Image processing method, transmit RF parameter determination method, magnetic resonance imaging system, and image processing apparatus

Abstract

To improve the image quality.SOLUTION: An image processing method comprises: obtaining first and second sets of image data generated by performing a series of MRI sequences using first and second different transmit RF parameters; calculating, from the first and second sets of image data, a correlation map indicative of correlation between the first and second sets of image data; and correcting the first set of image data using the correlation map to obtain a corrected image.SELECTED DRAWING: Figure 3

Description

[Technical field]

[0001] The embodiments disclosed herein and in the drawings relate to an image processing method, a transmit RF parameter determination method, a magnetic resonance imaging system, and an image processing device. The embodiments disclosed herein generally relate to a method, system, and computer program product for acquiring a corrected magnetic resonance imaging (MRI) image, and in one embodiment, a method, system, and computer program product for acquiring a corrected MRI image by performing a series of MRI scan sequences using at least two RF transmit parameters to acquire first and second image data sets in a single MRI examination. [Background technology]

[0002] When imaging a patient using an MRI system, the patient enters the gantry of the MRI system and undergoes a series of scan sequences (see FIG. 1, from scan sequence 1 to scan sequence n After the patient has completed a series of scan sequences, the patient exits the MRI system.

[0003] Known MRI systems are optionally calibrated for each patient before the start of the examination, as shown in the optional pre-scan sequence 102 of FIG. 1, to determine a reference RF transmit gain level (RFL (RF level): reference RFL value). During the pre-scan process, the RFL is calculated for at least one entire slice of a set of 2D slices, thereby calculating an RFL that represents the average of the voxels in the slice. The reference RFL value is then used to calculate transmit gain values ​​to achieve the desired B1 amplitude for all subsequent RF pulses in the MRI scan sequence. As will be appreciated by those skilled in the art, the B1 amplitude of an RF pulse is directly related to the flip angle of the RF pulse. Also, the flip angle of the RF pulse is the primary determinant of MR signal strength and image contrast. That is, the RFL value has a known direct relationship with the B1 magnetic field (transmit magnetic field) and the flip angle. By measuring the B1 magnetic field, the RFL can be calculated directly and easily. Generally, if the RFL is too low, "undertipping" occurs. Under-chipping is a condition in which data is acquired at a flip angle that is lower than expected. Conversely, if the RFL is too high, "over-chipping" occurs. Over-chipping is a condition in which data is acquired at a flip angle that is higher than expected. Both under-chipping and over-chipping can result in poor quality MRI images (e.g., black dots in the image).

[0004] However, even within a single MRI image, non-uniform B1 (ΔB1) can result in non-uniform image signals. Non-uniform transmit fields can result in non-constant flip angles and signal variations. Such effects have been observed at high field strengths (e.g., 3 T and above) where there is a large B1 variation due to dielectric effects. Variations in effective B1 levels result in dark or bright areas in the image compared to areas of the image with the desired B1 level. In conventional spin echo or fast spin echo imaging, the expected flip angle of the excitation pulse is π / 2 and the expected flip angle of the refocusing pulse is π. In this example, the dark areas in the image are due to under-chipping or over-chipping, where the flip angle is below or above the expected flip angle, respectively. In conventional spin echo or fast spin echo imaging, the image signal is expressed as sin 3 Since it is roughly proportional to ((π / 2)*(1+ΔB1)), for example, in the case of -30% under-chipping, the internal term (argument) of sin is (1+ΔB1=1-0.3=0.7), so the signal = sin 3 (0.7π / 2) = 0.71. For example, in the case of +20% overchipping, the internal term (argument) of sin is (1 + ΔB1 = 1 + 0.2 = 1.2), and the signal = sin 3 (1.2π / 2)=0.86. In the field-echo imaging example, the relationship between signal, flip angle, and tissue T1 relaxation is more complex. Thus, in field-echo imaging, dark (low signal) and bright (high signal) regions can appear due to the variation of the flip angle with ΔB1. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] U.S. Patent No. 8,198,891 [Patent Document 2] U.S. Patent No. 9,977,106 [Non-patent literature]

[0006] [Non-Patent Document 1] NEHRKE et al., “DREAM-A Novel Approach for Robust, Ultrafast, Multislice B1 Mapping”, Magnetic Resonance in Medicine, 2012, 10 pages [Non-Patent Document 2] INSKO et al., “Mapping of the Radiofrequency Field”, Journal of Magnetic Resonance, Series A, Volume 103, Issue 1, June 1,1993, 4 pages [Non-Patent Document 3] VASANAWALA et al., “Linear Combination Steady-State Free Precession MRI”, Magnetic Resonance in Medicine, 2000, 9 pages [Non-Patent Document 4] BANGERTER et al., “Analysis of Multiple-Acquisition SSFP”, Magnetic Resonance in Medicine, 2004, 10 pages [Non-Patent Document 5] YARNYKH, “Actual Flip-Angle Imaging in the Pulsed Steady State: A Method for Rapid Three-Dimensional Mapping of the Transmitted Radiofrequency Field”, Magnetic Resonance in Medicine, 2007, 9 pages [Non-Patent Document 6] ORZADA et al., “RF Excitation Using Time Interleaved Acquisition of Modes (TIAMO) to Address B1 Inhomogeneity in High-Field MRI”, Magnetic Resonance in Medicine, 2010, 7 Pages [Non-Patent Document 7] SACOLICK et al., “B1 Mapping by Bloch-Siegert Shift”, Magnetic Resonance in Medicine, 2010, 8 pages [Non-Patent Document 8] ORZADA et al., “Time-Interleaved Acquisition of Modes: An Analysis of SAR and Image Contrast Implications”, Magnetic Resonance in Medicine, 2012, 9 pages [Non-Patent Document 9] FOXALL, “Frequency-Modulated Steady-State Free Precession Imaging”, Magnetic Resonance in Medicine, 2002, 7pages Summary of the Invention [Problem to be solved by the invention]

[0007] One of the problems to be solved by the embodiments disclosed in this specification and the drawings is to improve image quality. However, the problems to be solved by the embodiments disclosed in this specification and the drawings are not limited to the above problem. Problems corresponding to the effects of each configuration shown in the embodiments described later can also be positioned as other problems. [Means for solving the problem]

[0008] An image processing method according to an embodiment acquires a first image data set and a second image data set generated by executing a series of MRI sequences using first transmit RF parameters and second transmit RF parameters that are different from each other, calculates a correlation map indicating a correlation between the first image data set and the second image data set from the first image data set and the second image data set, and acquires a corrected image by correcting the first image data set using the correlation map. [Brief description of the drawings]

[0009] [Figure 1] FIG. 1 illustrates a patient examination including an optional pre-scan and a series of scan sequences. [Diagram 2] FIG. 2 is a schematic block diagram of an MRI system according to one embodiment of the present disclosure. [Diagram 3] FIG. 3 is a flow chart illustrating an example of a process for acquiring a corrected MRI image by performing spatially resolved B1 amplitude-based adjustment. [Figure 4A] FIG. 4A shows images of spatially resolved B1 amplitude data acquired during determination of a B1 amplitude distribution that may occur when imaging two human thighs with an initial RF transmit gain level (RFL) as the initial RF parameter. [Figure 4B] FIG. 4B shows MRI images of the two thighs of FIG. 4A taken with initial RFL. [Figure 4C] FIG. 4C shows MRI images of the two thighs of FIGS. 4A and 4B taken while gradually increasing the RFL above the initial RFL of FIG. 4B. [Figure 4D] FIG. 4D shows MRI images of the two thighs of FIGS. 4A and 4B taken while gradually increasing the RFL above the initial RFL of FIG. 4B. [Figure 4E] FIG. 4E shows MRI images of the two thighs of FIGS. 4A and 4B taken while gradually increasing the RFL above the initial RFL of FIG. 4B. [Figure 4F]FIG. 4F shows MRI images of the two thighs of FIGS. 4A and 4B taken while gradually increasing the RFL above the initial RFL of FIG. 4B. [Figure 4G] FIG. 4G shows MRI images of the two thighs of FIGS. 4A and 4B taken while gradually increasing the RFL above the initial RFL of FIG. 4B. [Figure 4H] FIG. 4H shows MRI images of the two thighs of FIGS. 4A and 4B taken while gradually increasing the RFL above the initial RFL of FIG. 4B. [Figure 4I] FIG. 4I shows a grayscale image of the spatially resolved B1 amplitude data of FIG. 4A. [Figure 4J] FIG. 4J is an annotated grayscale image of the spatially resolved B1 amplitude data of FIG. 4I showing two dark regions representing B1 amplitude data below the desired average value. [Figure 5A] FIG. 5A is a map showing a distribution of spatially resolved B1 amplitude values ​​illustrating the variation of B1 amplitude in a portion of an imaged object. [Figure 5B] FIG. 5B is an annotated image of a map showing the distribution of spatially resolved B1 amplitude values ​​from FIG. 5A, illustrating a circular region of interest selected for use in determining the amount of correction to the initial RF imaging parameters. [Figure 5C] FIG. 5C shows a grayscale image of the spatially resolved B1 amplitude map of FIG. 5A. [Figure 5D] FIG. 5D is an annotated grayscale image of a map showing the distribution of spatially resolved B1 amplitude values ​​from FIG. 5A, showing a circular region of interest selected for use in determining the amount of correction to the initial RF imaging parameters. [Figure 6A] FIG. 6A is a diagram showing an example of a map showing the distribution of B1 amplitude values ​​of the subject's thigh. [Figure 6B] FIG. 6B illustrates a B1 amplitude projection corresponding to the projection direction indicated by line 605 in FIG. 6A. [Figure 6C] FIG. 6C illustrates a B1 amplitude projection corresponding to the projection direction represented by line 610 in FIG. 6A. [Figure 6D] FIG. 6D illustrates a B1 amplitude projection corresponding to the projection direction represented by line 615 in FIG. 6A. [Figure 6E] FIG. 6E illustrates a B1 amplitude projection corresponding to the projection direction represented by line 620 in FIG. 6A. [Figure 6F] FIG. 6F is a diagram showing an example of the B1 grayscale map of FIG. 6A. [Figure 7A] FIG. 7A is a data flow diagram showing an example of the flow of data corresponding to steps 330 and 335 of FIG. 3 for generating a corrected image according to the first embodiment. [Figure 7B] FIG. 7B is a data flow diagram showing an example of the flow of data for generating a corrected image according to the second embodiment. [Figure 8A] FIG. 8A is an illustration of a corrected image (compare with FIG. 4B) generated by the first method of combining multiple images of the same resolution acquired at different RFLs. [Figure 8B] FIG. 8B is a diagram of an original image (taken at a different RFL than FIG. 4B) used to generate the corrected image of FIG. 8A. [Figure 8C] FIG. 8C is a diagram of an original image (taken at a different RFL than FIG. 4B) used to generate the corrected image of FIG. 8A. [Figure 9A] FIG. 9A is an illustration of a corrected image (compare with FIG. 4B) generated by a second method of combining multiple images acquired at different RFLs and different resolutions. [Figure 9B] FIG. 9B is a diagram of the first and second original images taken at different RFLs and at higher and lower resolutions, respectively, from FIG. 4B and used to generate the corrected image of FIG. 9A. [Figure 9C] FIG. 9C is a diagram of the first and second original images taken at different RFLs and at higher and lower resolutions, respectively, from FIG. 4B, and used to generate the corrected image of FIG. 9A. [Figure 10A] FIG. 10A is an illustration of a downsampled image of a full resolution image obtained using a first RFL different from the RFL of FIG. 4B. [Figure 10B]FIG. 10B is an illustration of a low-resolution image acquired using a second RFL different from the RFL of FIG. 4B. [Figure 10C] FIG. 10C is a diagram of a RatioMap of the images of FIGS. 10A and 10B. [Figure 11A] FIG. 11A is a map showing the distribution of amplitude values ​​for the human pelvis. [Figure 11B] FIG. 11B is a diagram showing an example of a ratio map of the pelvis of the person of FIG. 11A. [Figure 11C] FIG. 11C is an illustration of an image of a human pelvis acquired using an RFL such that the average B1 level of the resulting images is at a target level of 1.0. [Figure 11D] FIG. 11D is an illustration of a corrected image obtained by the methods described herein. [Figure 11E] FIG. 11E is an image of a human pelvis acquired with an RFL such that the average B1 level of the resulting images is 0.8 times the target level of 1.0. [Figure 11F] FIG. 11F is an image of a human pelvis acquired with an RFL such that the average B1 level of the resulting images is 1.40 times the target level of 1.0. [Figure 11G] FIG. 11G shows an amplitude grayscale map corresponding to the map showing the distribution of amplitude values ​​of FIG. 11A. [Figure 11H] FIG. 11H is an annotated image of the amplitude grayscale map of FIG. 11G. [Figure 12] FIG. 12 is an illustration of two leg images corrected by the first ratio-based correction function. [Figure 13] FIG. 13 is a graph showing the relationship between B1 value and the resulting signal used in a fast spin echo environment at the first RFL. [Figure 14A] FIG. 14A is a graph showing the relationship between B1 value and the resulting signal used in a fast spin echo environment at the second and third RFL. [Figure 14B] FIG. 14B is a graph showing the ratios of the curves of FIG. 14A. [Figure 14C] FIG. 14C is a graph showing the relationship between estimated B1 values ​​and the ratio curve of FIG. 14B. [Figure 14D] FIG. 14D is a graph showing a correction function applied to an image based on the estimated B1 value obtained by applying the ratio of FIG. 14C. [Figure 15A] FIG. 15A shows an original image of two human thighs before correction. [Figure 15B] FIG. 15B shows the image of FIG. 15A corrected using a ratio-based correction function described herein. [Figure 16A] FIG. 16A shows an original image of a human pelvis before correction. [Figure 16B] FIG. 16B shows a corrected image of the image of FIG. 16A using the ratio-based correction function described herein. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0010] Several embodiments are illustrated by way of example in the referenced figures of the drawings. The embodiments and figures disclosed herein are to be considered illustrative rather than restrictive. The scope of the technology and subsequent claims is not to be limited to the examples shown and described herein.

[0011] Also, the embodiments are described primarily with respect to certain processes, systems, and computer program products provided in certain examples. However, the processes, systems, and computer program products may also operate effectively in other examples. The terms "embodiment," "one embodiment," "another embodiment," and the like may refer to the same or different embodiments. The embodiments are described with respect to methods and configurations that include certain components. However, such methods and configurations may include more or fewer components than those shown. Also, the arrangement and type of components may be modified without departing from the scope of the disclosure.

[0012] Also, the embodiments are described in the context of methods having certain steps, by way of example, although the methods and compositions may still operate validly by adding steps or performing the steps in a different order consistent with the example embodiments. Thus, the present disclosure is not intended to be limited to the embodiments described, but rather to be accorded the widest scope consistent with the principles and features described herein and is limited only by the appended claims.

[0013] Furthermore, when a range of values ​​is given, each value between the upper and lower limits of the range, as well as any other stated or intervening values ​​within the range, are included in the disclosure. When a stated range includes both an upper and lower limit, the range excluding either of the upper and lower limits is also included. Unless otherwise stated, terms used herein have their plain and ordinary meaning as understood by those of ordinary skill in the art. Any definitions are intended to aid the reader in understanding this disclosure and are not intended to modify or limit the meaning of such terms unless otherwise stated.

[0014] Reference will now be made to the drawings, in which like reference numbers indicate the same or equivalent parts throughout the drawings. FIG. 2 is a block diagram showing the overall configuration of an MRI system 1 according to a first embodiment. The MRI system (MRI apparatus) 1 of the first embodiment includes a gantry 100, a control cabinet 300, a console 400, and a bed 500.

[0015] The gantry 100 includes a static magnetic field magnet 10, a gradient coil assembly 11, and a whole body (WB) coil 12. These components are housed in a cylindrical housing. The bed 500 includes a bed body 50 and a top plate 51. The MRI system 1 further includes at least one RF coil 20 disposed in the vicinity of the object.

[0016] The control cabinet 300 includes three gradient coil power supplies 31 (X-axis gradient coil power supply 31x, Y-axis gradient coil power supply 31y, and Z-axis gradient coil power supply 31z), an RF receiver 32, an RF transmitter 33, and a sequence controller 34.

[0017] The static magnetic field magnet 10 of the gantry 100 has a generally cylindrical shape and generates a static magnetic field inside a bore. The bore is a space formed in a cylindrical structure and serves as an imaging region for an object (e.g., a patient). The static magnetic field magnet 10 has a superconducting coil inside. The superconducting coil is cooled to an extremely low temperature by liquid helium. In an excitation mode, the static magnetic field magnet 10 generates a static magnetic field by supplying a current to the superconducting coil from a static magnetic field power supply (not shown). The static magnetic field magnet 10 then transitions to a persistent current mode, and the static magnetic field power supply is disconnected. The static magnetic field magnet 10 transitioned to the persistent current mode continues to generate a strong static magnetic field for a long period of time, such as one year. The static magnetic field magnet 10 may be configured as a permanent magnet.

[0018] The gradient coil assembly 11 also has a substantially cylindrical shape and is fixed inside the static magnetic field magnet 10. The three-channel gradient coil assembly 11 has an X-axis gradient coil 11x, a Y-axis gradient coil 11y, and a Z-axis gradient coil 11z (not shown individually). The X-axis gradient coil 11x receives current supply from an X-axis gradient coil power supply device 31x to generate a gradient magnetic field Gx in the X-axis direction. The Y-axis gradient coil 11y receives current supply from a Y-axis gradient coil power supply device 31y to generate a gradient magnetic field Gy in the Y-axis direction. The Z-axis gradient coil 11z receives current supply from a Z-axis gradient coil power supply device 31z to generate a gradient magnetic field Gz in the Z-axis direction.

[0019] The bed body 50 of the bed 500 can move the top board 51 vertically and horizontally. The bed body 50 moves the top board 51, on which an object is placed, to a predetermined height before imaging. Then, when imaging the object, the bed body 50 moves the top board 51 horizontally to move the object into the bore.

[0020] The whole-body coil 12 has a substantially cylindrical shape surrounding the object, and is fixed inside the gradient coil assembly 11. The WB coil 12 applies an RF pulse transmitted from an RF transmitter 33 to the object, and receives a magnetic resonance (MR) signal emitted from the object due to excitation of hydrogen nuclei. The whole-body coil 12 may be referred to as a transmit coil.

[0021] The RF coil 20 receives MR signals emitted from an object at a position near the object. The RF coil 20 includes, for example, a plurality of coil members. There are various RF coils 20 for the head, chest, spine, lower limbs, whole body, etc., according to the anatomical imaging part of the object. Among the various RF coils, FIG. 2 shows an RF coil 20 for chest imaging.

[0022] The RF transmitter 33 transmits RF pulses to the WB coil 12 based on instructions from the sequence controller 34. The RF receiver 32 detects MR signals received by the WB coil 12 and / or the RF coil 20, converts the detected MR signals into digital signals, and transmits the raw data acquired by the detection to the sequence controller 34.

[0023] The sequence controller 34 scans an object by driving the gradient coil power supply 31, the RF transmitter 33, and the RF receiver 32 under the control of the console 400. The sequence controller 34 receives raw data from the RF receiver 32 by performing a scan, and transmits the received raw data to the console 400.

[0024] The sequence controller 34 includes a processing circuit (not shown). The processing circuit is configured as, for example, a processor that executes a predetermined program, or hardware such as a field programmable gate array (FPGA) or an application specific integrated circuit (ASIC).

[0025] The console 400 is configured as a computer comprising a processing circuit 40 , a memory 41 , a display 42 and an input interface 43 .

[0026] The memory 41 is a recording medium including a read-only memory (ROM) and a random access memory (RAM) in addition to an external memory device such as a hard disk drive (HDD) or an optical disk device. The memory 41 stores various programs executed by the processor of the processing circuit 40, as well as various data and information.

[0027] The display 42 is a display device such as a liquid crystal display panel, a plasma display panel, or an organic EL panel.

[0028] The input interface 43 includes various devices for an operator to input various data and information, and is composed of, for example, a mouse, a keyboard, a trackball, and / or a touch panel.

[0029] The processing circuit 40 is, for example, a circuit equipped with a central processing unit (CPU) and / or a special-purpose or general-purpose processor. The processor executes various programs stored in the memory 41 to realize various functions as described below. The processing circuit 40 can be configured by hardware such as an FPGA and an ASIC. The various functions as described below can also be realized by such hardware. The processing circuit 40 can also realize various functions by combining hardware processing and software processing based on its own processor and programs. The processing circuit 40 may be referred to as an RF transmit coil controller.

[0030] The console 400 controls the entire MRI system 1 by these components. Specifically, the console 400 accepts imaging conditions such as the type of pulse sequence, various information, and an imaging start command input by a user such as a medical imaging technician via an input interface 43 having a mouse and keyboard. The processing circuitry 40 causes the sequence controller 34 to execute a scan according to the input imaging conditions, and reconstructs an image based on the raw data, i.e., the digital MR signal, transmitted from the sequence controller 34. The reconstructed image is displayed on the display 42 and stored in the memory 41.

[0031] In the configuration of the MRI system 1 shown in FIG. 2, the control cabinet 300, the gantry 100 and the bed 500 (ie, all components except the console 400) constitute an imaging section or scanner.

[0032] 3 is a flow chart illustrating an exemplary method 305. Method 305 includes an example procedure for acquiring a corrected MRI image by acquiring at least two sets of image data (e.g., a first image data set and a second image data set) using at least two different transmit RF parameters (e.g., a first RF parameter and a second transmit RF parameter). Method 305 includes main steps 325, 330, 335 and several optional steps (shown within dashed boxes), which are described in detail below.

[0033] In step 325, at least two transmit RF parameters are acquired (e.g., retrieved from a memory or storage device (by itself or as part of a scan sequence for a series of MRI sequence acquisitions)) for imaging the same portion of the object to be imaged (e.g., for imaging the same slice of the patient). The at least two transmit RF parameters for acquiring the first and second image data sets, respectively, are, for example, a first and second RFL, a first and second RF gain, a first and second RF phase, a first and second RF shimming setting, etc. Specifically, the at least two transmit RF parameters are a first and second RF transmit gain level that are different from each other. Furthermore, the at least two transmit RF parameters may include any combination of these parameters (e.g., gain 1 and phase 1 for acquiring a first image data set for a slice and gain 2 and phase 2 for acquiring a second image data set for the same slice). As used herein, first and second "phases" refer to the phase between real and imaginary components of the RF signal generated by the transmitter, as opposed to a periodic pattern of phases.

[0034] From step 325, control proceeds to step 330 where at least two image data sets are acquired that were generated by running a series of MRI sequences using at least two transmit RF parameters. After step 330, control proceeds to step 335 where a first set of the at least two image data sets is corrected to generate a corrected MRI data image. This concludes the method 305.

[0035] The at least two transmit RF parameters can be obtained by a number of different techniques. In a first technique, the method 305 obtains the at least two transmit RF parameters by steps 310 to 320 of the figure. In step 310, the MRI system 1 (or a remote system in communication and cooperation with the MRI system 1) obtains spatially-resolved B1 amplitude measurements of an object (e.g., a person being imaged) (e.g., at an initial RFL set according to multiple MRI sequences acquired for the object). For example, acquisition processing of spatially resolved B1 amplitude data (spatially resolved B1 amplitude measurements) may include, but is not limited to, performing Bloch-Siegert Shift (BSS) processing, Actual Flip Angle Imaging (AFI) method, Dual Refocusing Echo Acquisition Mode (DREAM), Double Angle Method (DAM) processing, Phase-sensitive (PS) processing, Saturation Recovery (SR) processing, etc. Spatially resolved B1 amplitude measurements may be acquired as 1D projections, 2D slices, or 3D volumes. A subset of measurements may be used rather than the complete set of measurements.

[0036] The spatially resolved B1 amplitude measurements correspond, for example, to a B1 map. The B1 map may be expressed as a one-dimensional B1 map showing the distribution of B1 in one dimension, a two-dimensional B1 map showing the distribution of B1 in two dimensions, or a three-dimensional B1 map showing the distribution of B1 in three dimensions. The one-dimensional B1 map corresponds to the one-dimensional projection. The two-dimensional B1 map corresponds to the 2D slice. The three-dimensional B1 map corresponds to the 3D volume.

[0037] The processing circuitry 40 obtains spatially resolved B1 amplitude measurements obtained using at least one RF parameter, such as a flip angle, a shimming setting, or a gain level of an RF transmission.

[0038] The acquisition of the spatially resolved B1 amplitude measurements is not limited to the acquisition by the processing circuitry 40, and may be acquired by the sequence controller 34. The processing circuitry 40 acquires the spatially resolved B1 amplitude measurements (B1 values) using a Bloch-Siegert shift sequence as described above. The processing circuitry 40 may acquire the spatially resolved B1 amplitude measurements by Actual Flip Angle imaging as described above. The processing circuitry 40 also acquires 1D projections, 2D slices, or 3D volumes.

[0039] As used herein, "acquire" means to receive from the memory 41 or storage system of the MRI system 1, or entirely from an external source. Also, "acquire" means that a portion of information is acquired from one source and combined with additional information or modified so that the combined / modified information is accessible to the MRI system 1 (e.g., stored in the memory 41 or storage system of the MRI system 1). For example, a series of MRI sequences may be acquired from one source (memory 41 or an external device) and combined with additional information, such as a spatially resolved B1 amplitude measurement sequence. Similarly, a series of MRI sequences may be acquired from one source (memory 41 or an external device) and modified to change at least one initial RF imaging parameter.

[0040] FIG. 4A illustrates spatially resolved B1 amplitude data acquired (e.g., by the BSS or AFI method of B1 mapping) during the determination of a possible B1 amplitude distribution when imaging two human thighs at an initial RFL level (i.e., an RFL where the expected average B1 value for the entire imaging region is 1.0). As shown in FIG. 4A, a complete B1 amplitude map (B1 map) corresponding to a particular cross-section of an object (e.g., a person) to be imaged is acquired, although the acquired B1 amplitude map for the cross-section may be less than complete. The B1 amplitude data may be acquired during a portion of a pre-scan where B1 shimming calibration is performed, so as not to increase the overall processing time of the MRI process. Alternatively, a fast projection method requiring a short scan time (e.g., about 60 to 80 ms) may be utilized without B1 shimming calibration. In yet another embodiment, the B1 amplitude data may be acquired after a portion of a pre-scan where B1 shimming calibration is performed, so as to match the B1 amplitude data as closely as possible to the pre-acquisition of spatially resolved B1 amplitude measurements. In another embodiment, complete or partial B1 amplitude maps for other cross-sections may be acquired.

[0041] FIG. 4B shows MRI images of the two thighs of FIG. 4A imaged with initial RFL. FIG. 4B includes a symbol "-" added to the upper right of the left thigh (left thigh) to highlight the shading on the left thigh. This shading corresponds to the blue and green portions of the B1 map of FIG. 4A and indicates that the B1 amplitude data does not reach the B1 target value of 1.0 (shown in yellow in FIG. 4A) and is below the B1 target value causing signal distortion due to under-chipping. The green portion corresponds to about 80% of the target B1, and the blue portion corresponds to about 60% of the target B1. The amplitude values ​​decrease toward the center of the blue portion and increase away from the center of the blue portion. The blue and green portions of FIG. 4A correspond to the dark portions of the B1 amplitude grayscale map of FIG. 4I. These areas are highlighted on the annotated B1 amplitude map of FIG. 4J. In the B1 map of Figure 4J, the outer white ellipse corresponds to the green region with a B1 value of approximately 80% of the target B1, and the inner white ellipse corresponds to the blue region with a B1 value of approximately 60% of the target B1. In Figure 4J, the amplitude values ​​decrease toward the center of the inner ellipse and increase away from the center of the inner ellipse.

[0042] 4C-4H show MRI images of the two thighs of FIGS. 4A and 4B, imaged at RFLs gradually higher than the initial RFL (i.e., 1.05, 1.10, 1.15, 1.20, 1.25, 1.30 times the initial RFL). As is evident from FIG. 4C, when imaged at 1.05 times the RFL of FIG. 4B (i.e., 1.05 times), the portion of the image containing signal distortion due to under-chipping in FIG. 4B shows an improvement (i.e., reduced shading) compared to FIG. 4B, but the shading still remains. As the RFL increases to 1.10× (1.10 times), 1.15× (1.15 times), the under-chipping effect continues to decrease until about 1.20× (1.20 times). At about 1.20× (1.20 times), the shading starts to disappear (indicated by the symbol “=”) in FIG. 4F. However, increasing the RFL produces over-chipping in the right thigh from at least an RFL of 1.15× (Figure 4E) and deteriorates other parts of the image (indicated by the symbol "+"). In Figures 4G and 4H, even with a higher power RFL, over-chipping is seen in at least two areas (indicated by the symbol "+").

[0043] Returning to the description of method 305 of FIG. 3, once the spatially resolved B1 amplitude measurements have been acquired in step 310, control proceeds to step 315, where a spatially-resolved analysis is performed on the acquired spatially resolved subsets of B1 amplitude data to address at least one of under-chipping and over-chipping issues that may result from an initial RFL of the acquired series of MRI sequences. The spatially-resolved analysis on the acquired spatially resolved B1 amplitude data results in at least two B1 data subsets, as described in more detail below. The subsets may include, but are not limited to, a first subset representing a first B1 parameter range, such as locations of low B1 values, a second subset representing a second B1 parameter range, such as locations of high B1 values, and the like. In one embodiment, the received spatially resolved B1 amplitude measurements are quantized and optionally filtered before performing histogramming to generate the at least two subsets. In the histogramming, for example, the acquired B1 amplitude data is quantized to different values ​​and used to calculate a histogram.

[0044] Performing at least two spatially resolved analyses may, for example, correspond to obtaining a first and a second subset of spatially resolved B1 amplitude measurements using a histogram, as described above, where the histogram corresponds, for example, to a frequency distribution of B1 values.

[0045] Prior to the histogram calculation, filtering may be performed on the B1 amplitude data. For example, B1 amplitude values ​​below a certain threshold are removed since they correspond to air, not tissue, which is of interest to the image reviewer (e.g., a medical professional such as a radiologist). Similarly, values ​​above a statistical threshold (e.g., 4 standard deviations) from the mean B1 amplitude value are removed since they are likely anomalous data. Additionally, a position-based mask may be applied to the data to remove any values ​​outside the mask region. The mask region may be automatically selected by the system based on factors such as the type of region being imaged (e.g., thigh, prostate, skull, pelvis, etc.) and / or patient-specific parameters (e.g., patient weight, which affects the size / location of the region being imaged relative to the center of the region being imaged), obtained via input from an operator of the MRI system 1 or stored in memory 41. Alternatively, the B1 amplitude data may be quantized after filtering.

[0046] The occurrence of the quantized data values ​​across the set of B1 amplitude values ​​being processed is determined from the histogram and processing based on the calculated information is performed in step 320. As will be explained in more detail below, the set of B1 amplitude values ​​being processed is derived from at least one 1D projection, partial or complete slice (i.e., two-dimensional data), or partial or complete three-dimensional volume. In other words, the plurality of B1 amplitude values ​​being processed is derived from a B1 map.

[0047] In a first embodiment of step 320, at least two extreme value subsets in the B1 amplitude data are obtained by performing a spatially resolved process, which is a process of spatially resolved analysis. For example, a minimum X% (e.g., 10% or 15%) of the B1 amplitude data is selected as a first subset and used to increase the average B1 value generated during imaging (herein, "X" is considered as the threshold percentage of the first extreme value). As a second subset, the high extreme values ​​are selected instead of the low extreme values ​​(e.g., a maximum Y% such as 10% or 15%) and used to decrease the average B1 value generated during imaging.

[0048] Obtaining the first and second subsets corresponds, for example, to using the histogram to obtain a first percent of the smallest spatially resolved B1 amplitude measurements as the first subset and a second percent of the largest spatially resolved B1 amplitude measurements as the second subset.

[0049] Based on the data from the at least two subsets, the image processing method determines at least two transmit RF parameters to be used in producing the corrected MRI image. In such an embodiment, the first transmit RF parameter is the predetermined target volume (B1 target , e.g., 1.0). target The value is B1 target The values ​​are modified by dividing them by the average value of the first subset. In an exemplary embodiment, the first subset is a histogram (the average value is B1 low region Therefore, on average, the B1 amplitude value is the minimum of X% of the B1 amplitude value of the target The initial RFL that generates the value (RFL initial New low-range RFL (NewRFL) low region ) is calculated using the following formula (1).

[0050] New RFL low region =RFL initial ×(B1 target ) / (B1 low region ) (Formula 1)

[0051] As a result of such a change, the B1 of the dark hole-like regions increases, but some regions may become over-chipped.

[0052] In another embodiment, to avoid over-chipping, B1 target The value is initially reduced by a reduction factor (RFx, e.g., 0.85) to avoid excessively changing the image quality due to very low B1 values ​​in the low B1 region. In such an embodiment, the new low region RFL is calculated by the following equation (2):

[0053] New RFL low region =RFL initial ×(RFx×B1 target ) / (B1 low region ) (Formula 2)

[0054] In yet another embodiment, the new low region RFL value is limited by a predefined maximum percentage (maxfactor), for example using a maximum function (max(x,y)) whose result is less than or equal to x. Such an embodiment is described by equation (3) below.

[0055] New RFL low region =max(RFL initial ×maxfactor,RFL initial ×(B1 target ) / (B1 low region )) (Formula 3)

[0056] The second transmit RF parameter is B1 target The value is modified (or determined) by dividing the value by the average value of the second subset. In an exemplary embodiment, the second subset is a histogram (the average value is B1 high region Therefore, on average, the B1 amplitude is X% of the maximum. target The initial RFL that generates the value (RFL initial New high-range RFL (NewRFL) high region ) is calculated using the following formula (4):

[0057] New RFL high region =RFL initial ×(B1 target ) / (B1 high region ) (Equation 4)

[0058] Such a change results in a lower B1 in over-chipped areas, but some areas may become under-chipped.

[0059] In another embodiment, to reduce underchipping, B1 targetThe value is initially increased by a modification factor (MFx, e.g., 1.2) to avoid overly altering the image quality due to very high B1 values ​​in the high B1 region. In such an embodiment, the new high region RFL is calculated as shown in equation (5) below.

[0060] New RFL high region =RFL initial ×(RFx×B1 target ) / (B1 high region ) (Equation 5)

[0061] In yet another embodiment, the new high region RFL value is limited by a predetermined minimum percentage (minfactor), for example using a minimum function (min(x,y)) whose result is at least x. Such an embodiment is described by equation (6) below.

[0062] New RFL high region =min(RFL initial ×minfactor,RFL initial ×(B1 target ) / (B1 high region )) (Formula 6)

[0063] Therefore, processing circuitry 40 selects the first transmit RF parameter and the second transmit RF parameter by performing at least two spatially resolved analyses on the spatially resolved B1 amplitude measurements. For example, processing circuitry 40 selects the first transmit RF parameter using the first subset and the second transmit RF parameter using the second subset.

[0064] After determining the at least two transmit RF parameters (e.g., the first and second new RFLs) to be used for the series of MRI sequences, the determined at least two transmit RF parameters are stored in a memory or storage device (by itself or as part of a scan sequence for acquiring the series of MRI sequences). In one embodiment, steps 310 to 335 are performed within one MRI system 1. Alternatively, in another embodiment, some steps (e.g., steps 315 and 320) are performed outside the MRI system 1, and the MRI system 1 retrieves the stored at least two RF parameter values ​​from memory in step 325 (e.g., when acquiring the corresponding series of MRI sequences).

[0065] The analysis of the received spatially resolved B1 amplitude data is not limited to the above-described processing of steps 315 and 320. In alternative embodiments, the histogram processing is replaced by a number of alternative processing methods, including:

[0066] In a first alternative embodiment, the histogram of the complete slice is replaced by a region of interest process based on analysis (e.g., visual or automated inspection) of one-, two-, or three-dimensional spatially resolved B1 amplitude data. FIG. 5A shows a spatially resolved B1 amplitude map representing the change in B1 amplitude across a two-dimensional portion of the imaged object that can be obtained by the process of step 310. In an embodiment in which the B1 map of FIG. 5A is presented to the MRI operator, the MRI operator can select at least one region of interest on the B1 map using the input interface 43. FIG. 5B shows an example of a region of interest drawn on the B1 amplitude map by the MRI operator to indicate the region of interest (i.e., the system will select values ​​to use in the statistical process and determine high and / or low regions for initial RFL modification).

[0067] FIG 5C illustrates a spatially resolved B1 amplitude grayscale map representing the change in B1 amplitude across a two-dimensional portion of an imaged object that may be obtained by the process of step 310 and used in place of the B1 map of FIG 5A. In embodiments in which the B1 map of FIG 5C is presented to the MRI operator, the MRI operator may select at least one region of interest on the B1 map using the input interface 43. FIG 5D illustrates an example of a region of interest drawn on the B1 amplitude map by the MRI operator to indicate the region of interest (i.e., the system will select values ​​to use in statistical processing and determine high and / or low regions for initial RFL changes).

[0068] FIG. 5B is an annotated image of the spatially resolved B1 amplitude map (B1 map) of FIG. 5A showing a circular region of interest drawn or selected by the MRI operator. The circular region of interest is used to define a mask indicating where to select values ​​to determine the amount of correction applied to the initial RF imaging parameters. Shapes other than circular may be used to identify the two-dimensional region of interest. The shape, shape location and size may be automatically selected by the system based on factors such as the type of region being imaged (e.g., thigh, prostate, skull, pelvis, etc.) and / or patient-specific parameters (e.g., patient weight, which affects the size / location of the imaged region relative to the center of the imaged region). In such a configuration, the two-dimensional region of interest, even along a 1D projection, is used as a mask to filter out unwanted data outside the two-dimensional region of interest.

[0069] Alternatively, the MRI system 1 may collect the B1 data as 1D projection data. The geometric orientation of the 1D projection may be automatically selected by the system based on factors such as the type of region being imaged (e.g., thigh, prostate, skull, pelvis, etc.) and / or patient-specific parameters (e.g., patient weight, which affects the size / location of the region being imaged relative to the center of the imaged region).

[0070] In an embodiment utilizing an automated examination, the method utilizes at least one of (1) an artificial neural network trained to generate at least two correction factors from the spatially resolved B1 amplitude data and (2) a priori knowledge of expected locations of low / high B1 regions for automatic selection of at least one region of interest. For example, based on the region of the object to be scanned, the locations of the regions of interest exhibiting low / high B1 values ​​can be stored in memory 41 along with factors that may change the location of the regions of interest (e.g., type of region imaged and / or patient-specific parameters). For example, due to physical dielectric effects and the approximately cylindrical shape of most patients' legs, the approximate locations of low B1 (ΔB1<1) and high B1 (ΔB1>1) regions for all people of the same size are nearly identical. Thus, the locations of these regions are certain and known a priori. Some modification of the regions is possible based on a previous positioning image (e.g., a scout image). Positioning images are typically collected in an MR examination to guide the scan plan in the protocol.

[0071] 6A-6F show how the selection of 1D projection directions can improve the sensitivity of B1 measurements. The projection data shown in Fig. 6B-6E correspond to different projection directions (605, 610, 615, 620) of the B1 map in Fig. 6A and the B1 grayscale map in Fig. 6F. The horizontal axis of each of Fig. 6B-6E corresponds to 128 spatial locations of the projections of the respective projection directions (605, 610, 615, 620), measured with FOV=45 cm as shown. The projection direction 610 of the B1 projection data shown in Fig. 6C is the preferred direction because it allows observation of areas of low B1 amplitude (e.g., the anterior left thigh). The dip between the black circle (●) and the white circle (○) of the projection data in Fig. 6C corresponds to the average B1 amplitude value obtained approximately between the black circle and the white circle in Fig. 6A and 6F, where the average value is obtained perpendicular to the projection direction. The same projection data from projection direction 610 may also be used to identify a second subset of regions with high B1 amplitude. Alternatively, a second projection direction may be selected to identify the second subset of B1 data. However, in this example, only a single direction 610 is required to identify the first and second subsets. Other projection directions (605, 615, 620 shown in FIGS. 6B, 6D, and 6E, respectively) show how the majority of the B1 low / high regions are not observable due to the averaging of B1 amplitude along the dimension projected perpendicular to each line. Also, as can be seen from FIGS. 6B-6E, in at least one embodiment, an air threshold of 0.6 may be selected.

[0072] The horizontal axis of Fig. 6B shows the spatial position of the destination of projection 605. The horizontal axis of Fig. 6C shows the spatial position of the destination of projection 610. The horizontal axis of Fig. 6D shows the spatial position of the destination of projection 615. The horizontal axis of Fig. 6E shows the spatial position of the destination of projection 760.

[0073] Although the analysis of all (non-air) B1 amplitude values ​​for a particular 1D projection is described, any sub-part of the 1D projection may be used. For example, one projection may contain multiple drops in B1 values ​​(called holes). Also, calculations that do not distinguish between multiple holes may combine values ​​from multiple holes during processing (e.g., when determining the minimum X% of non-air values). Some imaging regions may be processed based on information from two holes, while in other embodiments, only the minimum Z% in the hole containing the deeper hole is used.

[0074] In yet another embodiment, the B1 amplitude data can be limited to a particular region of a 2D slice or 3D volume by the process described herein. For example, the process can use the entire data set to determine the minimum (non-air) value and select only values ​​that are within a certain distance (in 2D or 3D) from the minimum value to be used to calculate the new RFL. Similarly, a region growing method can be applied to the 2D slice or 3D volume to determine all values ​​within a threshold that correspond to the hole with the lowest adjacent data value.

[0075] As described with respect to steps 330 and 335 of FIG. 3, the MRI system 1 acquires at least two MRI image data sets, i.e., a first RF parameter (e.g., NewRFL high region MRI image data using the RFL1 corresponding to the above) and MRI image data using a second RFL parameter (e.g., NewRFL low regionand RFL2, which corresponds to the determined extreme value, are obtained and a corrected image is generated based on these data sets. As will be described in more detail below, in alternative embodiments, the RFL1 and RFL2 values ​​may be determined in a manner other than using the determined extreme values. For example, the RFL1 and RFL2 values ​​may be obtained from a database of RFL values. The RFL values ​​may be constant for all imaging procedures or may vary based on run-time parameters such as the type of region being imaged (e.g., thigh, prostate, skull, pelvis, etc.), the type of imaging performed on the region, and / or patient-specific parameters (e.g., patient weight, which affects the size / location of the imaged region relative to the imaging region center). In alternative embodiments, the RFL1 and RFL2 values ​​are manually selected and entered by a user. In fact, in general, the system may utilize any two values ​​for RFL1 and RFL2, as long as they are not the same value. One of the two values ​​may be greater than or less than the RFL that gives a mean B1 value of 1.0. In at least one embodiment, these values ​​are selected without performing a pre-scan.

[0076] As shown in the data flow diagram of FIG. 7A, a first image data set (collectively referred to as Image 1) is acquired using RFL1, and a second image data set (collectively referred to as Image 2) is acquired using RFL2. In a first embodiment, Image 1 and Image 2 are acquired at the same resolution (as shown in FIGS. 8B and 8C), and then combined to generate a corrected image by pixel-wise Maximum Intensity Projection (MIP) on the entire combined image. Other image combination functions such as averaging (mean, median, or mode) of pixel values ​​may also be used. Three or more sets of image data may also be acquired, each corresponding to a different transmit RF parameter, and each image data set is used to calculate the corrected image.

[0077] As shown in FIG. 8A, the corrected image includes at least one area of ​​reduced darkness compared to the dark area (indicated by the white arrow) in the original image. The examples of original images are two images both taken by full sampling acquisition, but the images may be acquired by different acquisition processes (e.g., full sampling or high-speed acquisition). For example, Image 2 may be accelerated according to the Sensitivity Encoding (SENSE) method, the Generalized Autocalibrating Partial Parallel Acquisition (GRAPPA) method, the Array coil Spatial Sensitivity Encoding (ARC) method, Compressed Sensing (CS), or Machine Learning Reconstruction (MLR) method. Also, for example, Image 2 may be acquired using an acceleration factor of 2 or 3. Image 1 and Image 2 may be accelerated by the same or different acceleration methods and acceleration factors.

[0078] In another embodiment, image 1 and image 2 are not acquired at the same resolution, but at higher and lower resolutions, respectively, as shown in Figures 9B and 9C. For example, image 2 has a lower resolution than image 1 (i.e., larger spatial voxels are used to generate image 2 than image 1). By using larger voxels, the additional scan time to acquire the second image is reduced. However, since the low-resolution image is smeared compared to the high-resolution image, MIP blending of the two images results in a blurred image, as shown in Figure 9A. To address the blurring issue, it is preferable to perform a resolution adjustment on at least one of the images before performing the image blending shown in Figures 7A and 7B.

[0079] The processing circuit 40 calculates a correlation map indicating the correlation between the first image data set and the second image data set from the first image data set and the second image data set. The correlation indicates the similarity of pixel values, i.e., B1 values, for each pixel. The correlation map indicates, for example, a ratio between the first image data set and the second image data set (hereinafter, referred to as a ratio map). Note that the correlation map is not limited to the ratio map, and may be a map based on the difference between the first image data set and the second image data set. In the following, for the sake of concrete explanation, the correlation map will be explained as a ratio map based on the ratio between pixel values ​​in the first image data set and pixel values ​​in the second image data set.

[0080] In one embodiment, a low-resolution image of image 1 (shown in FIG. 10A) 1(lowres) ) is generated, for example, by truncating the k-space data from the original image data of image 1. Next, image 1(lowres) (Figure 10A) and Image 2 (Figure 10B, for clarity, image 2(lowres) For every pixel (x,y) of the image (which may also be referred to as a pixel-by-pixel map), a ratioMap is calculated according to equation (7) below:

[0081]

number

[0082] Other thresholds than 1.0 can also be used to calculate the RatioMap. An image of the ratio map generated from Figures 10A and 10B is shown in Figure 10C. 2(lowres) (x,y) < image 1(lowres) Image, without the (x,y) contribution 1(lowres) and images 2(lowres)Other non-linear combinations of are also possible. Using RatioMap(x,y), the corrected image can be generated according to the following equation (8). As shown in equation (8), the processing circuit 40 obtains the corrected image by correcting the first image data set using the correlation map. Specifically, the processing circuit 40 corrects the first image data set using the ratio map to obtain the corrected image. Also, as shown in equations (7) and (8), the processing circuit 40 corrects the first image data set using the correlation map when the correlation map is greater than a threshold value.

[0083] image corrected (x,y)=Image 1×RBCF(RatioMap(x,y)) (Formula 8)

[0084] 7B shows a data flow diagram illustrating the image correction method described above. In one embodiment, the Ratio-Based Correction Function (RBCF) is simply unity gain (multiply by 1.0).

[0085] image 2(lowres) Image 1 to Image 2 1(lowres) In another embodiment using a different imaging resolution, rather than creating a low-resolution image of image 2 (i.e., a low-resolution image of image 1), an approximate high-resolution image of image 2 is created using at least one trained neural network. In this first embodiment, a first neural network is trained by receiving a low-resolution image of image 2 as input and a high-resolution version of the low-resolution image as a target output. After training, low-resolution image 2 is "upsampled" by the first neural network and may be subjected to the same type of image synthesis process as the process described above for synthesizing two full-resolution images.

[0086] In a second embodiment using a different imaging resolution, a second neural network is trained to perform a combined process of upsampling and image synthesis. In such an embodiment, the input for training the second neural network is a new lower range RFL (NewRFL low range ) based on full resolution imagery and a new high range RFL (NewRFL high range ) The target output for training the second neural network is NewRFL low range and NewRFL high range After training, the second neural network, NewRFL, low range Based on full resolution images and NewRFL high range The corrected image can be created directly from the low-resolution image based on

[0087] As shown in Figures 7A and 7B, other embodiments of the image correction process utilize more than two images (e.g., three or more images). These images may all be full resolution images or at least one high resolution image combined with at least one low resolution image. For the three image embodiment, the three images are collected with NewRFL values ​​calculated to bring a particular region closer to a target B1 value (e.g., 1.0). By using more image data sets, the image quality can be improved since the region of interest in each image will have a higher (better) signal.

[0088] In an embodiment of such three images, one maximum-resolution image is acquired based on RFL1, and two low-resolution images are acquired based on RFL2 and RFL3. RFL2 and RFL3 are selected such that (1) RFL2 < RFL1 holds and RFL2 is selected to have good chipping in an area including the over-chipping of RFL1, and (2) RFL3 > RFL1 holds and RFL3 is selected to have good chipping in an area including the under-chipping of RFL1. A corresponding RatioMap is created for (RFL2 / RFL1) and (RFL3 / RFL1) which are always at a ratio of at least 1.0. Then, the images are synthesized with each other to generate a corrected image according to (1) Corrected = (RFL1 × Ratio2 + RFL1 × Ratio3) / 2 or (2) Corrected = MIP[RFL1 × Ratio2; RFL1 × Ratio3].

[0089] Similar to the above-described example of the neural network for two images, a neural network can be trained to create a corrected image from three (or more) images including at least one maximum-resolution image and at least one low-resolution image. For example, the processing circuit 40 acquires a first image dataset at a first resolution and a second image dataset at a second resolution different from the first resolution. Then, the processing circuit 40 creates a correlation map from the first image dataset and the second image dataset at a resolution lower than the first resolution and the second resolution. At this time, the processing circuit 40 may correct the first image dataset using a neural network that receives the correlation map calculated at the low resolution.

[0090] FIG. 11A is a B1 amplitude color map of a human pelvis. FIG. 11B is a ratio map showing the pixel ratio (greater than 1) between FIG. 11F and 11E for RFL2 (FIG. 11F) of 1.4 and RFL1 (FIG. 11E) of 0.80. As is evident from FIG. 11D, the corrected pelvis image is more uniform in brightness compared to the conventional image of FIG. 11C with the average B1 value of the target B1 value. FIG. 11G is a grayscale image of the B1 amplitude color map of FIG. 11A. In FIG. 11H, the grayscale image of the B1 amplitude map is annotated similarly to the annotated B1 map of FIG. 4J. The outer white ellipse corresponds to the green region with a B1 value of approximately 80% of the target B1, and the inner white ellipse corresponds to the blue region with a B1 value of approximately 60% of the target B1. In FIG. 11H, the amplitude values ​​decrease toward the center of the inner ellipse and increase away from the center of the inner ellipse.

[0091] As described above with respect to FIG. 3, at least two image data sets are acquired by performing a series of MRI sequences with at least two transmit RF parameters. The RFL1 and RFL2 values ​​may be obtained from a database of RFL values. The RFL values ​​may be constant for all imaging procedures or may vary based on run-time parameters such as the type of region being imaged (e.g., thigh, prostate, skull, pelvis, etc.), the type of imaging performed on the region, and / or patient-specific parameters (e.g., patient weight, which affects the size / location of the imaged region relative to the imaging region center). In another embodiment, the RFL1 and RFL2 values ​​are manually selected and entered by a user. In fact, the system may utilize any two values, as long as RFL1 and RFL2 are not generally the same value. One of these values ​​may be greater than or less than the RFL that gives a mean B1 value of 1.0. In at least one embodiment, the values ​​are selected without performing a pre-scan. As noted above, the two image data sets do not have to be of the same resolution.

[0092] As described above, a ratio map is calculated from at least two collected image data sets as described in relation to equation (7) and is available for image correction using the ratio-based correction function (RBCF) as described in relation to equation (8). As shown in FIG. 7B, the ratio map may be at the same resolution as one or more of the image data sets or at a lower resolution than the image data. As an alternative to the unity gain-ratio-based correction function (RBCF) described above, an exponential correction factor (ECF, e.g., 1.5) may be used for the RBCF based on the following equation: For example, the processing circuitry 40 corrects the first image data set using the correlation map exponentially raised to the exponential correction factor ECF as shown in equation (9).

[0093] image corrected (x,y) = Image1 × (RatioMap(x,y) ECF ) (Formula 9)

[0094] More generally, a CorrelationMap represents the pixel-by-pixel correlation of two images acquired based on different RF parameters, and a RatioMap is an embodiment of a CorrelationMap. After the CorrelationMap is calculated, the images are corrected by a correlation-based correction function (CBCF, e.g., RBCF). The CorrelationMap can be calculated by various functions, including but not limited to (a) the difference between the two images, (b) the difference obtained by subtracting the logarithm of each image, etc.

[0095] Figure 12 shows an image of two legs corrected with the ratio-based correction function of Equation (9) at an ECF value of 1.5. As can be seen, such a configuration addresses the effects of local under- and over-chipping in the image. Although the image quality is improved, there are still cases of over-correction (indicated by the up arrow) and under-correction (indicated by the down arrow). Other functions based on unity gain and exponential correction factors are also possible.

[0096] Although equation (9) gives a first ratio-based correction function, other correction functions can be implemented. Such correction functions are based on a signal model that is based on the physical relationship between B1 and the image signal. Such a relationship (shown in FIG. 13) gives the signal=sin 3 Describe the spin-echo or fast-echo signal according to the formula (when the RFL is chosen so that the average B1 field is 1.0): Signal = sin(B1 x θ). In other embodiments, signal = sin(B1 x θ) may be used to model other types of pulse sequences, such as field-echo. Other effects (e.g., T1 relaxation effects) may be included in the model.

[0097] By modeling the effect of different RFLs (RFL1 and RFL2) on B1, different curves representing the resulting B1 distributions are obtained, as shown in FIG. 14A. From the two B1 distributions in FIG. 14A, a ratio is calculated over the entire B1 distribution range, as shown in FIG. 14B. The transpose of this ratio, shown in FIG. 14C, can be used as a ratio-to-B1 mapping function (e.g., implemented as a look-up table (e.g., stored in memory 41)) to determine an estimate of B1 for a particular pixel based on the ratio of values ​​obtained at at least two RFLs (i.e., RFL1 and RFL2). The ratio-to-B1 mapping function is then applied to the inverse of the B1 signal model (FIG. 14A) to calculate a ratio-based correction function (RBCF). The resulting ratio-based correction function (RBCF) is shown in FIG. 14D. This ratio-based correction function can be used in the system of FIG. 7B to pixel-by-pixel correct the image data set for RFL1 by multiplying pixel values ​​in the first image data set (corresponding to RFL1) by a ratio-based correction function that derives the ratio map from the first and second image data sets. That is, the correction values ​​required to correct at least the first image data set are generated by applying the calculated ratio map to a ratio-based correction function that models the inverse of the B1 model that represents how the first and second image data sets were acquired. Other RatioMaps can be calculated as shown in FIG. 7B. Corrected images can be created by correcting different portions of one or more images with multiple ratio-based correction functions. Such a configuration is useful for images using very high B0 magnetic fields (e.g., greater than 3T). That is, the processing circuitry 40 corrects the first image data set by applying the correlation map to a correlation-based correction function (inverse model) that models the inverse of the B1 model that represents how the first and second image data sets were acquired.

[0098] FIG. 15A shows an original image of two human thighs before correction. By applying the ratio-based correction function on a pixel-by-pixel basis (based on the ratio map of Equation (7) obtained for the image), the corrected image shown in FIG. 15B can be obtained. As shown in FIG. 15B, the dark areas due to under-chipping in the original image are brightened and more in keeping with the rest of the image without the addition of over-chipping. Similarly, FIG. 16A shows an original image of a human pelvis before correction. By applying the ratio-based correction function on a pixel-by-pixel basis (based on the ratio map of Equation (7) obtained for the image), the corrected image shown in FIG. 16B can be obtained. As shown in FIG. 16B, the dark areas due to under-chipping in the original image are brightened and more in keeping with the rest of the image without the addition of over-chipping.

[0099] In addition to other configurations and methods described herein, further configurations and methods are set forth in the non-limiting parentheticals set forth below, including, but not limited to, the following:

[0100] (1) An image processing method including, but not limited to, acquiring first and second image data sets generated by performing a series of MRI sequences using different first and second transmit RF parameters, calculating a correlation map from the first and second image data sets, and acquiring a corrected image by correcting the first image data set using the calculated correlation map.

[0101] (2) calculating a correlation map from the first and second image data sets includes calculating a ratio map from the first and second image data sets; 2. The method of claim 1, wherein correcting the first image data set using the calculated correlation map comprises correcting the first image data set using the calculated ratio map to obtain the corrected image.

[0102] (3) The method of (2), wherein correcting the first image data set using the calculated correlation map includes, but is not limited to, correcting the first image data set using the calculated correlation map raised to an exponential correction function.

[0103] (4) A method according to any one of (1) to (3), wherein the different first and second transmit RF parameters are different first and second RF transmit gain levels.

[0104] (5) obtaining spatially resolved B1 amplitude measurements; and A method according to any one of (1) to (3), further comprising, but not limited to, selecting the different first and second transmit RF parameters by performing at least two spatially resolved analyses on the acquired spatially resolved B1 amplitude measurements.

[0105] (6) performing the at least two spatially resolved analyses includes, but is not limited to, obtaining first and second subsets of spatially resolved B1 amplitude measurements using a histogram; The method of (4), wherein selecting the different first and second transmit RF parameters includes, but is not limited to, selecting the first transmit RF parameter by using a first subset of the spatially resolved B1 amplitude measurements and selecting the second transmit RF parameter by using a second subset of the spatially resolved B1 amplitude measurements.

[0106] (7) The method of (6), wherein using the histogram to obtain first and second subsets of spatially resolved B1 amplitude measurements includes, but is not limited to, using the histogram to obtain (a) a first percentile of the smallest spatially resolved B1 amplitude measurements as a first subset of the spatially resolved B1 amplitude measurements, and (b) a second percentile of the largest spatially resolved B1 amplitude measurements as a second subset of the spatially resolved B1 amplitude measurements.

[0107] (8) The method of (5), wherein acquiring the spatially resolved B1 amplitude measurements includes, but is not limited to, at least one of acquiring the spatially resolved B1 amplitude measurements using a Bloch-Siegert shift sequence and acquiring the spatially resolved B1 amplitude measurements by Actual Flip Angle (AFP) imaging.

[0108] (9) A method according to any one of (5) to (8), wherein obtaining the spatially resolved B1 amplitude measurements includes, but is not limited to, obtaining 1D projections.

[0109] (10) A method according to any one of (5) to (8), wherein acquiring the spatially resolved B1 amplitude measurements includes, but is not limited to, acquiring 2D slices.

[0110] (11) A method according to any one of (5) to (8), wherein obtaining the spatially resolved B1 amplitude measurements includes, but is not limited to, obtaining a 3D volume.

[0111] (12) The method of (1), wherein correcting the first image dataset using the calculated correlation map includes, but is not limited to, correcting the first image dataset by applying the calculated correlation map to a correlation-based correction function that models an inverse of a B1 model that represents how the first and second image datasets were acquired.

[0112] (13) Correcting the first image data set using the calculated correlation map includes, but is not limited to, correcting the first image data set using the calculated correlation map if the calculated correlation map is greater than a threshold. A method according to any one of (1) to (12).

[0113] (14) Acquiring the first and second image data sets may include, but is not limited to, acquiring the first image data set at a first resolution and acquiring the second image data set at a second resolution different from the first resolution; Calculating a correlation map from the first and second image data sets includes, but is not limited to, calculating the correlation map at a resolution lower than the first and second resolutions. A method according to any one of (1) to (13).

[0114] (15) The method of (14), wherein correcting the first image dataset using the calculated correlation map includes, but is not limited to, correcting the first image dataset using a neural network that receives the correlation map calculated at a low resolution.

[0115] (16) A transmit RF parameter determination method including, but not limited to, acquiring spatially resolved B1 amplitude measurements, performing at least two spatially resolved analyses on the acquired spatially resolved B1 amplitude measurements to acquire at least two subsets of spatially resolved B1 amplitude measurements, determining at least two transmit RF parameters to be used for a series of MRI sequences based on the at least two subsets of the acquired spatially resolved B1 amplitude measurements, and storing the at least two transmit RF parameters.

[0116] (17) The method according to (16), further comprising, but not limited to, performing a series of MRI sequences using the at least two transmit RF parameters.

[0117] (18) The method of (16) or (17), wherein the different first and second transmit RF parameters are different first and second RF transmitter gain levels.

[0118] (19) Performing the at least two spatially resolved analyses includes, but is not limited to, obtaining first and second subsets of spatially resolved B1 amplitude measurements using a histogram; The method of any one of (16) to (18), wherein selecting the different first and second transmit RF parameters includes, but is not limited to, selecting the first transmit RF parameter by using a first subset of the spatially resolved B1 amplitude measurements and selecting the second transmit RF parameter by using a second subset of the spatially resolved B1 amplitude measurements.

[0119] (20) The method of (19), wherein using the histogram to obtain first and second subsets of spatially resolved B1 amplitude measurements includes, but is not limited to, using the histogram to (a) obtain a smallest first percent of the spatially resolved B1 amplitude measurements as a first subset of the spatially resolved B1 amplitude measurements, and (b) obtain a largest second percent of the spatially resolved B1 amplitude measurements as a second subset of the spatially resolved B1 amplitude measurements.

[0120] (21) A method according to any one of (16) to (20), wherein obtaining the spatially resolved B1 amplitude measurements includes, but is not limited to, obtaining the spatially resolved B1 amplitude measurements using a Bloch-Siegert shift sequence.

[0121] (22) The method of any one of (16) to (20), wherein obtaining the spatially resolved B1 amplitude measurements includes, but is not limited to, obtaining the spatially resolved B1 amplitude measurements by Actual Flip Angle (AFP) imaging.

[0122] (23) A method according to any one of (16) to (22), wherein obtaining the spatially resolved B1 amplitude measurements includes, but is not limited to, obtaining 1D projections.

[0123] (24) The method of any one of (16) to (22), wherein acquiring the spatially resolved B1 amplitude measurements includes, but is not limited to, acquiring 2D slices.

[0124] (25) A method according to any one of (16) to (22), wherein obtaining the spatially resolved B1 amplitude measurements includes, but is not limited to, obtaining a 3D volume.

[0125] (26) A magnetic resonance imaging system including, but not limited to, a gantry, at least one RF transmit coil, and an RF transmit coil controller for performing any one of the procedures described in (1) to (25).

[0126] (27) A computer program product, including, but not limited to, a non-transitory computer-readable storage medium communicatively connected to a computer processor, the computer instructions including computer instructions that, when executed by the computer processor, cause the computer processor to perform any one of the steps set forth in (1) through (25).

[0127] Although some embodiments have been described, these embodiments are presented as examples and are not intended to limit the scope of the invention. The novel method, device, and system can be embodied in various other forms, and various omissions, substitutions, and modifications can be made without departing from the spirit of the invention. These embodiments and their modifications are within the scope of the invention and its equivalents as defined in the claims, as well as the scope and spirit of the invention.

[0128] When the technical idea of ​​the embodiment is realized in a transmission RF parameter determination method, the transmission RF parameter determination method includes: acquiring spatially resolved B1 amplitude measurements, performing at least two spatially resolved analyses on the acquired spatially resolved B1 amplitude measurements to acquire at least two subsets of spatially resolved B1 amplitude measurements, determining at least two transmission RF parameters to be used in a series of MRI sequences based on the acquired at least two subsets of spatially resolved B1 amplitude measurements, and storing the at least two transmission RF parameters. The procedures and effects of various processes in this transmission RF parameter determination method are the same as those described in the embodiment, and therefore will not be described again.

[0129] When the technical idea of ​​the embodiment is realized by an image processing device, the image processing device has a processing circuit that acquires a first image data set and a second image data set generated by executing a series of MRI sequences using a first transmission RF parameter and a second transmission RF parameter different from each other, calculates a correlation map indicating the correlation between the first image data set and the second image data set from the first image data set and the second image data set, and acquires a corrected image by correcting the first image data set using the correlation map. The image processing device corresponds to, for example, a console 400. The procedures and effects of various processes related to this image processing device are similar to those described in the embodiment, and therefore will not be described.

[0130] According to at least one of the above-described embodiments, image quality can be improved. For example, according to the present embodiment, the resolutions of a plurality of image data sets acquired using a plurality of preset RFLs are aligned to generate a correlation map (ratio map), and a corrected image can be generated by multiplying the image data set with a higher resolution among the plurality of image data sets by the generated ratio map. As a result, according to the present embodiment, a high-quality image with reduced over-chipping and under-chipping can be generated.

[0131] Although some embodiments have been described, these embodiments are presented as examples and are not intended to limit the scope of the invention. These embodiments can be implemented in various other forms, and various omissions, substitutions, modifications, and combinations of embodiments can be made without departing from the spirit of the invention. These embodiments and their modifications are included in the scope of the invention and its equivalents as described in the claims, as well as in the scope and spirit of the invention.

[0132] Regarding the above-described embodiments, the following supplementary notes are disclosed as one aspect and optional features of the invention. (Appendix 1) 1. An image processing method comprising: acquiring a first image data set and a second image data set generated by executing a series of MRI sequences using first transmit RF parameters and second transmit RF parameters that are different from each other; calculating a correlation map indicating a correlation between the first image data set and the second image data set from the first image data set and the second image data set; and correcting the first image data set using the correlation map to acquire a corrected image. (Appendix 2) The correlation map may be a ratio map based on a ratio between pixel values ​​in the first image dataset and pixel values ​​in the second image dataset, and the corrected image may be obtained by correcting the first image dataset using the ratio map. (Appendix 3) Correcting the first image data set with the ratio map may comprise correcting the first image data set with the correlation map raised to an exponential correction factor. (Appendix 4) The first transmit RF parameter and the second transmit RF parameter may be a first RF transmit gain level and a second RF transmit gain level that are different from each other. (Appendix 5) The first transmit RF parameter and the second transmit RF parameter may be selected by obtaining spatially resolved B1 amplitude measurements and performing at least two spatially resolved analyses on the spatially resolved B1 amplitude measurements. (Appendix 6) Performing the at least two spatially resolved analyses may include obtaining a first subset and a second subset of spatially resolved B1 amplitude measurements using a histogram, and selecting the first transmit RF parameter and the second transmit RF parameter may include selecting the first transmit RF parameter using the first subset and selecting the second transmit RF parameter using the second subset. (Appendix 7) Obtaining the first subset and the second subset may correspond to using the histogram to obtain a smallest first percent of the spatially resolved B1 amplitude measurements as the first subset and a largest second percent of the spatially resolved B1 amplitude measurements as the second subset. (Appendix 8) The spatially resolved B1 amplitude measurements may be acquired by Bloch-Siegert shift sequences or Actual Flip Angle (AFP) imaging. (Appendix 9) Correcting the first image dataset may include correcting the first image dataset by applying the correlation map to a correlation-based correction function that models an inverse of a B1 model that represents a manner in which the first image dataset and the second image dataset were acquired. (Appendix 10) Correcting the first image data set may include correcting the first image data set using the correlation map if the correlation map is greater than a threshold. (Appendix 11) The first image dataset may be acquired at a first resolution and the second image dataset may be acquired at a second resolution different from the first resolution, and the correlation map may be calculated from the first image dataset and the second image dataset at a resolution lower than the first resolution and the second resolution. (Appendix 12) Correcting the first image data set may include correcting the first image data set using a neural network that receives the correlation map calculated at the lower resolution. (Appendix 13) 11. A method for determining transmit RF parameters, comprising: acquiring spatially resolved B1 amplitude measurements; performing at least two spatially resolved analyses on the acquired spatially resolved B1 amplitude measurements to acquire at least two subsets of spatially resolved B1 amplitude measurements; determining at least two transmit RF parameters to be used for a series of MRI sequences based on the at least two subsets of the acquired spatially resolved B1 amplitude measurements; and storing the at least two transmit RF parameters. (Appendix 14) A magnetic resonance imaging system comprising: a gantry; at least one RF transmit coil; and an RF transmit coil controller, wherein the RF transmit coil controller acquires a first image data set and a second image data set generated by executing a series of MRI sequences using first transmit RF parameters and second transmit RF parameters that are different from each other, calculates a correlation map from the first image data set and the second image data set indicating a correlation between the first image data set and the second image data set, and acquires a corrected image by correcting the first image data set using the correlation map. (Appendix 15) an image processing device comprising: a processing circuit for acquiring a first image data set and a second image data set generated by executing a series of MRI sequences using first transmit RF parameters and second transmit RF parameters different from each other; calculating a correlation map indicating a correlation between the first image data set and the second image data set from the first image data set and the second image data set; and correcting the first image data set using the correlation map to acquire a corrected image. [Explanation of symbols]

[0133] 1. Magnetic Resonance Imaging (MRI) System 10 Static magnetic field magnet 11 Gradient Coil Assembly 12 Whole body (WB) coil 20 RF coil 31 Gradient coil power supply 32 RF Receiver 33 RF transmitter 34 Sequence Controller 40 Processing circuit 41 Memory 42 Display 43 Input Interface 50 Bed body 51 Tabletop 100 Gantry 300 Control Cabinet 400 Console 500 berths

Claims

1. acquiring a first image data set and a second image data set generated by performing a series of MRI sequences using different first and second transmit RF parameters; calculating a correlation map from the first image data set and the second image data set, the correlation map indicating a correlation between the first image data set and the second image data set; obtaining a corrected image by correcting the first image data set using the correlation map; An image processing method comprising:

2. the correlation map being a ratio map based on a ratio of pixel values ​​in the first image data set to pixel values ​​in the second image data set, The corrected image is obtained by correcting the first image data set using the ratio map. The image processing method according to claim 1 .

3. and correcting the first image data set using the ratio map comprises correcting the first image data set using the correlation map exponentially raised to a correction factor. The image processing method according to claim 2 .

4. the first transmit RF parameter and the second transmit RF parameter are a first RF transmit gain level and a second RF transmit gain level that are different from each other; The image processing method according to claim 1 .

5. Obtain spatially resolved B1 amplitude measurements; selecting the first transmit RF parameter and the second transmit RF parameter by performing at least two spatially resolved analyses on the spatially resolved B1 amplitude measurements; The image processing method of claim 1 further comprising:

6. performing the at least two spatially resolved analyses includes obtaining a first subset and a second subset of spatially resolved B1 amplitude measurements using a histogram; selecting the first transmit RF parameter and the second transmit RF parameter includes selecting the first transmit RF parameter using the first subset and selecting the second transmit RF parameter using the second subset. The image processing method according to claim 5.

7. obtaining the first and second subsets corresponds to obtaining, using the histogram, a smallest first percent of the spatially resolved B1 amplitude measurements as the first subset and a largest second percent of the spatially resolved B1 amplitude measurements as the second subset. The image processing method according to claim 6.

8. The spatially resolved B1 amplitude measurements are obtained by Bloch-Siegert shift sequences or Actual Flip Angle (AFP) imaging. The image processing method according to claim 5.

9. and correcting the first image data set includes correcting the first image data set by applying the correlation map to a correlation-based correction function that models an inverse of a B1 model that represents a manner in which the first image data set and the second image data set were acquired. The image processing method according to claim 1 .

10. and correcting the first image data set includes correcting the first image data set using the correlation map if the correlation map is greater than a threshold. The image processing method according to claim 1 .

11. the first image data set is acquired at a first resolution; the second image data set is acquired at a second resolution different from the first resolution; the correlation map is calculated from the first image data set and the second image data set at a resolution lower than the first resolution and the second resolution. The image processing method according to any one of claims 1 to 10.

12. and correcting the first image data set includes correcting the first image data set using a neural network that receives the correlation map calculated at the lower resolution. The image processing method according to claim 11.

13. Obtain spatially resolved B1 amplitude measurements; performing at least two spatially resolved analyses on the acquired spatially resolved B1 amplitude measurements to obtain at least two subsets of spatially resolved B1 amplitude measurements; determining at least two transmit RF parameters to be used in a series of MRI sequences based on at least two subsets of the acquired spatially resolved B1 amplitude measurements; storing the at least two transmit RF parameters; A method for determining transmit RF parameters comprising:

14. The gantry, At least one RF transmit coil; and an RF transmit coil controller; Equipped with The RF transmit coil controller includes: acquiring first and second image data sets generated by performing a series of MRI sequences using different first and second transmit RF parameters; calculating a correlation map from the first image data set and the second image data set, the correlation map indicating a correlation between the first image data set and the second image data set; obtaining a corrected image by correcting the first image data set using the correlation map; Magnetic resonance imaging system.

15. acquiring a first image data set and a second image data set generated by performing a series of MRI sequences using different first and second transmit RF parameters; calculating a correlation map from the first image data set and the second image data set, the correlation map indicating a correlation between the first image data set and the second image data set; a processing circuit for correcting the first image data set using the correlation map to obtain a corrected image; An image processing device comprising:

Images