JP5324880B2 - CT apparatus and metal shape extraction method - Google Patents

Description

  The present invention relates to a CT apparatus that uses radiation such as X-rays, and more particularly to a technique for reducing metal artifacts by extracting an accurate metal shape.

  A CT (Computed Tomography) apparatus uses X-rays as a light source, and obtains a cross-sectional image of a subject by image processing. This image processing has X-ray attenuation characteristics corresponding to the atomic number of the subject substance. It is assumed to have. When the object contains a metal, since the atomic number of the metal is large, the X-ray hardly transmits the metal, and the correct transmission amount cannot be measured. As a result, a correct cross-sectional image cannot be obtained, a lot of line-like noise is generated around the metal, and a cross section other than the metal portion cannot be reproduced correctly. Such noise is called a metal artifact. The reason why the metal artifact is generated is that the metal absorbs X-rays, so that the change in the X-ray absorption rate of the subject does not appear in the projection data, and the nonlinear characteristics are exhibited. Metal artifacts occur in all CTs for industrial use and medical use, but particularly in dental treatment, a large number of metals having a large atomic number are used, so the effect is significant. For this reason, severe metal artifacts occur in CT for dental treatment. When metal artifacts occur, a correct cross-sectional image of the subject cannot be obtained, often causing troubles in diagnosis and the like.

  Conventionally, an X-ray CT image reconstruction method disclosed in Patent Document 1 is known as a technique for reducing metal artifacts in an X-ray CT apparatus. FIG. 18 is a flowchart showing a processing flow of the X-ray CT image reconstruction method disclosed in Patent Document 1. In the X-ray CT image reconstruction method disclosed in Patent Document 1, X-rays are projected onto a subject from a number of directions, and a projection image (sinogram) representing the X-ray absorption rate is obtained (step S1000). Then, the value of a pixel having an appropriate threshold value or more on the sinogram, that is, the value of a pixel considered to be a metal is set to “0” (step S1001), and the sinogram after the processing of step S1001 is backprojected to generate a backprojected image. Obtain (step S1002). Subsequently, a region having a value of “0” in the backprojection image is recognized as a metal region (step S1003), and a backprojection image in which an appropriate X-ray attenuation value is given to the metal region is generated. Projection processing is performed on the back-projected image on a computer to obtain a projected image (sinogram) (step S1004). Then, the sinogram obtained in step S1004 is synthesized (added) to the sinogram obtained in step S1000 (step S1005), and the synthesized sinogram is reconstructed to obtain a cross-sectional image (reconstructed image) (step S1006). ).

  In the X-ray CT image reconstruction method disclosed in Patent Document 1, the value of the region that seems to be metal is set to “0” in the processing of step S1001, and the metal region is removed, but the threshold processing is performed completely. Therefore, in the image backprojected in the process of step S1002, there is a possibility that an area having a value of 0 is originally mixed with metal, and a non-metal area is erroneously recognized as a metal area in the process of step S1003. There was a possibility. In order to perform the processing of step S1003 without misrecognition, the threshold processing of step S1001 needs to be complete. However, it is difficult to set the threshold appropriately, and an area having a value of 0 is originally confused with metal. As there is a possibility, it is difficult to extract the shape of the correct metal region.

  Further, in the X-ray CT image reconstruction method disclosed in Patent Document 1, since the metal region and the other region are synthesized on the sinogram, the sinogram is obtained after appropriately setting the attenuation factor of the metal region. If the attenuation ratio of the metal region is not appropriate, there is a problem that the reconstructed image becomes an image in which contrast is insufficient or noise is conspicuous.

JP 2006-167161 A

  As described above, in the conventional X-ray CT apparatus, if the shape of the metal is complicated, it is difficult to accurately extract the shape, and therefore the sinogram region affected by the metal cannot be interpolated. There was a problem that it was difficult to remove.

  The present invention has been made to solve the above-described problems, and an object thereof is to provide a CT apparatus and a metal shape extraction method capable of extracting an accurate metal shape.

The CT apparatus of the present invention is disposed so as to face the radiation irradiating unit with the radiation irradiating means for irradiating the subject with radiation while rotating around the subject, A radiation imaging means for detecting radiation transmitted through the subject while rotating around the object, and collecting projection one-dimensional data from the radiation imaging means for each rotation angle of the radiation irradiation means and the radiation imaging means, Sinogram generation means for generating a sinogram in which projection one-dimensional data is arranged in order of rotation angle, first binarization processing means for binarizing the sinogram, and binarization by the first binarization processing means Back projection processing means for generating a back projection image by simply back projecting a subsequent sinogram into a two-dimensional space in calculation for each rotation angle, and second binarization processing means for binarizing the back projection image If, before Symbol the second of 2 Projection processing means for obtaining a sinogram by projecting the backprojected image after binarization by the conversion processing means, a metal region is specified based on the sinogram generated by the projection processing means, and the sinogram generation means generates and interpolation processing means for said interpolating the identified metal region on the sinogram, a reconstruction processing means for generating a reconstructed image by reconstructing the calculation of the sinogram after interpolation the interpolation processing means is generated, before Symbol it is characterized in further comprising a synthesizing process means for synthesizing the backprojection image wherein a reconstructed image reconstruction processing unit has generated the second binarization means is generated.

Further, the CT apparatus of the present invention is disposed so as to face the radiation irradiating means across the subject with radiation irradiating means for irradiating the subject with radiation while rotating around the subject, Radiation imaging means for detecting radiation transmitted through the subject while rotating around the subject, and one-dimensional projection data is collected from the radiation imaging means for each rotation angle of the radiation irradiation means and the radiation imaging means. 2 by a sinogram generating means for generating a sinogram in which each projection one-dimensional data is arranged in order of rotation angle, a first binarization processing means for binarizing the sinogram, and the first binarization processing means. A back projection processing means for generating a back projection image by simply back projecting the binarized sinogram into a two-dimensional space in calculation for each rotation angle, and a second binarization for binarizing the back projection image Processing means, Serial second binarization means is characterized by binarizing the inverse projection image the maximum value of the inverse projection image as a threshold value.

Further, the metal shape extraction method of the present invention includes a radiation irradiating means for irradiating the subject with radiation while rotating around the subject, and radiation transmitted through the subject while rotating around the subject. From the projection one-dimensional data acquisition procedure for acquiring the projection one-dimensional data by the radiation imaging means to be detected, and the projection one-dimensional data acquired by the radiation imaging means for each rotation angle of the radiation irradiation means and the radiation imaging means, A sinogram generation procedure for generating a sinogram in which projection one-dimensional data is arranged in rotation angle order, a first binarization processing procedure for binarizing the sinogram, and binarization by the first binarization processing procedure A backprojection processing procedure for generating a backprojection image by simply backprojecting a subsequent sinogram into a two-dimensional space in calculation for each rotation angle, and a second binarization processing procedure for binarizing the backprojection image and, A projection processing procedure for obtaining a sinogram backprojection image after binarization by serial second binarization procedure was projected on the calculation, to identify the metal area based on the sinogram produced by this projection procedure The interpolation processing procedure for interpolating the specified metal region on the sinogram generated by the sinogram generation procedure and the interpolated sinogram generated by the interpolation processing procedure are reconstructed by calculation to generate a reconstructed image characterized in that it comprises a reconstruction procedure, a synthetic procedure for combining the pre-Symbol backprojection image generated by the reconstruction image generated by the reconstruction procedure the second binarization processing procedure for It is what.

Further, the metal shape extraction method of the present invention includes a radiation irradiating means for irradiating the subject with radiation while rotating around the subject, and radiation transmitted through the subject while rotating around the subject. From the projection one-dimensional data acquisition procedure for acquiring the projection one-dimensional data by the radiation imaging means to be detected, and the projection one-dimensional data acquired by the radiation imaging means for each rotation angle of the radiation irradiation means and the radiation imaging means, A sinogram generation procedure for generating a sinogram in which projection one-dimensional data is arranged in rotation angle order, a first binarization processing procedure for binarizing the sinogram, and binarization by the first binarization processing procedure A backprojection processing procedure for generating a backprojection image by simply backprojecting a subsequent sinogram into a two-dimensional space in calculation for each rotation angle, and a second binarization processing procedure for binarizing the backprojection image And Seen, the second binarization processing procedure is characterized in that binarizing the inverse projection image the maximum value of the inverse projection image as a threshold value.

  According to the present invention, for each rotation angle of the radiation irradiating means and the radiation imaging means, one-dimensional projection data is collected from the radiation imaging means to generate a sinogram, the sinogram is binarized, and the binarized sinogram is rotated. A backprojected image is generated by performing simple backprojection on a two-dimensional space for each moving angle and binarizing the backprojected image, whereby a backprojected image of only metal can be obtained. As a result, according to the present invention, it is possible to extract a more accurate metal shape than before, and to reduce metal artifacts.

  In the present invention, the binarized backprojection image is projected on the calculation to obtain a sinogram, a metal region is identified based on the sinogram, and the metal region identified on the sinogram generated by the sinogram generation means , And a reconstructed image is generated by reconstructing the sinogram after the interpolation by calculation to obtain a cross-sectional image of the subject in which no metal exists.

  In the present invention, the reconstruction image generated by the reconstruction processing unit and the backprojection image generated by the second binarization processing unit are combined to reduce the number of metal artifacts and accurately form the metal. A cross-sectional image of the superimposed subject can be obtained.

Hereinafter, embodiments of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram showing a configuration of an X-ray CT apparatus according to an embodiment of the present invention.
The X-ray CT apparatus is disposed so as to face the X-ray irradiation apparatus 1 with the subject sandwiched between the X-ray irradiation apparatus 1 that irradiates the subject with X-rays while rotating around the subject. X-ray imaging apparatus 2 that detects X-rays that have passed through the subject while rotating around the object, control apparatus 3 that controls X-ray irradiation apparatus 1 and X-ray imaging apparatus 2, and X-ray imaging apparatus 2 An image processing device 4 that processes the obtained image to obtain a cross-sectional image of the subject, and a display device 5 that displays the cross-sectional image of the subject generated by the image processing device 4.

FIG. 2 is a diagram for explaining X-ray projection processing by the X-ray irradiation apparatus 1 and X-ray imaging processing by the X-ray imaging apparatus 2.
The X-ray irradiation apparatus 1 includes an X-ray tube 10 that irradiates an X-ray 101 to a patient 100 to be examined (hereinafter referred to as a subject), and an X-ray tube having a Z axis passing through the subject 100 as a rotation axis. A driving mechanism (not shown) that drives the X-ray tube 10 so that 10 rotates around the subject 100.

  The X-ray imaging apparatus 2 is disposed so as to face the X-ray tube 10 with the subject 100 interposed therebetween, a detector 20 that detects the X-ray 101 emitted from the X-ray tube 10, and a rotation axis about the Z axis. And a driving mechanism (not shown) for driving the detector 20 so that the detector 20 rotates around the subject 100. The detector 20 includes a large number of detection elements 200 arranged one-dimensionally along the channel direction 102 shown in FIG.

  FIG. 3 is a flowchart showing the operation of the X-ray CT apparatus of the present embodiment. First, the control device 3 controls the X-ray irradiation device 1 and the X-ray imaging device 2 so that the X-ray tube 10 and the detector 20 rotate around the subject 100. Since X-rays radiated from the X-ray tube 10 enter the detector 20 directly or through the subject 100, projection data indicating the X-ray transmittance is obtained from the output of the detector 20. Here, since the detector 20 includes a plurality of detection elements 200 arranged one-dimensionally, one-dimensional projection data (hereinafter referred to as projection one-dimensional data) is obtained. The image processing apparatus 4 collects the projection one-dimensional data for each rotation angle of the X-ray tube 10 and the detector 20 (step S100 in FIG. 3). In practice, one-dimensional projection data is not collected continuously for each rotation angle, but one-dimensional projection data is collected for each predetermined rotation step angle.

  FIG. 4 is a block diagram illustrating a configuration example of the image processing apparatus 4. The image processing apparatus 4 collects projection one-dimensional data from the X-ray imaging apparatus 2 for each rotation angle of the storage unit 40 for storing data and the X-ray irradiation apparatus 1 and the X-ray imaging apparatus 2, and outputs each projection one-dimensional data. A sinogram generation unit 41 that generates a sinogram in which data are arranged in order of rotation angles, a binarization processing unit (first binarization processing unit) 42 that binarizes the sinogram, and a binogram-processed sinogram. A backprojection processing unit 43 that generates a backprojection image by performing simple backprojection on a two-dimensional space in calculation for each moving angle, and a binarization processing unit that binarizes the backprojection image (second binarization processing) Part) 44, a projection processing unit 45 for projecting a binarized backprojection image in calculation and obtaining a sinogram, a metal region is specified based on the sinogram generated by the projection processing unit 45, and a sinogram generation unit 41 Interpolates the specified metal region on the sinogram generated by An interpolation processing unit 46, a reconstruction processing unit 47 that reconstructs the interpolated sinogram generated by the interpolation processing unit 46 to generate a reconstructed image, and a reconstructed image generated by the reconstruction processing unit 47. The composition processing unit 48 is configured to synthesize the back projection image generated by the binarization processing unit 44.

  The projection one-dimensional data collected from the detector 20 is stored in the storage unit 40. Each piece of projection one-dimensional data stored in the storage unit 40 corresponds to each detection element 200 and indicates the X-ray transmittance. The sinogram generation unit 41 converts the projection one-dimensional data so that each piece of projection one-dimensional data stored in the storage unit 40 changes from a value indicating X-ray transmittance to a value indicating X-ray absorption. (FIG. 3, step S101). The sinogram generation unit 41 performs such conversion processing on each projection one-dimensional data collected for each rotation angle.

  Since the projection one-dimensional data stored in the storage unit 40 is projection one-dimensional data using a fan beam spreading in a fan shape from the X-ray tube 10, it is necessary to convert the projection one-dimensional data into a projection one-dimensional data using a parallel beam. Therefore, the sinogram generation unit 41 converts the one-dimensional projection data by the fan beam from the plurality of projection one-dimensional data to the projection one-dimensional data obtained by parallelizing the beam in the same direction by interpolation, and then the converted projection one-dimensional data. These are arranged in parallel in the order of the rotation angles to generate a two-dimensional projection image (sinogram) (step S102 in FIG. 3). The sinogram generated by the sinogram generation unit 41 is stored in the storage unit 40.

  Assume that the X-ray tube 10 and the detector 20 rotate around the subject 100 as shown in FIG. In FIG. 5, the cross section of the subject 100 is drawn so as to become closer to white as the X-ray absorption rate increases. Reference numeral 103 denotes a metal present inside the subject 100. Since the X-ray absorption rate by the metal 103 is the highest, the metal 103 is drawn in white. Reference numeral 104 denotes the periphery of the subject 100, and the X-ray absorption rate is the lowest.

  A sinogram obtained from the subject 100 as shown in FIG. 5 is as shown in FIG. The X-ray absorption characteristics at the position A on the sinogram in FIG. 6 are shown in gray, and the X-ray absorption characteristics at the position B in black are shown in FIG. The vertical axis in FIG. 6 and the horizontal axis in FIG. 7 correspond to the position of each detection element 200 of the detector 20. Here, the detector 20 is composed of 1024 detection elements 200. The detector 20 outputs data indicating the X-ray transmittance. In FIGS. 6 and 7, as described in step S <b> 101, the data indicating the X-ray transmittance is absorbed by the X-ray. The sinogram after conversion into data indicating the rate is shown. FIG. 6 shows that the closer to white, the higher the X-ray absorption rate. On the sinogram shown in FIG. 6, a region where a sine curve is drawn is a region corresponding to the metal 103 in FIG. 5.

  Next, the binarization processing unit 42 performs threshold processing on each pixel of the sinogram generated by the sinogram generation unit 41 so that the threshold value is “1” and the threshold value is “0” to binarize the sinogram. (Step S103 in FIG. 3). The threshold value here is set to a value such that when binarized, the metal present in or around the subject 100 becomes almost completely “1”. Therefore, the threshold value may be set to a value slightly lower than the value of the metal X-ray absorption rate. There is a possibility that a region other than metal may be included in the region having the value “1”. However, since a region other than metal can be deleted by the processing described later, a region other than metal is included in the region having the value “1”. It does not matter. The binarized sinogram is stored in the storage unit 40.

  Subsequently, the backprojection processing unit 43 generates a backprojected image by simply backprojecting the binarized sinogram into a two-dimensional space for each rotation angle (step S104 in FIG. 3). The simple back projection is a calculation that performs the reverse process of acquiring the projection data of the subject 100 while the X-ray tube 10 and the detector 20 rotate. This back projection image corresponds to a cross-sectional image of the subject 100.

  8A to 8D are diagrams for explaining the principle of simple back projection. When the X-ray radiated from the X-ray tube 10 is detected by the detector 20 in the state shown in FIG. 8A, projection one-dimensional data 80 as shown in FIG. 8B is obtained. As described above, when the X-ray tube 10 and the detector 20 rotate around the subject 100, the projection one-dimensional data 80 is obtained for each rotation angle. When these projection one-dimensional data 80 are arranged in parallel in the order of rotation angles, a sinogram is completed. Next, as shown in FIG. 8C, when the projection one-dimensional data 80 is back-projected into the two-dimensional space for each rotation angle, and the back-projected images are superimposed and added, the result is shown in FIG. Such a backprojected image 81 is obtained. The generated backprojected image is stored in the storage unit 40.

  The binarization processing unit 44 performs threshold processing on each pixel of the backprojection image generated by the backprojection processing unit 43 so that the threshold value is “1” and less than the threshold value is “0”. The value is converted (step S105 in FIG. 3). If all the metal regions can be extracted as “1” by binarization in step S103, the back projections from all the angles overlap, so that the metal regions have the maximum value (the X-ray absorption rate is the maximum in the back projection image). ), And the region other than the metal included in the binarized sinogram takes a value lower than the maximum value.

  Therefore, basically, if the binarization is performed using the maximum value of the backprojection image as a threshold value, the metal region can be accurately extracted. However, since the pixel value is actually slightly lower than the maximum value at the metal edge portion, it is desirable that the binarization threshold value be slightly lower than the maximum value of the backprojection image. By the binarization processing in step S105, a back projection image of only metal can be obtained. The binarized backprojected image generated by the binarization processing unit 44 is stored in the storage unit 40.

  Next, the projection processing unit 45 projects the binarized backprojection image by calculation to obtain a projection image (sinogram) (step S106 in FIG. 3). This projection processing uses the backprojected image after binarization as a cross-sectional image of the subject, and calculates the same processing as acquiring the projection data of the subject 100 while the X-ray tube 10 and the detector 20 rotate. This is what you do. As in the actual projection, projection one-dimensional data is obtained for each rotation angle, and a sinogram can be generated by arranging these projection one-dimensional data in parallel in the rotation angle order.

  Since the sinogram here is for specifying a linear interpolation area to be described later, the X-ray absorptance and binarization of the metal area whose value is “1” in the binarized backprojection image. The X-ray absorptance of the non-metal region whose value is “0” in the back projection image later may be any value, the metal region is white (maximum absorptance), and the non-metal region is black What is necessary is just to project so that (absorption rate is the lowest). The sinogram generated by the projection processing unit 45 is stored in the storage unit 40.

  FIG. 9A is an example of a cross-sectional image of the subject 100 in which the metal 103 exists. FIG. 9B shows an ideal sinogram obtained from such a subject 100. When the processing of steps S100 to S102 is performed on the subject 100 as shown in FIG. 9A, a sinogram as shown in FIG. 9C is obtained, and the binarization processing of step S103 is performed on this sinogram. When this is done, a sinogram as shown in FIG. 9D is obtained. Further, when the sinogram of FIG. 9D is simply back-projected (step S104), a back-projected image as shown in FIG. 9E is obtained, and the binarization process of step S105 is performed on the back-projected image. Then, a back projection image as shown in FIG. 9F is obtained. When the projection process of step S106 is performed on the backprojected image of FIG. 9F, a sinogram as shown in FIG. 10 is obtained. The region in which the sinogram sine curve is drawn is substantially the same as the region in which the sine curve is drawn in FIG. 9B, and it can be seen that the metal region can be accurately extracted.

Next, the interpolation processing unit 46 specifies a metal region based on the sinogram generated by the projection processing unit 45 in step S106, and corresponds to the specified metal region on the sinogram generated by the sinogram generation unit 41 in step S102. The area is interpolated along the channel direction (step S107 in FIG. 3).
As interpolation processing, linear interpolation, polynomial interpolation, weblet interpolation, and the like are known, but any interpolation processing may be employed.

  FIG. 11 is a diagram for explaining a linear interpolation process of a sinogram and is a diagram showing an example of a sinogram generated by the sinogram generation unit 41. As described above, in the sinogram generated by the projection processing unit 45, the metal region is white and the region other than the metal is black, so the interpolation processing unit 46 can easily identify the metal region. Further, in the sinograms of FIGS. 6, 9 </ b> B to 9 </ b> D, 10, and 11, the vertical direction is the channel direction (the direction in which the detection elements 200 of the detector 20 are arranged).

  As shown in FIG. 11, the linear interpolation includes a value of the boundary 110 between the metal region and the non-metal region on the sinogram generated by the sinogram generation unit 41 and a value of the boundary 111 facing the boundary 110 along the channel direction. Is used to interpolate the value of the pixel between the boundary 110 and the boundary 111. More precisely, the values of the boundaries 110 and 111 between the metal region and the non-metal region are the pixels of the most recent pixel among the non-metal regions existing outside the boundary along the channel direction. Value. By such linear interpolation, a sinogram in which no metal exists can be generated. A sinogram after linear interpolation is shown in FIG. The sinogram after linear interpolation is stored in the storage unit 40.

  The reconstruction processing unit 47 computes the reconstructed image by projecting the sinogram after linear interpolation generated by the interpolation processing unit 46 into a two-dimensional space by calculation for each rotation angle by a known image reconstruction method. Generate (step S108 in FIG. 3). There are various image reconstruction methods. Generally, the FBP (Filtered Back Projection) method is used in terms of processing speed.

  13A and 13B are diagrams for explaining the principle of the FBP method. Unlike the simple backprojection described above, the FBP method filters projection one-dimensional data, and reverses the filtered projection one-dimensional data 130 into a two-dimensional space for each rotation angle as shown in FIG. Projected and back-projected images are superimposed and added to obtain a reconstructed image 131 as shown in FIG. Specifically, each image obtained by converting the projection one-dimensional data into a frequency space by FFT (Fast Fourier Transform), applying a filtering process on the frequency space, back-projecting back to the real space by inverse FFT, and back-projecting Are added together. In this way, a reconstructed image (cross-sectional image) in which no metal is present can be obtained by reconstructing a sinogram in which no metal is present. The reconstructed image generated by the reconstruction processing unit 47 is stored in the storage unit 40.

  The composition processing unit 48 generates a final reconstructed image by superimposing the reconstructed image generated by the reconfiguration processing unit 47 and the backprojected image generated by the binarization processing unit 44 for each corresponding pixel. (Step S109 in FIG. 3). By synthesizing the reconstructed image in which no metal exists and the back projection image in which only the metal is present, a reconstructed image in which the metal exists (cross-sectional image of the subject 100) can be obtained. At the time of this composition, a predetermined luminance value is assigned to the metal region of the backprojected image. Since it is only necessary to know that the metal region is a metal in the cross-sectional image of the subject 100, the luminance value assigned to the metal region whose value is “1” in the backprojection image may be an arbitrary value. The reconstructed image generated by the composition processing unit 48 is stored in the storage unit 40.

  The display device 5 displays the reconstructed image generated by the composition processing unit 48 (step S110 in FIG. 3). This completes the operation of the X-ray CT apparatus.

  FIGS. 14A to 14I are diagrams for explaining the linearity of the CT algorithm. 14A is a diagram showing an example of a cross-sectional image of a subject, FIG. 14B is a diagram showing a sinogram obtained by projecting the cross-sectional image of FIG. 14A, and FIG. 14C is a diagram. It is a figure which shows the reconstruction image obtained by reconstructing the sinogram of 14 (B). FIG. 14D shows an example of a cross-sectional image of a metal, FIG. 14E shows a sinogram obtained by projecting the cross-sectional image of FIG. 14D, and FIG. 14F shows FIG. It is a figure which shows the reconstruction image obtained by reconstructing the sinogram of (E). 14G is a diagram in which the cross-sectional image of the subject in FIG. 14A and the metal cross-sectional image in FIG. 14D are superimposed, and FIG. 14H is the cross-sectional image in FIG. FIG. 14 (I) is a diagram showing a reconstructed image obtained by reconstructing the sinogram of FIG. 14 (H).

  14A to 14I add the sinogram of FIG. 14B and the sinogram of FIG. 14E to generate the sinogram of FIG. 14H, and reconstruct the sinogram. 14 (C) and the reconstructed image of FIG. 14 (F) and the reconstructed image of FIG. 14 (F) are added to each other in FIG. 14 (I) approximated to the cross-sectional image of FIG. 14 (G). It shows that a reconstructed image is obtained.

  FIG. 15A is a diagram illustrating an example of the binogram binarized by the processing of step S103 of the present embodiment, and is a diagram illustrating the binarization of the sinogram illustrated in FIG. When the sinogram of FIG. 15A is simply backprojected by the process of step S104, a backprojected image as shown in FIG. 15B is obtained. When the backprojection image of FIG. 15B is binarized by the process of step S105, a backprojection image as shown in FIG. 15C is obtained, and a backprojection image of only metal can be obtained.

  When the backprojected image of FIG. 15C is projected by the process of step S106, a sinogram as shown in FIG. 15D is obtained. When the sinogram shown in FIG. 11 is linearly interpolated by the process of step S107, a sinogram as shown in FIG. 15E is obtained, and a sinogram from which the influence of metal is removed can be obtained. When the sinogram of FIG. 15E is reconstructed by the process of step S108, a reconstructed image as shown in FIG. 15F is obtained, and a reconstructed image without the presence of metal can be obtained. Further, when the reconstructed image of FIG. 15F and the backprojected image of FIG. 15C are superimposed by the process of step S109, a reconstructed image as shown in FIG. 16 is obtained, and a reconstructed image on which metal is superimposed. Can be obtained.

  FIG. 17 is a diagram showing a reconstructed image obtained by a conventional X-ray CT apparatus. This reconstructed image is an image obtained from the same subject as in FIGS. 11, 15A to 15F, and FIG. As is apparent from FIG. 17, the conventional X-ray CT apparatus generates metal artifacts, whereas the reconstructed image of the present embodiment shown in FIG. 16 can significantly reduce metal artifacts. I understand that.

  As described above, in the present embodiment, the sinogram is binarized, the binarized sinogram is calculated and simply back-projected into a two-dimensional space for each rotation angle, and a back-projected image is generated. By binarizing the image, a back projection image of only metal can be obtained. As a result, in the present embodiment, metal artifacts can be reduced as compared with the conventional case, and an accurate metal shape can be extracted. In the present embodiment, the binarized backprojection image is projected on the calculation to obtain a sinogram, a metal region is identified based on the sinogram, and is identified on the sinogram generated by the sinogram generator 41. A cross-sectional image of the subject in which no metal is present can be obtained by interpolating the metal region and reconstructing the interpolated sinogram by calculation to generate a reconstructed image. Further, in the present embodiment, the reconstruction image generated by the reconstruction processing unit 47 and the back projection image generated by the binarization processing unit 44 are combined to reduce the shape of the metal with less metal artifacts. A cross-sectional image of the subject superimposed with can be obtained.

  Since the metal artifact becomes a strong noise on the reconstructed image, it has a great adverse effect on the image processing using the reconstructed image. For example, when positioning using the phase-only method, similarity measurement, etc. are performed using a reconstructed image, it is clear that it will be a big problem, so metal artifact reduction is an image processing using images obtained by CT It is thought that it will bring about a great effect. In recent years, research on observation methods inside an object by transmission using radiation other than X-rays has progressed. Specific examples of radiation include ultrasonic waves, terahertz light, radio waves (millimeter waves), and the like. In these cases, the same problem is expected to occur with respect to the substance that blocks the transmission, and therefore the present invention is considered to be effective even when such radiation is used.

  In the present embodiment, as an example of the X-ray CT apparatus, a case where a wide-angle fan beam is used and a subject is fixed and a single cross section is measured is described. It is not limited to other methods such as a helical scan method that continuously scans while moving the table, a multi-slice CT in which a one-dimensional detector is arranged in multiple layers, and a cone beam CT that uses a two-dimensional detector. Can also be applied.

  The image processing apparatus 4 according to the present embodiment can be realized by a computer having a CPU, a storage device, and an external interface. The CPU executes processing as described in the present embodiment in accordance with a program stored in the storage device.

  The present invention can be applied to CT image processing technology for medical use and industrial use.

It is a block diagram which shows the structure of the X-ray CT apparatus which concerns on embodiment of this invention. It is a figure explaining the X-ray projection process by the X-ray projection apparatus and the X-ray imaging process by an X-ray imaging device in embodiment of this invention. It is a flowchart which shows operation | movement of the X-ray CT apparatus which concerns on embodiment of this invention. It is a block diagram which shows the structural example of the image processing apparatus of the X-ray CT apparatus which concerns on embodiment of this invention. It is a figure which shows one typical example of a subject. It is a figure which shows the sinogram obtained from the subject of FIG. It is a figure which shows the X-ray absorption characteristic of the cross section of the sinogram of FIG. It is a figure for demonstrating the principle of the simple back projection in embodiment of this invention. It is a figure which shows one example of the cross-sectional image of the subject in embodiment of this invention, the sinogram obtained from a subject, the binogram, the back projection image, and the back projection image after binarization. It is a figure which shows the sinogram obtained by projecting the back projection image after the binarization of FIG. It is a figure explaining the linear interpolation process of the sinogram in embodiment of this invention. It is a figure which shows the sinogram after linearly interpolating the sinogram of FIG. It is a figure explaining the principle of FBP method. It is a figure explaining the linearity of CT algorithm. Example of binarized sinogram in embodiment of the present invention, backprojected image, binarized backprojected image, sinogram obtained by projecting backprojected image, sinogram after linear interpolation, and after linear interpolation It is a figure which shows the reconstructed image which reconstructed the sinogram of. FIG. 16 is a diagram showing a reconstructed image in which the reconstructed image of FIG. 15 and the binarized backprojected image are superimposed. It is a figure which shows the reconstruction image obtained by the conventional X-ray CT apparatus. It is a flowchart which shows the flow of a process of the conventional X-ray CT image reconstruction method.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 ... X-ray irradiation apparatus, 2 ... X-ray imaging device, 3 ... Control apparatus, 4 ... Image processing apparatus, 5 ... Display apparatus, 10 ... X-ray tube, 20 ... Detector, 40 ... Memory | storage part, 41 ... Sinogram generation , 42..., Binarization processing unit (first binarization processing unit), 44... Binarization processing unit (second binarization processing unit), 43... Back projection processing unit, 45. , 46 ... interpolation processing unit, 47 ... reconstruction processing unit, 48 ... synthesis processing unit, 100 ... subject, 101 ... X-ray, 200 ... detection element.

Claims (4)

Radiation irradiating means for irradiating the subject with radiation while rotating around the subject;
A radiation imaging unit that is disposed so as to face the radiation irradiating unit across the subject, and that detects radiation transmitted through the subject while rotating around the subject;
A sinogram generating means for collecting projection one-dimensional data from the radiation imaging means for each rotation angle of the radiation irradiating means and the radiation imaging means, and generating a sinogram in which the projection one-dimensional data are arranged in the order of the rotation angles;
First binarization processing means for binarizing the sinogram;
Back projection processing means for generating a back projection image by simply back projecting the sinogram after binarization by the first binarization processing means into a two-dimensional space on a calculation basis for each rotation angle;
Second binarization processing means for binarizing the backprojected image ;
Projection processing means for projecting the backprojected image after binarization by the second binarization processing means to obtain a sinogram;
Interpolation processing means for identifying a metal area based on the sinogram generated by the projection processing means, and interpolating the identified metal area on the sinogram generated by the sinogram generation means;
Reconstruction processing means for reconstructing the interpolated sinogram generated by the interpolation processing means on calculation to generate a reconstructed image;
A CT apparatus comprising: a synthesis processing unit that synthesizes the reconstructed image generated by the reconstruction processing unit and the backprojected image generated by the second binarization processing unit . Radiation irradiating means for irradiating the subject with radiation while rotating around the subject;
A radiation imaging unit that is disposed so as to face the radiation irradiating unit across the subject, and that detects radiation transmitted through the subject while rotating around the subject;
A sinogram generating means for collecting projection one-dimensional data from the radiation imaging means for each rotation angle of the radiation irradiating means and the radiation imaging means, and generating a sinogram in which the projection one-dimensional data are arranged in the order of the rotation angles;
First binarization processing means for binarizing the sinogram;
Back projection processing means for generating a back projection image by simply back projecting the sinogram after binarization by the first binarization processing means into a two-dimensional space on a calculation basis for each rotation angle;
Second binarization processing means for binarizing the backprojected image,
The CT apparatus characterized in that the second binarization processing means binarizes the backprojection image using a maximum value of the backprojection image as a threshold value . Projected one-dimensional data by radiation irradiating means for irradiating the subject with radiation while rotating around the subject and radiation imaging means for detecting radiation transmitted through the subject while rotating around the subject. Projection one-dimensional data acquisition procedure for acquiring
A sinogram generation procedure for generating, from the projection one-dimensional data acquired by the radiation imaging means for each rotation angle of the radiation irradiating means and the radiation imaging means, a sinogram in which the projection one-dimensional data are arranged in order of rotation angles;
A first binarization processing procedure for binarizing the sinogram;
A back projection processing procedure for generating a back projection image by simply back projecting the binogram after binarization according to the first binarization processing procedure into a two-dimensional space on a calculation basis for each rotation angle;
A second binarization processing procedure for binarizing the backprojected image ;
A projection processing procedure for obtaining a sinogram by projecting the backprojected image after binarization by the second binarization processing procedure on calculation;
An interpolation processing procedure for identifying a metal region based on the sinogram generated by the projection processing procedure and interpolating the identified metal region on the sinogram generated by the sinogram generation procedure;
Reconstruction processing procedure for reconstructing the calculated sinogram generated by this interpolation processing procedure to generate a reconstructed image, and
A metal shape extraction method comprising: a synthesis processing procedure for synthesizing the reconstructed image generated by the reconstruction processing procedure and the backprojected image generated by the second binarization processing procedure . Projected one-dimensional data by radiation irradiating means for irradiating the subject with radiation while rotating around the subject and radiation imaging means for detecting radiation transmitted through the subject while rotating around the subject. Projection one-dimensional data acquisition procedure for acquiring
A sinogram generation procedure for generating, from the projection one-dimensional data acquired by the radiation imaging means for each rotation angle of the radiation irradiating means and the radiation imaging means, a sinogram in which the projection one-dimensional data are arranged in order of rotation angles;
A first binarization processing procedure for binarizing the sinogram;
A back projection processing procedure for generating a back projection image by simply back projecting the binogram after binarization according to the first binarization processing procedure into a two-dimensional space on a calculation basis for each rotation angle;
A second binarization processing procedure for binarizing the backprojected image,
The second binarization processing procedure binarizes the backprojection image using a maximum value of the backprojection image as a threshold value .