Magnetic resonance imaging apparatus and image generating method

Abstract

To accurately remove noise of a separation image obtained by separating spatially-overlapping signals by using a PI method in an image acquired by radial sampling by using a plurality of reception coils using MRI.SOLUTION: A magnetic resonance imaging apparatus measures complex blade images that spatially overlap each other for each blade by radial sampling from nuclear magnetic resonance signals received by a plurality of reception coils, calculates a blade image by separating spatially-overlapping signals for each blade by using sensitivity information of the plurality of reception coils, and removes noise mixed into the blade image by repetition calculation on the basis of a g-factor map expressing noise characteristics for each blade.SELECTED DRAWING: Figure 5

Description

[Technical field]

[0001] The present invention relates to magnetic resonance imaging (hereinafter referred to as "MRI") devices, and more particularly to noise reduction in radial sampling. [Background technology]

[0002] An MRI device is a device that measures NMR signals generated by the nuclear spins that constitute the tissues of a subject, particularly a human body, and creates two-dimensional or three-dimensional images of the morphology and function of the head, abdomen, limbs, etc. During imaging, the NMR signals are given different phase encodings by a gradient magnetic field and are frequency encoded and measured as time-series data. The measured NMR signals are placed at measurement points determined by the encoding of the measurement space (k-space) whose axis is the encoding direction, and the k-space data is reconstructed into an image by two-dimensional or three-dimensional Fourier transform.

[0003] As a method for acquiring data of measurement points in k-space (sampling method), in addition to Cartesian coordinate system sampling that measures along an axis, there are non-Cartesian coordinate system sampling such as radial sampling that measures radially from the origin of k-space. Radial sampling has the advantage that it can disperse the effects of body movement and is less susceptible to the effects of motion artifacts because it acquires data near the origin of k-space for each measurement.

[0004] On the other hand, one of the technical challenges in MRI is the shortening of imaging time. One method for shortening imaging time is to measure a small number of points in k-space and restore unmeasured points through signal processing. If non-fully sampled k-space data is directly reconstructed through a Fourier transform, spatially overlapping signals will appear in the resulting image. A method for acquiring spatially overlapping signals using multiple receiving coils and separating the signals by utilizing the difference in sensitivity distribution between the receiving coils is called the parallel imaging method (hereinafter referred to as the PI method), and various methods have been proposed depending on the separation method, such as the SENSE method, the GRAPPA method, and the CAIPIRINHA method.

[0005] The PI method can also be applied to the radial sampling described above. For example, Patent Document 1 and Non-Patent Document 1 disclose an imaging method in which the number of phase encodes is reduced and the PI method is applied in a sampling method known as the propeller method, in which phase encoding is added for each radial sampling angle.

[0006] It is known that the signal-to-noise ratio (SNR) of an image obtained by the PI method is inversely proportional to the geometry factor (g-factor), which is an index related to the arrangement of multiple receiving coils used for imaging. The g-factor is an index that takes a value of 1 or more, and when the difference in sensitivity between the receiving coils is small, the g-factor increases and the SNR decreases. Various methods have been proposed to prevent the decrease in SNR due to the increase in the g-factor. Representative methods include a noise removal method using regularization (e.g., Non-Patent Document 2), and a method that further improves the method using regularization by utilizing the fact that there is a correlation in the noise between images (separated images) separated by aliasing removal processing and adding this to the constraints of noise removal, thereby improving the accuracy of noise removal (e.g., Patent Document 2). [Prior art documents] [Patent documents]

[0007] [Patent Document 1] JP 2004-34418 A [Patent Document 2] JP 2019-42444 A [Non-patent literature]

[0008] [Non-Patent Document 1] Yuchou Chang, James G. Pipe, John P. Karis, Wende N. Gibbs, Nicholas R. Zwart and Michael Schaer, The effects of SENSE on PROPELLER imaging, Magnetic Resonance in Medicine, 74, 6, (1598-1608), (2014). [Non-Patent Document 2] King KF et al., “SENSE Image Quality Improvement Using Matrix Regularization”, In Proceedings of the 9th Annual Meeting of ISMRM, Glasgow, Scotland, 2001, p. 1771. Summary of the Invention [Problem to be solved by the invention]

[0009] In radial sampling, multiple bundles of echo signals, generally called blades, are acquired, and measurement data is collected while changing the arrangement angle of the blades. In this case, a method called gridding is used to combine the measurement data of each blade and fill the k-space. The above-mentioned conventional noise removal method is not designed taking into account the SNR of the image obtained after such gridding, so the effect of noise removal when applied to radial sampling is inevitably limited. In addition, the method of Patent Document 2 is a technology derived based on the knowledge that there is a correlation between noise in separated images for data measured in k-space in a Cartesian coordinate system, and this technology cannot be applied directly to data obtained by non-Cartesian coordinate system sampling.

[0010] The present invention aims to provide a noise removal method specialized for noise in images obtained by non-Cartesian coordinate system sampling using the PI method, in particular radial sampling, and to provide an MRI apparatus equipped with radial sampling using the PI method and a noise removal means specialized for it. [Means for solving the problem]

[0011] In the present invention, when generating an image from the nuclear magnetic resonance signals received by multiple receiver coils by radial sampling using the sensitivity distribution of the multiple receiver coils, a constraint condition reflecting the noise characteristics of each blade is set and noise is removed by repeated calculations, thereby enabling effective noise removal from images obtained by radial sampling.

[0012] That is, the MRI apparatus of the present invention comprises a measurement unit that applies a high frequency magnetic field and a gradient magnetic field to a subject placed in a static magnetic field space, and measures measurement data consisting of nuclear magnetic resonance signals generated from the subject using a plurality of receiving coils, a control unit that controls the operation of the measurement unit, and a calculation unit that processes the measurement data and creates an image, wherein the control unit divides the k-space in which the measurement data is placed into a plurality of blades and controls the measurement unit to perform measurement by thinning out the phase encoding for each blade, and the calculation unit has an image generation unit that generates an image by the PI method using the measurement data and the sensitivity distribution of the plurality of receiving coils, and a noise removal unit that removes noise from the reconstructed image by iterative calculation using noise-related constraints.

[0013] The "noise-related constraint" is a constraint that reflects the noise characteristics of each blade, and for example, a g-factor map calculated for each blade can be used.

[0014] Furthermore, the image generating method of the present invention is a method for generating an image by using a magnetic resonance imaging device equipped with a plurality of receiving coils to divide k-space into a plurality of blades, and processing measurement data for each blade measured by thinning out the phase encoding for each blade, and includes the steps of: generating an image for each blade by calculation of a parallel imaging method using the measurement data of each blade and the sensitivity distribution of the plurality of receiving coils; performing a noise removal process on the image for each blade by repeated calculations; converting the image for each blade after noise removal into data of measurement space, then gridding and relocating it to the measurement points in the k-space; and reconstructing an image using the k-space data after gridding.

[0015] Furthermore, the image generating method of the present invention is a method for generating an image by dividing k-space into a plurality of blades using a magnetic resonance imaging device equipped with a plurality of receiving coils, and processing measurement data for each blade measured by thinning out the phase encoding for each blade, and includes the steps of: generating an image for each blade by calculation of a parallel imaging method using the measurement data of each blade and the sensitivity distribution of the plurality of receiving coils; converting the image for each blade into data of measurement space, then gridding and relocating the image to the measurement points in the k-space; reconstructing an image using the k-space data after gridding; and performing noise removal processing on the reconstructed image by repeated calculations.

[0016] The step of performing the noise removal process includes a step of calculating a g-factor map for each blade, and further calculating an integrated g-factor by integrating the g-factor maps for each blade, and the noise removal process is performed by using the integrated g-factor as a constraint condition for the repeated calculation. Effect of the Invention

[0017] Radial sampling, which is resistant to body movement, can obtain high-quality images with noise removed. This reduces noise and improves the number of times the speed is increased by the PI method, shortening the imaging time. In addition, the shorter the imaging time, the less the MRI is affected by body movement, so shortening the imaging time can secondarily suppress the effects of body movement, improving the diagnostic accuracy. In addition, if the imaging parameters are adjusted to increase the spatial resolution by the amount of the shortened imaging time, the spatial resolution can be increased with the same imaging time, improving the diagnostic accuracy. [Brief description of the drawings]

[0018] [Figure 1] FIG. 1 is a block diagram showing a schematic configuration of an MRI apparatus according to an embodiment of the present invention. [Diagram 2] 1A is an external view of an MRI apparatus to which the present invention is applied; (a) is an MRI apparatus using a vertical magnetic field; (b) is an MRI apparatus using a horizontal magnetic field; and (c) is an MRI apparatus with an increased sense of openness. [Diagram 3] FIG. 1 is a diagram showing an example of an imaging flow. [Figure 4] Diagram showing the pulse sequence of 2D-GrE with radial sampling [Diagram 5] FIG. 1 is a diagram showing a configuration of a calculation unit according to a first embodiment; [Figure 6] FIG. 1 is a diagram showing a processing flow of a calculation unit according to a first embodiment; [Figure 7] FIG. 1 is a diagram showing an outline of image generation and noise processing in the first embodiment; [Figure 8] A diagram illustrating separation blade images before and after separation and the blade image. [Figure 9] FIG. 13 is a diagram showing a processing flow of a noise removal unit according to the first embodiment; [Figure 10] An example of a UI for noise removal processing [Figure 11] FIG. 13 is a diagram showing the configuration of a calculation unit according to a second embodiment. [Figure 12] FIG. 13 is a diagram showing a processing flow of a calculation unit according to a second embodiment; [Figure 13] FIG. 13 is a diagram showing an outline of image generation and noise processing in the second embodiment; [Figure 14]FIG. 13 is a diagram showing a processing flow of a noise removal unit according to a second embodiment; DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0019] An embodiment of an MRI apparatus to which the present invention is applied will be described.

[0020] [MRI Equipment Overview] First, an embodiment of an MRI apparatus to which the present invention is applied will be described with reference to Fig. 1. As shown in Fig. 1, this MRI apparatus 10 is roughly divided into a measurement unit 100 that measures nuclear magnetic resonance signals generated from a subject 101, a calculation unit 200 that performs image reconstruction, correction and other calculations using the nuclear magnetic resonance signals measured by the measurement unit 100, and a control unit 300 that controls the operation of the measurement unit 100 and the calculation unit 200. Note that in the embodiment shown in Fig. 1, the functions of the calculation unit 200 and the control unit 300 are configured to be realized by a single computer 500, but these may be separate means.

[0021] The measurement unit 100 comprises a static magnetic field coil 102 that generates a static magnetic field in the space in which the subject 101 is placed, a transmitting unit (105, 107) that transmits high-frequency magnetic field pulses to the subject 101 placed in the static magnetic field, a receiving unit (106, 108) that receives a nuclear magnetic resonance signal generated by the subject, and a gradient magnetic field coil 103 that applies a magnetic field gradient to the static magnetic field generated by the static magnetic field coil 102 in order to provide positional information to the nuclear magnetic resonance signal.

[0022] The static magnetic field coil 102 is composed of a normal conducting or superconducting static magnetic field coil, a static magnetic field generating magnet, etc., and depending on the direction of the static magnetic field generated, there are vertical magnetic field type, horizontal magnetic field type, etc., and the shape of the coil and the appearance of the entire device differ depending on the type. Figures 2(a) to (c) show the appearance of MRI devices of these different types. This embodiment can be applied to any of the MRI devices shown in the figures.

[0023] The transmitting section includes a transmitting high-frequency coil 105 (hereinafter simply referred to as a transmitting coil) that transmits a high-frequency magnetic field to a measurement region of the subject 101, and a transmitter 107 that includes a high-frequency oscillator, an amplifier, and the like. The receiving section includes a receiving high-frequency coil 106 (hereinafter simply referred to as a receiving coil) that receives a nuclear magnetic resonance signal generated from the subject 101, and a receiver 108 that includes a quadrature detection circuit, an A / D converter, and the like. In this embodiment, the receiving coil is made up of a plurality of channels (small receiving coils), each of which is connected to a quadrature detection circuit and an A / D converter that constitute the receiver 108. The nuclear magnetic resonance signal received by the receiver 108 is passed to the calculation section 200 as a complex digital signal.

[0024] The gradient magnetic field coil 103 has three sets of gradient magnetic field coils that apply gradient magnetic fields in the x-direction, y-direction, and z-direction, respectively, and each is connected to a gradient magnetic field power supply unit 110. The measurement unit 100 may also include a shim coil 104 that adjusts the static magnetic field distribution and a shim power supply unit 109 that drives it.

[0025] The measurement unit 100 further includes a sequence control device 111 that controls the operations of the shim power supply unit 109, the gradient magnetic field power supply unit 110, the transmitter 107, and the receiver 108, and controls the timing of application of the shim coils, the gradient magnetic field, and the high frequency magnetic field, and the reception of the nuclear magnetic resonance signal. The control time chart is called a pulse sequence, which is set in advance according to the measurement, and is stored in a storage device or the like provided in the computer 500.

[0026] The computer 500 includes the control unit 300 and the calculation unit 200, and controls the overall operation of the MRI apparatus 100, and performs various calculation processes on the received nuclear magnetic resonance signals. The computer 500 is an information processing device equipped with a CPU, a memory, a storage device, etc., and is connected to a display 510, an external storage device 508, an input device 509, etc.

[0027] The display 510 is an interface that displays the results obtained by the arithmetic processing to the operator. The input device 509 is an interface that allows the operator to input conditions, parameters, etc. necessary for the measurement and arithmetic processing performed in this embodiment. A user can input measurement parameters, such as the speed factor in the PI method, via the input device 509. The external storage device 508, together with the storage device inside the computer 500, holds data used in various arithmetic processing executed by the computer 500, data obtained by the arithmetic processing, input conditions, parameters, etc.

[0028] The control unit 300 includes a measurement control unit 310 that controls the operation of the measurement unit 100 via the sequence control device 111, a display control unit 320 that controls the display on the display 510, and the like. The measurement control unit 310 determines a specific imaging sequence including a sampling method based on a pulse sequence selected by an imaging method and imaging parameters set via an input device 509, and sets the sequence control unit 111. In this embodiment, the measurement control unit 310 measures the k-space for each measurement unit called a blade, collects measurement data while changing the angle of the blade in the k-space, and controls the number of phase encodes in the blade to be less than the original number of phase encodes according to the spatial resolution of the k-space. The method of reducing the number of phase encodes (thinning rate) may be determined in advance, or the user may set the thinning rate or the reciprocal thereof, that is, the speed multiplication number, as an imaging parameter. The speed multiplication number is equal to or less than the number of receiving coils.

[0029] The calculation unit 200 includes a gridding processing unit 210 that synthesizes data of grid points (measurement points) in k-space from measurement data, an image generating unit 220 that creates an image (called a blade image) by separating spatially overlapping signals using the sensitivity distribution of the receiving coil for each blade, and a noise removing unit 230 that performs noise removal processing on the separated blade images or an image obtained by synthesizing them. Specific processing contents of the calculation unit 200 will be described later.

[0030] The functions of each unit included in the computer 500 can be realized as software incorporated in the computer 500, and are realized by the CPU loading the program (software) held in the storage device into the memory and executing it. Various data used in the processing of each function and various data generated during the processing are stored in the storage device or the external storage device 508. Furthermore, at least one of the various functions realized by the computer 500 may be realized by an information processing device independent of the MRI apparatus 10 and capable of transmitting and receiving data to and from the MRI apparatus 10. Furthermore, all or a part of the functions may be realized by hardware such as an ASIC (Application Specific Integrated Circuit) or an FPGA (Field-programmable gate array).

[0031] [Overview of MRI system operation] Next, an overview of the operation of the MRI apparatus (mainly the computer 500) of this embodiment will be described.

[0032] The flow of operations in this embodiment is shown in FIG. First, settings of an imaging sequence and imaging conditions by a user are accepted via the input device 509 (S301). Radial sampling is set for the imaging sequence, and an imaging method (PI method) for measuring spatially overlapped signals for each blade is selected and set in order to shorten the imaging time. Spatially overlapped signals refer to signals that are overlapped from different positions in real space without being encoded by a gradient magnetic field, and include signals that are spatially overlapped by undersampling (so-called signals including aliasing). When these imaging conditions, etc. are set as an examination protocol, the conditions, etc. set in the examination protocol are read.

[0033] The measurement control unit 310 operates the sequence control device 111 in accordance with a pulse sequence set based on parameters input by the user, and measures nuclear magnetic resonance signals (echo signals) under predetermined conditions (S302).

[0034] As an example of a pulse sequence by radial sampling used by the measurement control unit 310, an example of a pulse sequence of 2D-GrE is shown in FIG. 4. In this figure, RF, Gs, G R1 , G R2 , A / D respectively represent the radio frequency magnetic field, slice gradient magnetic field, gradient magnetic field 1, gradient magnetic field 2, and analog / digital conversion. Here, as an example, imaging is performed with a thinning rate of 1 / 2 (speed rate of 2).

[0035] First, a slice gradient magnetic field pulse 401 is applied and a radio frequency magnetic field (RF) pulse 402 is irradiated to excite magnetization of a predetermined slice in the subject 101. Next, a rephase slice gradient magnetic field pulse 403 is applied to converge the phase of the magnetization dispersed by the application of the slice gradient magnetic field pulse 401, and a gradient magnetic field pulse G R1 (404), G R2 While applying the signal (405), a nuclear magnetic resonance signal (echo) is acquired and subjected to analog / digital conversion (406).

[0036] The measurement control unit 310 executes the above steps as R1 , G R2 The echoes required to obtain an image of one blade are measured by repeatedly changing the intensity of the echoes at a repetition time TR. In order to perform measurements at twice the speed, the echoes in the blade are thinned out by skipping every other encode in the blade. This reduces the imaging time by half. By executing such a pulse sequence, measurement space data is acquired for each receiving coil and for each blade.

[0037] The image generating unit 220 performs an inverse Fourier transform on the measurement space data for each blade in the complex space to obtain image space data, and then separates the spatially overlapping signals for each blade using the image data (unseparated image) for each blade and the sensitivity distribution of multiple receiving coils to calculate the blade image (S303). The sensitivity distribution of the receiving coils can be obtained in advance by a preliminary measurement. The calculation of the PI method that separates images using the sensitivity distribution of the receiving coils can be broadly divided into processing in the measurement space (e.g., the GRAPPA method) and processing in the image space (e.g., the SENSE method), and either of them may be adopted.

[0038] The gridding processing unit 210 performs a Fourier transform on the separated blade images from which noise has been removed, and then arranges the measurement space data for each blade at the measurement points in the k-space by gridding to obtain k-space data (S305). The image generating unit 220 performs an inverse Fourier transform on the gridded k-space data to generate an image. That is, overlapping signals are separated, and a single image in which each blade image is synthesized is obtained (S306).

[0039] The noise removal unit 230 removes noise from the image by repeated calculations (S304 or S307). The noise removal is performed on the blade image generated in S303 or the image generated in S306, using the g-factor map calculated for each blade or a g-factor map obtained by integrating the g-factor maps. The noise removal is achieved by performing repeated calculations to minimize noise under a predetermined constraint condition, similar to noise removal using a nonlinear filter including known Total Variation regularization or sparse regularization, but the noise removal unit 230 of this embodiment performs repeated calculations that include the g-factor map for each blade or the integrated g-factor map as a constraint condition.

[0040] The display control unit 320 displays the generated image on the display 510, or records the information in the external storage device 508 (S308).

[0041] According to this embodiment, by calculating a g-factor map for each blade and adding it to the constraints in the repeated noise removal calculations, noise that depends on the g-factor in PI method images can be effectively removed even when the PI method is applied to radial sampling, which is a non-Cartesian coordinate system sampling.As a result of both the sampling method and the shortened imaging time of the PI method, it is possible to obtain images with high suppression of body movement artifacts and reduced noise.

[0042] The following describes the details of the process for removing noise in two embodiments that differ from each other. Note that the configuration of the MRI apparatus and the outline of the process described above are common to both embodiments and will be referred to as appropriate.

[0043] First Embodiment In this embodiment, noise is removed from each blade image before the final image is obtained by gridding. In addition, a method for separating spatially overlapping signals for each blade using the SENSE method is adopted as the calculation of the PI method.

[0044] The configuration of the calculation unit 200 (image generation unit) of this embodiment is shown in Fig. 5. As shown in the figure, the image generation unit 220 includes a complex blade image calculation unit 221, a blade image calculation unit 222, and a reconstruction unit 223 that generates a final image, and the noise removal unit 230 includes a plurality of constraint units (231 to 233) that set constraint conditions in the iterative calculation, a g-factor calculation unit 234 that calculates a g-factor map used in the iterative calculation, and a iterative calculation unit 235 for executing the iterative calculation.

[0045] The function and operation of each part of the calculation part 200 of this embodiment will be described below with reference to Fig. 6 and Fig. 7. Fig. 6 is a diagram showing the flow of the operation, and Fig. 7 is a diagram showing the data obtained at each step.

[0046] [Complex blade image calculation: S601] In S302 of FIG. 3, the complex blade image calculation unit 221 arranges the echo signals measured by the multiple receiving coils for each blade in a measurement space with the blade angle direction and the phase encoding direction in the blade perpendicular thereto as axes to obtain measurement space data (FIG. 7: 701), and calculates a complex blade image (FIG. 7, 702) for each blade by performing an inverse Fourier transform on this measurement space data. Here, since the coordinate system of the measurement space rotates for each blade, the coordinate system of the obtained complex blade image also has a rotated shape similar to the relationship of the measurement space for each blade, as shown in FIG. 7. In addition, since the phase encoding is thinned out, each blade image is, for example, as shown in FIG. 8, an image 810 in which both sides (parts indicated by diagonal lines) 801 and 802 in the phase encoding direction of the subject 800 are folded back to the opposite sides (the right side part is turned to the left, and the left side part is turned to the right) and spatially overlap. In the figure, images of two receiving coils C1 and C2 are shown.

[0047] [Blade image calculation: S602] The blade image calculation unit 222 separates the spatially overlapped complex blade images 702 using the sensitivity distribution of multiple receiving coils and the SENSE method. Here, taking one blade as an example, the signal of the separated blade image at position n (n is an integer from 1 to N: N is the number of overlapping images, i.e., the speed multiplication number: in this example, 2) is expressed as ρ n , the sensitivity of receiving coil m (m is an integer from 1 to M) at position n is C mn Then, the blade image signal S generated from the receiving coil m is m is expressed by equation (1).

[0048]

number

[0049] If the vector and matrix in equation (1) are vector S, matrix C, and vector ρ, respectively, the signal vector ρ (the signal of the blade image after separation) can be calculated from equation (2).

[0050]

number

[0051] For example, when signals measured at twice the speed (N=2) in the SENSE method are separated by using two receiving coils (M=2), vector ρ is a 2×1 vector, sensitivity matrix C is a 2×2 matrix, and vector S is a 2×1 vector. By using equation (2), it is possible to separate spatially overlapping images ρ1 and ρ2 as shown in FIG. 8. In other words, image 820 can be obtained. Here, the separated individual images ρ1 and ρ2 are called separated images (or separated blade images), and image 820 after separation is simply called a blade image. When the above calculations are applied to all blades, the blade image ρ of each blade is bl can be calculated by the formula (3). Note that bl represents the blade number (bl is an integer from 1 to BL), and BL blade images (FIG. 7, 703) can be calculated.

[0052]

number

[0053] In formula (3), C bl represents the matrix for each blade obtained by transforming the sensitivity matrix C into the coordinate system of each blade, and the superscript H indicates that it is a complex transpose matrix. S bl is a matrix representing the signal of the blade image.

[0054] [Noise Reduction: S603] The noise removal unit 230 performs a noise removal process by performing repeated calculations (iterative processing) so as to minimize a function defined as a constraint condition. In this embodiment, a case will be described in which the constraint conditions for the repeated calculations are that the blade image before noise removal and the blade image after noise removal are substantially equal (hereinafter referred to as image constraint before and after noise removal), a constraint using a g-factor (a constraint that ensures that the areas with a large amount of noise in the image before noise removal and the image after noise removal are not too far apart) (hereinafter referred to as noise-related constraint), and a constraint that the noise of an image obtained by mapping the blade image into a sparse space is substantially equal to zero (hereinafter referred to as sparse space constraint).

[0055] The noise removal (S603) including the generation of the constraint conditions will be described in detail below. FIG. 9 shows the process flow of the noise removal. First, the g-factor calculation unit 234 calculates a g-factor map for each blade to be used as a weight for the repeated calculation (S6031). The image constraint unit 231 before and after noise removal generates a function representing the image constraint before and after noise removal (S6032), the noise constraint unit 233 generates a function representing the noise constraint using the g-factor for each blade (S6033), and the sparse space constraint unit 232 generates a function representing the sparse space constraint (S6034). The order of steps S6032 to S6034 may be any order, or they may be performed in parallel. Finally, the functions generated by the constraint units 231 to 233 are combined to perform repeated calculation processing for each blade (S6035), and the noise processing 704 shown in FIG. 7 is performed. Next, each process will be described in detail.

[0056] [G-factor map for each blade: S6031] The g-factor calculation unit 234 calculates the sensitivity matrix C for each blade used in the formula (3). bl and its complex transpose matrix C H bl Using the noise correlation matrix Ψ between the receiving coils, the g-factor map for each blade is calculated using equation (4).

number

[0057] [Function of image constraints before and after noise removal: S6032] The pre- and post-noise-removal image constraint unit 231 generates a function E1(I1, I2, bl) defined by equation (5), where the blade images before noise removal are ρ1 and ρ2, and the blade images after noise removal are I1 and I2, respectively.

[0058]

number

[0059] [Function for noise-related constraints: S6033] Next, the noise constraint unit 232 generates a function E2(I1, I2, bl) defined by equation (6).

number

[0060] [Sparse space based constraint determination: S6034] Next, the sparse spatial constraining unit 233 generates a function E3(I1, I2, bl) defined by equation (7).

number

[0061] In formula (7), Φ is a sparse space mapping operator that maps an image to a sparse space. The sparse space mapping operator can be a well-known one such as a Wavelet transform or a Curvelet transform, but here, for example, a Wavelet transform is used. In addition, ||||1 represents the L1 norm. A (i.bl) represents the weight image of each blade. Weight image A (i.bl) For example, g-factor can be used as the weight image. (i.bl) is not limited to this, and weighting may not be applied.

[0062] The function E3 (I1, I2, bl) in equation (7) represents a constraint condition (hereinafter referred to as sparse space constraint) for making an image mapped into a sparse space by the Wavelet transform into a further sparse image by the L1 norm.

[0063] [Repeat calculation: S6035] The iterative calculation unit 235 removes noise by iterative calculation processing based on the constraints generated by the three constraint units 231 to 233. That is, the function E defined by the formula (8) for each blade is total Calculate the denoised separated blade images I1 and I2 for each blade by minimizing (I1, I2, bl).

[0064]

number

[0065] [Griding Blade: S604] The gridding processing unit 210 performs a Fourier transform on the noise-removed blade image (Figure 7, 707) to obtain measurement space data, and then grids the measurement space data of each blade with different angles into the original k-space and places the data in k-space (Figure 7, 708).

[0066] [Image generated by: S605] The image generator 220 (reconstructor 223) generates an image by performing an inverse Fourier transform on the gridded k-space data (FIG. 7, 709). The generated image is displayed on the display 508, or stored in the storage device 510 or transferred to a device other than the MRI device, as in step S308 in FIG. 3.

[0067] According to the MRI apparatus and image processing method of the present embodiment, in an image acquired using a plurality of receiving coils, a noise removal process is performed in the image space for each blade image as a process, thereby making it possible to obtain a high-quality image from which noise has been effectively removed in radial sampling measurements, and to improve the accuracy of diagnosis using the MRI apparatus. In particular, in the noise removal for each blade image, by repeatedly performing calculations using the g-factor map for each blade as a constraint condition, it is possible to prevent deviation from the original image or excessive smoothing due to noise removal, and to improve the accuracy of noise removal for the blade image.

[0068] In this embodiment, three types of constraint conditions, i.e., image constraint before and after noise removal, noise constraint, and sparse space constraint, are used as constraint conditions for repeated noise removal calculations, but the constraint conditions are not limited to a combination of these three types as long as they include an additional constraint on the g-factor map for each blade, and can be modified to use, for example, a combination of two types of constraints, i.e., before and after noise image removal constraint, or a noise constraint and a sparse space constraint, or two or more types of sparse space constraint. Furthermore, it is also possible to add the correlation of noise in separated blade images as a constraint condition (noise correlation constraint), as disclosed in Patent Document 2.

[0069] The combination of these constraint conditions may be selected by the user, or the number of constraint conditions, combinations, or the regularization parameter (λ) and weight (A) when combined may be used as simple indices that can be selected by the user. For example, since it is known that the degree of noise removal varies by changing the regularization parameter (λ), the degree of noise removal may be selected by the user. FIG. 10 shows an example of a UI screen for user selection. In this example, for example, a UI for selecting the degree of noise (denoising strength) is set on an imaging parameter setting screen. Depending on whether "High", "Middle", or "Low" is selected, the weights can be made different, and the constraints to be combined or the number of constraints can be changed.

[0070] Second Embodiment Unlike the first embodiment in which noise removal is performed for each blade image, this embodiment is an embodiment in which the blade image is gridded, an image is generated by inverse Fourier transform, and then noise is removed using a separately created integrated g-factor map. Note that, hereinafter, the image generated by gridding the blade image and performing Fourier transform is referred to as a gridded image.

[0071] The configuration of the calculation unit 200 of this embodiment is shown in Fig. 11. In Fig. 11, the same elements as in Fig. 5 are designated by the same reference numerals, and duplicated explanations will be omitted. As shown in Fig. 11, the calculation unit 200 of this embodiment additionally includes an integrated g-factor calculation unit 236 that integrates the g-factors for each blade to generate one g-factor map (referred to as an integrated g-factor map).

[0072] Hereinafter, the operation of each part of the calculation unit 200 will be described with reference to Fig. 12 and Fig. 13. In Fig. 12, steps that perform the same processes as those in the first embodiment shown in Fig. 6 are indicated by the same reference numerals, and detailed description thereof will be omitted.

[0073] As shown in Fig. 13, a complex blade image 702 is calculated by Fourier transform using measurement space data 701 obtained for each receiving coil and each blade (S601). Next, a blade image 703 is generated using the sensitivity distribution of the receiving coil transformed into the image space for each blade (S602). The blade image 703 is inverse Fourier transformed to return it to measurement space data 714 (S606). After that, the measurement space data of each blade is gridded to obtain k-space data 715 (S607), and a gridding image 716 is generated (S608).

[0074] Thereafter, the noise removal unit 230 performs noise processing on the gridding image 716. In the noise processing, first, the integrated g-factor calculation unit 236 creates a g-factor map 705 for each blade, and then integrates them to create an integrated g-factor map 717 (S609). Next, the noise removal unit 230 performs noise removal processing by repeated calculations (iterative processing) using the gridding image 716 and the integrated g-factor map (S610).

[0075] The calculation of the g-factor map for each blade is the same as in the first embodiment (equation (4)), and the sensitivity matrix C blThe method for integrating the g-factors for each blade into one map is arbitrary, but a method based on the Monte Calro method as described in Non-Patent Document 1, a method described in the following Non-Patent Document, or other methods may be used. [Non-patent document 3] Robson PM, Grant AK, Madhuranthakam AJ, Lattanzi R, Sodickson DK, McKenzie CA. Comprehensive quantification of signal-to-noise ratio and g-factor for image-based and k-space-based parallel imaging reconstructions. Magn Reson Med 2008;60:895-907.

[0076] The technology disclosed in Non-Patent Document 1 uses a g-factor map as a performance evaluation of high-speed imaging that combines radial scan and SENSE, and is not used for noise removal.

[0077] The method of noise removal by the noise removal unit 230 is basically the same as that in the first embodiment, and noise removal is performed by repeated calculations using the noise removal before and after constraints, the g-factor constraint, and the sparse space constraint. FIG. 14 shows a process flow for removing noise in this embodiment.

[0078] [Constraint determination based on images before and after noise removal: S6101] The pre- and post-noise-removal constraint unit 231 generates a function E1(I) defined by equation (8), where each gridding image before noise removal is ρ and each gridding image after noise removal is I.

[0079]

number

[0080] [Noise-related constraint determination: S6102] Next, the noise constraint unit 232 generates a function E2(I) defined by equation (10).

[0081]

number

[0082] [Sparse space based constraint determination: S6103] Next, the sparse spatial constraint unit 233 generates a function E3(I) defined by equation (11).

[0083]

number

[0084] [Repeat calculation: S6104] The iterative calculation unit 235 removes noise by iterative calculation processing based on the constraints generated by the three constraint units 231 to 233. That is, the function E defined by the formula (12) is total A noise-removed image is calculated by minimizing (I).

[0085]

number

[0086] Through the above steps S6101 to S6104, the noise removal step S610 is completed, and a post-noise removed image 718 is obtained as shown in FIG.

[0087] According to the MRI apparatus and image processing method of the present embodiment, in an image acquired using a plurality of receiving coils, noise removal is performed by repeated calculations using the noise correlation of the gridding image and the integrated g-factor map as constraints, thereby improving the accuracy of noise removal of the gridding image, allowing a high-quality image to be calculated, and improving diagnostic accuracy. Furthermore, according to the present embodiment, by performing repeated calculations using constraints necessary for noise reduction together with the noise correlation constraint, it is possible to prevent deviation from the original image or excessive smoothing due to noise removal in radial sampling measurements.

[0088] In this embodiment, three constraint conditions were set and noise processing was performed by repeated calculations, but the number and combination of constraint conditions can be changed as appropriate, and the regularization parameters and weights of each constraint condition can be changed according to the denoising strength specified by the user, as in the first embodiment.

[0089] In the above embodiment, two-dimensional imaging (FIG. 3) has been taken as an example; however, the present invention can also be applied to three-dimensional imaging in the same manner as in the case of two-dimensional imaging by using three-dimensional maps as the reception sensitivity map and the g-factor map. [Explanation of symbols]

[0090] 10: MRI apparatus, 100: measurement unit, 101: subject, 102: static magnetic field coil, 103: gradient magnetic field coil, 104: shim coil, 105: transmission coil, 106: reception coil, 107: transmitter, 108: receiver, 109: shim power supply unit, 110: gradient magnetic field power supply unit, 111: sequence control device, 200: calculation unit, 210: gridding processing unit, 220: image generation unit, 221: complex blade image calculation unit, 222: blade image calculation unit, 230: noise removal unit, 300: control unit, 310: measurement control unit, 320: display control unit, 500: calculator, 508: external storage device, 509: input device, 510: display

Claims

1. a measurement unit that applies a high frequency magnetic field and a gradient magnetic field to a subject placed in a static magnetic field space, and measures measurement data consisting of nuclear magnetic resonance signals generated from the subject using a plurality of receiving coils; A control unit that controls the operation of the measurement unit; a calculation unit that processes the measurement data and creates an image, The control unit divides a k-space in which measurement data is arranged into a plurality of blades, and controls the measurement unit to perform measurement by thinning out the phase encoding for each blade; the calculation unit includes an image generation unit that generates an image by a parallel imaging method using the measurement data and a sensitivity distribution of the plurality of receiving coils, and a noise removal unit that performs noise removal on the reconstructed image by repeated calculations using a constraint condition that reflects noise characteristics of each blade as a constraint condition regarding noise, The noise removal unit is provided with a g-factor calculation unit that calculates a g-factor map for each blade as a constraint condition reflecting the noise characteristics of each blade, and a constraint function using the g-factor map is set as a constraint condition regarding the noise.

2. 2. The magnetic resonance imaging apparatus according to claim 1, A magnetic resonance imaging device characterized in that the image generation unit has a blade image calculation unit that calculates an image for each blade using measurement data for each blade and the sensitivity distribution of the multiple receiving coils, and the noise removal unit performs noise removal on the blade image calculated for each blade using a g-factor map for each blade.

3. 2. The magnetic resonance imaging apparatus according to claim 1, A magnetic resonance imaging device characterized in that the image generation unit has a blade image calculation unit that calculates an image for each blade using measurement data for each blade and the sensitivity distribution of the multiple receiving coils, and the noise removal unit performs noise removal on the blade image calculated for each blade.

4. 4. The magnetic resonance imaging apparatus according to claim 3, A magnetic resonance imaging device characterized in that the image generation unit includes a gridding unit that rearranges measurement data to measurement points in k-space, and a reconstruction unit that reconstructs an image using the k-space data after gridding, and the gridding unit converts the images for each blade processed by the noise removal into measurement space and performs gridding.

5. 2. The magnetic resonance imaging apparatus according to claim 1, a noise removal unit that removes noise from the image generated by the reconstruction unit through repeated calculations using constraints on noise, the noise removal unit removing noise from the image generated by the reconstruction unit; and a noise removal unit that removes noise from the image generated by the reconstruction unit through repeated calculations using constraints on noise.

6. 6. The magnetic resonance imaging apparatus according to claim 5, The noise removal unit is provided with an integrated g-factor map calculation unit that calculates a g-factor map for each blade and calculates an integrated g-factor map using the g-factor for each blade, and a constraint function using the integrated g-factor map is set as a constraint condition regarding the noise.

7. 2. The magnetic resonance imaging apparatus according to claim 1, The magnetic resonance imaging apparatus according to claim 1, wherein the noise removal unit uses two or more types of constraint conditions including a constraint condition related to the noise as a constraint condition for the iterative calculation.

8. A method for generating an image by dividing a k-space into a plurality of blades using a magnetic resonance imaging apparatus equipped with a plurality of receiving coils, and processing measurement data for each blade obtained by thinning out phase encoding for each blade, comprising the steps of: generating an image of each blade by a parallel imaging method using the measurement data of each blade and the sensitivity distribution of the plurality of receiving coils; performing a noise reduction process on the image of each blade by repeated calculations; A step of converting the image for each blade after noise removal into data in a measurement space, and then gridding and relocating the data to the measurement points in the k-space; and performing image reconstruction using the gridded k-space data; The image generation method is characterized in that the step of performing the noise removal process includes calculating a g-factor map for each blade, setting a constraint function using the g-factor map, and performing noise removal on the image of each blade by repeated calculations using the constraint function using the g-factor map as a constraint condition reflecting the noise characteristics of each blade.

9. A method for generating an image by dividing a k-space into a plurality of blades using a magnetic resonance imaging apparatus equipped with a plurality of receiving coils, and processing measurement data for each blade obtained by thinning out phase encoding for each blade, comprising the steps of: generating an image of each blade by a parallel imaging method using the measurement data of each blade and the sensitivity distribution of the plurality of receiving coils; A step of converting the image for each blade into data in a measurement space, and then gridding and relocating the data to the measurement points in the k-space; reconstructing an image using the gridded k-space data; and performing a noise removal process on the reconstructed image by repeated calculations using a constraint condition reflecting the noise characteristics of each blade as a constraint condition on noise, The noise removal step includes a step of calculating a g-factor map for each blade, and further calculating an integrated g-factor by integrating the g-factor maps for each blade, and the noise removal process is performed using the integrated g-factor as a constraint condition for repeated calculations as a constraint condition reflecting the noise characteristics of each blade.

Images