JP2009195584A - Image processor and medical imaging apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an art favorably providing a compensation result of strain and displacement of an image even if there is a slice having a large contrast difference between a reference image and an object image when positioning different types of images of the same object site consisting of a plurality of slices. <P>SOLUTION: When finding such a conversion coefficient that an overlap with a reference image formed by transforming coordinates and binarizing a binarized object image becomes the maximum, this image processor finds conversion coefficients using a conventional method in the respective slices and compensates the conversion coefficients of each slice based on the correlation between the conversion coefficient and a predetermined slice sequence. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to an image processing technique of a medical image photographing apparatus, and more particularly to a technique for correcting positional deviation and distortion of a medical image when aligning a plurality of medical images.

  For example, in a medical image processing apparatus such as a magnetic resonance imaging apparatus (MRI apparatus) or an X-ray CT apparatus, a plurality of images may be acquired with different imaging conditions for the same target region. In some cases, images of the same target part acquired by different types of medical image processing apparatuses are combined.

  Even in the case of images taken of the same target region, the positions of the images may differ if the types of medical image processing apparatuses are different. Even in the same medical image photographing processing apparatus, the distortion direction and contrast may be different for each image by changing the photographing condition.

  In particular, single shot spin echo (SE) -echo planar imaging (EPI) is widely used to avoid motion artifacts in diffusion weighted imaging (DWI) and diffusion tensor imaging (DTI), which are photographed using an MRI apparatus. However, in the DWI-SE-EPI, a strong diffusion gradient magnetic field (MPG) is used to emphasize the attenuation of the MR signal due to diffusion. The image may be distorted by a non-uniform magnetic field induced by eddy currents caused by the MPG. In addition, since the distortion of the image depends on the timing, size, and direction of application of MPG, it varies depending on the shooting.

  In the DWI, in general, based on a plurality of images (target images) photographed by changing the MPG direction and images (reference images) obtained without applying MPG, an ADC (Applied Diffusion Coefficient) or an isotropic diffusion weighted image is used. Is used for diagnosis. A positional shift due to the above-described image distortion may occur between the target image and the reference image, and this positional shift causes an artifact in the ADC image calculated from these images, leading to a decrease in resolution. . In general, in DTI, imaging is performed a plurality of times while changing the application direction of MPG, and then a tensor analysis is mathematically performed to form an imaging target. However, as described above, since different distortions are generated in images obtained by changing the application direction of MPG, accurate tensor analysis cannot be performed. There is parallel imaging as a means for reducing the influence of eddy currents. However, image distortion remains.

  Some image distortions caused by MPG as described above are processed in an image space and corrected (see, for example, Patent Document 1). In the technique disclosed in Patent Document 1, a binary image of each of the reference image and the target image is created, a conversion coefficient that maximizes the overlap of the two binarized images is obtained, and the target image is determined using the conversion coefficient. to correct. The conversion coefficient calculates the square of the pixel value difference from each point of the reference image while moving the pixel of the binarized target image by coordinate conversion, and the square of the pixel value difference of the entire image Is calculated as a matrix that minimizes the value obtained.

JP 2007-167374 A

  The technique described in Patent Document 1 can automatically process image distortion between images having different image contrasts and correction of misalignment caused by the images at high speed. However, for example, in the DTI of the head, b = 0 (b is a gradient magnetic field factor indicating the degree of signal attenuation due to diffusion, and b = 0 indicates that no MPG is applied near the base of the skull or the top of the head. The contrast difference between the reference image and the target image to which MPG is applied is large. Thus, when applied to correction between images having a large contrast difference between the reference image and the target image, there may be a difference in the shape of the target extracted by binarization, and an accurate conversion matrix May not be obtained, and as a result, image distortion and displacement may not be corrected satisfactorily.

  The present invention has been made in view of the above circumstances, and a slice having a large contrast difference between a reference image and a target image when performing alignment between different types of images of the same target region including a plurality of slices. It is an object of the present invention to provide a technique for obtaining an image distortion and misalignment correction result satisfactorily.

  In order to achieve the above object, the present invention obtains the conversion coefficient so that the overlap with the binarized reference image is maximized by coordinate transformation of the binarized target image. After obtaining the conversion coefficient using the conventional method, the conversion coefficient of each slice is further corrected based on the correlation between the conversion coefficient and a predetermined slice order.

  Specifically, an image processing apparatus that processes an image captured by a medical image capturing apparatus that captures and images a predetermined region of a subject with a plurality of slices, and is obtained by capturing the same target region. In order to align the reference image and the target image, the image correction unit corrects the positional deviation and / or image distortion of the target image, and the image correction unit determines the pixel position of the target image for each slice. Conversion coefficient calculation means for calculating a conversion coefficient of a conversion equation to be converted, and conversion coefficient correction means for correcting the conversion coefficient for each slice based on a correlation between a predetermined slice order and the conversion coefficient. An image processing apparatus is provided.

  According to the present invention, when performing alignment between different types of images of the same target region composed of a plurality of slices, even if there is a slice having a large contrast difference between the reference image and the target image, the image is satisfactorily obtained. Correction results of distortion and misalignment can be obtained.

<< First Embodiment >>
Hereinafter, a first embodiment to which the present invention is applied will be described with reference to the drawings. Note that components having the same function are denoted by the same reference symbols throughout the drawings for describing the embodiments of the invention, and the repetitive description thereof is omitted. In the present embodiment, a case where a magnetic resonance imaging (MRI) apparatus is used as a medical image diagnostic apparatus for acquiring a reference image and a target image will be described as an example.

  First, an overall outline of an example of the MRI apparatus of the present embodiment will be described. FIG. 1 is a block diagram showing the overall configuration of the MRI apparatus of this embodiment. The MRI apparatus 100 obtains a tomographic image of a subject using the NMR phenomenon, and includes a static magnetic field generation system 2, a gradient magnetic field generation system 3, a transmission system 5, a reception system 6, and a signal processing system 7. And a sequencer 4 and a central processing unit (CPU) 8.

  The static magnetic field generation system 2 generates a uniform static magnetic field in the direction perpendicular to the body axis in the space around the subject 1 if the vertical magnetic field method is used, and in the body axis direction if the horizontal magnetic field method is used. Therefore, a permanent magnet type, normal conduction type or superconducting type static magnetic field generation source is arranged around the subject 1.

  The gradient magnetic field generation system 3 includes a gradient magnetic field coil 9 wound in the three-axis directions of X, Y, and Z in the coordinate system (stationary coordinate system) of the MRI apparatus 100 and a gradient magnetic field power source that drives each gradient magnetic field coil. 10, and the gradient magnetic field power supply 10 of each coil is driven in accordance with a command from the sequencer 4 to be described later, so that gradient magnetic fields Gx, Gy, and Gz are applied to the subject 1 in the three axial directions of X, Y, and Z. Apply to. In general, a slice direction gradient magnetic field pulse (Gs) is applied in any one direction of X, Y, and Z to set a slice plane (imaging section) for the subject 1, and the phase encode direction gradient magnetic field is applied to the remaining two directions. A pulse (Gp) and a frequency encoding direction gradient magnetic field pulse (Gf) are applied, and position information in each direction is encoded in the echo signal.

  The transmission system 5 applies a high-frequency magnetic field pulse (hereinafter referred to as “RF pulse”) to cause nuclear magnetic resonance to occur in the nuclear spins of atoms constituting the biological tissue of the subject 1. A modulator 12, a high frequency amplifier 13, and a high frequency coil (transmission coil) 14a on the transmission side are provided. The high-frequency pulse output from the high-frequency oscillator 11 is amplitude-modulated by the modulator 12 at a timing according to a command from the sequencer 4, and after the amplitude-modulated high-frequency pulse is amplified by the high-frequency amplifier 13, it is arranged close to the subject 1. The electromagnetic wave (RF pulse) is applied to the subject 1 by being supplied to the transmission coil 14a.

  The receiving system 6 detects an NMR signal (echo signal) emitted by nuclear magnetic resonance of nuclear spins constituting the biological tissue of the subject 1. The receiving system 6 receives a high-frequency coil (receiving coil) 14 b on the receiving side and a signal amplifier 15. And a quadrature phase detector 16 and an A / D converter 17. The response electromagnetic wave (NMR signal) of the subject 1 induced by the electromagnetic wave applied from the transmission coil 14 a is detected by the reception coil 14 b arranged close to the subject 1. The detected NMR signal is amplified by the signal amplifier 15 and then divided into two orthogonal signals by the quadrature phase detector 16 at the timing according to the command from the sequencer 4, and each is digitally converted by the A / D converter 17. It is converted into a quantity and sent to the signal processing system 7.

  The sequencer 4 controls the application of the RF pulse and the gradient magnetic field pulse according to a predetermined pulse sequence. It operates under the control of the CPU 8 and sends various commands necessary for collecting tomographic image data of the subject 1 to the transmission system 5, the gradient magnetic field generation system 3, and the reception system 6. The pulse sequence is a time chart of a combination of on / off timing, amplitude, etc., such as an RF pulse and a gradient magnetic field pulse, which is created in advance according to the purpose of imaging and stored as a program in a memory (not shown) or the like. . The CPU 8 controls the sequencer 4 according to the pulse sequence.

  The signal processing system 7 performs various data processing and display and storage of processing results. The CPU 8, a storage device 18 such as a ROM and a RAM, an external storage device 19 such as an optical disk and a magnetic disk, and a display device 20. It consists of. When data from the receiving system 6 is input to the CPU 8, the CPU 8 executes processing such as signal processing and image reconstruction, and displays the tomographic image of the subject 1 as a result on the display device 20, and also a storage device. 18 or the external storage device 19.

  The operation unit 25 receives various control information of the MRI apparatus 100 itself and various control information of processing performed by the signal processing system 7, and includes a trackball or mouse 23 and a keyboard 24. The operation unit 25 is disposed close to the display device 20, and the operator interactively inputs information necessary for various processes of the MRI apparatus 100 through the operation unit 25 while looking at the display device 20.

  In FIG. 1, the transmission coil 14 a and the gradient magnetic field coil 9 are opposed to the subject 1 in the static magnetic field space of the static magnetic field generation system 2 in which the subject 1 is inserted. If it is a horizontal magnetic field system, it is installed so as to surround the subject 1. The receiving coil 14 b is installed so as to face the subject 1 or surround the subject 1.

  At present, the radionuclide to be imaged by the MRI apparatus is a hydrogen nucleus (proton) which is a main constituent material of the subject as widely used clinically. In the MRI apparatus, the information about the spatial distribution of proton density and the spatial distribution of the relaxation time of the excited state is imaged, so that the form or function of the human head, abdomen, limbs, etc. can be imaged two-dimensionally or three-dimensionally. To do.

  In the present embodiment, the MRI apparatus 100 captures the same target region with a plurality of slices under the first condition and the second condition, and the respective first image group (reference image group) and second image. A group (target image group) is obtained. Then, in order to align the reference image and the target image, image distortion and / or displacement correction (hereinafter, referred to as distortion correction together) is performed on the target image for each slice.

  FIG. 2 is a diagram for explaining an overview of image processing according to the present embodiment. Here, an example in which the number of slices is n, a reference image is acquired without applying MPG by DTI, and the target image is acquired by applying MPG by DTI is the target region of the head region. I will explain it. In order to align with the acquired reference images SL1, SL2,... SLn, distortion correction is performed on the target images SL′1, SL′2,. “1, SL” 2,... SL ”n).

The distortion correction is performed by calculating a coordinate conversion formula for each slice of the target image and using the coordinate conversion formula. In this embodiment, since the diffusion-weighted image is targeted, the main misalignment and image distortion have shift (movement), expansion / contraction, and shear (shear) components. For this reason, it can convert and correct | amend by affine transformation. That is, the coordinate point (x, y) is converted to the coordinate point (x ′, y ′) by the conversion formula represented by the following formula (1) and corrected.
x ′ = ax + by + e
y ′ = cx + dy + f (1)
Each coefficient (conversion coefficient) of the coordinate conversion formula is calculated using corresponding slices (SL1 and SL′1,... SLn and SL′n in FIG. 2) between the reference image and the target image. That is, the corresponding slices of the reference image and the target image are binarized, and calculation is performed so that the target image overlaps the reference image most for each slice.

  However, as described above, in the case of a slice having a large contrast difference between the reference image and the target image, a difference occurs in the shape of the target object extracted by binarization. Is likely to occur. For example, when the same target region is imaged with a plurality of slices as in the present embodiment, a slice having a large contrast difference locally may occur depending on the imaging conditions. For example, in the case of photographing the head, slices with a large contrast difference are likely to occur at the base of the skull or the top of the head. Therefore, in the present embodiment, the calculated conversion coefficient is corrected using the correlation between the conversion coefficient and a predetermined slice order, thereby improving the accuracy of the conversion coefficient of a slice having a large contrast difference.

  In order to realize the above processing, the MRI apparatus 100 of the present embodiment includes an image processing unit 200 that performs these processes. FIG. 3 is a configuration diagram of the image processing unit 200 of the present embodiment. As shown in the figure, the image processing unit 200 according to the present embodiment receives a reference image and a target image captured by the MRI apparatus 100 from the storage device 18 or the external storage device 19, and holds the memory 210, and distortion correction. And a display device 230 for displaying a calculation result, and an input device 240 for receiving various instructions such as calculation conditions. Note that the memory 210, the arithmetic unit 220, the display device 230, and the input device 240 may be shared by the storage device 18 and / or the external storage device 19, the CPU 8, the display device 20, and the operation unit 25 of the MRI apparatus 100, respectively. Good. The image processing unit 200 may be constructed on an independent information processing apparatus that can transmit and receive image data to and from the MRI apparatus 100.

  In order to realize the above processing, the image processing unit 200 of the present embodiment includes an image binarization processing unit 231, a conversion coefficient calculation processing unit 232, a conversion coefficient correction processing unit 233, and a distortion correction processing unit 234. Prepare. Each of these functions is realized by the arithmetic unit 20 executing a program stored in the memory 210. Details of each function will be described below.

  The image binarization processing unit 231 binarizes the image acquired by the MRI apparatus 100. The processing by the image binarization processing unit 231 is basically the same as the method described in Patent Document 1. That is, a threshold value is determined using a discriminant analysis method, and a predetermined value or 0 is set for each pixel according to the determined threshold value.

  For the image binarized by the image binarization processing unit 231, the conversion coefficient calculation processing unit 232 determines a conversion coefficient for each slice of the same slice of the reference image and the target image. This is basically the same as the method described in Patent Document 1. That is, the conversion coefficient of the coordinate conversion formula is obtained so that the target image overlaps the reference image most. In the present embodiment, since the diffusion weighted image is targeted, the coordinate conversion formula is the above-described formula (1), and six conversion coefficients to be obtained are a, b, c, d, e, and f. Also, the transform coefficient calculation processing unit 232 records each obtained transform coefficient in association with information (slice information) specifying each slice.

  The conversion coefficient correction processing unit 233 performs a conversion coefficient correction process that excludes the influence of the error due to the contrast difference as described above from the conversion coefficients calculated for each slice by the conversion coefficient calculation processing unit 232. The conversion coefficient correction process is performed by linear regression of the conversion coefficient for each slice. Details of the conversion coefficient correction processing will be described later.

  The distortion correction processing unit 234 corrects the target image using each conversion coefficient corrected by the conversion coefficient correction processing unit 233. The correction method is basically the same as the method of Patent Document 1. That is, the coordinate conversion formula (formula (1)) having the conversion coefficient calculated by the conversion coefficient correction processing unit 233 is applied to the target image. For each point (each pixel) of the target image, coordinate conversion is performed by the coordinate conversion formula (1), and the pixel value of the point before coordinate conversion is set to the pixel value of the point (pixel) after conversion. However, since the coordinate value of the point (pixel) after conversion does not become an integer, the pixel value that becomes an integer is obtained by interpolation. This interpolation is performed using a known interpolation method.

  Next, conversion coefficient correction processing by the conversion coefficient correction processing unit 233 will be described. The conversion coefficient correction process is performed for each conversion coefficient (six conversion coefficients described above in the present embodiment) for each slice recorded by the conversion coefficient calculation processing unit 232. In the present embodiment, the conversion coefficients for each slice are plotted in the order of predetermined slice information. Then, it is determined whether or not each plotted conversion coefficient is within a predetermined threshold range. An approximate curve is obtained by regression using only the conversion coefficient determined to be within the range. Then, the conversion coefficient determined to be within the range is used as it is, and the conversion coefficient determined to be out of the range is changed to a value on the approximate curve.

  The threshold value is set based on a generally assumed misregistration and / or image distortion amount in the diffusion weighted image. This value is determined empirically considering the characteristics of the device. For example, shift (c, f) is ± 5 pixels, expansion / contraction (a, e) is 0.8 and 1.2 (1 / pixel), and share (b, d) is −0.2 and +0.2 (1 / Pixel), the upper and lower limits of the coefficients a, b, c, d, e, and f are set as threshold values.

  As the predetermined order of slice information, the order of slices in the slice direction (slice position order), the measurement order of slices (excitation order), and the like can be considered.

  A linear function can be used as the approximate curve. As described above, the position shift and image distortion of the diffusion weighted image are due to the influence of the eddy current. Magnetic field inhomogeneity due to eddy currents is considered to be spatially dominated by the zeroth and first order components, and image distortion caused by this has a first order relationship in the slice direction or slice excitation order. .

  Next, details of the conversion coefficient correction processing by the conversion coefficient correction processing unit 233 after plotting will be described with reference to the drawings. FIG. 4 is a process flow of the conversion coefficient correction process of the present embodiment. Hereinafter, in the present processing flow, a case where a among the six coefficients of the above formula (1) is corrected will be described as an example of the conversion coefficient. The same processing is performed for the remaining coefficients. The number of slices is M, and slice numbers m (m = 1 to M) are assigned to each slice in accordance with a predetermined slice information order (plot order). Also, the conversion coefficient corresponding to each slice is assumed to be am, the approximate line is l1, and the value corresponding to the slice number m on the approximate line l1 is l1m.

  First, the counter mc for counting the slice number m is set to 1 (step S301). Next, it is determined whether or not the conversion coefficient amc of m = mc is within the threshold range (step S302). If it is outside the threshold range, it is invalid (step S303), and if it is within the threshold range, it is valid (step S304). Thereafter, it is determined whether the processing has been completed for all the conversion coefficients (mc = M?) (Step S305). If not, the value of mc is incremented by 1 (step S306), and the process returns to step S302.

  If all the conversion coefficients have been processed, the approximate straight line l1 is determined using only the effective conversion coefficient am (step S307). The invalidated conversion coefficient am is replaced with the value l1m on the approximate straight line l1 (step S308).

  The above procedure is performed for each conversion coefficient (a, b, c, d, e, f) to correct each conversion coefficient. Then, distortion correction of each slice is performed by a coordinate conversion formula using the corrected conversion coefficient.

  As described above, according to the present embodiment, a conversion coefficient that is considered to be an unreasonable value is excluded using a threshold value, and a conversion coefficient having a reasonable value is calculated from the conversion coefficient of a slice. Replace with the value on the approximate line. That is, the conversion coefficient having an unreasonable value is corrected by using the correlation between the conversion coefficient having a reasonable value and a predetermined slice order. In addition, the influence of variations in conversion coefficients can be reduced by setting the threshold value. Therefore, even if there is a slice having a large contrast difference between the reference image and the target image, a coordinate conversion formula having an appropriate conversion coefficient can be obtained, and the accuracy of distortion correction can be improved.

  By performing distortion correction of the diffusion weighted image that is the basis of the diffusion tensor image using the corrected conversion coefficient of the present embodiment, the result of the diffusion tensor analysis is improved. For example, in the color map, when the distortion correction of the target image is performed using the conversion coefficient that is not subjected to the correction processing of the present embodiment, even if the correction is performed for each slice, the contour shift between the skull base and the slice near the top of the head. However, if the conversion coefficient that has been subjected to the correction process of the present embodiment is used, the deviation of the contour is improved.

  If the excitation order of slices is used as the order of slice information, it is possible to reduce the effect of eddy currents accumulated when the repetition time during imaging is short.

  In this embodiment, the effective conversion coefficient is configured to be used as it is, but the effective conversion coefficient may be replaced with a value on an approximate line.

<< Second Embodiment >>
Next, a second embodiment to which the present invention is applied will be described. A configuration different from the first embodiment of the present embodiment is a conversion coefficient correction process by a conversion coefficient correction processing unit. Therefore, other configurations are basically the same as those in the first embodiment, and therefore only the transform coefficient correction process will be described here.

  Also in this embodiment, the transform coefficient correction process is performed for each transform coefficient (six transform coefficients) for each slice. In the present embodiment, the conversion coefficients for each slice are plotted in the order of predetermined slice information. Then, an approximate curve is obtained from the plotted conversion coefficients. For each conversion coefficient, a difference from a value on the approximate curve is obtained, and it is determined whether each conversion coefficient is within a predetermined effective range. The conversion coefficient within the effective range is used as it is, and the conversion coefficient outside the effective range is changed to a value on the approximate curve.

  Also in this embodiment, a linear function can be used as the approximate curve. In addition, as the predetermined order of slice information, the order of slice positions, the order of measurement, and the like can be considered.

  In addition, the effective range can be set as a range of a constant multiple of the standard deviation of each conversion coefficient with the approximate curve value as the center. For example, 0.7 can be used as the constant.

  Next, details of the conversion coefficient correction processing by the conversion coefficient correction processing unit 233-2 (not shown) of the present embodiment after plotting will be described with reference to the drawings. FIG. 5 is a process flow of the conversion coefficient correction process of this embodiment. Hereinafter, in the present processing flow, a case where a among the six coefficients of the above formula (1) is corrected will be described as an example of the conversion coefficient. The same processing is performed for the remaining coefficients. The number of slices is M, and slice numbers m (m = 1 to M) are assigned to each slice according to a predetermined slice information order. Further, each conversion coefficient is am, the approximate line is l2, and the value corresponding to the slice number m on the approximate line l2 is l2m.

  An approximate straight line l2 is determined using all the conversion coefficients am (step S401). Next, for each slice, the difference km between the calculated conversion coefficient am and the point l2m on the approximate straight line l2 is calculated, the standard deviation σ of km is obtained, and the effective range Nσ multiplied by a constant N is calculated (step) S402).

  A counter mc for counting slice numbers is set to 1 (step S403). Next, it is determined whether or not the absolute value (| mc−l2mc |) of the difference from the value on the approximate straight line l2 of the conversion coefficient amc of m = mc is within the effective range Nσ (step S404). If it is outside the valid range Nσ, it is invalid (step S405), and if it is within the range, it is valid (step S406). Thereafter, it is determined whether the processing has been completed for all the conversion coefficients (mc = M?) (Step S407). If not, the value of mc is incremented by 1 (step S408), and the process returns to step S404.

  If all the conversion coefficients have been processed, the invalid conversion coefficient am is replaced with the value l2m on the approximate line l2 (step S409).

  The above procedure is performed for each conversion coefficient (a, b, c, d, e, f) to correct each conversion coefficient. Then, distortion correction of each slice is performed by a coordinate conversion formula using the corrected conversion coefficient.

  As described above, according to the present embodiment, a conversion coefficient that is far from the value on the approximate line calculated from the conversion coefficient of each slice is replaced with a value on the approximate line. That is, the conversion coefficient having an unreasonable value is corrected by using the correlation between the conversion coefficient having a reasonable value and a predetermined slice order. In addition, by setting the effective range, it is possible to reduce the influence due to the variation of the conversion coefficient. Therefore, even if there is a slice having a large contrast difference between the reference image and the target image, a coordinate conversion formula having an appropriate conversion coefficient can be obtained, and the accuracy of distortion correction can be improved.

  By performing distortion correction of the diffusion weighted image that is the basis of the diffusion tensor image using the corrected conversion coefficient of the present embodiment, the result of the diffusion tensor analysis is improved. For example, in the color map, when the distortion correction of the target image is performed using the conversion coefficient that is not subjected to the correction processing of the present embodiment, even if the correction is performed for each slice, the contour shift between the skull base and the slice near the top of the head. However, if the conversion coefficient that has been subjected to the correction process of the present embodiment is used, the deviation of the contour is improved.

  In addition, you may comprise so that the conversion factor correction process of this embodiment may be repeated a predetermined number of times. By repeating, variation in conversion coefficients can be suppressed.

  In this embodiment, the effective conversion coefficient is configured to be used as it is, but the effective conversion coefficient may be replaced with a value on an approximate line.

<< Third Embodiment >>
Next, a third embodiment to which the present invention is applied will be described. A configuration different from the first embodiment of the present embodiment is a conversion coefficient correction process by a conversion coefficient correction processing unit. Therefore, other configurations are basically the same as those in the first embodiment, and therefore only the transform coefficient correction process will be described here.

  Also in this embodiment, the transform coefficient correction process is performed for each transform coefficient (six transform coefficients) for each slice. In the present embodiment, the conversion coefficients for each slice are plotted in the order of predetermined slice information. The process after plotting is a combination of the conversion coefficient correction process of the first embodiment and the conversion coefficient correction process of the second embodiment.

  That is, it is determined whether or not each plotted conversion coefficient is within a predetermined threshold range. An approximate curve is obtained by linear regression using only the conversion coefficient determined to be within the range. Then, the conversion coefficient determined to be within the range is used as it is, and the conversion coefficient determined to be out of the range is replaced with a value on the approximate curve. The conversion coefficient determined to be within the range and the conversion coefficient after replacement are plotted again in the order of slice information to obtain a new approximate curve. Then, the difference between each conversion coefficient and the value on the approximate curve is obtained, and it is determined whether each conversion coefficient is within a predetermined effective range. The conversion coefficient within the effective range is used as it is, and the conversion coefficient outside the effective range is changed to a value on the approximate curve.

  In the present embodiment, the threshold value, the approximate curve, the order of slice information, and the effective range are the same as those in the first and second embodiments.

  Next, details of the conversion coefficient correction processing by the conversion coefficient correction processing unit 233-3 (not shown) of the present embodiment will be described. FIG. 6 is a process flow of the conversion coefficient correction process of the present embodiment. Hereinafter, in the present processing flow, a case where a among the six coefficients of the above formula (1) is corrected will be described as an example of the conversion coefficient. The same processing is performed for the remaining coefficients. In addition, the number of slices is M, slice numbers m (m = 1 to M) are assigned to the slices in accordance with a predetermined order of slice information, and each conversion coefficient is am. Further, the approximate lines are l1 and l2, and the values corresponding to the slice number m on the approximate lines l1 and l2 are l1m and l2m, respectively.

  First, the counter mc for counting the slice number m is set to 1 (step S501). Next, it is determined whether or not the conversion coefficient amc of m = mc is within the threshold range (step S502). If it is outside the threshold range, it is invalid (step S503), and if it is within the threshold range, it is valid (step S504). Thereafter, it is determined whether the processing has been completed for all the conversion coefficients (mc = M?) (Step S505). If not, the value of mc is incremented by 1 (step S506), and the process returns to step S502.

  When the processing has been completed for all the conversion coefficients, linear regression is performed using only the effective conversion coefficient am, and the first approximate straight line l1 is determined (step S507). The invalidated conversion coefficient am is replaced with the value l1m on the first approximate straight line l1 (step S508).

  Next, a linear regression is performed on all conversion coefficients of the effective conversion coefficient am and the conversion coefficient replaced with l1m to determine the second approximate line l2 (step S509). Next, for each slice, a difference km between the calculated conversion coefficient am and the point l2m on the approximate curve is calculated, a standard deviation σ of km is obtained, and an effective range Nσ multiplied by a constant N is calculated (step S510). ).

  A counter mc for counting slice numbers is set to 1 (step S511). Next, it is determined whether or not the absolute value (| mc−l2mc |) of the difference from the value on the approximate straight line l2 of the conversion coefficient amc of m = mc is within the effective range Nσ (step S512). If it is outside the valid range Nσ, it is invalid (step S513), and if it is within the range, it is valid (step S514). Thereafter, it is determined whether the processing has been completed for all the conversion coefficients (mc = M?) (Step S515). If not, the value of mc is incremented by 1 (step S516), and the process returns to step S512.

  If all the conversion coefficients have been processed, the invalid conversion coefficient am is replaced with the value l2m on the second approximate straight line l2 (step S517).

  The above procedure is performed for each conversion coefficient (a, b, c, d, e, f) to correct each conversion coefficient. Then, distortion correction of each slice is performed by a coordinate conversion formula using the corrected conversion coefficient.

  Examples of the first approximate straight line l1 and the second approximate straight line l2 in the conversion coefficient correction processing of the present embodiment are shown in FIG. Here, an example of the conversion coefficient f (shift) which is the most important distortion component in the diffusion weighted image is shown. FIG. 7A is an example of the first approximate straight line 11, the horizontal axis is the slice order, and the vertical axis is the value of the conversion coefficient f. FIG. 7B is an example of the second approximate straight line l2, where the horizontal axis is the slice order and the vertical axis is the value of the conversion coefficient f.

  As described above, according to the present embodiment, a conversion coefficient that is considered to be an unreasonable value is excluded using a threshold value, and a conversion coefficient having a reasonable value is calculated from the conversion coefficient of a slice. Replace with the value on the approximate line. Further, an approximate straight line is calculated from the conversion coefficients of each slice after the irrational value is excluded, and a conversion coefficient far away from the approximate straight line is replaced with a value on the approximate straight line. In this way, the conversion coefficient with an unreasonable value is corrected twice using the correlation between the conversion coefficient with a reasonable value and a predetermined slice order. Further, the influence of variations in conversion coefficients can be reduced by setting the threshold value and the effective range. Therefore, the conversion coefficient can be corrected with higher accuracy than the methods of the first and second embodiments. Therefore, the accuracy of distortion correction can be increased.

  In the conversion coefficient correction process of the present embodiment, the process from step 509 to step S517 may be repeated a predetermined number of times in order to suppress variations in the conversion coefficient.

  In the present embodiment, the first approximate straight line 11 may be calculated using all the conversion coefficients without determining whether the conversion coefficient is valid or invalid by the threshold value. Further, in step S508 and step S517, all conversion coefficients may be replaced with values on the approximate straight lines l1 and l2.

  Note that, in the above-described embodiment, the case where positional displacement and image distortion occur in two directions, the x direction and the y direction, has been described as an example. However, in echo planar imaging (EPI), distortion hardly occurs in the frequency encoding direction. Therefore, the coordinate conversion may be performed only in one direction of the x direction or the y direction.

  In each of the above embodiments, the case where there is one type of target image has been described as an example, but there may be a plurality of target images. For example, the reference image is an image to which MPG is not applied in diffusion weighted imaging, and the target image is a plurality of images to which MPG is applied in different directions. Even in such a case, each target image is corrected using a coordinate conversion formula having a conversion coefficient obtained by the above processing.

  In each of the above embodiments, the case where both the reference image and the target image are acquired by the MRI apparatus has been described as an example. However, the medical image diagnostic apparatus that acquires these images is not limited to the MRI apparatus. Further, the reference image and the target image may be acquired by different medical image diagnostic apparatuses.

  Further, in each of the above embodiments, the threshold value and the effective range are set in advance, but may be configured so that the operator can input when correcting the conversion coefficient. In this case, a function for displaying the conversion coefficients plotted in the slice order on the display device 230 and a GUI that allows the operator to set the threshold and the effective range while viewing the display are prepared. Note that not only the threshold value and the effective range, but also the slice used when calculating the approximate line, the slice order, the number of repetitions of the transform coefficient correction process, and the like may be set using the GUI.

  By preparing the GUI, it is possible to finely adjust the conversion coefficient correction process, and a more accurate conversion coefficient can be obtained. Further, by configuring so that the slice to be used can be selected, calculation time can be shortened when the number of slices is large.

It is a block diagram which shows the whole structure of the MRI apparatus of 1st embodiment. It is a figure for demonstrating the outline | summary of the image processing of 1st embodiment. It is a block diagram of the image processing part of 1st embodiment. It is a processing flow of the conversion coefficient correction process of 1st embodiment. It is a processing flow of the conversion coefficient correction process of 2nd embodiment. It is a processing flow of the conversion coefficient correction process of 3rd embodiment. It is a figure for demonstrating the approximate line of 3rd embodiment.

Explanation of symbols

1: subject, 2: static magnetic field generation system, 3: gradient magnetic field generation system, 4: sequencer, 5: transmission system, 6: reception system, 7: signal processing system, 8: central processing unit (CPU), 9: Gradient magnetic field coil, 10: Gradient magnetic field power supply, 11: High frequency transmitter, 12: Modulator, 13: High frequency amplifier, 14a: High frequency coil (transmitting coil), 14b: High frequency coil (receiving coil), 15: Signal amplifier, 16 : Quadrature detector, 17: A / D converter, 18: storage device, 19: external storage device, 20: display device, 23: trackball or mouse, 24: keyboard, 25: operation unit, 100: MRI device , 200: image processing unit, 210: memory, 220: arithmetic unit, 230: display device, 240: input device, 231: image binarization processing unit, 232: conversion coefficient calculation processing unit, 233: conversion coefficient correction processing unit 2 34: Distortion correction processing unit

Claims (15)

An image processing apparatus that processes an image captured by a medical image capturing apparatus that captures and images a predetermined region of a subject with a plurality of slices,
In order to align the reference image obtained by photographing the same target region and the target image, the image correction means for correcting the positional deviation and / or image distortion of the target image,
The image correcting means includes
Conversion coefficient calculating means for calculating a conversion coefficient of a conversion equation for converting a pixel position of the target image for each slice;
An image processing apparatus comprising: a conversion coefficient correction unit that corrects the conversion coefficient for each slice based on a correlation between a predetermined slice order and the conversion coefficient. The image processing apparatus according to claim 1,
The transform coefficient correcting unit specifies a correlation between the slice order and the transform coefficient using transform coefficients within a predetermined upper and lower threshold range, and converts the transform coefficient outside the range to the correlation An image processing apparatus comprising threshold correction means for correcting using a relationship. The image processing apparatus according to claim 1 or 2,
The transform coefficient correcting means includes effective range correcting means for correcting a transform coefficient outside a predetermined range specified from the correlation between the slice order and the transform coefficient using the correlation. Image processing device. The image processing apparatus according to claim 3,
The effective range correction unit calculates a correlation between the slice order and the previous period conversion coefficient using the conversion coefficient corrected by the threshold correction unit, and a conversion coefficient outside a predetermined range specified from the correlation Is corrected using the correlation. The image processing apparatus according to claim 3 or 4,
The image processing apparatus, wherein the conversion coefficient correction unit repeats the correction by the effective range correction unit a predetermined number of times. The image processing apparatus according to any one of claims 1 to 5,
The predetermined slice order is an arrangement order according to a slice position. The image processing apparatus according to any one of claims 1 to 5,
The image processing apparatus according to claim 1, wherein the predetermined slice order is a slice excitation order. The image processing apparatus according to any one of claims 1 to 7,
The image processing apparatus, wherein the correlation is approximated by a linear function. The image processing apparatus according to any one of claims 3 to 8,
The image processing apparatus according to claim 1, wherein the predetermined range is defined by a constant multiple of a standard deviation of a difference from a function value specifying the correlation of each conversion coefficient. The image processing apparatus according to any one of claims 1 to 9,
The image processing apparatus, wherein the reference image and the target image are captured by different medical image capturing apparatuses. The image processing apparatus according to any one of claims 1 to 10,
The medical imaging apparatus is a magnetic resonance imaging apparatus,
The reference image is an image taken by diffusion weighted imaging without applying MPG (Motion Probing Gradient),
The image processing apparatus, wherein the target image is an image captured by applying the MPG by diffusion weighted imaging. The image processing apparatus according to any one of claims 1 to 10,
The medical imaging apparatus is a magnetic resonance imaging apparatus,
The reference image is an image photographed by applying MPG (Motion Probing Gradient) by diffusion weighted imaging,
The target image is an image captured by applying the MPG in a direction different from the direction applied when capturing the reference image by diffusion weighted imaging and an image captured without applying the MPG. An image processing apparatus.   A medical image photographing apparatus comprising the image processing apparatus according to claim 1. Computer
An image correction unit that corrects misalignment and / or image distortion of the target image in order to align the reference image obtained by photographing the same target region with a plurality of slices and the target image,
Conversion coefficient calculating means for calculating a conversion coefficient of a conversion equation for converting a pixel position of the target image for each slice;
A program for functioning as an image processing unit comprising: a conversion coefficient correction unit that corrects the conversion coefficient for each slice based on a correlation between a predetermined slice order and the conversion coefficient. An image processing method for processing an image captured by a medical image capturing apparatus that captures and images a predetermined area of a subject with a plurality of slices,
In order to align the reference image obtained by photographing the same target region and the target image, the image processing step for correcting the positional deviation and / or image distortion of the target image,
The image processing step includes
A conversion coefficient calculation step of calculating a conversion coefficient of a conversion equation for converting a pixel position of the target image for each slice;
An image processing method, comprising: a conversion coefficient correction step for correcting the conversion coefficient for each slice based on a correlation between a predetermined slice order and the conversion coefficient.

Images