Magnetic resonance imaging apparatus and computer program product

Abstract

The present invention relates to a magnetic resonance imaging apparatus and a sensitivity distribution calculation program. When obtaining a sensitivity distribution in k-space, spectral leakage is suppressed by unfolding data that serves as the basis for obtaining the sensitivity distribution like a mirror image, thereby stably calculating the sensitivity distribution. When obtaining a sensitivity distribution in k-space, image data that serves as the basis for obtaining the sensitivity distribution is unfolded like a mirror image, and the unfolded image is transformed into k-space data to calculate the frequency components (frequency space data) of the sensitivity distribution. A region corresponding to the original image data is cut out from the calculated frequency space data to obtain the sensitivity distribution.

Description

Technical Field

[0001] The present invention relates to a magnetic resonance imaging apparatus (hereinafter referred to as an MRI apparatus), and particularly to a technique for calculating a sensitivity distribution in k-space. Background Art

[0002] There is a technique (Non-Patent Document 1, Non-Patent Document 2) of receiving data in an interleaved manner using a multi-channel receiving coil and reconstructing an image using information on the sensitivity distribution of the coil. In order to perform this image reconstruction, it is necessary to correctly calculate the sensitivity distribution of the coil, but there is a problem that it is difficult to correctly calculate the sensitivity distribution in the boundary portion of the subject to be detected. This is because: in order to perform stable reconstruction, it is necessary to calculate the sensitivity distribution so that a little extrapolation can be performed even in a region where the subject to be detected does not exist, but this extrapolation is difficult; at the boundary of the subject to be detected, the received data that is the basis for calculating the sensitivity becomes small, the SNR becomes low, and the error becomes large. In order to solve this problem, there is a technique (Patent Document 1, Patent Document 2) of obtaining the sensitivity distribution in k-space (frequency domain).

[0003] In the operation of obtaining the sensitivity distribution in k-space, an image of each channel and an image that does not depend on the sensitivity distribution (reference image) are used, and the sensitivity is obtained by deconvolving the frequency component of the sensitivity distribution convolved in k-space in the reference image in k-space. In such an operation in k-space, the sensitivity distribution in the boundary region where the signal is small is also obtained using the surrounding region where the signal is large, and it is possible to prevent the sensitivity distribution from becoming extremely incorrect in the boundary region. In addition, based on the assumption that the sensitivity distribution does not change significantly, it is possible to make the size of the kernel obtained by deconvolution small, and it is possible to stably obtain a sensitivity distribution in which large changes are restricted and natural.

[0004] Prior Art Documents

[0005] Patent Documents

[0006] Patent Document 1: JP Patent No. 6820876 Specification

[0007] Patent Document 2: US Patent No. 10353023 Specification

[0008] Non-Patent Documents

[0009] Non-Patent Document 1: Klaas P. Pruessmann, et al. Magnetic Resonance in Medicine 42: 952-962 (1999)

[0010] Non-Patent Document 2: Mark A. Griswold, et al., Magnetic Resonance in Medicine 47: 1202-1210 (2002)

[0011] In the techniques for obtaining the sensitivity distribution by operations in the k-space as disclosed in Patent Documents 1 and 2, countermeasures against spectral leakage are required. If the countermeasures against spectral leakage are insufficient, in regions with small signals, the sensitivity distribution will vibrate or become unstable. In order to reduce the influence of spectral leakage, there is a method of padding zeros to the data for calculating the sensitivity distribution in the image space to expand it to a given FOV. However, if the FOV is expanded by padding zeros, there is a problem that the degree of freedom of the solution increases and large vibrations occur in the signal-free region.

[0012] Although vibrations can be suppressed by reducing the kernel size during convolution, if the kernel size is reduced, the corresponding frequency range becomes narrower, and there is a problem that high frequencies can no longer be represented. Summary of the Invention

[0013] In order to solve the above problems, the present invention calculates the sensitivity distribution while suppressing spectral leakage by expanding the data that is the basis for calculating the sensitivity distribution like a mirror image.

[0014] That is, the MRI apparatus of the present invention includes: a measurement unit having a reception coil composed of a plurality of channels, and measuring nuclear magnetic resonance signals of a subject for each channel of the reception coil; and an image calculation unit that generates an image of the subject using the sensitivity distribution of each channel of the reception coil and the channel images generated from the nuclear magnetic resonance signals of each channel measured by the measurement unit. The image calculation unit includes a sensitivity distribution calculation unit that calculates the sensitivity distribution of each channel of the reception coil in the k-space using an image for sensitivity distribution obtained for the sensitivity distribution. The sensitivity distribution calculation unit has an expanded image creation unit that performs one or more inversion processes on at least a part of the image for sensitivity distribution, and generates an expanded image that combines the generated inverted image and the original image for sensitivity distribution. The sensitivity distribution calculation unit calculates the sensitivity distribution using the expanded image.

[0015] In the present invention, the "sensitivity distribution" includes not only the sensitivity distribution in the image space and the k-space (frequency space), but also information equivalent to the sensitivity distribution such as the weight coefficients in the k-space used in the image reconstruction method (GRAPPA method) of Non-Patent Document 2. The present invention can also be applied when obtaining such sensitivity distribution information.

[0016] Advantages of the Invention

[0017] According to the present invention, when obtaining the sensitivity distribution in k-space, the sensitivity distribution can be obtained more stably. In addition, while ensuring stable acquisition of the sensitivity distribution, the kernel size can be increased, and the high-frequency components of the sensitivity distribution can also be represented. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 is a functional block diagram of an MRI apparatus.

[0019] Figure 2 is a diagram showing the overall structure of the MRI apparatus according to the embodiment.

[0020] Figure 3 is a diagram showing the imaging order of the first embodiment.

[0021] Figure 4 is a diagram showing an example of a pulse sequence for calculating the sensitivity distribution in the first embodiment.

[0022] Figure 5 is a diagram showing the order of calculating the sensitivity distribution.

[0023] Figure 6 is a functional block diagram of the sensitivity distribution calculation unit.

[0024] Figure 7 is a diagram showing Figure 5 the details of the slice cutting process of the flow of.

[0025] Figure 8 For (A) of, it is a diagram showing the details of the process of expanding into a mirror image, and (B) is a diagram showing the details of the process of cutting out from the mirror image.

[0026] Figure 9 is a diagram showing the process of the first embodiment.

[0027] Figure 10 is a diagram showing the effect of the first embodiment.

[0028] Figure 11 is a diagram explaining the process of Modification 1 of the first embodiment.

[0029] Figure 12 is a diagram showing the generation of the expanded image of Modification 3 of the first embodiment.

[0030] Figure 13 is a diagram showing the generation of the expanded image of Modification 4 of the first embodiment.

[0031] Figure 14 is a diagram showing the generation of the expanded image of the second embodiment.

[0032] Figure 15This is a diagram showing the process of the third embodiment.

[0033] Figure 16 This is a diagram showing an example of the user interface.

[0034] Explanation of reference numerals

[0035] 10 Measuring unit, 20 Arithmetic unit, 21 Sensitivity distribution calculation unit, 23 Image generation unit, 30 Control unit, 40 User interface, 211 Layer cutting unit, 212 Expanded image creation unit, 213 Fourier transform unit, 214 Sensitivity distribution frequency component calculation unit, 215 Mirror cutting unit, 216 Mask image multiplication unit. Detailed implementation mode

[0036] First, the implementation mode of the MRI apparatus common to each of the embodiments of the present invention described later will be described.

[0037] The MRI apparatus of the present embodiment is configured as Figure 1 shown, and includes: a measuring unit 10 that acquires nuclear magnetic resonance signals; an arithmetic unit (image arithmetic unit) 20 that performs operations such as image reconstruction using the nuclear magnetic resonance signals (hereinafter referred to as MR signals) acquired by the measuring unit; and a control unit 30 that controls the operations of each unit constituting the MRI apparatus, such as the measuring unit 10 and the arithmetic unit 20. The control unit 30 may be a part of the function of the arithmetic unit 20. Further, the MRI apparatus can include a user interface device 40, and the user interface device 40 includes devices for the user to input instructions for starting / ending imaging, setting imaging conditions, etc. to the measuring unit 10 and the control unit 30, or for presenting the processing results of the arithmetic unit 20 to the user.

[0038] The measuring unit 10 has the same structure as a general MRI apparatus, and includes: a static magnetic field generation unit 11 that generates a uniform magnetic field in the space where the subject is placed; a gradient magnetic field generation unit 12 that applies a magnetic field gradient to the static magnetic field; a transmission unit 13 that irradiates a high-frequency magnetic field pulse (hereinafter referred to as an RF pulse) that excites the nuclei of the atoms constituting the tissue of the subject, typically protons; and a reception unit 14 that receives the nuclear magnetic resonance signals generated from the subject.

[0039] Specifically, as Figure 2 shown, the static magnetic field generation unit 11 includes a static magnetic field coil 102 and a shim coil 104 for correcting its magnetic field uniformity and its power supply unit 113, the gradient magnetic field generation unit 12 includes a gradient magnetic field coil 103 and its power supply unit 113, the transmission unit 13 includes a transmission coil 105 and a transmitter 107 that transmits an RF signal to the transmission coil 105, and the reception unit includes a reception coil 106 and a receiver 108 that demodulates the MR signals received by the reception coil 106.

[0040] The static magnetic field generation unit 11 can use a static magnetic field coil 102 of a normal conducting magnet type or a superconducting magnet type, or a permanent magnet. In addition, depending on the orientation of the magnetic field, there are a vertical magnetic field type, a horizontal magnetic field type, or a type in which the orientation of the magnetic field is inclined with respect to the horizontal direction, etc., and any one of them can be adopted.

[0041] The gradient magnetic field coil 103 is shown as a single block in Figure 2 , but it is composed of gradient magnetic field coils in three mutually orthogonal axis directions and is respectively connected to their respective power supply units 112. By applying the gradient magnetic field of each axis, the excitation region of the object to be detected can be selected, or position information can be given to the MR signal.

[0042] The transmit coil 105 is generally fixed in the static magnetic field space together with the shim coil 104 and the gradient magnetic field coil 103.

[0043] The receive coil 106 is shown as a single block in Figure 2 , but it is composed of a receive coil having a plurality of channels and has a function of receiving for each channel. As the multi-channel receive coil, known multi-channels such as a receive coil formed by combining sub-coils corresponding to each channel, a receive coil in which a plurality of loop coils are arranged in a given configuration, etc. can be used. Here, the sub-coils or the loop coils respectively correspond to the respective receive coils. In addition, as the receive coil 106, a body coil that also serves as a transmit coil can be used, and reception is performed by switching.

[0044] Under the control of the control unit 30, the measurement unit 10 causes the transmit coil 105 and the gradient magnetic field coil 103 to operate in accordance with the imaging conditions specified by the user and a given pulse sequence, irradiates an RF pulse to the object to be detected 101 in the static magnetic field, and the receive coil 106 receives the MR signal generated from the object to be detected thereby. The control of the measurement unit 10 based on the pulse sequence among the controls of the control unit 30 is achieved by the sequence control device 114. As long as there is no special description, the imaging method performed by the measurement unit 10 can adopt known MRI techniques.

[0045] The operation unit 20 includes: a sensitivity distribution calculation unit 21 that calculates the sensitivity distribution of the receiving coil using the nuclear magnetic resonance signal measured by the measurement unit 10; and an image generation unit 23 that generates an image using the sensitivity distribution calculated by the sensitivity distribution calculation unit 21 and the k-space data as measurement data. In the present embodiment, the sensitivity distribution calculation unit 21 uses the sensitivity distribution calculation data, obtains the frequency component of the sensitivity distribution of each channel of the receiving coil through operations in the k-space, and uses this frequency component to calculate the coil sensitivity distribution as real-space data or the weight in the k-space equivalent to the sensitivity distribution (collectively referred to as the sensitivity distribution). At this time, inversion processing and unfolded image creation processing, which will be described in detail later, are performed on the image used for the sensitivity distribution, and the unfolded image is used to calculate the sensitivity distribution. The image generation unit 23 performs image reconstruction based on the parallel imaging method using the calculated sensitivity distribution or the weight in the k-space.

[0046] The functions of the operation unit 20 can be realized, for example, by loading and executing a program that determines the order of sensitivity distribution calculation and image generation on a computer equipped with a CPU or GPU and a memory. Among them, part or all of the operations can also be configured by hardware such as an ASIC (Application Specific Integrated Circuit) or an FPGA (Field Programmable Gate Array), or realized by a computer different from the MRI device or an operation unit on the cloud.

[0047] The control unit 30 includes a sequence control device 114 that controls the measurement unit 10 according to a given pulse sequence, controls the operations of the measurement unit 10 and the operation unit 20. When the operation unit 20 is implemented by the CPU of a computer, part or all of this function can be realized by loading a control program in the same CPU, and part or all can also be realized by other CPUs or hardware.

[0048] In Figure 2 In the structure shown, the functions of the operation unit 20 and the control unit 30 implemented by software are embedded in the computer 109, and the computer 109 is connected to a display 110, an input device 115, an external storage device 111, etc. of the user interface device 40.

[0049] An overview of the shooting order in the case of performing parallel imaging using the MRI device of the present embodiment will be described. The shooting of parallel imaging, as Figure 3 shown, includes steps of measurement (S1001), sensitivity distribution calculation (S1002), signal separation processing (S1003), and image display (S1004).

[0050] In measurement step S1001, a multi-channel receiving coil is used, and each receiving coil measures the NMR signal and acquires image data for each channel. The NMR signal becomes k-space data of a given matrix size determined by phase encoding and the number of samples. However, in parallel imaging, high-speed measurement is performed by performing undersampling (skipping measurements) on the signal.

[0051] In the sensitivity distribution calculation step S1002, the sensitivity distribution of each channel is calculated. In the calculation of the sensitivity distribution, reference image data excluding the influence of the sensitivity distribution and the image data of each channel are used. These images for calculating the sensitivity distribution can be acquired by performing measurements separately from the measurement (formal measurement) in step S1001, and the data for the sensitivity distribution can also be acquired during the formal measurement. In the case of having a wide-area receiving coil such as a body coil, a wide-area receiving coil can also be used to acquire a reference image. In the MRI apparatus of the present embodiment, using these images for the sensitivity distribution, the sensitivity distribution of each channel or the weight coefficient in k-space is calculated by an operation in k-space. At this time, the inversion process is repeatedly performed on the image (original image) generated from the data for the sensitivity distribution to generate an unfolded image, and the sensitivity distribution is calculated in the frequency domain using this as the data for the sensitivity distribution. The details of the inversion process and the sensitivity distribution calculation will be described later.

[0052] In the signal separation processing step S1003, by performing undersampling, the signals that overlap in the image space are separated (unfolded) using the sensitivity distribution to generate an image without aliasing. The signal separation processing can be performed using known parallel imaging operations such as the SENSE method, the SMASH method, and the GRAPPA method.

[0053] In the image display step S1004, the image of the subject generated by the signal separation processing is displayed on the display 110 or the like.

[0054] According to the MRI apparatus of the present embodiment, instead of expanding the FOV of the sensitivity distribution image by zero-padding as in the conventional method to solve the problem of spectral leakage, the problem of spectral leakage is solved by using the unfolded image (mirror image) of the data for the sensitivity distribution, and the problem that the sensitivity distribution vibrates in the signal-free region is solved, so that a stable sensitivity distribution can be obtained. Furthermore, according to the present embodiment, the kernel size can be made relatively large during the deconvolution operation in the frequency domain, and a sensitivity distribution in the frequency domain with a wide bandwidth can be obtained.

[0055] The reasons are briefly described below.

[0056] When photographing a imaging area with a multi-channel coil, since the sensitivity is higher the closer it is to the coil, generally, if the sensitivity distribution becomes higher at one end of the FOV, it will become lower at the opposite end and become discontinuous at the FOV end. If Fourier transform (transformation from image data to k-space data) is performed in this state, spectral leakage will occur, and the sensitivity distribution will have a wide range of frequencies. In the existing method, to address this problem, the sensitivity distribution is extended by zero-padding the FOV of the image to twice its size, etc., and an arbitrary sensitivity distribution is allowed in the zero-padded area, so that the FOV ends can be connected, and thus a response is made. In this case, there will be an arbitrary sensitivity distribution in the zero-padded area, the degree of freedom of the solution increases, and large values exist throughout the entire kernel, and the sensitivity distribution in the zero-padded area and the background area will vibrate significantly.

[0057] In the present embodiment, instead of extending the sensitivity distribution by zero-padding the image, a mirror image is created to be an expanded image, thereby being able to eliminate the discontinuity at the FOV end while suppressing the degree of freedom of the solution, and being able to suppress the vibration of the sensitivity distribution in the background area.

[0058] The vibration of the sensitivity distribution in the background area is also suppressed by reducing the kernel size in the existing method, but reducing the kernel size can no longer correspond to high frequencies, and errors will occur. In the present embodiment, since the vibration can be suppressed, the kernel size can be made relatively large, and a sensitivity distribution that also represents high frequencies can be obtained.

[0059] Next, an embodiment that further specifies the function of the sensitivity distribution calculation unit will be described. In the following embodiments, the embodiment of the sensitivity distribution calculation method will be described centering on the function of the sensitivity distribution calculation unit 21, and the functions of the measurement unit 10 and the image generation unit 23 will also be mentioned as needed. In addition, regarding Figure 1 、 Figure 3 For the structures common to the respective embodiments shown in, appropriately refer to these drawings.

[0060] <First Embodiment>

[0061] In the present embodiment, the case of calculating a two-dimensional (2D) sensitivity distribution will be described. To calculate the 2D sensitivity distribution, the process of inverting the sensitivity distribution data (the sensitivity distribution image serving as the basis) obtained by the sequence according to the sensitivity distribution is repeated to generate a 2D expanded image, which is used in the calculation of the sensitivity distribution.

[0062] Refer again to Figure 3 to describe the processing of the MRI apparatus of the present embodiment. Here, as an example, the case of calculating a two-dimensional (2D) sensitivity distribution using the data (sensitivity distribution data) obtained by the three-dimensional (3D) sensitivity distribution sequence will be described.

[0063] First, in the measurement step ( Figure 3 : S1001), the measurement unit 10 performs a measurement for obtaining data for sensitivity distribution. In the present embodiment, data for sensitivity distribution is obtained by separately executing a pulse sequence for sensitivity distribution from a pulse sequence (formal imaging sequence) for formal imaging for obtaining an image of a subject to be detected. The data for sensitivity distribution is obtained by both a receive coil having a plurality of channels (hereinafter referred to as an array coil) and a body coil.

[0064] The type of the pulse sequence is not particularly limited. In Figure 4 , a pulse sequence 400 of a typical 3D-GrE (gradient echo) method is shown as an example of a sequence for sensitivity distribution. In the figure, the irradiation timing of an RF high-frequency magnetic field pulse 402 and the reception timing of a nuclear magnetic resonance signal 407 are shown in order from the top. Gs, Gp, and Gr respectively represent the polarity / intensity and application timing of gradient magnetic field pulses in three mutually orthogonal axial directions. Regarding the three-digit numbers, the same numbers indicate the same type of pulses. Regarding the numbers following the three-digit numbers with a hyphen, the same numbers indicate the pulses applied within the same repetition time. In addition, Gs is a slice selection gradient magnetic field, Gp is a phase encoding gradient magnetic field, and Gr is a readout gradient magnetic field. The longitudinal arrow shown beside the pulse indicates that the gradient magnetic field intensity changes whenever the pulse sequence repeats.

[0065] In the Figure 4 imaging sequence, first, an RF pulse 402 and a slice selection gradient magnetic field 401 are applied. After a given region is excited, a rephasing pulse 403 of a slice gradient magnetic field Gs and a phase encoding pulse 404 are applied. At this time, a dephasing pulse 405 of a readout gradient magnetic field Gr is applied, and a rephasing pulse with its polarity reversed is applied. At the same time, the nuclear magnetic resonance signal 407 is sampled. Finally, a rephasing pulse 409 is applied in the phase encoding direction. The process from the application of the RF pulse 402 to the application of the rephasing pulse 409 is repeated at a given repetition time TR, and the nuclear magnetic resonance signals required for image reconstruction are collected. The digital data of the sampled nuclear magnetic resonance signal becomes k-space data. In this sequence, 3D-data can be obtained. The 3D-data is obtained for each channel of the array coil and the body coil, respectively.

[0066] Regarding the formal imaging sequence and the sequence for sensitivity distribution, the types of the sequences may be the same or different. In the case of the sequence for sensitivity distribution, the number of encoding steps for slice encoding and phase encoding is small, and a low-resolution image for sensitivity distribution is obtained. Thus, when the sequence for sensitivity distribution is executed as a pre-scan different from the formal imaging sequence, the measurement time required for the pre-scan can be suppressed to a short time. However, for the low-resolution sensitivity distribution, it is inconsistent with the high-resolution formal imaging image in terms of resolution and field of view (FOV). Therefore, in order to apply the coil sensitivity distribution to image reconstruction, in the sensitivity distribution calculation process described later, a process for making it consistent with the formal imaging image is performed.

[0067] Next, in the sensitivity distribution calculation step ( Figure 3 : S1002), the sensitivity distribution calculation unit 21 calculates the coil sensitivity distribution applied to the image (object image) of the formal imaging using the data for sensitivity distribution (hereinafter referred to as Smap data) collected by the measurement unit 10. In this embodiment, since parallel imaging is performed, the coil sensitivity distribution is calculated for each of a plurality of receiving coils (channels).

[0068] In Figure 5 FIG. shows the details of the sensitivity distribution calculation step S1002. Generally speaking, the sensitivity distribution calculation includes a slice cutting process S151, an expanded image creation process S152, a transformation of the Smap data that is the expanded image into k-space data S153, a calculation of the frequency components of each receiving coil using the k-space data S154, a transformation of the frequency components (k-space) of the sensitivity distribution into real-space data S155, a process of cutting out the original region from the sensitivity distribution that is the expanded image S156, and further includes a process of multiplying by a mask image S157 as needed.

[0069] To implement these processes, the sensitivity distribution calculation unit 21, for example, as shown in the functional block diagram of Figure 6 , includes a slice cutting unit 211, an expanded image creation unit 212, a Fourier transform unit 213, a sensitivity distribution frequency component calculation unit 214, a mirror cutting unit 215, and a mask image multiplication operation unit 216. In addition, Figure 6 This is just an example of the sensitivity distribution calculation unit 21, and a part of the functional units can be appropriately omitted, or other functional units can be added. Hereinafter, as the operations of each unit, the details of the processes performed by the sensitivity distribution calculation unit 21 will be described.

[0070] [Slice Cutting Process S151]

[0071] The slice-out process is performed separately on the Smap data received by the array coil (Smap data having the same number as the number of channels) and the Smap data received by the body coil. In addition, since the Smap data obtained by the body coil is used as a reference in the subsequent sensitivity distribution calculation, it is referred to as Ref data (Ref image data with respect to the image space) in the following description.

[0072] In Figure 7 FIG. shows the details of the slice-out process S151.

[0073] In the slice-out process S151, first, the Fourier transform unit 213 performs Fourier transform on the 3D-Smap data and the 3D-Ref data respectively. After transforming from k-space data into image data (real-space data) (S1511), the slice-out unit 211 cuts out 2D image data (slice) from the 3D image data as real-space data (S1512). At this time, the 2D image data of the cut-out slice is cut out in such a way that its FOV becomes the same as the FOV of the object image (the image obtained according to the formal imaging sequence). Generally, in the sensitivity distribution measurement, since the sensitivity distribution image is acquired with a FOV larger than that of the object image, the FOV of the sensitivity distribution image is made to coincide with the FOV of the object image. However, regarding the matrix size of the image, the sensitivity distribution image is a low-resolution image smaller than the object image.

[0074] [Expanded image creation process S152]

[0075] The expanded image creation unit 212 repeatedly performs an inversion process on the 2D-Smap image data and the 2D-Ref image data obtained by the slice-out unit 211 through the slice-out process S151 to create an expanded image. In Figure 8 (A) of shows the details of the inversion process. In Figure 8 (A) of, the image 1A shown in the upper left is the 2D-Smap image data 801 obtained by the slice-out process S151 and is set as the original image. Initially, this original image 1A is copied to be line-symmetric on the lower side to create the first mirror image 1B. Next, the image (1A + 1B) formed by combining the original image 1A and the mirror image 1B is copied to be line-symmetric on the left side to create the second mirror image (1C + 1D), and the image (1A + 1B) and the second mirror image (1C + 1D) are combined to obtain the expanded image 802. The same process is also performed on the 2D-Ref image data.

[0076] [Fourier transform S153]

[0077] The Fourier transform unit 213 performs a Fourier transform on the expanded image obtained by mirror-expanding the expanded image created by the expanded image creation unit 212 through the expanded image creation process S152, to set it as data in the frequency space, i.e., k-space data.

[0078] [Frequency component calculation S154]

[0079] The sensitivity distribution frequency component calculation unit 214 calculates the frequency components of the sensitivity distribution based on the k-space data of the Smap image. The calculation of the frequency components of the sensitivity distribution can apply the method described in Patent Document 1. The following refers to Figure 9 to explain the method for calculating the frequency components of the sensitivity distribution.

[0080] In the expanded image creation process S152, the Smap image data created by mirror-expanding the data received by the i-th channel of the array coil is set as Si, and the Ref image data created by mirror-expanding the data received by the body coil is set as Sb.

[0081] In real space, if the vector at each position in real space is set as r, then the Smap image Si(r) of the i-th channel of the array coil is characterized by the product of the sensitivity distribution Ci(r) of each coil and the Ref image Sb(r) as shown in Equation (1) ( Figure 9 : 901).

[0082] (Mathematical formula 1)

[0083] Si(r) = Ci(r) × Sb(r) (1)

[0084] If both sides of Equation (1) are transformed into k-space data, the relationship of Equation (1) is characterized by the convolution integral of Equation (2) ( Figure 9 : 902).

[0085] (Mathematical formula 2)

[0086] Si’(k) = Ci’(k) * Sb’(k) (2)

[0087] In the formula, k is the k-space position vector, and Si’(k), Ci’(k), and Sb’(k) are the results obtained by transforming the real-space images of Si(r), Ci(r), and Sb(r) into k-space data, respectively (the same applies hereinafter).

[0088] As shown in Equation (2), Si’(k) is the result of convolving Sb’(k) and the frequency components Ci’(k) of the sensitivity distribution.

[0089] The above formula (2) can be characterized by the weight coefficient wi with a given kernel size as shown in formula (3) by ignoring the high-frequency regions of Si’(k) and Ci’(k).

[0090]

Mathematical formula 3

[0091]

[0092] In formula (3), kx and ky are the positions in k-space, bx and by are the positions in k-space within the kernel, and Bx and By are the kernel sizes. The kernel size is not particularly limited, but the smaller the kernel size, the larger the high-frequency components to be ignored. Generally, assuming a smooth variation according to the sensitivity distribution, small sizes such as 11×11 and 19×19 are set. Thus, even in the boundary region of the object to be detected, the data around it can be used to stably obtain the sensitivity distribution.

[0093] If the elements of Si’ and wi in formula (3) are arranged as vertical vectors and Sb’ is re-expressed as a matrix with the elements corresponding to each other in order, then formula (3) can be rewritten as formula (4), and the weight vector wi (i.e., the weight coefficient) can be calculated from the pseudo-inverse matrix shown in formula (5).

[0094]

Mathematical formula 4

[0095]

[0096]

Mathematical formula 5

[0097]

[0098] Here, H represents the adjoint matrix. Additionally, if the sizes of the Smap image are set to Nx and Ny respectively (where Nx and Ny are even numbers), then the size of the matrix Sb’ becomes (vertical, horizontal) = {(Nx - Bx + 1) × (Ny - By + 1), Bx × By} when the kernel size is odd.

[0099] This weight coefficient wi corresponds to the frequency components of the sensitivity distribution of the receiving coil (channel). Since the image that is the premise of the calculation is an unfolded image, it is not the frequency components of the sensitivity distribution itself ( Figure 9 : 903).

[0100] [Transformation S155 to real space]

[0101] After the Fourier transform unit 213 makes the obtained weight coefficient wi match the resolution of the formal imaging by zero-padding (zero-padding in k-space) or the like, it sets it as image data through inverse Fourier transform. The obtained image is as Figure 9As shown, S1 and Sb used when calculating the frequency components of the sensitivity distribution are the image 904 that unfolds the sensitivity distribution image.

[0102] [Mirror cut-out process S156]

[0103] The mirror cut-out section 215 cuts out from the image 904 after inverse Fourier transform the original region before mirroring where the original Smap image is unfolded (corresponding to the region of the Smap image 801), as shown in Figure 8 (B), to obtain the sensitivity distribution (sensitivity map) 905 as real-space data. Figure 8

[0104] [Mask multiplication operation S157]

[0105] After that, the mask image multiplication section 216 multiplies the sensitivity distribution image 905 after Fourier transform by a given mask image (not shown) in order to remove the noise around the sensitivity distribution. Therefore, the mask image multiplication section 216 creates a mask image based on the low-resolution image obtained as the sensitivity distribution image. The mask image can be created at any time as long as a slice is cut out from the 3D sensitivity image ( Figure 7 : 1512).

[0106] Through S151 to S157 described above, Figure 3 the sensitivity distribution calculation process S1002 ends. These processes are performed for each channel of the receive coil and for each slice position cut out in S151, and the sensitivity distribution of the quantity of slices is obtained for each channel.

[0107] After that, the process of generating an image using the sensitivity distribution calculated through the above processes and through parallel imaging operations ( Figure 3 : signal separation step S1003, image display step S1004) is as described above. However, in this embodiment, since the sensitivity distribution is obtained for each slice, if the image obtained according to the formal imaging sequence is a 3D image, the signal separation process is performed for each slice of the image at the corresponding slice position.

[0108] If the image obtained through formal imaging is a multi-slice image, in slice cut-out (S1512), the slice position that is the same as the slice of the formal imaging is cut out to calculate the sensitivity distribution, and for each slice of the corresponding multi-slice image, the signal separation process is performed using the sensitivity distribution at the corresponding slice position.

[0109] According to the present embodiment, when calculating the sensitivity distribution in the frequency domain, by using the image obtained by expanding the original image as the data for the sensitivity distribution, it is possible to suppress the vibration of the signal-free region generated in the sensitivity distribution obtained when directly using the original image, and the sensitivity distribution can be stably obtained. In addition, since vibration suppression can be effectively performed, the kernel size can be made relatively large, and thus the high-frequency components of the sensitivity distribution can also be represented, and the accuracy of the signal separation process in parallel imaging can be improved.

[0110] To confirm the effect of the present embodiment, the result of calculating the sensitivity distribution image by actually using an image of a human head is shown. In Figure 10 Figure (a) is the Smap image received by a certain channel, (b) is the sensitivity distribution image calculated by the sensitivity distribution calculation method of the present embodiment with a kernel size of 61×61, (c) is the image Smac(r) obtained by synthesizing each channel image of the Smap image using the sensitivity distribution by the following formula (6), and (d) is the absolute value of the difference between the result obtained by applying the sensitivity distribution of (b) to the synthesized image of (c) and the image of each channel of (a), which is the result of weighting the error of each channel's sensitivity distribution by the magnitude of the pixel value and becomes a reference for the magnitude of the error generated in image reconstruction. Figure 10 Here, the overline represents the complex conjugate.

[0111]

Mathematical formula 6

[0112]

[0113] On the other hand, (e), (f), and (g) are the results calculated by the existing method of zero-padding the FOV of the original image to about twice instead of the method of expanding into a mirror image of the present embodiment, corresponding to (b), (c), and (d) respectively. The kernel size is the same 61×61. Furthermore, in the existing method, the results of reducing the kernel size to 41×41 are shown in (h), (i), and (j).

[0114] It can be seen that compared with the sensitivity distribution (e) of the existing method, the vibration in the background region of the sensitivity distribution (b) of the present embodiment is less. In addition, since in the detected object boundary region, the error generated in the region indicated by the arrow in (g) does not occur in (d), it can be seen that the vibration in the detected object boundary region is also suppressed. Regarding the error in the region inside the detected object, it can be seen that they are equivalent after comparing (d) and (g).

[0115]

[0116] ​In the existing method, if the core size is reduced from 61×61 to 41×41 as shown in (h), the vibration of the sensitivity distribution can be suppressed to the same level as in (b), but the error in (j) generally becomes larger than that in (d).

[0117] The first embodiment of the MRI apparatus of the present invention has been described above, but the method for acquiring image data of the sensitivity distribution and the like is not limited to the method of the first embodiment, and various modifications can be made.

[0118] <Modification 1, Modification 2>

[0119] For example, in the first embodiment, the data received by the body coil is used as the Ref image data, but as shown in Figure 11 , instead of being received by the body coil, data obtained by synthesizing the Smap data of the array coil by the method described in Patent Document 1 (a synthesis method using a temporary coil sensitivity map) may be used (Modification 1).

[0120] In addition, in the first embodiment, the sequence for the sensitivity distribution is performed separately from the formal imaging sequence, but the sensitivity distribution calculation can also be performed using a part of the measurement data obtained in the formal imaging sequence (self-calibration) (Modification 2).

[0121] Furthermore, regarding the unfolded image creation process S152, the following modifications can be made.

[0122] <Modification 3>

[0123] In the first embodiment, in the unfolded image creation process S152, the image obtained by cutting out a slice from the 3D-Smap image data is directly used as a mirror image to obtain the unfolded image. However, in this modification, after padding zeros around the original images of the Smap image and the Ref image, the same inversion process as in the first embodiment is repeated to generate an unfolded image that unfolds the original image with zeros padded. Figure 12 shows an outline of the process of this modification.

[0124] In this modification, as shown in Figure 12 , zeros are padded around the original image 801 (1A) to obtain an image 1A0 with an extended FOV. In the zero-padding of the existing method, the image with zeros padded in this way is directly used in the calculation of the sensitivity distribution in the frequency domain. However, in this modification, the image 1A0 is subjected to the same process of duplicating it twice to be line-symmetric as in the first embodiment (S152) to generate an unfolded image (1A0 + 1B0 + 1C0 + 1D0) 803. The calculation of the frequency components of the sensitivity distribution using this unfolded image can be performed in the same manner as in the first embodiment.

[0125] According to this modification example, the mirror images can be smoothly combined by combining images with a little zero-padding around, and while maintaining the zero-padding effect (suppressing spectral leakage) of the existing method, the vibration in the signal-free region caused by the increased degree of freedom in calculating the sensitivity distribution can be suppressed.

[0126] <Modification Example 4>

[0127] In the first embodiment, the entire original image 1A is inverted to obtain the mirror image 1B, but it is also possible to combine them by inverting only a part of the image. Use Figure 13 to illustrate the processing of this modification example. In Figure 13 , for easy understanding, an image of 4×4 is shown as an example. The numbers 1 to 16 are the pixel numbers for distinction.

[0128] In the initial inversion process, only the pixels 5 to 12 (1B’) in the central region of the original image 1A composed of 16 pixels are copied and combined so that they are line-symmetric with respect to the pixels 1 to 4 of the original image 1A. Next, only the pixels in the central 2 columns of the combined image (1A + 1B’) are copied and combined so that they are line-symmetric with respect to the leftmost column pixels of the image (1A + 1B’), and the expanded image 804 is generated. The expanded image 804 formed in this modification example in this way is an image in which the upper left region (1A) in the expanded image 804 corresponds to the original image 801 and the matrix size is smaller than that of the expanded image in the first embodiment.

[0129] Using this expanded image 804 (Smap image and Ref image), the frequency components of the sensitivity distribution are calculated as the coefficient w1 in the k-space. The calculation method is the same as that in the first embodiment, but in this modification example, since the coefficient in the calculated k-space becomes a mirror image where the end regions do not overlap, similar to the image 804, only the range corresponding to the region 1A of the image 804 needs to be obtained. That is, the ranges of bx and by of the variable wi(bx, by) obtained in Equation (3) only become ranges of 0 or more as shown in Equation (7), and the scale of the calculation can be reduced. Figure 13

[0130]

Mathematical Formula 7

[0131]

[0132] In this way, according to this modification example, an expanded image is generated by combining images obtained by inverting only a part of the sensitivity distribution data, and it is used in the calculation of the sensitivity distribution in the frequency domain. Thus, in addition to the same effects as in the first embodiment, the effect of reducing the amount of calculation can also be obtained.

[0133] <Second Embodiment>

[0134] In the first embodiment, it was described that the formal imaging is multi-slice imaging and the 2D sensitivity distribution of each slice is obtained. However, in this embodiment, in 3D imaging, 3D data is acquired as the sensitivity distribution data, and the 3D sensitivity distribution is calculated.

[0135] In this embodiment, in the measurement step ( Figure 3 : S1001), both the formal imaging sequence and the sensitivity distribution sequence are 3D imaged to obtain 3D data as the sensitivity distribution data.

[0136] In the sensitivity distribution calculation step ( Figure 3 : S1002), the sensitivity distribution data (3D - Smap data) received by each channel of the receive coil and the sensitivity distribution data (3D - Ref data) received by the body coil are used. At this time, the 3D - image data after being transformed into the image space is expanded along each axis as Figure 14 shown to obtain the expanded image. That is, the original image (2A) 1401 is inverted in the x - axis direction to obtain a mirror image (2B) that is symmetric about the yz - plane. Next, the image 1402 obtained by combining the mirror image 2B and the original image 2A is inverted in the y - axis direction to obtain a mirror image (2C + 2D) that is symmetric about the xz - plane. The image 1403 obtained by combining this mirror image (2C + 2D) and the image 1402 is inverted in the z - axis direction to obtain a mirror image 1404 that is symmetric about the xy - plane. This mirror image 1404 is used as the expanded image for the subsequent sensitivity distribution calculation process ( Figure 5 S153 - S157).

[0137] For calculating the sensitivity distribution, an expression obtained by expanding each element of Equation (3) into three - dimensions can be used, and the frequency components of the sensitivity distribution are calculated in the same manner as in the first embodiment.

[0138] In this embodiment, regarding the generation of the expanded image, similar to Modification 3 and Modification 4 of the first embodiment, modifications such as padding zeros around the original image 1401 and combining by only inverting the central region of the original image 1401 can be adopted, and the same effects as Modification 3 and 4 can be obtained.

[0139] <Third Embodiment>

[0140] In the first embodiment and the second embodiment, for the signal separation process of each channel ( Figure 3On the premise of S1003), the sensitivity distribution of each channel is obtained. However, in the present embodiment, on the premise of parallel imaging using the GRAPPA method (the method described in Non-Patent Document 2) as an image generation method, information equivalent to the sensitivity distribution of each channel is calculated.

[0141] The processing of the present embodiment will be described by taking the case of 2D imaging as an example. In Figure 15 The processing flow of the present embodiment is shown. In Figure 15 Among them, the processing of the same content as Figure 5 is shown with the same reference numerals. As shown in the figure, in the present embodiment, slice extraction (S151) is performed on the 3D data obtained by the sequence according to the 3D-sensitivity distribution, and the unfolded images shown in the first embodiment and its modified examples 3 and 4 are used as the data for the sensitivity distribution used in the calculation (S152), and the unfolded images are transformed into k-space data (S153). These processes are the same as those in the first embodiment.

[0142] The GRAPPA method is a technique for reconstructing an image by interpolating non-measured lines in the k-space for the k-space data that has been undersampled. When interpolating non-measured lines, data of multiple measured lines including other coils are interpolated by weighted summation. The weights of the multiple measured lines correspond to the weights of each channel. In the present embodiment, the weights of the data of each channel are calculated as information equivalent to the sensitivity distribution (S158).

[0143] Specifically, in the calculation of the frequency component of the sensitivity distribution, instead of the formula (2) used in the frequency component calculation S154 of the first embodiment, the weight c i ' (k, j) of the i-th channel with respect to the j-th channel is obtained based on the following formula (8).

[0144] (Mathematical formula 8)

[0145] S i ' (k) = Σj c i ' (k, j) * S j ' (k) (8)

[0146] This weight c i ' (k, j) corresponds to the frequency component (the weight wi in the k-space) of the sensitivity distribution of each channel obtained in the first embodiment. In the first embodiment, the Fourier inverse transform is performed on this weight wi to obtain the sensitivity distribution in the image space. However, in the present embodiment, c i (k, j) is used as the weight for the measurement data in the k-space to interpolate non-measured lines, thereby performing image reconstruction.

[0147] The above has described the embodiments and modification examples of the MRI apparatus of the present invention. However, these are just examples, and elements can be added to or a part of the illustrated configuration and processes can be omitted.

[0148] In addition, in the above-described embodiments, the participation of the user has not been particularly mentioned. However, it is also possible for the user to specify whether to expand into a mirror image (whether it is the method described in the first embodiment) and whether to perform zero-padding, etc. (modification example 3 or modification example 4) via Figure 16 such a user interface.

[0149] The above has described the embodiments of the present invention. However, the function (sensitivity distribution calculation program) of the sensitivity distribution calculation unit 21 in these embodiments is not limited to the MRI apparatus. As long as it is a device (image processing apparatus) having a unit capable of performing sensitivity distribution calculation in the frequency space, it can be realized, and the function realized by such an image processing apparatus is also included in the present invention.

[0150] In addition, the present invention can be applied not only to parallel imaging in which k-space data are divided, but also to multi-slice selective excitation method (SMS) parallel imaging and multi-channel image synthesis (MAC synthesis) in the same way. In general MAC synthesis, since the sum of the squares of the sensitivity distributions is used, the phase information disappears. However, by using the sensitivity distribution obtained by the method of the present embodiment, an image (complex image) having phase information can be obtained.

Claims

1. A magnetic resonance imaging device, characterized in that, Comprising: A measurement unit having a receiving coil composed of a plurality of channels, and measuring nuclear magnetic resonance signals of a subject for each channel of the receiving coil; and An image calculation unit that generates an image of the subject by using the sensitivity distribution of each channel of the receiving coil and channel images generated from the nuclear magnetic resonance signals of each channel measured by the measurement unit, The image calculation unit includes a sensitivity distribution calculation unit that calculates the sensitivity distribution of each channel of the receiving coil in the k-space by using a sensitivity distribution image obtained for the sensitivity distribution, The sensitivity distribution calculation unit includes an expanded image creation unit that performs one or more inversion processes on at least a part of the sensitivity distribution image, and generates an expanded image that combines the sensitivity distribution image and the inverted image, The sensitivity distribution calculation unit calculates the sensitivity distribution by using the expanded image.

2. The magnetic resonance imaging apparatus according to claim 1, wherein The expanded image creation unit generates a mirror image obtained by inverting the entire image as the inverted image.

3. The magnetic resonance imaging apparatus according to claim 1, wherein The expanded image creation unit forms a partial mirror image obtained by inverting a region other than the end region of the image as the inverted image, and combines the partial mirror image with the end of the original sensitivity distribution image to generate the expanded image.

4. The magnetic resonance imaging apparatus according to claim 1, wherein The expanded image creation unit performs one or more inversion processes after expanding the periphery of the sensitivity distribution image by zero-padding.

5. The magnetic resonance imaging apparatus according to claim 1, wherein The sensitivity distribution calculation unit calculates the sensitivity distribution by using a region in the k-space corresponding to the region of the original sensitivity distribution image in the expanded image.

6. The magnetic resonance imaging apparatus according to claim 1, wherein The sensitivity distribution image is a three-dimensional image, The sensitivity distribution calculation unit further includes a slice extraction unit that extracts a two-dimensional sensitivity distribution image from the three-dimensional sensitivity distribution image, The sensitivity distribution calculation unit calculates the sensitivity distribution for each slice.

7. The magnetic resonance imaging apparatus according to claim 6, wherein The slice extraction unit extracts the two-dimensional sensitivity distribution image in such a manner that the FOV becomes the same as the FOV of the image of the subject created by the image calculation unit.

8. The magnetic resonance imaging apparatus according to claim 1, wherein The sensitivity distribution image is a three-dimensional image, The expanded image creation unit performs an inversion process on the three-dimensional image, And generates a three-dimensional expanded image obtained by expanding the three-dimensional image in three-dimensional directions.

9. The magnetic resonance imaging apparatus according to claim 1, wherein The sensitivity distribution image includes: A reference image; and A sensitivity distribution channel image generated from the nuclear magnetic resonance signals of each channel, The measurement unit further includes a wide - area receiving coil for measuring nuclear magnetic resonance signals for a reference image as the receiving coil.

10. The magnetic resonance imaging apparatus according to claim 1, wherein The sensitivity distribution image includes: A reference image; and A channel image for sensitivity distribution generated from nuclear magnetic resonance signals for each channel, The image operation unit further includes an image synthesis unit that synthesizes the channel images to generate the reference image.

11. The magnetic resonance imaging apparatus according to claim 1, wherein The sensitivity distribution calculation unit calculates the weight of the sensitivity distribution data for each channel with respect to all channels as the sensitivity distribution, The image operation unit performs image reconstruction based on the GRAPPA method on the nuclear magnetic resonance signals measured for each channel using the sensitivity distribution including the weights in the k - space for each channel.

12. A computer program product includes a sensitivity distribution calculation program that processes nuclear magnetic resonance signals measured for each channel of a magnetic resonance imaging apparatus having a receiving coil composed of a plurality of channels, and calculates the sensitivity distribution for each channel. The sensitivity distribution calculation program is characterized in that it causes a computer to execute the following steps: Read in a sensitivity distribution image including a reference image and channel images, or generate the sensitivity distribution image from nuclear magnetic resonance signals obtained by the magnetic resonance imaging apparatus; Perform one or more inversion processes on at least a part of the sensitivity distribution image, combine the generated inverted image and the original sensitivity distribution image, and generate unfolded images for the reference image and the channel images respectively; Perform an inverse Fourier transform on the unfolded images to obtain k - space data; And Calculate the sensitivity distribution for each channel based on the unfolded images in the k - space.

13. The computer program product according to claim 12, wherein The sensitivity distribution image is three - dimensional data, The sensitivity distribution calculation program further includes the following steps: Cut out a two - dimensional sensitivity distribution image from the three - dimensional data.

14. The computer program product according to claim 12, wherein The step of calculating the sensitivity distribution further includes the following steps: Remove the peripheral noise of the sensitivity distribution calculated for each channel.

Images