieves the preset spatial resolution; by constraining the sidelobe regions of the PSF, image artifacts are suppressed. This allows the invention to obtain a better signal-to-noise ratio while avoiding image blurring and artifacts caused by T2 attenuation, resulting in magnetic resonance images with a better signal-to-noise ratio, equally clear structural details, and fewer artifacts.

[0034] 3. This invention achieves efficient coordination and global optimization of multi-dimensional imaging targets. For the first time, this invention coordinates and optimizes multiple often conflicting imaging targets within a unified mathematical framework: it improves the signal-to-noise ratio while maintaining resolution without increasing scan time; it achieves a significant increase in signal-to-noise ratio even when SAR is strictly limited to a constant flip angle scheme of 50%, resolving the contradiction between high performance and scan safety; in abdominal imaging, by introducing liver-spleen contrast constraints, it successfully optimizes signal-to-noise ratio and resolution while preserving soft tissue contrast (>0.3) that meets diagnostic requirements, avoiding sacrificing image diagnostic value for the pursuit of a single metric.

[0035] 4. Robustness to real-world scanning environments. The experimental results of this invention are in high agreement with the predictions, demonstrating good robustness in B1 field inhomogeneity experiments. Compared to the Direct design, signal deviation is reduced by approximately 50%, and its robustness is comparable to the FFA-Fermi method, possessing the reliability for clinical applications. In an abdominal volunteer experiment, an average B1 field deviation of 13.5% was observed in the imaging area of ​​one participant, but the imaging results still showed an improvement in signal-to-noise ratio of over 100%, proving that the design of this invention has a certain resistance to B1 field inhomogeneity.

[0036] 5. This invention provides a universal, powerful, and highly promising sequence design paradigm. The optimization framework of this invention does not depend on a specific k-space ordering or trajectory. This framework can flexibly adjust the optimization objective and constraints according to different tissue relaxation parameters and imaging requirements, thus making it applicable to various magnetic resonance imaging (MRI) applications. Experiments have successfully applied it to center-ordered brain imaging and sequentially ordered abdominal imaging based on partial Fourier reconstruction undersampling, achieving equally excellent results. This demonstrates that this scheme, as a general design paradigm, has the great potential to be easily extended to other contrast weights, anatomical sites, and even non-Cartesian sampling, providing a powerful tool for solving a wider range of MRI sequence optimization problems. Attached Figure Description

[0037] Figure 1The figure shows the results of the brain imaging flip angle design and the characteristics of the k-space filtering function and its point spread function in Example 1.

[0038] Figure 2 The images shown in Example 1 are brain images with different contrasts designed using different flip angles. The image window is scaled by a ×5.0 ratio to display background noise.

[0039] Figure 3 The graph shows the theoretical improvement and the measured improvement results of brain imaging SNR at different contrast levels in Example 1.

[0040] Figure 4 This is a graph showing the measurement results of brain B1 non-uniformity in Example 1;

[0041] Figure 5 The figure shows the results of the abdominal imaging flip angle design and the characteristics of the k-space filtering function and its point spread function in Example 2.

[0042] Figure 6 The images shown are of the abdomen designed with different flip angles in Example 2. The image window is scaled by a ratio of ×6.7 to display background noise.

[0043] Figure 7 The figure shows the theoretical and measured improvement results of the abdominal imaging SNR in Example 2.

[0044] Figure 8 This is a graph showing the measurement results of abdominal B1 heterogeneity in Example 2. Detailed Implementation

[0045] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Many specific details are set forth in the following description to provide a thorough understanding of the present invention. However, the present invention can be practiced in many other ways different from those described herein, and those skilled in the art can make similar modifications without departing from the spirit of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below. Technical features in various embodiments of the present invention can be combined accordingly without mutual conflict.

[0046] In a preferred embodiment of the present invention, a method for optimizing fast echo sequence (FSE) parameters for magnetic resonance data acquisition is provided. This method jointly optimizes two sets of key parameters in the FSE sequence, whereby the two sets of optimized parameters are the flip angle sequence corresponding to the echo refocusing pulse sequence in the fast echo sequence. and design signal strength sequence The meaning of this will be explained in detail later.

[0047] The basic sequence structure of the FSE sequence to be optimized in this invention belongs to existing technology. When the FSE sequence is executed by a magnetic resonance imaging (MRI) device, after applying an echo refocusing pulse sequence with a flip angle, signals need to be acquired and k-space lines in k-space filled line by line. However, it should be noted that not every echo refocusing pulse requires signal acquisition and filling of a k-space line; the specific timing of acquisition can be adjusted according to the actual FSE sequence design. Furthermore, this invention also simultaneously introduces a k-space filter window function, that is, applying a window function to the originally acquired k-space data during the signal processing of image reconstruction. (k-space filter window function) It can be viewed as a sequence of design signal strengths. N represents the total number of k-space lines. As an index of the k-space line The first k-space line filled is denoted as the 0th line, and the last k-space line filled is denoted as the 1st line. The i-th k-space line, filled sequentially, corresponds to the design signal strength. Therefore, in the subsequent signal processing of k-space data, any first... For each k-space line filled with a signal, a weighting factor is set for signal filtering. This weighting factor For the first Design signal strength of k-space lines Compared with theoretical signal strength The ratio. Among them, the first Theoretical signal strength of k-space lines It is based on the flip angle sequence through the Extended Phase Graphs (EPG) algorithm. Simulation results show that the EPG algorithm is a current technology, and its input, besides the flip angle sequence, is... In addition, it is also necessary to include the physiological parameters of the imaging target (T1 and T2 relaxation times) and echo spacing, etc. The specific algorithm process will not be described in detail here.

[0048] For the aforementioned FSE sequences, the specific implementation process of the parameter optimization method of the present invention is as follows:

[0049] The flip angle sequence corresponding to the echo refocusing pulse sequence in the fast echo sequence and design signal strength sequence As parameters to be optimized, to maximize the design signal strength at the center of k-space. Sum of squared weighting factors corresponding to all k-space lines The ratio between the two values ​​is the optimization objective. Under image quality and security constraints, a numerical optimization algorithm is used to solve for the parameters to be optimized, thus obtaining the optimal flip angle sequence corresponding to the fast echo sequence. and the optimal design signal strength sequence .

[0050] It should be noted that the design signal strength at the center of k-space... yes The specific design signal strength is determined by the filling order of the k-space. The design signal strength corresponding to the k-space line filling the center of the k-space is the k-space strength. This invention can be applied to different k-space sampling trajectories, including but not limited to Cartesian sampling with center sorting and sequential sorting, and can also be extended to non-Cartesian sampling trajectories.

[0051] The optimization problem described above, consisting of the parameters to be optimized, the optimization objective, and the constraints, can be expressed as a formula:

[0052]

[0053] In the formula: N is the number of k-space lines used to fill k-space; It is a flip angle sequence. This is the flip angle corresponding to the i-th echo refocusing pulse in the fast echo sequence; This is the design signal intensity sequence used in the image reconstruction stage. This represents a k-space filter window function, consisting of N design signal strengths to be optimized. composition; The weighting factor is used in subsequent signal processing to perform weighted filtering on the signal of the i-th k-space line. To determine the theoretical signal strength for filling the i-th k-th space line, based on the flip angle sequence... Obtained through simulation using the extended phase diagram algorithm; The design signal strength is centered at k-space; Image quality constraint and Security constraint represent image quality constraint and security constraint, respectively.

[0054] Therefore, the optimization task of solving for the parameters to be optimized in this optimization problem can be expressed as:

[0055]

[0056] It should also be noted that the above optimization objective is not limited to a specific form; when solving using numerical optimization algorithms, it can be transformed into other equivalent forms. Let's assume the above optimization objective of this invention is simplified to... Then, the following linear transformations are all equivalent forms:

[0057] The scaling factor

[0058] The offset For any constant

[0059] The scaling factor Offset For any constant

[0060] The scaling factor

[0061] The scaling factor Offset For any constant

[0062] Of course, in addition to linear transformation, the above optimization objective can also be transformed into other equivalent forms through nonlinear monotonic transformation. All equivalent forms that can achieve the same optimization objective should also be understood as being equivalent to the above optimization objective of this invention and are also within the scope of protection of this invention.

[0063] Considering that conventional optimization algorithms generally use minimizing the objective function as the optimization objective, in the embodiments of the present invention, when using a numerical optimization algorithm to perform the solution, the above-mentioned maximization objective function can be equivalently transformed into minimizing the negative objective function by taking a negative number, and its formula is as follows:

[0064]

[0065] Therefore, existing optimization algorithms (such as the `minimize` function in the `scipy` package combined with the trust region algorithm) can be used to solve for the parameters to be optimized by minimizing the negative objective function under image quality and security constraints. The optimal solution obtained is the optimal flip angle sequence corresponding to the fast echo sequence. and the optimal design signal strength sequence .

[0066] In addition, it should be noted that the image quality constraints and security constraints in this invention need to be determined based on the actual magnetic resonance image quality requirements and the security requirements during the execution of the imaging sequence, and there are no restrictions on them.

[0067] As a preferred embodiment of the present invention, the image quality constraint can be one or more of image resolution constraint, image artifact constraint, and image contrast constraint. Specifically, the image resolution constraint ensures that the spatial resolution of the magnetic resonance image reconstructed based on k-space data meets a preset resolution requirement; the image artifact constraint ensures that the artifact evaluation index of the magnetic resonance image reconstructed based on k-space data meets a preset artifact control requirement; and the image contrast constraint ensures that the contrast between tissues of interest in the magnetic resonance image reconstructed based on k-space data meets a preset contrast requirement. Additionally, the safety constraint can be set to ensure that the energy deposition per unit time does not exceed a human safety threshold. The specific requirements and thresholds for each constraint can be determined according to actual optimization needs.

[0068] The aforementioned image quality and security constraints can be imposed on the reconstructed magnetic resonance images based on k-space data, but generally, constraints on the reconstructed images can be achieved more directly through weighted k-space.

[0069] Therefore, as a preferred embodiment of the present invention, the design signal strength sequence can be... After undergoing the inverse Fourier transform (IFT), it becomes the point spread function (PSF). Based on the PSF definition, image resolution constraints and image artifact constraints are directly controlled through weighted k-space. Specifically, image resolution constraints, image artifact constraints, image contrast constraints, and security constraints can each satisfy the following definitions:

[0070] 1) The image resolution constraint adopts the PSF resolution constraint, which is defined as follows: the full width of the point spread function at any height within the range of 1 / 3 to 2 / 3 of the peak height does not exceed a preset width value; more preferably, the image resolution constraint is defined as the full width of the point spread function at 1 / 2 of the peak height (i.e., the full width at half maximum (FWHM)) does not exceed a preset width value. This PSF resolution constraint can directly constrain the main lobe width of the reconstructed image, control the spatial resolution of the image, and ensure the image's detail resolution capability.

[0071] 2) Image artifact constraints employ PSF sidelobe constraints, defined as follows: the absolute size of the sidelobe of the point spread function or the relative size of the sidelobe to the main lobe does not exceed a preset size value, where the size type includes area or height; more preferably, the image artifact constraint is defined as the ratio of the area of ​​the first sidelobe of the point spread function to the area of ​​the main lobe not exceeding a preset ratio. This PSF sidelobe constraint can be used to control image artifacts and assist in controlling blur.

[0072] 3) The image contrast constraint is defined as follows: the relative difference in theoretical signal intensity of the two tissues of interest at their respective k-space centers is greater than a preset contrast value; more preferably, the image contrast constraint is set to the ratio obtained by dividing the difference in theoretical signal intensity of the two tissues of interest at their respective k-space centers by the sum of their respective theoretical signal intensity at their respective k-space centers by a value greater than the preset contrast value. It can be expressed by the formula as ,in and Let A and B be the theoretical signal intensities at the center of k-space for the two organizations of interest, respectively. This constraint ensures that the sum of the squares of all flip angles throughout the echo train is below a specific SAR safety limit, guaranteeing that the scan is harmless to humans.

[0073] 4) The safety constraint is defined as follows: the sum of the squares of the flip angles of all echo refocusing pulses does not exceed a preset safety threshold. This constraint is mainly applied in certain application scenarios where it is necessary to ensure the contrast between specific tissues. The k-space center signal can be constrained according to the tissue relaxation parameters (T1, T2) to ensure that the contrast of the target tissue meets clinical requirements.

[0074] However, it should be noted that the above-mentioned constraint definitions and types are only some preferred embodiments of the present invention and are not limitations on the present invention. The specific image quality constraints and security constraints selected can be adjusted according to actual optimization needs.

[0075] Furthermore, it should be noted that after constructing the constrained optimization problem, this invention can employ numerical optimization algorithms to solve this constrained optimization problem. Possible numerical optimization algorithms include, but are not limited to, trust domain algorithms, interior-point methods, differential evolution methods, and gradient descent methods. By solving this constrained nonlinear optimization problem using the above algorithms, globally or locally optimal flip angle sequences and weight factor sequences can be obtained. This method considers the influence of B1 field inhomogeneity in its design. After optimization, B1 field deviation simulations are performed on the flip angle sequence to evaluate its signal stability under ±15% deviation, ensuring the clinical applicability of the method. During magnetic resonance imaging (MRI) scanning, the optimized flip angle sequence is applied to a fast spin echo sequence. Radiofrequency excitation pulses are applied, followed by echo refocusing pulses sequentially according to the optimized flip angle sequence. In the signal processing stage, the optimized weight factor sequence can be used as a k-space filter window function to filter the acquired k-space data. An inverse Fourier transform is then performed on the filtered k-space data to obtain the final high-quality reconstructed image. The specific methods for MRI data acquisition, signal processing, and image reconstruction are further described in detail below.

[0076] In another embodiment of the present invention, a high signal-to-noise ratio magnetic resonance data acquisition method is further provided, the specific steps of which are as follows:

[0077] After pre-executing the aforementioned fast echo imaging parameter optimization method for magnetic resonance data acquisition, the optimal flip angle sequence corresponding to the fast echo sequence is obtained. and the optimal design signal strength sequence ;

[0078] The magnetic resonance imaging (MRI) device is controlled to run a fast echo sequence to scan the imaging target. During the operation, a radio frequency excitation pulse is applied first, followed by an optimized flip angle sequence. Echo refocusing pulses are applied sequentially and signals are acquired to fill the k-space lines in the k-space one by one. After the scan is completed, the k-space data corresponding to the imaging target is obtained.

[0079] It should be noted that this invention optimizes the optimal flip angle sequence in Fast Echo Sequence (FSE). Other sequence parameters can be set based on conventional empirical values ​​or parameters recommended by the equipment manufacturer. During the operation of the FSE sequence, the specific signal acquisition and k-space filling trajectory planning can follow conventional practices and are not restricted in this regard.

[0080] In another embodiment of the present invention, a magnetic resonance signal processing method is further provided, which specifically involves: after obtaining the k-space data corresponding to the imaging target according to the above-described high signal-to-noise ratio magnetic resonance data acquisition method, continuing to process the optimally designed signal intensity sequence obtained through optimization. The signals in all k-space lines are filtered sequentially. The original signal strength of the k-space line is multiplied by the signal strength sequence of the k-space line in the optimal design. The corresponding weighting factor, that is, the weighting factor of the collected first... The signal data on each k-space line is multiplied by the corresponding filter weight. This completes the filtering process. After weighted filtering of the entire k-space data, the filtered k-space data is saved and can then be used for subsequent magnetic resonance image reconstruction.

[0081] It should be noted that the reconstruction algorithm for magnetic resonance images based on k-space data can employ the traditional inverse Fourier transform method, as well as deep learning reconstruction and other emerging and cutting-edge methods; there are no restrictions on which method to use.

[0082] In the above embodiments of the present invention, a joint optimization model with PSF as the core indicator is proposed, abandoning the traditional fixed k-space window function and achieving an improvement in image signal-to-noise ratio. The present invention constructs a unified multi-objective optimization framework, which can selectively integrate multiple key requirements such as maximizing signal-to-noise ratio (objective function), accurately controlling resolution and artifacts (PSF constraint), ensuring scan safety (SAR constraint), and maintaining diagnostic contrast (contrast constraint) into a single mathematical problem for collaborative solution. Moreover, the present invention is a universal design method; the optimization process of this scheme is independent of a specific k-space trajectory and can be applied to various FSE sequence variants, including center sorting, sequential sorting, and even non-Cartesian sampling.

[0083] It should be noted that the fast echo imaging parameter optimization method, high signal-to-noise ratio magnetic resonance data acquisition method, and magnetic resonance signal processing method shown in the above embodiments can essentially be implemented in the form of computer programs or software functional modules running on computer electronic devices.

[0084] Therefore, based on the same inventive concept, the present invention provides a computer program product, including a computer program / instruction, which, when executed by a processor, can implement the aforementioned fast echo imaging parameter optimization method for magnetic resonance data acquisition, or can implement the aforementioned high signal-to-noise ratio magnetic resonance data acquisition method, or can implement the aforementioned magnetic resonance signal processing method.

[0085] Similarly, based on the same inventive concept, the present invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, can implement the aforementioned fast echo imaging parameter optimization method for magnetic resonance data acquisition, or the aforementioned high signal-to-noise ratio magnetic resonance data acquisition method, or the aforementioned magnetic resonance signal processing method.

[0086] Similarly, based on the same inventive concept, the present invention provides a computer electronic device, which includes a memory and a processor;

[0087] The memory is used to store computer programs;

[0088] The processor is configured to, when executing the computer program, implement the aforementioned fast echo imaging parameter optimization method for magnetic resonance data acquisition, or implement the aforementioned high signal-to-noise ratio magnetic resonance data acquisition method, or implement the aforementioned magnetic resonance signal processing method.

[0089] It is understood that the aforementioned storage media and memory may include random access memory (RAM) or non-volatile memory (NVM), such as at least one disk storage device. Furthermore, the storage media may also be various media capable of storing program code, such as USB flash drives, external hard drives, magnetic disks, or optical discs.

[0090] It is understood that the processors mentioned above can be general-purpose processors, including central processing units (CPUs), network processors (NPs), etc.; they can also be digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0091] It is understood that the aforementioned computer electronic devices can take the form of personal computers, local servers, network devices, or cloud servers.

[0092] It should also be noted that those skilled in the art will understand that, for the sake of convenience and brevity, the specific working process of the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here. In the embodiments provided in this application, the division of steps or modules in the system and method is merely a logical functional division, and there may be other division methods in actual implementation. For example, multiple modules or steps may be combined or integrated together, and a module or step may also be split.

[0093] The aforementioned high signal-to-noise ratio magnetic resonance data acquisition method, in addition to relying on electronic devices capable of executing computer programs, also requires the integration of magnetic resonance imaging equipment.

[0094] Therefore, in another embodiment of the present invention, a magnetic resonance imaging system is provided, which includes a magnetic resonance scanner and a control unit, wherein the control unit stores a computer program, which, when executed, controls the magnetic resonance scanner to implement the aforementioned high signal-to-noise ratio magnetic resonance data acquisition method.

[0095] It should be noted that a magnetic resonance scanner can be any device capable of magnetic resonance imaging. Its structure is existing technology and mature commercial products can be used, with no limit on the specific model.

[0096] In addition, as a preferred embodiment of the present invention, the above-mentioned magnetic resonance imaging system further includes a signal processing device, which processes the signal intensity sequence according to the optimized design. The signals in all k-space lines are filtered sequentially. The original signal strength of the k-space line is multiplied by the signal strength sequence of the k-space line in the optimal design. The corresponding weighting factors are used to complete the filtering, and then the filtered k-space data is subjected to inverse Fourier transform to reconstruct a high signal-to-noise ratio magnetic resonance image.

[0097] The aforementioned control unit and signal processing device can be implemented using a host computer or other computer electronic equipment capable of data processing and outputting control commands. The aforementioned control unit and signal processing device can be external devices independent of the control unit inherent in the magnetic resonance scanner itself. However, if the control unit inherent in the magnetic resonance scanner can meet the requirements, it can also be directly integrated and coupled into the control unit inherent in the magnetic resonance scanner, with the same control unit performing the processing of various types of data and communication control between magnetic resonance scanners; there are no restrictions on this.

[0098] To enable those skilled in the art to better understand the specific implementation and technical effects of the present invention, the following two embodiments demonstrate the application of the present invention in practical scenarios.

[0099] Example 1: T1-weighted, T2-weighted, and proton density-weighted 2D brain imaging

[0100] In this embodiment, taking a 2D brain imaging scenario as an example, the specific implementation process of the above-mentioned fast echo imaging parameter optimization method, high signal-to-noise ratio magnetic resonance data acquisition method, and magnetic resonance signal processing method for magnetic resonance data acquisition is demonstrated.

[0101] Step 1: Determine imaging parameters and constraints

[0102] In this embodiment, the imaging parameters are set as follows: field of view (FOV) = 240mm, resolution 0.9 × 0.9 mm. 2The phase encoding number (PE) is 256. A flip angle optimization method based on a fixed Fermi window is used as the baseline design method. The PSF function of the fixed Fermi window used in this method is measured, yielding a main lobe full width at half maximum (FWHM) of 2.57 and an FWHM area ratio of approximately 0.6. Gray matter is used as the target signal, with T1 = 1600 ms and T2 = 100 ms. Simulations are performed based on these tissue parameters. The FSE sequence used for imaging is set with ETL = 32 and ESP = 12 ms. TR = 5000 ms is set to ensure sufficient relaxation recovery time. Contrast is prepared using a preparatory pulse before excitation, and all contrast measurements are acquired using a central view with TE = 12 ms.

[0103] In addition, in this embodiment, two experimental groups were set up based on whether the sidelobe area ratio was strictly limited: an FFA-PSFs group with strictly limited sidelobe area ratio and an FFA-PSF group without strictly limited sidelobe area ratio. To ensure imaging quality and safety, the following image quality constraints 1) and 2), and safety constraint 3) were set during the optimization process.

[0104] Constraint 1): The full width at half maximum (FWHM) of the optimized PSF is no greater than the full FWHM of the fixed Fermi window, which is 2.57. The constraint formula is expressed as follows: ;

[0105] Constraint 2): The sidelobe area ratio of the PSF must meet the corresponding scheme restrictions. For the FFA-PSFs group with strict restrictions on the sidelobe area ratio, the half-width ratio should not exceed 0.6. For the FFA-PSF group with relaxed restrictions, the half-width ratio should not exceed 0.62. The constraint formula is expressed as follows: :

[0106] Constraint 3): The RF power deposition shall not exceed the preset SAR limit, that is, the sum of the squares of the flip angles in the echo chain shall not exceed half of the SAR value in the subsequent CFA scheme used for comparison. The constraint formula is expressed as follows: , This represents the flip angle of the i-th echo refocusing pulse in the CFA scheme.

[0107] Step 2: Establishment of the joint optimization mathematical model

[0108] Based on the above imaging parameter settings and constraints, the nonlinear joint constraint optimization problem for brain imaging can be summarized as follows:

[0109]

[0110]

[0111] In the formula: The design signal strength representing the center of the k-space is acquired using center sampling in this embodiment. This is the design signal strength of the first echo. The number of space lines N is equal to the length of the flip angle sequence; in this embodiment, N=32. To avoid imperfect refocusing pulses failing to reach 180 degrees and affecting imaging performance, the flip angle sequence is limited in the optimization problem. Each flip angle must be less than 150 degrees. During the optimization process, According to the flip angle sequence The Extended Phase Map (EPG) algorithm is used for calculation. The EPG algorithm model can obtain the signal amplitude corresponding to each echo under a given flip angle sequence.

[0112] The specific optimization process can be performed using the `minimize` function in the `scipy` package. Specifically, the trust region algorithm is used to solve this constrained nonlinear optimization problem. The optimization variable is a sequence of flip angles of length 32. and k-space filter window function (i.e., design signal strength sequence) The initial values ​​are set to the fixed Fermi window function used by the CFA and FFA-Fermi methods. After finally optimizing and solving to obtain the optimal solution, the optimal flip angle sequence corresponding to the fast echo sequence can be obtained. and the optimal design signal strength sequence .

[0113] Step 3: Sequence Execution, Signal Filtering, and Image Reconstruction

[0114] In the Siemens Prisma 3T sequence development environment, a 2D FSE sequence was programmed, with its echo train length set to 32 and the flip angle of the 32 refocusing RF pulses set to the optimized value. The scan was performed on the magnetic resonance scanner according to the preset parameters (FOV 240mm, PE=256, TR / TE=5000 / 12ms), and the optimal flip angle sequence was obtained. Radio frequency echo pulses are applied sequentially, and k-space data is acquired along one k-space line at each echo time. The contrast requirements for T1w and T2w are met by configuring different preparation pulses. To calculate the signal-to-noise ratio, data is acquired twice and averaged. The optimized k-space filter is then used. The acquired k-space data is weighted and filtered by multiplying the signal on the i-th k-space line. In Weighted filtering is performed on all 32 k-space lines to obtain filtered k-space data. An inverse Fourier transform is then applied to the filtered k-space data to reconstruct a brain image with the appropriate contrast.

[0115] To verify the technical effectiveness of this invention, multiple optimization designs and scanning schemes were implemented on five subjects (two women) using the same sequence parameters on the same scanner. In addition to the two schemes in this embodiment—the joint optimization design based on PSF constraints (i.e., the aforementioned FFA-PSFs) and the joint optimization design based on PSF constraints but relaxing the PSF sidelobe area ratio limit (i.e., the aforementioned FFA-PSF)—the following three comparative schemes were also implemented in this embodiment:

[0116] Comparison Scheme 1 (denoted as CFA): A constant flip angle design was adopted. In T1-weighted, T2-weighted, and proton density-weighted 2D brain imaging, the CFA scheme was designed with an initial flip angle of 135 degrees, followed by a rapid decrease in the flip angle and a maintenance of 95 degrees.

[0117] Comparison Scheme 2 (denoted as Direct): This scheme uses a flip angle design that directly generates the Fermi window shape signal. In this embodiment, the Fermi window is defined as follows:

[0118]

[0119] in, It is the echo ordinal number. It is the echo train length, in this embodiment The value is 32, and other parameters include... =0.9, =0.046.

[0120] Comparative scheme 3 (denoted as FFA-Fermi): adopts an advanced design based on fixed Fermi window optimization in the existing technology (for details, see Zhao L, Chang CD, Alsop D C. Controlling T2 blurring in 3DRARE arterial spin labeling acquisition through optimal combination of variable flip angles and k‐space filtering[J]. Magnetic resonance inmedicine, 2018, 80(4): 1391-1401.), and the Fermi window used is the same as that used in comparative scheme 2.

[0121] The above three comparison schemes, as well as the flip angle design results and PSF function under the FFA-PSFs and FFA-PSF schemes in this embodiment, are as follows: Figure 1 As shown. Figure 2 Images reconstructed using different schemes are shown. Figure 2 The SNR was calculated from the reconstructed volunteer brain images shown, and the results are as follows: Figure 3 As shown in the results, the FFA-PSF scheme in this embodiment provides a significantly higher signal-to-noise ratio than all the comparative schemes, more than twice that of the CFA scheme, which is in high agreement with the theoretical prediction, fully demonstrating the superior performance of the optimization framework of this invention in improving the signal-to-noise ratio. Figure 2 As can be seen from the reconstructed images shown, the background noise is highest in Comparison Scheme 1 (CFA). Figure 2 As shown by the middle arrow, the image noise is reduced compared to schemes 2 and 3, but the image reconstructed by the FFA-PSF scheme in this embodiment performs best in terms of overall image cleanliness, providing the best SNR index while effectively suppressing T2 blur.

[0122] Furthermore, to demonstrate the robustness of this invention to B1 field inhomogeneity, a saturated double-angle method (see Cunningham CH, Pauly JM, Nayak K S. Saturated double-angle method for rapid B1+ mapping[J]. Magnetic Resonance in Medicine: An Official Journal of the International Society for Magnetic Resonance in Medicine, 2006, 55(6): 1326-1333.) was used to acquire a B1 map. The deviations between the acquired signal strength, the final signal obtained according to the original filtering weights, and the designed acquired and filtered signals were simulated under conditions of a 15% increase and decrease in the B1 field value. The results are as follows: Figure 4 As shown in Table 1, the results indicate that the Direct method is more sensitive to B1 inhomogeneity, and the robustness of this invention to B1 field inhomogeneity is comparable to that of the FFA-Fermi method.

[0123] Table 1. Simulation results of robust performance against head B1 field inhomogeneity

[0124]

[0125] Example 2: 2D Abdominal Imaging

[0126] In this embodiment, taking a 2D abdominal imaging scenario as an example, the specific implementation process of the above-mentioned fast echo imaging parameter optimization method, high signal-to-noise ratio magnetic resonance data acquisition method, and magnetic resonance signal processing method for magnetic resonance data acquisition is demonstrated.

[0127] Step 1: Determine imaging parameters and constraints

[0128] In this embodiment, the imaging parameters are set as follows: field of view (FOV) = 252mm, resolution 1.6×1.6mm. 2 The phase encoding number (PE) is 160. A fixed Fermi window-based flip angle optimization method is used as the baseline design method. The fixed Fermi window function used in this method is measured, and the full width at half maximum (FWHM) is 3.54, with an area ratio of approximately 0.621. Liver tissue is used as the target signal for optimization, with T1 = 809 ms and T2 = 34 ms. Simulations are performed based on these tissue parameters. In the FSE sequence used for imaging, ETL = 90 and ESP = 6 ms are set. Simultaneously, TE = 60 ms and TR = 3000 ms are set to ensure sufficient relaxation recovery time for the tissue signal and to avoid prolonged breath-holding by the subject. To shorten the scan time, this embodiment uses 9 / 16 partial Fourier acquisition, where phase estimation is performed by the central 32 lines. Each excitation acquires complete single-image k-space data using a sequential view sampling method.

[0129] In addition, in this embodiment, two experimental groups were set up based on whether the sidelobe area ratio was strictly limited: an FFA-PSFs group with strictly limited sidelobe area ratio and an FFA-PSF group without strictly limited sidelobe area ratio. To ensure imaging quality and safety, the following image quality constraints 1), 2), and 3), and safety constraint 4) were set during the optimization process.

[0130] Constraint 1): The full width at half maximum (FWHM) of the optimized PSF is no greater than the full FWHM of the fixed Fermi window (3.54). The constraint formula is expressed as follows: ;

[0131] Constraint 2): For the FFA-PSFs group with strict restrictions on sidelobe area ratio, since the main lobe shape of the optimized Fermi window PSF function is relatively wide, the half-width ratio is set to no more than 0.65 to make the sidelobe area ratios approximately uniform. For the FFA-PSF group with relaxed restrictions, the half-width ratio is set to no more than 0.696. The constraint formula is expressed as follows: ;

[0132] Constraint 3): An additional liver-spleen contrast constraint is introduced, preserving soft tissue contrast that meets diagnostic requirements (liver and spleen contrast > 0.3). Spleen signal is calculated at T1 = 1328 ms and T2 = 61 ms. The constraint formula is expressed as follows: ,in and These represent the theoretical signal intensities of the liver and spleen at their respective k-space centers.

[0133] Constraint 4): The RF power deposition shall not exceed the preset SAR limit, that is, the sum of the squares of the flip angles in the echo chain shall not exceed half of the SAR value in the subsequent CFA scheme used for comparison. The constraint formula is expressed as follows: , This represents the flip angle of the i-th echo refocusing pulse in the CFA scheme.

[0134] Step 2: Establishment of the joint optimization mathematical model

[0135] Based on the above imaging parameter settings and constraints, the following nonlinear joint constraint optimization problem for abdominal imaging can be constructed:

[0136]

[0137]

[0138] In the formula: The design signal strength representing the center of the k-space is obtained using partial Fourier transform data acquisition in this embodiment. The design signal strength is for the 10th echo. The number of space lines N is equal to the length of the flip angle sequence; in this embodiment, N=90. To ensure sequence stability and avoid the impact of imperfect refocusing pulses failing to reach 180 degrees, the flip angle sequence is limited in the optimization problem. Each flip angle in the image needs to be less than 165 degrees, while limiting the flip angle to no less than 30 degrees to reduce the image's sensitivity to motion. During the optimization process, According to the flip angle sequence The Extended Phase Map (EPG) algorithm is used for calculation. The theoretical signal amplitude corresponding to each echo under a given flip angle sequence can be obtained through the EPG algorithm model.

[0139] The specific optimization process can be performed using the `minimize` function in the `scipy` package. Specifically, the trust region algorithm is used to solve this constrained nonlinear optimization problem. The optimization variable is a sequence of flip angles of length 90. and the corresponding k-space filter window function (i.e., the design signal strength sequence) The initial values ​​are set to the fixed Fermi window function used by the CFA and FFA-Fermi methods. After finally optimizing and solving to obtain the optimal solution, the optimal flip angle sequence corresponding to the fast echo sequence can be obtained. and the optimal design signal strength sequence .

[0140] Step 3: Sequence Execution, Signal Filtering, and Image Reconstruction

[0141] In the Siemens Prisma 3T sequence development environment, a 2D FSE sequence was programmed, with its echo train length set to 90 and the flip angle of the 90 refocusing RF pulses set to the optimized value. The scan was performed on the magnetic resonance scanner according to the preset parameters (FOV 252mm, PE=160, TR / TE=3000 / 60ms), and the optimal flip angle sequence was optimized during the scan. Radio frequency echo pulses are applied sequentially, and k-space data is acquired along one k-space line at each echo time. To calculate the signal-to-noise ratio, two acquisitions are performed within the same breath-hold and the average is taken. The optimized k-space filter is then used. The acquired k-space data is weighted and filtered by multiplying the signal on the i-th k-space line. In Weighted filtering was performed on all 90 k-space lines to obtain filtered k-space data. An inverse Fourier transform was then performed on the filtered k-space data to reconstruct the abdominal image.

[0142] To verify the technical effectiveness of this invention, multiple optimization designs and scanning schemes were implemented on five subjects (two women) using the same sequence parameters on the same scanner. In addition to the two schemes in this embodiment—the joint optimization design based on PSF constraints (i.e., the aforementioned FFA-PSFs) and the joint optimization design based on PSF constraints but relaxing the PSF sidelobe area ratio limit (i.e., the aforementioned FFA-PSF)—the following three comparative schemes were also implemented in this embodiment:

[0143] Comparison Scheme 1 (denoted as CFA): adopts a constant flip angle design. In 2D abdominal imaging, the CFA scheme is designed with an initial flip angle of 155 degrees and subsequent flip angles of 130 degrees.

[0144] Comparison Scheme 2 (denoted as Direct): This scheme uses a flip angle design that directly generates the Fermi window shape signal. In this embodiment, the Fermi window is defined as follows:

[0145]

[0146] in, It is the echo ordinal number. It is the echo train length, in this embodiment It is 90, and other parameters include =0.49, =0.142.

[0147] Comparative Scheme 3 (denoted as FFA-Fermi): It adopts an advanced design based on fixed Fermi window optimization in the prior art (see the FFA-Fermi scheme in Example 1 for details). The Fermi window used here is the same as the Fermi window used in Comparative Scheme 2.

[0148] The above three comparison schemes, as well as the flip angle design results and PSF function under the FFA-PSFs and FFA-PSF schemes in this embodiment, are as follows: Figure 5 As shown. Figure 6 The images show reconstructed using different schemes, derived from reconstructed images of the volunteer's abdomen. Figure 6 The SNR was calculated, and the CFA scheme showed a significant increase in noise level after filtering, so it was not included in the quantitative comparison of signal-to-noise ratio. The results for other schemes are as follows: Figure 7 As shown, the signal-to-noise ratio provided by the FFA-PSF scheme in this embodiment is significantly higher than that of all the comparative schemes, reaching more than twice that of the Direct scheme, which is in high agreement with the theoretical prediction, fully demonstrating the excellent performance of the optimization framework of this invention in improving the signal-to-noise ratio. Figure 6 The reconstructed images show that, in contrast scheme 2 (Direct), the background noise is the highest (e.g., Figure 2 (As indicated by the middle arrow). Compared with Scheme 3, the image noise is reduced, but the FFA-PSFs and FFA-PSF schemes in this embodiment show the best overall image cleanliness, effectively suppressing T2 blur while providing the best SNR index.

[0149] Similarly, to verify the robustness of this invention to B1 field inhomogeneity, the B1 map was acquired using the saturated dual-angle method, and the deviations of the acquired signal strength, the final signal obtained according to the original filtering weights, and the designed acquired and filtered signals were simulated under conditions of a 15% increase and decrease in the B1 field value. The results are as follows: Figure 8 As shown in Table 2, the results indicate that the robustness of this invention to B1 field inhomogeneities is comparable to the FFA-Fermi method and significantly superior to the Direct method.

[0150] Table 2. Simulation results of robust performance of abdominal B1 field inhomogeneity

[0151]

[0152] The embodiments described above are merely some preferred implementations of the present invention and are not intended to limit the invention. Those skilled in the art can make various changes and modifications without departing from the spirit and scope of the invention. Therefore, all technical solutions obtained through equivalent substitution or transformation fall within the protection scope of the present invention.

Claims

1. A method for optimizing fast echo imaging parameters for magnetic resonance data acquisition, characterized in that: When the fast echo sequence to be optimized is executed by the magnetic resonance device, after applying the echo refocusing pulse sequence with a flip angle, the signal needs to be acquired and the k-space lines in the k-space need to be filled one by one; the signal filled in each k-space line corresponds to a weighting factor for signal filtering, which is the ratio of the designed signal strength to the theoretical signal strength of the k-space line; The optimization method includes: Using the flip angle sequence and the design signal intensity sequence corresponding to the echo refocusing pulse sequence in the fast echo sequence as the parameters to be optimized, and taking the ratio between the design signal intensity at the center of k space and the sum of squares of the weight factors corresponding to all k space lines as the optimization objective, a numerical optimization algorithm is used to solve for the parameters to be optimized under image quality constraints and security constraints, so as to obtain the optimal flip angle sequence and the optimal design signal intensity sequence corresponding to the fast echo sequence.

2. The method for optimizing fast echo imaging parameters for magnetic resonance data acquisition as described in claim 1, characterized in that: The image quality constraints include one or more of image resolution constraints, image artifact constraints, and image contrast constraints; the image resolution constraint is that the spatial resolution of the magnetic resonance image reconstructed based on k-space data meets a preset resolution requirement; the image artifact constraint is that the artifact evaluation index of the magnetic resonance image reconstructed based on k-space data meets a preset artifact control requirement; and the image contrast constraint is that the contrast between tissues of interest in the magnetic resonance image reconstructed based on k-space data meets a preset contrast requirement. The security constraint is that the energy deposition per unit time does not exceed the energy threshold.

3. The method for optimizing fast echo imaging parameters for magnetic resonance data acquisition as described in claim 1, characterized in that: The inverse Fourier transform of the designed signal intensity sequence is used as the point spread function. One or more of the image resolution constraint, image artifact constraint, image contrast constraint, and security constraint satisfy the following definition: The image resolution constraint is defined as follows: the full width of the point spread function at any height within the range of 1 / 3 to 2 / 3 of the peak height does not exceed a preset width value; more preferably, the image resolution constraint is defined as the full width of the point spread function at 1 / 2 of the peak height does not exceed a preset width value; The image artifact constraint is defined as follows: the absolute size of the side lobe of the point spread function or the relative size of the side lobe to the main lobe does not exceed a preset size value, and the size type includes area or height; more preferably, the image artifact constraint is defined as the ratio of the area of ​​the first side lobe of the point spread function to the area of ​​the main lobe does not exceed a preset ratio. The image contrast constraint is defined as follows: the relative difference in theoretical signal intensity of the two tissues of interest at the center of k space is greater than a preset contrast value; more preferably, the image contrast constraint is set to the ratio obtained by dividing the difference in theoretical signal intensity of the two tissues of interest at the center of k space by the sum of the theoretical signal intensity of the two tissues of interest at the center of k space by a value greater than the preset contrast value. The safety constraint is defined as follows: the sum of the squares of the flip angles of all echo refocusing pulses does not exceed a preset safety threshold.

4. The method for optimizing fast echo imaging parameters for magnetic resonance data acquisition as described in claim 1, characterized in that: When performing the solution using a numerical optimization algorithm, the optimization objective is transformed into its equivalent form, which preferably transforms the maximization objective function into the minimization objective function.

5. A high signal-to-noise ratio magnetic resonance data acquisition method, characterized in that: After performing the fast echo imaging parameter optimization method for magnetic resonance data acquisition as described in claim 1 in advance, the optimal flip angle sequence and the optimal design signal intensity sequence corresponding to the fast echo sequence are obtained. The magnetic resonance imaging equipment is controlled to run a fast echo sequence to scan the imaging target. During the operation, a radio frequency excitation pulse is first applied, and then echo refocusing pulses are applied sequentially according to the optimized flip angle sequence. The signals are collected and filled into the k-space lines in the k-space one by one. After the scan is completed, the k-space data corresponding to the imaging target is obtained.

6. A magnetic resonance signal processing method, characterized in that, After obtaining the k-space data corresponding to the imaging target according to the high signal-to-noise ratio magnetic resonance data acquisition method of claim 5, the signals in all k-space lines are filtered sequentially according to the optimized weight factor obtained by optimization. The original filling signal intensity in each k-space line needs to be multiplied by the weight factor corresponding to the k-space line in the optimal design signal intensity sequence to complete the filtering. The filtered k-space data is saved for subsequent magnetic resonance image reconstruction and analysis.

7. A computer program product comprising a computer program / instructions, characterized in that, When the computer program / instruction is executed by the processor, it can implement the fast echo imaging parameter optimization method for magnetic resonance data acquisition as described in any one of claims 1 to 4, or the high signal-to-noise ratio magnetic resonance data acquisition method as described in claim 5, or the magnetic resonance signal processing method as described in claim 6.

8. A computer electronic device, characterized in that, Including memory and processor; The memory is used to store computer programs; The processor is configured to, when executing the computer program, implement the fast echo imaging parameter optimization method for magnetic resonance data acquisition as described in any one of claims 1 to 4, or implement the high signal-to-noise ratio magnetic resonance data acquisition method as described in claim 5, or implement the magnetic resonance signal processing method as described in claim 6.

9. A magnetic resonance imaging system, characterized in that, It includes a magnetic resonance scanner and a control unit, wherein the control unit stores a computer program, which, when executed, controls the magnetic resonance scanner to implement the high signal-to-noise ratio magnetic resonance data acquisition method as described in claim 5.

10. The magnetic resonance imaging system as described in claim 9, characterized in that, It also includes a signal processing device, which filters the signals in all k-space lines according to the optimized design signal intensity sequence. The original filling signal intensity in each k-space line needs to be multiplied by the weight factor corresponding to that k-space line in the optimal design signal intensity sequence to complete the filtering. Then, the filtered k-space data is reconstructed to obtain a high signal-to-noise ratio magnetic resonance image.

Images