JPH0649034B2 - Image reconstruction method in magnetic resonance imaging apparatus - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial field of application] The present invention relates to an MR imaging apparatus (hereinafter referred to as MRI) for obtaining a medically effective tomographic image from a subject using a nuclear magnetic resonance phenomenon.
The present invention relates to a device, and more particularly, to a method for correcting distortion included in a signal obtained by the device and reconstructing a high-resolution image.

[Conventional technology]

In the MRI apparatus, it is necessary to separate / identify the signal from the subject in correspondence with the position of each part of the subject. One of the methods for that purpose is to apply a gradient magnetic field to the target space to change the magnetic field strength corresponding to the position of each part of the subject, thereby changing the resonance frequency or the amount of phase shift of each part of the subject. There is a way to get information about. This method is, for example, Proc.IEE
Since it is described in detail in E.71.338 (1983), Proc. IEEE.70.1152 (1982), etc., it is omitted here. The MRI apparatus has various modifications and improvements originating from these, but in principle, the position information necessary for imaging is obtained by the means for applying the gradient magnetic field.

Now, in such a method, it is necessary to control the gradient magnetic field at high speed and with high accuracy. As a typical example, it is necessary to pass a current of several tens of amperes in a coil used for generating a gradient magnetic field with a rise time of about 1 ms, and the amplitude and pulse width of the obtained gradient magnetic field are required to have an accuracy of the order of 10 ^−3. To be done.

Three types of superconducting, normal conducting, and permanent magnets have been put into practical use as MRI apparatuses, and any of them has an eddy current effect as a factor that hinders the accuracy of the gradient magnetic field. For superconducting magnets, the vacuum container, for normal conducting magnets, the coil itself or for cooling, and for permanent magnets, the pole pieces and magnets themselves for improving uniformity are the conductive materials such as aluminum, copper, and iron. Is arranged near the gradient coil, and an eddy current is generated by the rapid magnetic flux change. The magnetic field due to the eddy current is generated so as to cancel the original gradient magnetic field, and further decreases with the passage of time. Therefore, in the conventional apparatus, an extra current is supplied in advance for the amount reduced by the eddy current, and the gradient magnetic field coil is overdriven to address this problem.

The gradient coil normally requires three directions of X, Y and Z for imaging. The amount of overdrive is adjusted independently according to the characteristics of each coil. For example, when the gradient magnetic field with respect to the trapezoidal current waveform shows the response of (b) as shown in FIG. 2 (a), the gradient magnetic field waveform as shown in (d) is obtained by adding the overdrive waveform (c) in advance. Adjust the waveform in (c).
This is performed independently for X, Y, and Z.

[Problems to be Solved by the Invention]

The above-mentioned conventional technique is premised on the fact that a linear relationship is established between the current flowing through the gradient magnetic field coil and the generated eddy current. However, in an actual MRI apparatus, non-linearity may occur, which causes various problems described later. The factors of non-linearity are considered to be changes in eddy current distribution due to Lorentzka between strong static magnetic field and eddy current, and magnetic saturation of magnetic materials such as iron.

FIG. 3 shows an example of an imaging sequence of the Fourier imaging method which is usually used. Here, RF is an irradiation pulse that causes an NMR phenomenon, G _Z is a slice gradient magnetic field for selecting a desired cross section, G _Y is a phase encoding gradient magnetic field for imaging, and G _X is a frequency encoding gradient for imaging. The magnetic field, F (t), is the obtained detection signal. Here, in actual measurement, the amplitude of G _Y is variously changed and the sequence of FIG. 3 is repeatedly executed. If a linear relationship is established between the current flowing in the coil and the eddy current, the obtained data are arranged in the K space (space in which the Fourier plane is expressed by wave number) at the intersections on the square matrix as shown by the solid line in FIG. Love. However, if there is a non-linearity with respect to the above-mentioned current, the eddy current goes toward the saturation direction in response to an increase in the amplitude of the gradient magnetic field current I _Y of _Y , and the above-mentioned overdrive becomes excessive, and the K map of the obtained data is y. The shape is stretched non-linearly in the direction. Further, the gradient magnetic field current I _X of _X is applied with the same amplitude during the repetition of the sequence, but is applied together with G _Y during the section ta in FIG. In this case, the magnetic field due to I _Y is I _X
And the equivalent G _X increases for the same reason as above, and the coordinate point on K _X moves in the x direction (in the increasing direction of time t) according to the amplitude of I _Y. Move non-linearly. The dotted line in FIG. 4 schematically shows the actual distortion of the K map obtained in this way.

The prior art described above does not consider the distortion of the K map due to such non-linearity of the eddy current, so that the reconstructed image is blurred and the image quality is deteriorated.

An object of the present invention is to provide an MRI apparatus of high image quality by reducing the blur on an image by correcting such a distortion of the K map.

[Means for Solving the Problems]

 The above object is achieved by the following means.

That is, based on the detection signal A (t, I _y ) accompanied by the distortion shown by the dotted line in FIG.
(k _x, k _y) calculated by interpolation calculation, newly obtained F '(k _x, k _y)
The Fourier transform of is used to obtain a reconstructed image with improved blur.

Here, an ideal coordinate point (k
The relationship between ( _x , k _y ) and the coordinate point (t, I _y ) obtained by measurement must be known. As a method of linearly estimating this relationship, a lattice-shaped phantom is actually measured by an MRI apparatus, and this relationship can be obtained from the array of peak points of the obtained data. Also, polynomial approximation is made with k _y = a ₀ + a ₁ · I _y + a ₂ I _y ² + ... (1) k _x = t + b ₀ + b ₁ I _y + b ₂ · I _y ² + ... (2) A method of experimentally obtaining the coefficients a _n and b _{n with} which the blur is minimized may be used.

Since the correspondence of mutual coordinates has sufficient reproducibility if the sequence is determined, a method of preparing a distortion pattern corresponding to the sequence in advance and performing recovery processing of a desired distortion pattern after obtaining measurement data is used. .

Example of Invention

An embodiment of the present invention will be specifically described below with reference to the drawings.

FIG. 5 is a block diagram showing the overall schematic configuration of the MRI apparatus of one embodiment of the present invention.

The MRI apparatus of this embodiment obtains a tomographic image of a subject by utilizing a nuclear magnetic resonance (NMR) phenomenon. As shown in FIG. 5, the static magnetic field generating magnet 10 and the central processing unit (C
PU) 11, a sequencer 12, a transmission system 13, a magnetic field gradient generation system 14, a reception system 15 and a signal processing system 16.

The static magnetic field generating magnet 10 is for generating a strong and uniform static magnetic field around the subject 1 in the body axis direction or in a direction orthogonal to the body axis, and has a certain spread around the subject 1. A magnetic field generating means of permanent magnet type, normal conducting type or superconducting type is arranged in the space.

The sequence 12 operates under the control of the CPU 11 and sends various commands necessary for collecting tomographic image data of the subject 1 to the transmission system 13, the magnetic field gradient generation system 14, and the reception system 15.

The transmission system 13 includes a high-frequency oscillator 17, a modulator 18, a high-frequency amplifier 19, and a high-frequency coil 20a on the transmission side. The high-frequency pulse output from the high-frequency oscillator 17 is output by the modulator 18 according to a command from the sequencer 12. Amplitude modulation is performed, and the high frequency pulse thus amplitude-modulated is supplied to the high frequency amplifier 1
After being amplified in 9, the electromagnetic waves are applied to the subject 1 by supplying them to the high frequency coil 20a arranged close to the subject 1.

The magnetic field gradient generation system 14 includes a gradient magnetic field coil 21 wound in three axial directions of X, Y, and Z, and a gradient magnetic field power source 22 that drives each coil.
Gradient magnetic field power supply 2 for each coil according to the instructions from
By driving 2 the gradient magnetic fields G _x , G _y and G _z in the triaxial directions of X, Y and Z are applied to the subject 1. The slice plane for the subject 1 can be set by the method of applying the gradient magnetic field. The receiving system 15
Is composed of a high-frequency coil 20b on the receiving side, an amplifier 23, a quadrature phase detector 24, and an A / D converter 25, and the electromagnetic wave (NMR Signal) is detected by the high-frequency coil 20b arranged close to the subject 1, and the A / D converter 2 is passed through the amplifier 23 and the quadrature phase detector 24.
5 is converted into a digital amount and is further input to the quadrature detector 2 at a timing according to an instruction from the sequencer 12.
Two series of collected data are sampled by 4, and the signals are sent to the signal processing system 16. The signal processing system 16 includes a CPU 11, a recording device such as a magnetic disk 26 and a magnetic tape 27, and a display 28 such as a CRT. The CPU 11 performs processing such as Fourier transform, correction coefficient calculation, and image reconstruction. ,
The signal intensity distribution of an arbitrary cross section or the distribution obtained by performing an appropriate calculation on a plurality of signals is imaged to display 28
It is supposed to be displayed in.

In FIG. 5, the high frequency coil 20a on the transmission side,
Radio frequency coil 20b and gradient magnetic field coil 21 on the receiving side
Are arranged in the magnetic field space of the static magnetic field generating magnet 10 arranged in the space around the subject 1.

Next, an embodiment of a method of correcting the above-mentioned distorted measurement data obtained by such an apparatus will be described with reference to FIG.

Here, as a typical example of distortion of measurement data, assume that the error amount of distortion increases according to a quadratic function. That is, the above equations (1) and (2) can be expressed as K _y = I _y + a · I _y ² (3) K _x = t + b · I _y ² (4), and for convenience, a And b are known. Further, K _x , K _y , t, and I _y are normalized digital values and are integers from −128 to 127.

Step 101: Input the measured data A (t, I _y ).

Step 102: The equation (4) is solved for t, the position of t corresponding to the correct position K _x in the x direction is calculated, and the integer part is L and the decimal part is m.

Step 103: The value of t obtained in Step 102 has a value below the decimal point, and there is no corresponding measurement data.
Therefore, the measurement data A (L, I _y ), A closest to the corresponding point
B (K _x , I _y ) is obtained by correcting the distortion in the x direction by an interpolation calculation using the value of m with (L + 1, I _y ).

The operations of steps 102 and 103 are repeated from -128 to 127 for I _y and K _x , respectively, and the two-dimensional all data B (K _x ,
I _y ). At this stage, the distortion in the y direction remains.

Step 104: (3) reacting a solved for I _y, calculates the value of I _y corresponding to the correct position K _y of the y direction, and the integer part of N, decimal part and p.

Step 105: Step 10 as well as step 103
Using N and p calculated in step 4, y is obtained from the interpolation data B (K _x , N) and B (K _x , N + 1) closest to the corresponding point.
The direction distortion correction data F ′ (K _x , K _y ) is obtained.

This step 105 is executed for K _{x from} -128 to 127, and further the operation including step 104 is performed as I _y.
About -128 to 127
Dimensional final total distortion correction data F ′ (K _x , K _y )
To get

Step 106: The obtained F ′ (K _x , K _y ) is converted into data on a square matrix without distortion, and this F ′ (K _x , K _y ) is transformed into a two-dimensional Fourier transform to obtain a blurred image. Obtain an improved image C (I, J).

This interpolation method is efficient because it interpolates two-dimensional data in two directions by one-dimensional array data. Although linear interpolation is used for ease of explanation, if a higher-order interpolation method such as the least-squares method or spline interpolation is used,
Errors due to interpolation can be reduced. Since these are generally known methods, description thereof will be omitted. A two-dimensional interpolation method may be directly used instead of the method of combining one-dimensional interpolation as in the present embodiment.

Although the distortion error parameters a and b are already known, as described above, it is practical to change a and b in advance to obtain a value that gives the best improvement in blurring, and use that value during actual measurement. is there.

In the present embodiment, only the quadratic error is defined by the equations (3) and (4). However, even if the polynomial approximation is performed as in the equations (1) and (2), the steps 102 and 104 in FIG. It can be easily inferred that this can be done by only changing the equations of t and I _y of.

In addition, in order to reduce the calculation time, steps 102 and 104
It is also possible to use a method in which L, m, N, and p in Table 1 are tabled and prepared in advance, and substantially only the interpolation calculation is executed.

There are various methods for measuring sequences of MRI, and the effect of eddy current may differ depending on the method. Therefore, it is also a practical method to tabulate the a, b or L, m, N, p for each method.

In the present embodiment, the distortion in both the x and y directions was corrected, but the method of correcting only one of them also has the effect of improving blur.

For convenience, the measurement data is described as a 256 × 256 matrix, but any other matrix can be applied.

Although the present embodiment is directed to the distortion correction for two-dimensional data, it can be easily inferred that it can be applied to three-dimensional measurement data.

〔The invention's effect〕

According to the present invention, it is possible to correct the distortion due to the non-linearity of the eddy current included in the measurement signal obtained from the MRI apparatus, and it is possible to obtain a high-quality image with improved blurring that occurs in the reconstructed image.

[Brief description of drawings]

FIG. 1 is a flow chart of an embodiment of the present invention, FIG. 2 is a supplementary explanatory view of a current waveform and a gradient magnetic field waveform, FIG. 3 is an embodiment of an imaging sequence of the Fourier imaging method, and FIG. 4 is measurement data. 5 is a block diagram of the MRI apparatus, and FIG. 5 is a block diagram of the distortion on the two-dimensional Fourier plane. 101 ... Input of measurement data, 102 ... Calculation of corresponding position in x direction, 103 ... Calculation of B ( _Kx , _Iy ), 104 ...
... corresponding position calculation in the y direction, 105 ... F '(K _x ,
Calculation of K _y ), 106 ... Two-dimensional Fourier transform.

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Masao Kuroda Inventor Masao Kuroda 2-1 New Juyo 22-1, Kashiwa City, Chiba Prefecture (72) Inventor Hirotaka Takeshima 2-12 Shinjuyo, Kashiwa-shi, Chiba Stocks (72) Inventor, Chikako Nakamura, 2-11 Shinjuyo, Kashiwa-shi, Chiba, Ltd. Akihiro Kunishima (56) References, Examiner, Institute of Technology Research Institute, Hitachi, Japan (56,209,154, JP, A) JP-A-63-216556 (JP, A)

Claims (5)

[Claims] 1. A static magnetic field, a gradient magnetic field and a high frequency magnetic field are generated according to a predetermined procedure to detect a magnetic resonance signal from a desired inspection region of an inspection object, and the eddy current non-linearity contained in the measurement signal is detected. An image reconstruction method in a magnetic resonance imaging apparatus, characterized in that an image is reconstructed by performing distortion correction processing on a Fourier plane. 2. The image reconstructing method according to claim 1, wherein the distortion correction processing on the Fourier plane is performed by a combination of interpolation processing of one-dimensional array data in different directions. 3. The image reconstructing method according to claim 1, wherein the distortion correction processing on the Fourier plane is performed by two-dimensional interpolation of two-dimensional array data. 4. The image reconstructing method according to claim 1, wherein the distortion correction processing on the Fourier plane is performed by interpolation processing of one-dimensional array data in only one direction. 5. The image according to claim 1, wherein in the distortion correction processing on the Fourier plane, coefficients expressing distortion are made into a table in advance, and the interpolation processing is performed by referring to the table. Reconstruction method.