Dynamic quantitative imaging method and imaging device based on magnetic resonance fingerprinting imaging

Abstract

This invention provides a dynamic quantitative imaging method and imaging device based on magnetic resonance fingerprinting. The method includes: scanning according to preset acquisition parameters to obtain the time series of magnetic resonance images of all voxels; dividing the time series corresponding to each voxel into multiple time windows; matching each time window with the dictionary of the corresponding time window; and obtaining a quantitative map of the time period corresponding to each time window based on the matching result. In other words, this invention divides the time series of each voxel acquired by scanning into multiple time windows in the time dimension, and each time window is equipped with a dictionary. By matching the dictionary of each time window, the quantitative processing result of each time window for each voxel is obtained. Therefore, a quantitative map is obtained for each time period corresponding to each time window, thereby forming a quantitative map with high temporal resolution. This not only allows for the acquisition of quantitative maps for each time period but also the acquisition of imaging change information based on time changes.

Description

Technical Field

[0001] This invention relates to the field of magnetic resonance imaging technology, and in particular to a dynamic quantitative imaging method and imaging device based on magnetic resonance fingerprint imaging. Background Technology

[0002] Quantitative parametric imaging is an important research direction in magnetic resonance imaging (MRI). Traditional quantitative methods involve using different experimental parameters and performing multiple scans to obtain relaxation signal curves under different parameters. The longitudinal relaxation time T1 and transverse relaxation time T2 are then estimated through curve fitting. However, this method is highly susceptible to imaging conditions, leading to large parameter estimation errors and lengthy quantitative sequences, which cannot meet the requirements of clinical medical applications.

[0003] In response, magnetic resonance fingerprinting (MRF) technology has ushered in a new era of quantitative parametric imaging. MRF acquires transient signals by varying the flip angle and repeat time (TR) in a pseudo-random manner. It also uses a dictionary of pre-calculated signal evolution curves corresponding to different parameters, comparing the acquired signal evolution curves with the dictionary to obtain the quantitative value of the corresponding voxel. Because the dictionary provides prior information about the voxels, MRF can perform high-magnification downsampling, thus achieving rapid quantitative imaging. However, since the temporal dimension of MRF data is typically used to calculate consistency with the dictionary's temporal dimension, the parametric maps obtained by MRF in this method are both static parametric maps, similar to those obtained by traditional quantitative methods, resulting in low temporal resolution.

[0004] Therefore, a new quantitative imaging method is urgently needed to solve the above-mentioned technical problems. Summary of the Invention

[0005] The purpose of this invention is to provide a dynamic quantitative imaging method and imaging device based on magnetic resonance fingerprint imaging, so as to solve the problem of how to obtain dynamic quantitative imaging with high temporal resolution.

[0006] To address the aforementioned technical problems, this invention provides a dynamic quantitative imaging method based on magnetic resonance fingerprinting, comprising:

[0007] Scanning is performed according to preset acquisition parameters to obtain time series of magnetic resonance images of all voxels;

[0008] The time series corresponding to each voxel is divided into multiple time sliding windows;

[0009] Match each time window segment with the dictionary corresponding to that time window;

[0010] Based on the matching results, obtain a quantitative graph of the time period corresponding to each time window segment.

[0011] Optionally, in the aforementioned dynamic quantitative imaging method based on magnetic resonance fingerprinting, before matching each time window segment with the corresponding dictionary of the time window, the following steps are included:

[0012] Based on the acquisition parameters corresponding to the time sliding window and a variety of preset quantitative parameter combinations, multiple evolution curves of the time sliding window are simulated.

[0013] The dictionary for the time sliding window is determined based at least on the correspondence between the evolution curve and the combination of quantitative parameters;

[0014] The quantitative parameter combination includes the initial state of the magnetization vector and a variety of quantitative parameters.

[0015] Optionally, in the dynamic quantitative imaging method based on magnetic resonance fingerprinting, the preset combination of multiple quantitative parameters includes at least one preset initial state of the magnetization vector.

[0016] Optionally, in the dynamic quantitative imaging method based on magnetic resonance fingerprinting, in the same voxel, the initial state of the magnetization vector of the time window in this segment is determined according to the matching result of the time window in the previous segment; and the initial state of the magnetization vector of the time window in this segment is used as the initial state of the magnetization vector in the preset combination of multiple quantitative parameters.

[0017] Optionally, in the dynamic quantitative imaging method based on magnetic resonance fingerprinting, in each voxel, the initial state of the magnetization vector of the first time sliding window is a thermal equilibrium state.

[0018] Optionally, in the dynamic quantitative imaging method based on magnetic resonance fingerprinting, the process of matching each time window segment with the corresponding dictionary of the time window includes:

[0019] The evolution curve corresponding to each time sliding window segment is compared with the evolution curve simulated in the dictionary of the time sliding window, and the quantitative parameter corresponding to the evolution curve in the dictionary with the highest similarity is taken as the quantitative value of the time sliding window.

[0020] Optionally, in the dynamic quantitative imaging method based on magnetic resonance fingerprinting, the matching result includes the quantitative value of each time segment in each voxel.

[0021] Optionally, in the aforementioned dynamic quantitative imaging method based on magnetic resonance fingerprinting, the process of obtaining the quantitative map of the time period corresponding to each time window segment according to the matching result includes:

[0022] The quantitative map for the time period is generated based on the quantitative values ​​of all the voxels within the same time period.

[0023] Optionally, in the dynamic quantitative imaging method based on magnetic resonance fingerprinting, the length of each time window segment is greater than or equal to the step size of the time window.

[0024] Based on the same inventive concept, the present invention also provides an imaging device for performing the dynamic quantitative imaging method based on magnetic resonance fingerprint imaging; wherein, the imaging device includes an information acquisition unit, an information processing unit, a dictionary matching unit, and an image generation unit;

[0025] The information acquisition unit is used to scan according to preset acquisition parameters to obtain the time series of magnetic resonance images of all voxels.

[0026] The information processing unit is used to divide the time series corresponding to each voxel into multiple time sliding windows;

[0027] The dictionary matching unit is used to match each segment of the time window with the dictionary of the corresponding time window;

[0028] The image generation unit is used to obtain a quantitative image of the time period corresponding to each time window segment based on the matching results.

[0029] In summary, this invention provides a dynamic quantitative imaging method and imaging device based on magnetic resonance fingerprinting. The method includes: scanning according to preset acquisition parameters to acquire the magnetic resonance image time series of all voxels; dividing the time series corresponding to each voxel into multiple time windows; matching each time window segment with a corresponding dictionary; and obtaining a quantitative map of the time period corresponding to each time window segment based on the matching results. It can be seen that this invention divides the time series of each voxel acquired by scanning into multiple time windows in the time dimension, and each time window segment is equipped with a corresponding dictionary. By performing dictionary matching on each time window segment, the quantitative processing result of each time window segment for each voxel can be obtained. Based on this, a quantitative map can be obtained for each time period corresponding to each time window segment, thereby forming a quantitative map with high temporal resolution. This not only allows for the acquisition of quantitative maps for each time period but also the acquisition of dynamic imaging change information based on time changes, thus improving the comprehensiveness of the acquired imaging information. Attached Figure Description

[0030] Those skilled in the art will understand that the accompanying drawings are provided to better understand the invention and do not constitute any limitation on the scope of the invention. Wherein:

[0031] Figure 1 This is a flowchart of the dynamic quantitative imaging method based on magnetic resonance fingerprinting in Embodiment 1 of the present invention.

[0032] Figure 2 This is a schematic diagram of the time series of magnetic resonance images in Embodiment 1 of the present invention.

[0033] Figure 3 This is a schematic diagram of the time-dividing sliding window in Embodiment 1 of the present invention.

[0034] Figure 4 This is a schematic diagram of the formation of a quantitative graph in Embodiment 1 of the present invention.

[0035] Figure 5 This is a flowchart illustrating the offline dictionary generation process in Embodiment 3 of the present invention.

[0036] Figure 6 This is a schematic diagram of the online dynamic dictionary generation process in Embodiment 4 of the present invention.

[0037] Figure 7 This is a schematic diagram of the time sliding window in Embodiment 4 of the present invention.

[0038] Figure 8 This is a schematic diagram of the imaging device in Embodiment 10 of the present invention.

[0039] In the attached image:

[0040] 100 - Information acquisition unit; 101 - Information processing unit; 102 - Dictionary matching unit; 103 - Image generation unit; 104 - Dictionary generation unit. Detailed Implementation

[0041] To make the objectives, advantages, and features of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be noted that the drawings are all in a very simplified form and are not drawn to scale, and are only used to facilitate and clearly illustrate the objectives of the embodiments of the present invention. Furthermore, the structures shown in the drawings are often part of the actual structures. In particular, different figures may emphasize different aspects and sometimes use different scales. It should also be understood that, unless specifically stated or indicated, the terms "first," "second," "third," etc., in the specification are only used to distinguish the various components, elements, steps, etc., in the specification, and are not used to indicate the logical or sequential relationships between the various components, elements, steps, etc.

[0042] <Example 1>

[0043] Please see Figure 1 This embodiment provides a dynamic quantitative imaging method based on magnetic resonance fingerprinting, including:

[0044] Step 1 S10: Scan according to the preset acquisition parameters to obtain the time series of magnetic resonance images of all voxels.

[0045] Step 2 S20: Divide the time series corresponding to each voxel into multiple time sliding windows.

[0046] Step 3 S30: Match each time window segment with the dictionary of the corresponding time window.

[0047] Step 4S40: Obtain a quantitative graph of the time period corresponding to each time window segment based on the matching results.

[0048] As can be seen, in this embodiment, the time series of each voxel acquired by scanning is divided into multiple time windows along the time dimension, and each time window is equipped with a corresponding dictionary. By performing dictionary matching on each time window segment, the quantitative processing result of each time window segment for each voxel can be obtained. Based on this, a quantitative map can be obtained for each time period corresponding to each time window segment, thereby forming a quantitative map with high temporal resolution. This not only allows for the acquisition of quantitative maps for each time period but also enables the acquisition of dynamic imaging change information based on time changes, thereby improving the comprehensiveness of the acquired imaging information.

[0049] The following is in conjunction with the appendix Figures 1-4 This embodiment provides a detailed explanation of the dynamic quantitative imaging method based on magnetic resonance fingerprinting.

[0050] Step 1 S10: Scan according to the preset acquisition parameters to obtain the time series of magnetic resonance images of all voxels.

[0051] It should be noted that each examined tissue exhibits a unique signal response to a given input sequence. Therefore, to measure multiple tissue characteristics in a single acquisition, MRF allows for pseudo-random variation of the acquisition parameters. Preferably, the acquisition parameters include: flip angle (FA), radio frequency phase, gradient phase shift, repetition time (TR), echo time (TE), and K-space sampling trajectory, etc. This embodiment is not limited to specific values ​​for the acquisition parameters and can be preset according to the examined tissue and imaging requirements.

[0052] Furthermore, after setting the acquisition parameters, a magnetic resonance imaging (MRI) device is used to acquire signals from the examined tissue to obtain a time series of MRI images of all voxels in the examined tissue. For example, please refer to [link to example]. Figure 2The time series of all acquired voxels spans from T1 to T500, and one magnetic resonance image is acquired at each time point, resulting in 500 images (I1 to I500). Each image represents the imaging of all voxels at the corresponding time point, i.e., the imaging of the examined tissue. It should be noted that a voxel refers to a small volume element of the examined tissue, defined by a specific size and coordinates, on the plane where imaging is desired.

[0053] Step 2 S20: Divide the time series corresponding to each voxel into multiple time sliding windows.

[0054] It is understood that the information covered by the time series of each voxel includes the evolution curve of the quantitative parameters of that voxel on the corresponding time axis. In this embodiment, the quantitative parameters include, but are not limited to, longitudinal relaxation time T1 and lateral relaxation time T2. The time series corresponding to each voxel is divided into multiple parts sequentially along the time axis, with each part serving as a time window. These multiple time windows are used to characterize the changes in the quantitative parameters of the voxel within different time periods. Preferably, each time window corresponds to the same time period, and this embodiment does not limit the number of time windows for each voxel, but it must be greater than or equal to 2.

[0055] For example, Figure 3 The diagram shows the time series of a voxel, where the time interval is T1 to Tn, and n is a positive integer. Assuming the length of each unit time in the time series is 1, then the length of each time window is 3, meaning every three units of time constitute one time window; and the step size of the time window movement is 2. Specifically, the three unit times corresponding to the first time window A1 are T1, T2, and T3; the second time window A2 moves two time intervals relative to the first time window A1, so the three unit times corresponding to the second time window A2 are T3, T4, and T5. Based on this, the three unit times corresponding to the third time window A3 are T5, T6, and T7; the three unit times corresponding to the fourth time window A4 are T7, T8, and T9. And so on, the voxel time series can be divided into multiple time windows. Preferably, the time window division rules for all voxels are the same, so that in subsequent processes, the time intervals corresponding to the same time window can be used to obtain a quantitative plot.

[0056] Step 3 S30: Match each time window segment with the dictionary of the corresponding time window.

[0057] Each time window in this embodiment is associated with a dictionary. The dictionary includes evolution curves simulated with different quantitative parameters under different initial magnetization vector states. The matched time window is assigned the quantitative parameters of the corresponding evolution curve in the dictionary as the closest quantitative value.

[0058] Furthermore, assuming the examined tissue is divided into 1000 voxels, and the time series corresponding to each voxel is divided into 100 time windows, then each time window within each voxel corresponds to a dictionary. During the matching process, each time window requires dictionary matching, so each voxel requires 100 matches, and 1000 voxels require 1.0 x 10^10 matches. 5 The matching process is repeated. In this embodiment, the matching order of each time window is not limited. All time windows corresponding to the same voxel can be matched simultaneously or sequentially, and all voxels can also be matched simultaneously or sequentially.

[0059] Step 4S40: Obtain a quantitative graph of the time period corresponding to each time window segment based on the matching results.

[0060] After all time windows for all voxels are matched, each time window for each voxel is assigned a quantitative value. Then, based on the quantitative values ​​and evolution curves of all voxels within the time period corresponding to the same time window, a corresponding quantitative plot is obtained. For example, please refer to [link to example]. Figure 4 Assume the examined tissue is divided into 1000 voxels, and the time series of each voxel is divided into m time windows, where m is a positive integer. Each time window undergoes dictionary matching to obtain the corresponding quantitative value Q. The time periods corresponding to each voxel's time windows A1 to Am are A1' to Am', respectively. Based on the quantitative values ​​Q of all voxels corresponding to each time period and their evolution curves, a quantitative map for that time period is obtained. For example, in time period A1', based on the quantitative values ​​Q(1,1) of voxel 1, Q(2,1) of voxel 2, ..., Q(1000,1) of voxel 1000, and their corresponding evolution curves, a quantitative map M1 corresponding to time period A1' can be obtained. Correspondingly, a quantitative map M2 corresponding to time period A2', a quantitative map M3 corresponding to time period A3', ..., and a quantitative map Mm corresponding to time period Am' are obtained. Therefore, the method provided in this embodiment can acquire m quantitative images of the examined tissue in a single acquisition, compared to the prior art which can only acquire one quantitative image per acquisition. Clearly, the method provided in this embodiment can acquire quantitative images with high temporal resolution to capture the changes in the examined tissue over different time periods, achieving dynamic quantitative imaging.

[0061] In summary, the dynamic quantitative imaging method based on magnetic resonance fingerprinting provided in this embodiment divides the time series of each voxel into multiple time windows. Dictionary matching is performed on each time window to obtain multiple matching results. Based on these multiple matching results, multiple quantitative maps based on time axis changes can be obtained, thereby extracting dynamic imaging changes in the examined tissue during the detection period, achieving dynamic quantitative imaging, and helping to improve the guiding value of magnetic resonance imaging for clinical treatment. For details not covered in this embodiment, please refer to other embodiments.

[0062] <Example 2>

[0063] This embodiment provides a dynamic quantitative imaging method based on magnetic resonance fingerprinting, comprising: based on the method described in Embodiment 1, before matching each time window segment with the dictionary of the corresponding time window, including: simulating multiple evolution curves of the time window according to the acquisition parameters corresponding to the time window and a preset combination of multiple quantitative parameters; determining the dictionary of the time window at least according to the correspondence between the evolution curves and the combination of quantitative parameters; wherein, the combination of quantitative parameters includes the initial state of the magnetization vector and a variety of quantitative parameters.

[0064] Furthermore, the dictionary is a set of signal evolutions, optionally generated using Bloch equations to simulate all possible tissue properties measurable within physiological limits. The dictionary is uniquely designed for each MRF sequence. Therefore, during dictionary generation, it is necessary to iterate through a certain step size based on the acquisition parameters corresponding to the time window and a variety of preset quantitative parameter combinations to simulate and generate multiple evolution curves. The quantitative parameter combinations include the initial state of the magnetization vector and various quantitative parameters. The initial state of the magnetization vector includes the initial state of the transverse magnetization vector Mxy and the initial state of the longitudinal magnetization vector Mz corresponding to the voxel after providing RF pulse excitation within the time window period. The quantitative parameters include, but are not limited to, the longitudinal relaxation time T1 and the transverse relaxation time T2. For example, the quantitative parameter combinations are a set N(Mxy, Mz, T1, T2). Changing the quantity of any element in set N forms a different quantitative parameter combination.

[0065] Furthermore, under the acquisition parameters corresponding to the time window, each combination of quantitative parameters will simulate a corresponding evolution curve. The dictionary is then constructed from the acquisition parameters, quantitative parameters, evolution curves, and the correspondence between the evolution curves and the quantitative parameters within the time window. Each combination of quantitative parameters constitutes an entry in the dictionary. Therefore, in this embodiment, each time window corresponds to a dedicated dictionary. Assuming the examined tissue is divided into 1000 voxels, and the time series of each voxel is divided into 100 time windows, then a 1.0 x 102 time window is generated. 5 One dictionary, for 1.0x10 5 Second match.

[0066] In summary, the dynamic quantitative imaging method based on magnetic resonance fingerprinting provided in this embodiment simulates a dictionary specific to each time window within each voxel, ensuring the feasibility of multiple matching and acquisition of multiple quantitative images. For details not covered in this embodiment, please refer to other embodiments.

[0067] <Example 3>

[0068] This embodiment provides a dynamic quantitative imaging method based on magnetic resonance fingerprinting, comprising: based on the methods provided in Embodiments 1 and 2, the preset multiple quantitative parameter combinations include at least one preset initial state of the magnetization vector. That is, the multiple quantitative parameter combinations are obtained by preset, and the initial state of the magnetization vector in the multiple quantitative parameter combinations can be many or the same.

[0069] Based on this, please refer to Figure 5This embodiment provides an offline dictionary generation method. First, based on preset acquisition parameters, a magnetic resonance imaging (MRI) device is used to acquire time series of MRI images of all voxels of the examined tissue. Second, time windows are determined for each voxel's time series according to time window partitioning rules. Then, based on each time window and the preset acquisition parameters, a dictionary simulation algorithm is used to traverse all combinations of quantitative parameters to generate multiple evolution curves, thereby obtaining the dictionary for each time window. Because the same quantitative parameter will form different evolution curves under different initial magnetization vector states, to ensure better robustness of the dictionary, this embodiment includes multiple initial magnetization vector states for various combinations of quantitative parameters, thus obtaining the evolution curves under multiple initial magnetization vector states to ensure the accuracy of subsequent matching. It should be noted that in this embodiment, the dictionary for each time window is generated before dictionary matching. Furthermore, after obtaining the preset acquisition parameters and time window partitioning rules, the dictionary generation step can be performed without waiting for the scan to complete. That is, scanning the examined tissue and generating the dictionary can be performed simultaneously.

[0070] After generating the dictionary for each time window, dictionary matching can be performed separately based on the corresponding dictionary and time window to obtain matching results. The dictionary matching process can be sequential or performed independently. For example, after dictionary generation, time window A1 is first matched with its corresponding dictionary to obtain matching result 1. Then, time window A2 is matched with its corresponding dictionary to obtain matching result 2. This process continues until time window Am is matched with its corresponding dictionary to obtain matching result m, thus completing the entire matching process.

[0071] Furthermore, this embodiment does not limit the generation order of each dictionary corresponding to each voxel. All dictionaries in all voxels can be generated synchronously, or all dictionaries in each voxel can be generated one by one and dictionary generation can be performed on each voxel at the same time. Alternatively, all dictionaries in each voxel can be generated one by one or simultaneously and dictionary generation can be performed on each voxel one by one.

[0072] In summary, the dynamic quantitative imaging method based on magnetic resonance fingerprinting provided in this embodiment achieves offline dictionary generation by pre-setting a combination of quantitative parameters. This allows for the generation of various dictionaries simultaneously with the scanning and acquisition of time series data, shortening the running time and improving imaging efficiency. For details not covered in this embodiment, please refer to other embodiments.

[0073] <Example 4>

[0074] This embodiment provides a dynamic quantitative imaging method based on magnetic resonance fingerprinting, comprising: based on the methods provided in Embodiments 1 and 2, in the same voxel, the initial state of the magnetization vector of the current time sliding window is determined according to the matching result of the previous time sliding window; and the obtained initial state of the magnetization vector of the current time sliding window is used as the initial state of the magnetization vector in the preset combination of multiple quantitative parameters. Obviously, the initial state of the magnetization vector in this embodiment is determined according to the matching result, eliminating the need for preset multiple cases and reducing the computational load of dictionary generation.

[0075] Based on this, please refer to Figure 6 and Figure 7 This embodiment provides an online dictionary generation method. Each dictionary is generated based on the matching result of the previous time window segment within the same voxel, thus generating a dictionary for the current time window segment. That is, the dictionary can be dynamically generated during the matching process. Specifically, firstly, according to preset acquisition parameters, a magnetic resonance imaging (MRI) device is used to acquire time series of MRI images of all voxels of the examined tissue. Secondly, each time window of the time series for each voxel is determined according to time window division rules. Then, a dictionary for each time window is dynamically generated. For example, within the same voxel, based on the acquisition parameters corresponding to the first time window A1 and multiple quantitative parameter combinations, a dictionary simulation algorithm is used to simulate multiple evolution curves to constitute the dictionary for the first time window A1. The initial state of the magnetization vector in the multiple quantitative parameter combinations corresponding to the first time window A1 is a preset state, preferably a thermal equilibrium state. Subsequently, the evolution curve of the first time window A1 acquired by scanning is matched with multiple evolution curves in the dictionary of the first time window A1 to obtain matching result 1. Based on the matching result 1, the initial state of the magnetization vector of the second time window A2 can be directly retrieved, i.e., directly read. Figure 7 The initial state of the magnetization vector at time T3 is shown. Based on this, the initial state of the magnetization vector at time T3 is used as the initial state of the magnetization vector in the dictionary generation process of the second time window A2, which is a combination of multiple quantitative parameters. Based on this, the dictionary of the second time window A2 is generated by simulation. Similarly, the evolution curve of the second time window A2 is matched with multiple evolution curves in the dictionary of the second time window A2 to obtain matching result 2. According to the matching result 2, the initial state of the magnetization vector at time T5 can be read as the initial state of the magnetization vector of the third time window. And so on, after the matching of the time window in the previous section is completed, the dictionary of the time window in this section is generated according to the matching result of the previous section, until the dictionary of the m-th time window Am is simulated and the matching result m is obtained.

[0076] Compared to the offline dictionary generation method in Embodiment 3, the dictionary generation process for each time window in this embodiment does not require traversing all initial states of the magnetization vectors; only one is needed. This reduces the computational load of dictionary generation, improves imaging efficiency, and enhances the accuracy of dictionary matching. Furthermore, the dynamic dictionary generation method is only applicable to each time window within the same voxel, divided sequentially along the time axis. Also, this embodiment does not limit the initial state of the magnetization vector used in the dictionary generation process of the first time window A1.

[0077] In summary, the dynamic quantitative imaging method based on magnetic resonance fingerprinting provided in this embodiment generates a dictionary for each time window dynamically and online. Based on the matching results of the previous time window, the initial state of the magnetization vector for the current time window can be directly obtained. Therefore, during the simulation calculation of the time window, it is not necessary to traverse all types of initial magnetization vector states; only the initial state of the magnetization vector for the current time window is used for calculation. This not only reduces the computational load and improves imaging efficiency but also helps to improve the accuracy of dictionary matching. For details not covered in this embodiment, please refer to other embodiments.

[0078] <Example 5>

[0079] This embodiment provides a dynamic quantitative imaging method based on magnetic resonance fingerprinting, including: based on the methods provided in Embodiments 1, 2 and 4, in each voxel, the magnetization vector of the first time sliding window is initially in thermal equilibrium.

[0080] As shown in Embodiment 4, during the generation of the dictionary using a dynamic online method, the initial state of the magnetization vector of the first time window needs to be preset manually. Preferably, the initial state of the first time window corresponding to each voxel is a thermal equilibrium state. That is, the transverse magnetization vector Mxy is 0, and the longitudinal magnetization vector Mz is M0. It should be noted that the thermal equilibrium state is the natural state of the examined tissue. When the examined tissue is placed in a magnetic resonance environment, it will naturally reach a thermal equilibrium state. Only after being excited by radio frequency will the magnetization vector state of the examined component shift. Therefore, the initial state of the first time window being a thermal equilibrium state is intended to indicate that there is a sufficient interval between the previous scan sequence and the method provided in this embodiment before execution, which can avoid interference from the previous scan sequence to the current scan and ensure the accuracy of the acquired data. For details not covered in this embodiment, please refer to other embodiments.

[0081] Example 6

[0082] This embodiment provides a dynamic quantitative imaging method based on magnetic resonance fingerprinting, including: based on the methods provided in Embodiment 1 and Embodiment 2, the process of matching each time window segment with the corresponding dictionary of the time window includes: comparing the evolution curve corresponding to each time window segment with the simulated evolution curve in the dictionary of the time window, and taking the quantitative parameter corresponding to the evolution curve in the dictionary with the highest similarity as the quantitative value of the time window.

[0083] Furthermore, the dictionary of the time sliding window includes the acquisition parameters, quantitative parameter combinations, evolution curves, and the correspondence between the quantitative parameter combinations and the evolution curves corresponding to the time sliding window. It can be understood that the acquisition parameters and the quantitative parameter combinations are the parameter conditions when the dictionary is generated through simulation, and the evolution curves are signal evolution curves generated based on these parameter conditions. Therefore, the acquisition parameters and their change trajectories, the initial state of the magnetization vector, the quantitative parameters, and the evolution curves in the dictionary of the time sliding window all have specific mapping relationships. Based on this, by comparing the evolution curves simulated in the dictionary with the evolution curves of the time sliding window acquired through scanning, the specific quantitative parameter corresponding to the evolution curve in the dictionary with the highest similarity is taken as the value of the quantitative parameter of the time sliding window, i.e., the quantitative value. For example, the dictionary includes X longitudinal relaxation evolution curves, and the longitudinal relaxation evolution curves acquired through scanning in the time sliding window are compared with these X evolution curves. Assuming that the Xth evolution curve has the highest similarity to the longitudinal relaxation evolution curve obtained by scanning in the time sliding window, the longitudinal relaxation time T1 corresponding to the Xth evolution curve is assigned to the time sliding window. Therefore, the longitudinal relaxation time T1 is one of the quantitative values ​​of the time sliding window, and the Xth evolution curve is identified as the longitudinal relaxation evolution curve of the time sliding window.

[0084] Based on this, dictionary matching can be performed on the time window of each voxel to obtain the matching result of each time window. For details not covered in this embodiment, please refer to other embodiments.

[0085] <Example 7>

[0086] This embodiment provides a dynamic quantitative imaging method based on magnetic resonance fingerprinting, including: based on the methods provided in Embodiments 1, 2, and 6, the matching result includes a quantitative value for each time segment in each voxel. Further, the matching result also includes an evolution curve from the dictionary corresponding to the quantitative value. Based on the quantitative values ​​and the evolution curve, etc., a feasibility guarantee is provided for the subsequent construction of a quantitative map. For details not covered in this embodiment, please refer to other embodiments.

[0087] <Example 8>

[0088] This embodiment provides a dynamic quantitative imaging method based on magnetic resonance fingerprinting, including: based on the methods provided in Embodiments 1, 2, 6 and 7, the process of obtaining a quantitative map of the time period corresponding to each time window according to the matching result includes: generating the quantitative map of the time period according to the quantitative values ​​of all voxels within the same time period.

[0089] Furthermore, each voxel's time window corresponds to a specific time segment within the entire time axis. Based on this, a quantitative map corresponding to each time segment can be generated using the quantitative values ​​of all voxels within that time segment and the corresponding evolution curves. Assuming there are m time segments, m quantitative maps are generated, resulting in multiple quantitative maps based on changes along the time axis. Compared to existing technologies, the method provided in this embodiment can acquire multiple quantitative maps with high temporal resolution, achieving dynamic quantification. Therefore, the changes in the examined tissue along the time axis can be accurately recorded by multiple quantitative maps, acquiring imaging change information while simultaneously acquiring imaging information, thus improving the guiding value of magnetic resonance imaging for clinical treatment. Further details not covered in this embodiment can be found in other embodiments.

[0090] <Example 9>

[0091] This embodiment provides a dynamic quantitative imaging method based on magnetic resonance fingerprinting, including: based on the method provided in Embodiment 1, the length of each time window segment is greater than or equal to the step size of the time window.

[0092] Preferably, the time sliding window described in this embodiment satisfies the following formula:

[0093] N = M + S(K-1); and N > M ≥ S;

[0094] Where N is the total time length of the time series; M is the time length of each time window segment; S is the step size of the time window; and K is the number of time windows. For example, when N = 1000, M = 20, and S = 10, the number of time windows K = 99. Alternatively, when N = 1000, M = S = 20, the number of time windows K = 50.

[0095] It is understandable that when the duration M of each time window is greater than or equal to the step size S of the time window, it can be ensured that the initial state of the magnetization vector for the next time window can be queried after each time window is matched, during the process of dynamically generating the dictionary online. For example, in a voxel, N = 1000, M = 20, S = 10; the time range of the first time window is T1 to T20, and the time range of the second time window is T10 to T30. And after the first time window is matched, the initial state of the magnetization vector at time T10 can be queried according to the matching result, and the initial state of the magnetization vector at time T10 is used as the initial state of the magnetization vector calculated by the dictionary simulation for the second time window. For another example, in a voxel, N = 1000, M = 20, S = 20; the time range of the first time window is T1 to T20, and the time range of the second time window is T21 to T40. After the first time window is matched, the initial state of the magnetization vector at time T20 can be retrieved based on the matching result. The final initial state of the magnetization vector at time T20 is equal to the initial state of the magnetization vector at time T21. Therefore, the initial state of the magnetization vector at time T20 can be considered equal to the initial state of the magnetization vector at time T21. Thus, the initial state of the magnetization vector at time T20 is used as the initial state of the magnetization vector calculated by the dictionary simulation for the second time window. Based on this, other time windows also obtain the initial state of the magnetization vector in the same way as the two examples above.

[0096] Of course, if the duration M of each time window is less than the step size S of the time window, the initial state of the magnetization vector of the next time window can also be calculated based on the evolution curve corresponding to the matching result of the previous time window. However, compared with the direct query method provided in this embodiment, this calculation method increases the computational load and has limited accuracy, affecting the accuracy of the dictionary. Therefore, the method provided in this embodiment limits the duration of each time window to be greater than or equal to the step size of the time window, which can directly obtain the initial state of the magnetization vector of the time window, helping to improve the accuracy of the dictionary and improve imaging efficiency. For details not covered in this embodiment, please refer to other embodiments.

[0097] <Example 10>

[0098] Please see Figure 8This embodiment provides an imaging device for executing any one of the aforementioned embodiments one to nine of the dynamic quantitative imaging method based on magnetic resonance fingerprint imaging. The imaging device includes an information acquisition unit 100, an information processing unit 101, a dictionary matching unit 102, and an image generation unit 103. The information acquisition unit 100 is used to scan according to preset acquisition parameters to acquire the magnetic resonance image time series of all voxels. The information processing unit 101 is used to divide the time series corresponding to each voxel into multiple time windows, and the length of each time window segment is greater than or equal to the step size of the time window. The dictionary matching unit 102 is used to match each time window segment with the corresponding dictionary of the time window. The image generation unit 103 is used to acquire a quantitative image of the time period corresponding to each time window segment based on the matching result.

[0099] Furthermore, the imaging device also includes a dictionary generation unit 104. The dictionary generation unit 104 is used to simulate multiple evolution curves of the time sliding window based on the acquisition parameters corresponding to the time sliding window and a preset combination of multiple quantitative parameters; and to determine the dictionary of the time sliding window based at least on the correspondence between the evolution curves and the combination of quantitative parameters; wherein the combination of quantitative parameters includes an initial state of the magnetization vector and multiple quantitative parameters. Further, the dictionary generation unit 104 can generate the dictionary using an offline generation method. That is, the preset combination of multiple quantitative parameters includes at least one preset initial state of the magnetization vector. In other words, the initial state of the magnetization vector in the preset combination of multiple quantitative parameters is pre-set, and the initial state of the magnetization vector can be one or more. Preferably, there are multiple initial states of the magnetization vector, and they are combined to form multiple combinations of quantitative parameters. The dictionary generation unit 104 simulates multiple evolution curves by traversing multiple combinations of quantitative parameters based on the acquisition parameters corresponding to the time sliding window. Alternatively, the dictionary generation unit 104 can generate the dictionary using a dynamic online generation method. That is, within the same voxel, the initial state of the magnetization vector of the current time window is determined based on the matching result of the previous time window; and the obtained initial state of the magnetization vector of the current time window is used as the initial state of the magnetization vector in the preset multiple quantitative parameter combinations. Therefore, in the process of forming the dictionary of the time window, the initial state of the magnetization vector in each quantitative parameter combination is the same, that is, determined according to the matching process of the previous time window. Specifically, in each voxel, the initial state of the magnetization vector of the first time window is a thermal equilibrium state.

[0100] Based on this, the dictionary matching unit 102 is used to compare the evolution curve corresponding to each time window segment with the simulated evolution curve in the dictionary of the time window, and use the quantitative parameter corresponding to the evolution curve in the dictionary with the highest similarity as the quantitative value of the time window. The matching result includes the quantitative value of each time segment in each voxel. Furthermore, the image generation unit 103 is used to generate the quantitative map of the time segment based on the quantitative values ​​of all voxels within the same time segment.

[0101] In summary, the information acquisition unit 100, information processing unit 101, dictionary generation unit 104, dictionary matching unit 102, and image generation unit 103 in the imaging device provided in this embodiment are used to respectively realize information acquisition, time window division, dictionary generation, dictionary matching, and quantitative map generation according to the dynamic quantitative imaging method based on magnetic resonance fingerprint imaging. Based on this, a quantitative map can be obtained for each time period corresponding to the time window, thereby forming a quantitative map spectrum with high temporal resolution. It can not only acquire quantitative maps for each time period but also acquire dynamic imaging change information based on time changes, thereby improving the comprehensiveness of the acquired imaging information. For details not covered in this embodiment, please refer to other embodiments.

[0102] <Example 11>

[0103] This embodiment provides a computer storage medium. The computer storage medium stores executable instructions, which, when executed by a processor, cause the processor to perform the steps in the dynamic quantitative imaging method based on magnetic resonance fingerprint imaging.

[0104] Based on this, the computer storage medium provided in this embodiment divides the time series of each voxel into multiple time windows when executing instructions. Dictionary matching is performed on each time window to obtain multiple matching results. Based on these multiple matching results, multiple quantitative maps based on time axis changes can be obtained, thereby extracting dynamic imaging changes of the examined tissue within the detection time period, achieving dynamic quantitative imaging, and helping to improve the guiding value of magnetic resonance imaging for clinical treatment. For details not covered in this embodiment, please refer to other embodiments.

[0105] <Example Twelve>

[0106] This embodiment provides a magnetic resonance imaging (MRI) device, including the imaging device and / or the computer storage medium. In other words, the MRI device may include only the imaging device described in Embodiment 10, or only the computer storage medium described in Embodiment 11, or both the imaging device described in Embodiment 10 and the computer storage medium described in Embodiment 11.

[0107] Based on this, the magnetic resonance imaging device provided in this embodiment divides the time series of each voxel into multiple time windows during dynamic quantitative imaging. Dictionary matching is performed on each time window to obtain multiple matching results. Based on these multiple matching results, multiple quantitative maps based on time axis changes can be obtained, thereby extracting dynamic imaging changes in the examined tissue within the detection period, achieving dynamic quantitative imaging, and helping to improve the guiding value of magnetic resonance imaging for clinical treatment. For details not covered in this embodiment, please refer to other embodiments.

[0108] It should be noted that the various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. Similar or identical parts between embodiments can be referred to mutually. In addition, different parts between embodiments can also be combined with each other, and this invention does not limit this.

[0109] Furthermore, it should be understood that although the present invention has been disclosed above with reference to preferred embodiments, these embodiments are not intended to limit the present invention. For any person skilled in the art, many possible variations and modifications can be made to the technical solutions of the present invention based on the disclosed technical content, or equivalent embodiments can be modified accordingly, without departing from the scope of the present invention. Therefore, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention, without departing from the content of the present invention, shall still fall within the scope of protection of the present invention.

Claims

1. A dynamic quantitative imaging method based on magnetic resonance fingerprinting, characterized in that, include: Scanning is performed according to preset acquisition parameters to obtain time series of magnetic resonance images of all voxels; The time series corresponding to each voxel is divided into multiple time sliding windows; Each time window segment is matched with a corresponding dictionary for the time window segment; wherein, each time window segment in each voxel is provided with a corresponding dictionary; Based on the matching results, obtain a quantitative graph of the time period corresponding to each time window segment; The matching result includes the quantitative value of each time period in each voxel, and the quantitative map of the time period is generated based on the quantitative values ​​of all voxels in the same time period.

2. The dynamic quantitative imaging method based on magnetic resonance fingerprinting according to claim 1, characterized in that, Before matching each of the time windows with the corresponding dictionary of the time windows, the process includes: Based on the acquisition parameters corresponding to the time sliding window and a variety of preset quantitative parameter combinations, multiple evolution curves of the time sliding window are simulated. The dictionary for the time sliding window is determined based at least on the correspondence between the evolution curve and the combination of quantitative parameters; The quantitative parameter combination includes the initial state of the magnetization vector and a variety of quantitative parameters.

3. The dynamic quantitative imaging method based on magnetic resonance fingerprinting according to claim 2, characterized in that, The preset combination of multiple quantitative parameters includes at least one preset initial state of the magnetization vector.

4. The dynamic quantitative imaging method based on magnetic resonance fingerprinting according to claim 2, characterized in that, Within the same voxel, the initial state of the magnetization vector of the time window in this section is determined based on the matching result of the time window in the previous section; and the initial state of the magnetization vector of the time window in this section is used as the initial state of the magnetization vector in the preset combination of multiple quantitative parameters.

5. The dynamic quantitative imaging method based on magnetic resonance fingerprinting according to claim 4, characterized in that, In each voxel, the magnetization vector of the first time sliding window is initially in thermal equilibrium.

6. The dynamic quantitative imaging method based on magnetic resonance fingerprinting according to claim 2, characterized in that, The process of matching each of the time windows with the corresponding dictionary of the time windows includes: The evolution curve corresponding to each time sliding window segment is compared with the simulated evolution curve in the dictionary of the time sliding window, and the quantitative parameter corresponding to the evolution curve in the dictionary with the highest similarity is taken as the quantitative value of the time sliding window.

7. The dynamic quantitative imaging method based on magnetic resonance fingerprinting according to claim 1, characterized in that, The length of each time window segment is greater than or equal to the step size of the time window.

8. An imaging device, characterized in that, The device is used to perform the dynamic quantitative imaging method based on magnetic resonance fingerprint imaging as described in any one of claims 1 to 7; wherein the imaging device includes an information acquisition unit (100), an information processing unit (101), a dictionary matching unit (102), and an image generation unit (103). The information acquisition unit (100) is used to scan according to preset acquisition parameters to obtain the magnetic resonance image time series of all voxels; The information processing unit (101) is used to divide the time series corresponding to each voxel into multiple time sliding windows; The dictionary matching unit (102) is used to match each segment of the time window with the dictionary of the corresponding time window; The image generation unit (103) is used to obtain a quantitative image of the time period corresponding to each time window segment based on the matching result.

Images