Magnetic resonance radio frequency shimming optimization method and device based on spatial weight and adaptive reweighting

Abstract

The invention discloses a magnetic resonance radio frequency shimming optimization method and device based on spatial weight and adaptive reweighting. The magnetic resonance radio frequency shimming optimization method based on spatial weight and adaptive reweighting comprises the following steps: acquiring excitation field distribution to be optimized; obtaining a target area excitation magnetic field matrix, a peripheral area excitation magnetic field matrix and a target function according to the to-be-optimized excitation field distribution; according to the target area excitation magnetic field matrix, the peripheral area excitation magnetic field matrix and a target function, through an iteration adaptive reweighting mechanism, dynamically updating a peripheral weight so as to generate a final channel complex weight; and transmitting the final channel complex number weight to a magnetic resonance emission system, thereby enabling the magnetic resonance emission system to realize radio frequency shimming according to the final channel complex number weight. Through the generated final channel complex weight, the peripheral response can be balanced while the performance of the target region is ensured, and the uniformity control of the B1 + field of each region is realized.

Description

Technical Field

[0001] This application relates to the field of radio frequency shimming optimization technology, and in particular to a magnetic resonance radio frequency shimming optimization method and apparatus based on spatial weighting and adaptive reweighting. Background Technology

[0002] Under high-field and ultra-high-field magnetic resonance imaging (MRI) conditions, the radiofrequency wavelength shortens significantly with increasing magnetic field strength, while the dielectric resonance effect is enhanced, leading to a more complex distribution of the electromagnetic field within the body. For imaging regions of deep targets such as the heart and liver, the excitation magnetic field B1... ﹢ The field often exhibits significant inhomogeneity, leading to fluctuations in image signal intensity and a decrease in local contrast, affecting the accuracy of quantitative measurements. Simultaneously, peripheral tissues (such as subcutaneous fat and chest wall) are prone to forming local high electric field regions, increasing the risk of local specific absorption rate SAR hotspots and posing a potential threat to imaging security.

[0003] Existing multi-channel parallel transmission RF shimming techniques can improve B1 in the target area to some extent. ﹢ While high-field MRI achieves good field uniformity, its optimization strategies often employ uniform weighting or single-target-region weighting, lacking flexible allocation and dynamic adjustment based on the importance of different anatomical regions. This often results in situations where, although good uniformity can be achieved within the ROI, the field intensity in the peripheral regions is either too high or too low. This can lead to excessive energy concentration in the peripheral regions, posing safety risks, or it can cause a decrease in overall image uniformity, thus limiting the widespread application of ultra-high-field MRI in clinical and research settings.

[0004] (1) The existing patent CN104515963A, “Magnetic Resonance Radio Frequency Shimming System”, uses multiple independent excitation sources and a selectable multi-channel / multi-source architecture to improve the degree of freedom of shimming control (more hardware / system level) to solve this problem. However, it focuses on system topology and channel number expansion; it is not a voxel-level spatial weighting or iterative reweighting optimization framework, and the cost is high due to hardware shimming.

[0005] (2) The prior art patent with patent number CN111712719A, "Active B1 of transmitting coil" ﹢ "Homing" proposes introducing an active B1 outside the main transmitting coil. ﹢ The shimming coil is supplied with radio frequency power during transmission to improve the imaging region B1. ﹢ Uniformity. This approach leans towards active shimming design at the hardware level. This solution requires an additional active B1 coil outside the main transmitting coil. ﹢ The shimming coil is powered and controlled separately during imaging. This increases the structural complexity, cost, and maintenance difficulty of the MRI system, and places high demands on the modification of existing equipment.

[0006] (3) The existing technology patent CN113219389A, entitled "Method, Device and Readable Storage Medium for Determining Magnetic Resonance Radio Frequency Mode", is mainly based on single-channel sensitivity pre-acquisition, constructing multiple emission modes and exciting them sequentially / alternately to alleviate the high field B1. ﹢ Non-uniformity and reduced SAR fall under the category of transmission mode-level RF shimming strategies. This patent constructs multiple transmission modes through single-channel sensitivity and relies on alternating excitation to improve uniformity. However, the optimized target does not differentiate between ROI and non-ROI, failing to simultaneously ensure the uniformity of deep targets and the security of the surrounding areas.

[0007] Therefore, it is desirable to have a technical solution to overcome or at least mitigate one of the aforementioned defects of the prior art. Summary of the Invention

[0008] The purpose of this application is to provide a magnetic resonance radio frequency shimming optimization method based on spatial weights and adaptive reweighting to overcome or at least mitigate one of the above-mentioned defects of the prior art.

[0009] To achieve the above objectives, this application provides a magnetic resonance radio frequency shimming optimization method based on spatial weights and adaptive reweighting, wherein the magnetic resonance radio frequency shimming optimization method based on spatial weights and adaptive reweighting includes: Obtain the excitation field distribution to be optimized; The excitation magnetic field matrix of the target region, the excitation magnetic field matrix of the outer region, and the objective function are obtained based on the excitation field distribution to be optimized. Based on the excitation magnetic field matrix of the target region, the excitation magnetic field matrix of the peripheral region, and the objective function, the peripheral weights are dynamically updated through an iterative adaptive reweighting mechanism, thereby generating the final channel complex weights. The final channel complex weights are passed to the magnetic resonance emission system, thereby enabling the magnetic resonance emission system to achieve radio frequency shimming based on the final channel complex weights.

[0010] Optionally, obtaining the excitation field distribution to be optimized includes: B1 is obtained through experimental data collection or simulation calculation. ﹢ Field response data; According to B1 ﹢ Field response data acquisition multi-channel B1 ﹢ Response matrix of field response data Where N is the total number of pixels or voxels. Let C be the number of channels, and C be the complex field.

[0011] Optionally, obtaining the target mask, outer peripheral mask, target region excitation magnetic field matrix, and remaining region excitation magnetic field matrix based on the excitation field distribution to be optimized includes: Based on anatomical structure, voxels are divided into core target regions and peripheral regions, and target masks are defined. Peripheral mask Count the number of pixels , ; From the response matrix Extract the excitation magnetic field matrix of the target region. , Remaining region excitation magnetic field matrix Set the ideal excitation magnetic field matrix for the target region. Ideal excitation magnetic field matrix in the outer peripheral region ; Based on the target region's excitation magnetic field matrix Calculate the uniformity weight of the target region ; Based on the target region's excitation magnetic field matrix , Remaining region excitation magnetic field matrix Calculate the remaining area control weights .

[0012] Optionally, the objective function formula is as follows: ;in, To generate a magnetic field matrix for the target region, To generate a magnetic field matrix for the remaining region; To generate an ideal magnetic field matrix for the target region, The ideal excitation magnetic field matrix for the outer peripheral region; Weights for uniformity in the target region; The remaining region is controlled by weights.

[0013] Optionally, the step of dynamically updating the peripheral weights through an iterative adaptive reweighting mechanism to generate the final channel complex weights includes: Set initial iteration parameters: Set the iteration number k=0, and the initial peripheral weights... = ; Repeat the following iterations until the iteration condition is met: According to the response matrix Adaptively update the peripheral weights to obtain the (k+1)th peripheral weight. ; Based on the (k+1)th peripheral weight Adjust the objective function and solve for the complex weights; wherein the iteration condition is: When the change in the energy ratio of the complex weights in two consecutive iterations is lower than the threshold, the optimization iteration stops, and the last calculated complex weight is output as the final channel complex weight.

[0014] Optionally, the step of basing the response matrix Adaptively update the peripheral weights to obtain the (k+1)th peripheral weight. include: Calculate the current region field strength norm: use the channel excitation vector from the k-th iteration. Substitute into the matrix to calculate the field strength norm of the target region. , peripheral regional field strength norm ; Dynamically update peripheral weights: according to the iterative formula Perform weight updates to obtain the (k+1)th peripheral weight. .

[0015] Optionally, the step of basing the weights on the (k+1)th peripheral perimeter is... Adjusting the objective function and solving for the complex weights includes: Use the updated Adjust the objective function ;

[0016] An amplitude-phase separation strategy is employed for optimization: In the k-th round of the outer loop, the amplitude-phase separation strategy is first calculated. Extract its phase This iteration is fixed Only the amplitude coefficient is optimized. At this point, the complex weight form is ; Solve the optimized solution using nonlinear least squares or conjugate gradient methods. We obtain the complex weights after the kth round of optimization.

[0017] This application also provides a magnetic resonance radio frequency shimming optimization device based on spatial weights and adaptive reweighting, the magnetic resonance radio frequency shimming optimization device based on spatial weights and adaptive reweighting includes: An excitation field distribution acquisition module is used to acquire the excitation field distribution to be optimized. The iterative optimization data acquisition module is used to acquire the excitation magnetic field matrix of the target region, the excitation magnetic field matrix of the outer region, and the objective function based on the excitation field distribution to be optimized. The final channel complex weight generation module is used to dynamically update the peripheral weights based on the target region excitation magnetic field matrix, the peripheral region excitation magnetic field matrix, and the objective function through an iterative adaptive reweighting mechanism, thereby generating the final channel complex weights. The transmitting module is used to transmit the final channel complex weights to the magnetic resonance transmitting system, so that the magnetic resonance transmitting system can achieve radio frequency shimming according to the final channel complex weights.

[0018] The final channel complex weights generated by the magnetic resonance radio frequency shimming optimization method based on spatial weighting and adaptive reweighting in this application can balance the peripheral response while ensuring the performance of the target region, achieving B1 in each region. ﹢ The field uniformity is controlled, and the overall transmission efficiency is guaranteed, achieving high robustness, high efficiency, and safe controllability of multi-channel RF field shimming optimization. Attached Figure Description

[0019] Figure 1 This is a flowchart illustrating a magnetic resonance radio frequency shimming optimization method based on spatial weights and adaptive reweighting according to an embodiment of this application.

[0020] Figure 2 This is a schematic diagram of a 5T ultra-high field MRI body emission coil model.

[0021] Figure 3 This is a schematic diagram of the magnetic field distribution for different radio frequency shimming methods. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions in the embodiments of this application will be described in more detail below with reference to the accompanying drawings. In the drawings, the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The described embodiments are some, but not all, embodiments of this application. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this application, and should not be construed as limiting this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application. The embodiments of this application will be described in detail below with reference to the accompanying drawings.

[0023] In the description of this application, it should be understood that the terms "center", "longitudinal", "lateral", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limiting the scope of protection of this application. Example 1:

[0024] like Figure 1The magnetic resonance radio frequency shimming optimization method based on spatial weights and adaptive reweighting, as shown, includes: Step 1: Obtain the excitation field distribution to be optimized; Step 2: Obtain the excitation magnetic field matrix of the target region, the excitation magnetic field matrix of the outer region, and the objective function based on the excitation field distribution to be optimized; Step 3: Based on the excitation magnetic field matrix of the target region, the excitation magnetic field matrix of the peripheral region, and the objective function, the peripheral weights are dynamically updated through an iterative adaptive reweighting mechanism to generate the final channel complex weights. Step 4: The final channel complex weights are passed to the magnetic resonance emission system, so that the magnetic resonance emission system can achieve radio frequency shimming according to the final channel complex weights.

[0025] In this embodiment, obtaining the excitation field distribution to be optimized includes: B1 is obtained through experimental data collection or simulation calculation. ﹢ Field response data; According to B1 ﹢ Field response data acquisition multi-channel B1 ﹢ Response matrix of field response data Where N is the total number of pixels or voxels. Let C be the number of channels, and C be the complex field.

[0026] In this embodiment, obtaining the target mask, peripheral mask, target region excitation magnetic field matrix, and remaining region excitation magnetic field matrix based on the excitation field distribution to be optimized includes: Based on anatomical structure, voxels are divided into core target regions and peripheral regions, and target masks are defined. Peripheral mask Count the number of pixels , ; From the response matrix Extract the excitation magnetic field matrix of the target region. , Remaining region excitation magnetic field matrix Set the ideal excitation magnetic field matrix for the target region. Ideal excitation magnetic field matrix in the outer peripheral region ; Based on the target region's excitation magnetic field matrix Calculate the uniformity weight of the target region ; Based on the target region's excitation magnetic field matrix , Remaining region excitation magnetic field matrix Calculate the remaining area control weights .

[0027] In this embodiment, the objective function formula is as follows: ;in, To generate a magnetic field matrix for the target region, To generate a magnetic field matrix for the remaining region; To generate an ideal magnetic field matrix for the target region, The ideal excitation magnetic field matrix for the outer peripheral region; Weights for uniformity in the target region; The remaining region is controlled by weights.

[0028] In this embodiment, the step of dynamically updating the peripheral weights through an iterative adaptive reweighting mechanism to generate the final channel complex weights includes: Set initial iteration parameters: Set the iteration number k=0, and the initial peripheral weights... = ; Repeat the following iterations until the iteration condition is met: According to the response matrix Adaptively update the peripheral weights to obtain the (k+1)th peripheral weight. ; Based on the (k+1)th peripheral weight Adjust the objective function and solve for the complex weights; wherein the iteration condition is: When the change in the energy ratio of the complex weights in two consecutive iterations is lower than the threshold, the optimization iteration stops, and the last calculated complex weight is output as the final channel complex weight.

[0029] In this embodiment, the step of using the response matrix... Adaptively update the peripheral weights to obtain the (k+1)th peripheral weight. include: Calculate the current region field strength norm: use the channel excitation vector from the k-th iteration. Substitute into the matrix to calculate the field strength norm of the target region. , peripheral regional field strength norm ; Dynamically update peripheral weights: according to the iterative formula Perform weight updates to obtain the (k+1)th peripheral weight. .

[0030] In this embodiment, the step of using the (k+1)th peripheral weight... Adjusting the objective function and solving for the complex weights includes: Use the updated Adjust the objective function ; An amplitude-phase separation strategy is employed for optimization: In the k-th round of the outer loop, the amplitude-phase separation strategy is first calculated. Extract its phase This iteration is fixed Only the amplitude coefficient is optimized. At this point, the complex weight form is ; Solve the optimized solution using nonlinear least squares or conjugate gradient methods. We obtain the complex weights after the kth round of optimization.

[0031] This application also provides a magnetic resonance radio frequency shimming optimization device based on spatial weights and adaptive reweighting. The device includes an excitation field distribution acquisition module, an iterative optimization data acquisition module, a final channel complex weight generation module, and a transmission module. The excitation field distribution acquisition module is used to acquire the excitation field distribution to be optimized; The data acquisition module for iterative optimization is used to obtain the excitation magnetic field matrix of the target region, the excitation magnetic field matrix of the outer region, and the objective function based on the excitation field distribution to be optimized. The final channel complex weight generation module is used to dynamically update the peripheral weights based on the target region excitation magnetic field matrix, the peripheral region excitation magnetic field matrix, and the objective function through an iterative adaptive reweighting mechanism, thereby generating the final channel complex weights. The transmitting module is used to pass the final channel complex weights to the magnetic resonance transmitting system, so that the magnetic resonance transmitting system can achieve radio frequency shimming according to the final channel complex weights. Example 2:

[0032] Multi-channel data acquisition or simulation Response matrix, ,in In pixels / voxels This represents the number of channels.

[0033] Spatial weights are constructed to divide the imaging voxels into core target regions and peripheral regions, and a target mask is defined. (Select target area) and peripheral mask (Remaining area), corresponding number of pixels , (The corresponding pixel count refers to the number of voxels (pixels) selected in the mask. N1 is the total number of voxels in the target area, and N2 is the total number of voxels in the outer area.)

[0034] Set target weights and peripheral initial weight parameters , towards the target and peripheral region weight After normalization adjustment, the B1+ modality matrix of each region is multiplied by a weighting coefficient to form a weighted response matrix and a target vector, thus completing the spatial weighting modeling.

[0035]

[0036] The excitation magnetic field matrix of the target region (size is) ); The excitation magnetic field matrix for the remaining region (size is) ); The ideal excitation magnetic field matrix for the target region (size is) ); The ideal excitation magnetic field matrix for the remaining region (size is) ); The uniformity weight of the target region is used to control... With ideal magnetic field The matching degree is usually set to 1, but it can be adjusted according to actual needs.

[0037] Weights are assigned to the remaining regions to constrain non-target regions. The excitation magnetic field should be optimized to avoid high field strength in the background region, which could cause safety issues or interference.

[0038] For the target region, the optimization objective is to make the magnetic field in the target region as uniform as possible, and the calculation formula is as follows: ; Since the uniformity in the objective function is typically the squared error, the squared term can balance the order of magnitude and prevent overfitting. It's very large, use it directly. This may cause optimizations to ignore other items, so it is necessary to amplify its impact.

[0039] The peripheral region weights are set to be adjustable to ensure differentiated contributions from different optimization objective functions. To avoid the optimization results becoming overly dependent on the initial settings due to fixed weights, an iterative adaptive reweighting mechanism is introduced. Adaptive reweighting can be performed based on methods such as region energy ratio (ratio update, i.e., target pixel ratio) or hotspot-driven amplification (based on residual / threshold). This example uses the ratio update method for weighting, but residual optimization can also be used. The peripheral region weights are calculated as follows: ; This is a custom control coefficient for the remaining area. Number of pixels in the target area With the number of pixels in the remaining area The ratio of . The penalty for deviations in the field amplitude of the outer perimeter region can be set to (0~1). If the remaining region pixels ( (Larger than the target area) Use directly This will lead to the remaining region's error dominating the optimization. Therefore, by... Adjusting the ratio to balance the two components avoids an excessive number of voxels in the peripheral region affecting the contribution of the core region, thus achieving adaptive reweighting that dynamically adjusts the weights based on the actual amount of data.

[0040] Let the first The weight for the next iteration is ,but ;

[0041] For the first The channel excitation vector of the next iteration ensures that the peripheral weights can be dynamically adjusted according to the relative strength between the current target region and the peripheral B1+ during the optimization process.

[0042] Constructing the objective function: used to calculate the objective function value of the optimization problem. ; The complex weight w of each channel is calculated using nonlinear least squares or conjugate gradient methods. If both amplitude and phase are freely optimized simultaneously, the problem becomes strongly nonconvex, prone to divergence or local extrema. Therefore, this invention employs an alternating phase-fixed strategy, calculating a "reference solution" in the k-th round of the outer loop. The specific method is as follows: You can choose any publicly available and implementable path; this reference solution is... The closed-form solution of a standard weighted least squares algorithm under the objective is of the following form:

[0043] Alternatively, a phase substitution strategy based on least magnitude squares can be used: ,

[0044] Extract the phase from the reference solution. ; Construct a phase matrix and fix the phase on the diagonal so that subsequent optimizations are performed only in the amplitude dimension.

[0045]

[0046] The amplitude parameters of each channel Multiply by a phase factor Thus, they are combined into complex weights:

[0047] This completes the rewriting of the objective function to be only about Convex quadratic target:

[0048] This approach updates only the amplitude dimension in each iteration, avoiding rapid phase jitter. It brings the amplitude in the target region closer to the ideal value while controlling the amplitude in the outer region, significantly improving numerical stability and convergence efficiency.

[0049] After each outer loop, the weighted residual energy of the objective function is calculated. Or based on the energy ratio of the target area to the surrounding area. In the outer loop, the energy change between two adjacent iterations is monitored to satisfy... When the optimization converges, the final complex weight w is output and used to drive the magnetic field.

[0050] The effectiveness of the method in this invention has been verified through simulation experiments. For example... Figure 2 As shown, a 5T ultra-high field MRI body emission coil model was established, and a human electromagnetic model was loaded. Taking the heart section as an example, the magnetic field distribution of the heart section was obtained.

[0051] Subsequently, the excitation magnetic field data on the cross-section of the heart were extracted and imported into the MATLAB platform for post-processing.

[0052] In the MATLAB environment, the original multi-channel B1 was first processed. ﹢ The field data is reconstructed to form the excitation field distribution to be optimized. Then, three excitation strategies are set respectively: Unshielded field conditions: Traditional circular polarization excitation was used as a control; Conventional shimming method: Employs the traditional least squares optimization strategy for amplitude; The method of this application is to introduce a spatial weighting factor and an adaptive reweighting iteration mechanism during the optimization process to dynamically adjust the relative importance of the target region and non-target regions.

[0053] To ensure fairness in comparing different methods, all protocols were evaluated within the same ROI region (the cardiac target area indicated by the dashed circle). Quantitative evaluation used the following two indicators: Coefficient of Variation (CV): This refers to the coefficient of variation within the ROI (B1).﹢ The ratio of the standard deviation to the mean of a field characterizes the uniformity of the field distribution. RF excitation efficiency: defined as B1 within the ROI ﹢ The ratio of the field average to the square root of the transmit power characterizes energy efficiency.

[0054] See Figure 3 Experimental results show that the method of this application can ensure the uniformity of the target ROI region while taking into account the distribution of the peripheral field intensity, and significantly improve the radio frequency excitation efficiency. Through iteration, the weight parameters for screening the field uniformity of this invention with better performance are: .

[0055] The results show that Figure 3 The left image (original field distribution) shows the ROI region (target region) defined by the circular dashed circle, B1 ﹢ The field distribution is significantly uneven. Conventional / traditional shimming optimization methods improve magnetic field uniformity, but a weak magnetic field distribution still exists in the peripheral region. The uniformity outside the target region (CV_other) is also calculated. The traditional shimming method has CV_other = 0.6, while the method of this invention has CV_other = 0.43. The right figure (the method of this invention) employs spatial weighting and an adaptive reweighting mechanism to achieve equalization of the field distribution in the target region and the peripheral magnetic field, improving both uniformity and radio frequency excitation efficiency.

[0056] This application explicitly distinguishes between the target region (ROI) and the surrounding region and constructs adjustable spatial weights. This invention can effectively control field distribution mismatch in the surrounding region while ensuring the imaging quality of the target region, thus reducing safety risks. Furthermore, an adaptive reweighting mechanism is introduced to dynamically update the weights based on the energy ratio of the ROI and the surrounding region or local hotspot conditions. This reduces the sensitivity to the initial weight settings and significantly improves the robustness and stability of the optimization results.

[0057] In terms of optimization strategy, this application adopts a phase-fixed, amplitude-iterative optimization approach, decomposing the complex non-convex problem into subproblems with better convergence, effectively improving the iterative convergence speed and numerical stability, and making the target region B1... ﹢ The field distribution can gradually approach an ideal uniform state. At the same time, this application can take into account both field shimming within the ROI and response control in the peripheral region during the optimization process, ensuring that safety is guaranteed while improving imaging uniformity.

[0058] Therefore, this application has the advantages of stable optimization effect, fine regional control and strong adaptability. It is especially suitable for multi-channel parallel transmission systems. The ROI and peripheral region can be flexibly set according to different anatomical regions and imaging tasks, which facilitates its application in various ultra-high field MRI systems.

[0059] This application establishes mask functions for ROIs and non-ROIs, and forms spatial weight coefficients using a voxel-weighted approach. These weight coefficients not only reflect the relative importance of different regions in the optimization objective, but also ensure the comparability of contributions from different regions through voxel number normalization. Compared with traditional methods, this method effectively avoids overfitting on small ROIs and the dilution of weights in large ROIs, thus achieving more reasonable spatial balance control.

[0060] This application proposes an adaptive reweighting mechanism that dynamically adjusts the weights of different regions during iterative optimization. After each optimization iteration, based on the objective function residual and local field distribution deviation, the energy difference or uniformity index between the ROI and non-ROI is calculated in real time, and the weight coefficients are updated accordingly. This mechanism enables fine-grained control of the radio frequency field distribution, improving the convergence and robustness of the algorithm under complex anatomical structures.

[0061] This application introduces an iterative strategy that separates amplitude and phase during the optimization process. In each outer loop iteration, the phase is fixed, and only the amplitude is optimized to gradually approximate the target field distribution. In the next outer loop iteration, the phase is updated as needed, and amplitude optimization is resumed. This approach effectively reduces the non-convexity and instability of the optimization problem, significantly improving the iterative convergence speed and the stability of the solution.

[0062] This application can effectively solve and improve the problem of uneven distribution of excitation magnetic field during magnetic resonance imaging (MRI). It can be applied to any anatomical region (such as liver, heart, brain, breast, abdomen, etc.), achieving a balance between the uniformity of the target area and the control of the peripheral response, and has broad application value.

[0063] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and not to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A magnetic resonance radio frequency shimming optimization method based on spatial weights and adaptive reweighting, characterized in that, The magnetic resonance radio frequency shim optimization method based on spatial weights and adaptive reweighting includes: Obtain the excitation field distribution to be optimized; The excitation magnetic field matrix of the target region, the excitation magnetic field matrix of the outer region, and the objective function are obtained based on the excitation field distribution to be optimized. Based on the excitation magnetic field matrix of the target region, the excitation magnetic field matrix of the peripheral region, and the objective function, the peripheral weights are dynamically updated through an iterative adaptive reweighting mechanism, thereby generating the final channel complex weights. The final channel complex weights are passed to the magnetic resonance emission system, thereby enabling the magnetic resonance emission system to achieve radio frequency shimming based on the final channel complex weights.

2. The magnetic resonance radio frequency shimming optimization method based on spatial weights and adaptive reweighting as described in claim 1, characterized in that, The process of obtaining the excitation field distribution to be optimized includes: B1 is obtained through experimental data collection or simulation calculation. ﹢ Field response data; According to B1 ﹢ Field response data acquisition multi-channel B1 ﹢ Response matrix of field response data Where N is the total number of pixels or voxels. Let C be the number of channels, and C be the complex field.

3. The magnetic resonance radio frequency shimming optimization method based on spatial weights and adaptive reweighting as described in claim 2, characterized in that, The process of obtaining the target mask, outer peripheral mask, target region excitation magnetic field matrix, and remaining region excitation magnetic field matrix based on the excitation field distribution to be optimized includes: Based on anatomical structure, voxels are divided into core target regions and peripheral regions, and target masks are defined. Peripheral mask Count of pixels , ; From the response matrix Extract the excitation magnetic field matrix of the target region. , Remaining region excitation magnetic field matrix Set the ideal excitation magnetic field matrix for the target region. Ideal excitation magnetic field matrix in the outer peripheral region ; Based on the target region's excitation magnetic field matrix Calculate the uniformity weight of the target region ; Based on the target region's excitation magnetic field matrix , Remaining region excitation magnetic field matrix Calculate the remaining area control weights .

4. The magnetic resonance radio frequency shimming optimization method based on spatial weights and adaptive reweighting as described in claim 3, characterized in that, The objective function formula is as follows: ;in, To generate a magnetic field matrix for the target region, To generate a magnetic field matrix for the remaining region; To generate an ideal magnetic field matrix for the target region, The ideal excitation magnetic field matrix for the outer peripheral region; Weights for uniformity in the target region; The remaining region is controlled by weights.

5. The magnetic resonance radio frequency shimming optimization method based on spatial weights and adaptive reweighting as described in claim 4, characterized in that, The step of dynamically updating the peripheral weights through an iterative adaptive reweighting mechanism to generate the final channel complex weights includes: Set initial iteration parameters: Set the iteration number k=0, and the initial peripheral weights... = ; Repeat the following iterations until the iteration condition is met: According to the response matrix Adaptively update the peripheral weights to obtain the (k+1)th peripheral weight. ; Based on the (k+1)th peripheral weight Adjust the objective function and solve for the complex weights; wherein the iteration condition is: When the change in the energy ratio of the complex weights in two consecutive iterations is lower than the threshold, the optimization iteration stops, and the last calculated complex weight is output as the final channel complex weight.

6. The magnetic resonance radio frequency shimming optimization method based on spatial weights and adaptive reweighting as described in claim 5, characterized in that, According to the response matrix Adaptively update the peripheral weights to obtain the (k+1)th peripheral weight. include: Calculate the current region field strength norm: use the channel excitation vector from the k-th iteration. Substitute into the matrix to calculate the field strength norm of the target region. , peripheral regional field strength norm ; Dynamically update peripheral weights: according to the iterative formula Perform weight updates to obtain the (k+1)th peripheral weight. .

7. The magnetic resonance radio frequency shimming optimization method based on spatial weights and adaptive reweighting as described in claim 6, characterized in that, The weighting based on the (k+1)th peripheral weight Adjusting the objective function and solving for the complex weights includes: Use the updated Adjust the objective function ； An amplitude-phase separation strategy is employed for optimization: In the k-th round of the outer loop, the amplitude-phase separation strategy is first calculated. Extract its phase This iteration is fixed Only the amplitude coefficient is optimized. At this point, the complex weight form is ; Solve the optimized solution using nonlinear least squares or conjugate gradient methods. We obtain the complex weights after the kth round of optimization.

8. A magnetic resonance radio frequency shimming optimization device based on spatial weighting and adaptive reweighting, characterized in that, The magnetic resonance radio frequency shim optimization device based on spatial weights and adaptive reweighting includes: An excitation field distribution acquisition module is used to acquire the excitation field distribution to be optimized. The iterative optimization data acquisition module is used to acquire the excitation magnetic field matrix of the target region, the excitation magnetic field matrix of the outer region, and the objective function based on the excitation field distribution to be optimized. The final channel complex weight generation module is used to dynamically update the peripheral weights based on the target region excitation magnetic field matrix, the peripheral region excitation magnetic field matrix, and the objective function through an iterative adaptive reweighting mechanism, thereby generating the final channel complex weights. The transmitting module is used to transmit the final channel complex weights to the magnetic resonance transmitting system, so that the magnetic resonance transmitting system can achieve radio frequency shimming according to the final channel complex weights.

Images