Correction device, method, and program

Abstract

To provide a correction device capable of acquiring a physically proper corrected image with reduced artifacts without performing measurement with two types of tube voltages, method and program.SOLUTION: A correction device 400 includes: a projection image acquisition part 410 for acquiring a measured projection image; a virtual energy setting part 420 for setting virtual energy different from measured energy, the X-ray energy applied when the measured projection image is acquired; a dual energy decomposition part 460 for calculating a projection image during irradiation with virtual energy X-ray under assumed conditions and performing dual energy decomposition; a consistency index evaluation part 470 for specifying an optimum condition among conditions assumed by evaluating a temporarily corrected image obtained by the dual energy decomposition using a consistency index; and a corrected image generation part 480 for generating a projection image when the optimum condition is used, as a corrected image.SELECTED DRAWING: Figure 3

Description

[Technical Field]

[0001] The present invention relates to a correction device, method, and program for correcting artifacts. [Background technology]

[0002] CT images can sometimes exhibit severe artifacts, such as metal artifacts. The causes include inconsistencies (beam hardening) resulting from using continuous X-rays for measurement while assuming monochromatic X-rays for reconstruction, low counting ability due to significant absorption of incident X-rays by metals resulting in almost no X-ray detection by the detector, and the contribution of scattered radiation that is not assumed during reconstruction.

[0003] Conventionally, hardware or software measures have been taken to address these issues. Hardware measures include using monochromatic X-rays for measurement and placing a grid in front of the detector to reduce scattered radiation. Software measures include reconstructing the image using a successive method while making assumptions about the reconstructed image.

[0004] However, when scattered radiation is generated by causes other than X-ray absorption due to the photoelectric effect, the above measures may not be sufficient to reduce artifacts. Therefore, a method is sometimes employed in which the projection images measured at high and low energies are separated into the photoelectric effect component and the Compton scattering component at the projection image level and then resolved (see Patent Documents 1 and 2). This method is called dual-energy resolving, and it reduces artifacts by using the separated projection images.

[0005] On the other hand, a method is known for correcting the intensity distribution of a projection image that has become inconsistent due to some error factor, such that the consistency condition is satisfied, which states that the sum of the intensity of the projection image does not depend on the angle (see Non-Patent Literature 1). This consistency condition is called the Helgason-Ludwig condition. Measured projection images do not satisfy this conservation law regarding the linear absorption coefficient due to various factors, but this method determines the parameters of the correction function to satisfy it. [Prior art documents] [Patent Documents]

[0006] [Patent Document 1] Japanese Patent Publication No. 2009-261942 [Patent Document 2] Japanese Patent Publication No. 2014-000409 [Non-patent literature]

[0007] [Non-Patent Document 1] Tobias Wurfl, Nicole Maas, Frank Dennerlein, Andreas K. Maier, “A new calibration-free beam hardening reduction method for industrial CT”, 8th Conference on Industrial Computed Tomography, Wels, Austria (iCT 2018) [Overview of the Initiative] [Problems that the invention aims to solve]

[0008] In the dual-energy resolution method described above, it is necessary to irradiate the sample with X-rays at two different tube voltages and obtain a measured projection image. However, measuring at two different tube voltages requires not only changing the voltage but also implementing sufficient X-ray leakage countermeasures in the apparatus, making its introduction difficult. For example, if you want to measure at a tube voltage of 225kV, the hardware needs to be reinforced to prevent X-ray leakage.

[0009] Furthermore, simply performing corrections to satisfy only the Helgason-Ludwig condition can sometimes reverse the intensity of the reconstructed image. This is likely because attempting to eliminate multiple error factors all at once can lead to the selection of a physically unsuitable answer.

[0010] This invention has been made in view of these circumstances, and aims to provide a correction device, method, and program that can obtain a physically valid corrected image with reduced artifacts without measuring with two different tube voltages. [Means for solving the problem]

[0011] (1) To achieve the above objective, the present invention is a correction device for correcting artifacts in CT images, and is characterized by comprising: a projection image acquisition unit that acquires a measured projection image; a virtual energy setting unit that sets a virtual energy different from the measured energy, which is the energy of the X-rays irradiated at the time the measured projection image is acquired; a dual energy decomposition unit that calculates a projection image under assumed conditions when the X-rays of the virtual energy are irradiated and performs dual energy decomposition; a consistency index evaluation unit that evaluates the provisional correction image obtained by the dual energy decomposition using a consistency index to identify the optimal condition among the assumed conditions; and a correction image generation unit that generates a projection image using the optimal condition as a correction image.

[0012] (2) Furthermore, the correction device described in (1) above is characterized in that the consistency index is an index that indicates the degree to which the Helgason-Ludwig condition is satisfied.

[0013] (3) In addition, in the correction device described in (1) or (2) above, the virtual energy setting unit is characterized in that it sets the virtual energy higher than the measured energy.

[0014] (4) Furthermore, in the correction device described in (3) above, the virtual energy setting unit is characterized in that it sets the virtual energy such that the difference from the measured energy is greater than or equal to a predetermined value.

[0015] (5) Further, in the correction device according to any one of (1) to (4) above, the dual energy decomposition unit is characterized in that a polynomial is used as the functional form of the conversion function used when calculating the projection image by irradiating X-rays of the virtual energy.

[0016] (6) Further, in the correction device according to any one of (1) to (5) above, the dual energy decomposition unit is characterized in that a linear absorption coefficient model according to the measurement conditions when acquiring the actually measured projection image is used.

[0017] (7) Further, in the correction device according to any one of (1) to (6) above, the consistency index evaluation unit is characterized in that a particle swarm optimization method is used for setting the parameters of the conversion function among the assumed conditions.

[0018] (8) Further, in the correction device according to any one of (1) to (7) above, the virtual energy setting unit is characterized in that it uses the setting of the incident X-ray distribution from the received user.

[0019] (9) Further, the correction method of the present invention is a correction method for correcting artifacts in a CT image, and includes steps of acquiring an actually measured projection image, setting a virtual energy different from the actually measured energy which is the energy of the X-rays irradiated when acquiring the actually measured projection image, calculating a projection image under assumed conditions when irradiating X-rays of the virtual energy and performing dual energy decomposition, specifying an optimal condition among the assumed conditions by evaluating the virtual correction image obtained by the dual energy decomposition with a consistency index, and generating a projection image when using the optimal condition as a correction image.

[0020] (10) Further, the correction program of the present invention is a correction program for correcting artifacts in a CT image, and includes a process of acquiring a measured projection image, a process of setting a virtual energy different from the measured energy which is the energy of X-rays irradiated when the measured projection image is acquired, a process of calculating a projection image under assumed conditions at the time of irradiation with X-rays of the virtual energy and performing dual-energy decomposition, a process of specifying an optimal condition among the assumed conditions by evaluating a provisional correction image obtained by the dual-energy decomposition using a consistency index, and a process of generating a projection image obtained when using the optimal condition as a correction image, and is characterized by causing a computer to execute the processes.

Brief Description of the Drawings

[0021] [Figure 1] (a) and (b) are schematic views showing the integral of the linear absorption coefficient in dual-energy decomposition and projection at an angle k, respectively. [Figure 2] It is a schematic view showing the configuration of the system of the present invention. [Figure 3] It is a block diagram showing the configuration of the processing device of the present invention. [Figure 4] It is a flowchart showing the correction method of the present invention. [Figure 5] It is a schematic view showing an example of a GUI. [Figure 6] (a) and (b) are reconstructed images using the projection images before and after correction when imaging a composite sample C1, respectively. [Figure 7] It is a reconstructed image using the projection image after correction based only on the Helgason-Ludwig consistency condition.

Embodiments of the Invention

[0022] Next, embodiments of the present invention will be described with reference to the drawings. For ease of understanding of the description, the same reference numerals are assigned to the same components in each drawing, and duplicate descriptions are omitted.

[0023] [Principle] A CT scanner irradiates a sample with cone-shaped or parallel beams of X-rays from any angle, and a detector acquires the distribution of the X-ray absorption coefficient, i.e., a projection image. To irradiate X-rays from any angle, a CT scanner is configured to rotate either the sample stage relative to a fixed X-ray source and detector, or to rotate the gantry, which is an integrated unit of the X-ray source and detector.

[0024] By rotating the gantry in this way, the sample is projected from various angles, and the distribution of the sample's linear absorption coefficient f can be estimated from the intensity of the projected image. Then, the three-dimensional distribution of the linear absorption coefficient is determined from the two-dimensional projected image. This is called reconstruction. Reconstruction is basically performed by back-projection of the projected image. When X-rays are irradiated onto a sample, scattered radiation is generated by causes other than the photoelectric effect, which causes artifacts in the reconstructed image. To reduce these artifacts, the intensity distribution detected by the detector is considered first.

[0025] Assuming that continuous X-rays are a collection of N finite monochromatic X-rays, the incident X-ray intensity can be replaced by the intensity obtained by summing the intensities Db (b=1,2,…,N) of each monochromatic X-ray. Furthermore, the nonlinearity between the X-ray transmission distance and X-ray attenuation can be expressed by summing the attenuation of each monochromatic X-ray over its total energy. The intensity I detected at each detector pixel is expressed by equation (1) below as the sum of the intensities of each monochromatic X-ray attenuated by the material.

[0026]

number

[0027] Assuming that the intensity attenuation of X-rays is due to the photoelectric effect and Compton scattering, a composite equation using two line integrals, as shown in equation (2), is obtained for the coordinates (including angles) of each detector. The values ​​of these line integrals correspond to projections of the distribution of the line absorption coefficient and density distribution at a certain reference energy E0. The energy dependence of the photoelectric effect and Compton scattering is expressed by the power factor and Klein-Nishina factor, respectively. If these projections can be extracted, they can be treated as corrected values.

[0028]

number

[0029] If the intensity obtained at two different tube voltages, high-energy and low-energy, is known, then two equations containing two unknowns can be set up, as shown in equations (3) and (4). By solving this system of equations, a projection image resolved by the scattering effect can be obtained. This is dual-energy resolution.

[0030]

number

[0031] Equation (4) is a simplified representation of equation (3). This is a system of nonlinear equations with an unknown line integral value, and can be solved using Newton's method for multiple variables. Figure 1(a) shows a schematic diagram of the dual energy decomposition. The dual energy decomposition shown in Figure 1(a) can output intensity distributions due to either the photoelectric effect only or the Compton effect only, given an input intensity distribution measured at a low energy tube voltage.

[0032]

number

[0033] In this invention, one of the intensities obtained with two different tube voltages is obtained by actual measurement, and the other is calculated by substituting it into a conversion function G (for example, equation (5)) that includes parameter a. However, parameter a is not a single value, but a set of multiple values ​​like a vector a=(a1,…,a n ) means. Here, I Low This is the measured projection image, I High It is treated as a virtual projection, but the reverse is also acceptable.

[0034]

number

[0035] By using equations (4) and (5) above, a hypothetical projection image can be obtained that is based on only one intensity attenuation factor, such as the photoelectric effect. Since dual energy decomposition is performed using the calculated projection image, this method should also be called pseudo-dual energy decomposition. For multiple parameters a, multiple hypothetical projection images can be obtained. Then, the consistency index of each hypothetical projection image is calculated, and the parameter a that best fits the consistency condition is determined. Alternatively, equation (6), which uses the value P obtained by logarithmically transforming the intensity I in equation (5), can be used instead of equation (5).

[0036]

number

[0037] If the sample is within the measurement field of view, there are no extraneous error factors, and X-ray absorption occurs ideally, the sum of the intensity of the projected image should be independent of the angle. This is called consistency. The intensity detected by the detector is equivalent to the line integral of the line absorption coefficient along the line connecting the detection position and the radiation source. By integrating this value with respect to the detector coordinate s, the sum of the line absorption coefficients of the sample can be obtained. When the measurement is performed using the cone beam method, the range is limited to the central cross-section Z=0. However, in reality, consistency is broken due to some error factors. Therefore, a plausible projected image can be obtained by determining a transformation function that satisfies the consistency condition of the projected image to a certain extent.

[0038] As a consistency index, for example, the NRMSD (Normalized Root Mean Square Deviation), an index for the Helgason-Ludwig consistency condition as shown in equation (7), can be used. Figure 1(b) is a schematic diagram showing the integral of the linear absorption coefficient in the projection of angle k. Φ PDE This is a function that transforms the projected image into a corrected projected image. As the integral of the projected image (the sum of the linear absorption coefficients contained in the cross-section) approaches an invariant, NRMSD approaches zero.

[0039]

number

[0040] M k This is obtained by integrating the projection image obtained by quasi-dual energy decomposition of the measured projection image at angle k in the CU direction. In reality, since measurements are taken with a two-dimensional detector, only the central portion is used in the CV direction. Average j (M j ) is the integral averaged with respect to angle, and j is independent of the outer k. Note that the CV direction is parallel to the axis of rotation, and the CU direction is perpendicular to it on the detector.

[0041] Once the parameter 'a' that best fits the consistency criteria is determined, the previously assumed conditions are adopted as the optimal correction conditions, and the projected image is calculated under these conditions. Then, a reconstructed image is generated based on the corrected series of projected images. In this way, a corrected image with reduced artifacts and physical validity can be obtained.

[0042] [Overall System] Figure 2 is a schematic diagram showing the configuration of system 100. System 100 includes a CT scanner 200, a processing unit 300, an input device 510, and an output device 520. The processing unit 300 is equipped with a correction device 400 and is connected to the CT scanner 200.

[0043] As shown in Figure 2, the CT apparatus 200 is configured to allow the sample to rotate relative to the X-ray source 260 and the detector 270. The CT apparatus 200 is not limited to this configuration, and may also be configured to rotate a gantry in which the X-ray source and detector are integrated. Furthermore, the CT apparatus 200 may be either a parallel beam apparatus or a cone beam apparatus.

[0044] The processing unit 300 is connected to the CT scanner 200 and controls the CT scanner 200 and processes the acquired data. The correction device 400 corrects the projection image. The processing unit 300 may be a PC terminal or a server on the cloud. The input device 510 is, for example, a keyboard or mouse and provides input to the processing unit 300. The output device 520 is, for example, a display and shows the projection image, etc.

[0045] As shown in Figure 2, the correction device 400 may be configured as a part of the processing unit 300, or the correction device 400 and the processing unit 300 may be configured as an integrated unit. The correction device 400 may also be provided as a separate component from the processing unit 300.

[0046] [CT device] As shown in Figure 2, the CT apparatus 200 includes a rotation control unit 210, a sample stage 250, an X-ray source 260, a detector 270, and a drive unit 280. X-ray CT imaging is performed by rotating the sample stage 250, which is installed between the X-ray source 260 and the detector 270. Alternatively, the X-ray source 260 and the detector 270 may be installed in a gantry (not shown), and the gantry may be rotated relative to a sample fixed to the sample stage 250.

[0047] The CT scanner 200 drives the sample stage 250 at the timing instructed by the processing unit 300 and acquires a projection image of the sample. The measurement data is transmitted to the processing unit 300. The CT scanner 200 is suitable for use with precision industrial products such as semiconductor devices, but it can be applied not only to industrial equipment but also to animal equipment.

[0048] The X-ray source 260 directs X-rays towards the detector 270. The detector 270 has a light-receiving surface that receives X-rays and can measure the intensity distribution of X-rays transmitted through the sample using a large number of pixels. The rotation control unit 210 rotates the sample stage 250 at a speed set during CT imaging using the drive unit 280.

[0049] [Processing device] Figure 3 is a block diagram showing the configuration of the processing unit 300. The processing unit 300 is composed of a computer consisting of a CPU (Central Processing Unit), ROM (Read Only Memory), RAM (Random Access Memory), and memory connected to a bus. The processing unit 300 is connected to the CT device 200 to receive information.

[0050] The processing unit 300 comprises a measurement data storage unit 310, a device information storage unit 320, a reconstruction unit 330, and an output control unit 340. Each unit can send and receive information via a control bus L. The processing unit 300 also includes a correction device 400 as part of its structure. The input device 510 and the output device 520 are connected to the CPU via appropriate interfaces.

[0051] The measurement data storage unit 310 stores measurement data acquired from the CT scanner 200. The measurement data includes rotation angle information and the corresponding projection image. The device information storage unit 320 stores device information acquired from the CT scanner 200. The device information includes the device name, beam shape, measurement geometry, scanning method, etc. The reconstruction unit 330 reconstructs a CT image from the series of projection images. The output control unit 340 displays the reconstructed CT image on the output device 520. This allows the user to confirm the corrected CT image. The user can also give instructions or specifications to the processing device, correction device, etc., based on the CT image.

[0052] [Correction device] The correction device 400, as part of the processing unit 300, is comprised of a computer equipped with a CPU, ROM, RAM, and memory. The correction device 400 may be directly connected to the CT device 200 or connected to the CT device 200 via the processing unit. The correction device 400 may also receive information from the CT device or from the processing unit 300.

[0053] The correction device 400 comprises a projection image acquisition unit 410, a virtual energy setting unit 420, an incident X-ray distribution setting unit 430, a linear absorption coefficient model setting unit 440, a transformation function model setting unit 450, a dual energy decomposition unit 460, a consistency index evaluation unit 470, and a corrected image generation unit 480. Each unit can send and receive information via the control bus L. If the correction device 400 and the processing unit 300 have different configurations, the input device 510 and output device 520 are also connected to the CPU of the correction device 400 via an appropriate interface. In this case, the input device 510 and output device 520 may be different from those connected to the processing unit 300.

[0054] The projection image acquisition unit 410 acquires the measured projection images from the measurement data storage unit 310 as the images to be corrected. The series of projection images obtained from a single CT scan are the images to be corrected. This is particularly effective when there are strong artifacts in the projection images. For the series of projection images, the optimal correction conditions may be determined for each image, or the optimal correction conditions may be determined for a representative projection image and applied to the other projection images.

[0055] The virtual energy setting unit 420 sets a virtual energy that is different from the measured energy. The measured energy is the energy of the X-rays irradiated when the measured projection image was acquired. The virtual energy is a hypothetical energy that is different from the measured energy. For example, when the measured energy is 100kV, the virtual energy is set to 200kV.

[0056] It is preferable to set the virtual energy higher than the measured energy. This allows for correction to reduce artifacts using dual energy decomposition even in devices that cannot use high energy due to configuration limitations. It is also preferable to set the virtual energy so that the difference from the measured energy is greater than or equal to a predetermined value. An example of a predetermined value is 60kV. By setting a sufficient difference, a stable solution can be obtained during optimization, and the correction calculation becomes easier. In principle, it is also possible to set the virtual energy lower than the measured energy.

[0057] The incident X-ray distribution setting unit 430 sets the incident X-ray distribution when irradiating with virtual energy X-rays. Preferably, the incident X-ray distribution setting unit 430 uses the incident X-ray distribution setting received from the user. This enables analysis according to the measurement conditions and allows for obtaining an optimally corrected image. The incident X-ray distribution of the measured energy is automatically determined by the tube voltage and filter conditions at the time of measurement.

[0058] The linear absorption coefficient model setting unit 440 sets the linear absorption coefficient model used for dual-energy resolution. Basically, the photoelectric effect and Compton scattering models are sufficient, but depending on the application and measurement conditions, models such as Rayleigh scattering may also be used. The functional form of the linear absorption coefficient used for dual-energy resolution can be set according to the measurement conditions when acquiring the measured projection image. For example, if the sample contains only two types of material, such as soft tissue and bone, the model can be set using the energy dependence of the linear absorption coefficient of each material. In this way, even under special conditions, an appropriate model can be used to perform analysis according to the situation.

[0059] The transformation function model setting unit 450 sets a transformation function model that converts the intensity distribution at the time of measurement to the intensity distribution at the time of X-ray irradiation estimation. Preferably, the function form of the transformation function is a polynomial. This simplifies calculations and yields physically valid correction results. Besides polynomials, functions with positive coefficients, composed of linear and exponential functions, can also be used. The parameters in the transformation function are assumed and evaluated based on the consistency of the calculated provisional correction image. Each setting unit may be configured according to user instructions, or it may be configured automatically. Furthermore, settings may be adjusted according to the situation, or they may be set to predetermined values.

[0060] The dual energy resolution unit 460 calculates the projection image under assumed conditions when X-ray irradiation is performed with a virtual energy. In doing so, it assumes parameters of the transformation function. Then, it performs dual energy resolution using the intensity distribution calculated by the transformation function to calculate the intensity distribution under assumed conditions as a provisionally corrected image.

[0061] The consistency index evaluation unit 470 identifies the optimal conditions by evaluating the obtained provisionally corrected images using a consistency index. Evaluating with a consistency index means calculating the consistency index of the obtained provisionally corrected images and determining the parameter 'a' that best fits the consistency conditions. Preferably, the consistency index is an index that indicates the degree to which the Helgason-Ludwig condition is satisfied. This makes it possible to obtain a physically valid corrected image from among the provisionally corrected images generated under the assumed conditions.

[0062] For setting the parameters of the transformation function, it is preferable to use the particle swarm optimization method. The particle swarm optimization method treats the positions of the particles as candidate solutions, and each particle updates its position while maintaining its own optimal position and the optimal position for the entire swarm. This allows for the selection of the overall optimal solution without getting stuck in local optima, and an optimally corrected image can be obtained. Gradient methods and simplex methods can also be used as optimization methods.

[0063] The correction image generation unit 480 generates a corrected image based on the projection image obtained using optimal conditions. When optimal conditions are obtained, it is preferable to generate a projection image of the entire detection surface under those conditions. If it is possible to reuse a provisional correction image that has already been generated, it may be reused. In this way, a physically valid correction image with reduced artifacts can be obtained without measuring with two different tube voltages.

[0064] [Measurement method] The sample is placed in the CT scanner 200, and a projection image is acquired while irradiating the sample with X-rays by repeatedly moving the rotation axis and projecting X-rays under predetermined conditions. The CT scanner 200 transmits device information such as the scanning method and the acquired projection image as measurement data to the processing unit 300 or correction unit 400. Since a virtually high-energy projection image is created by numerical calculations after the measurement, this measurement only needs to be performed with one type of voltage.

[0065] [Correction Method] FIG. 4 is a flowchart showing a correction method. First, data of projection images to be corrected among the measurement data is acquired (step S1). Then, a virtual energy different from the actually measured energy is set and the intensity distribution of incident X-rays is set (step S2), and a model of the linear absorption coefficient is set (step S3). In this way, a virtual energy incident X-ray distribution, a model of the linear absorption coefficient, etc. are set in advance. Also, a conversion function model for converting the intensity distribution at the actually measured energy into the intensity distribution at a different energy is acquired (step S4).

[0066] The parameter a of the acquired conversion function model is set (step S5), and dual-energy decomposition is performed on the intensity distribution when the actually measured projection data is imaged to solve a set of simultaneous equations. As a result, the intensity distribution at the virtual energy is obtained as the solution (step S6). For example, the parameter a = (a1,..., a n ) can start from an initial setting where a1 = 1 and the other a n are almost zero.

[0067] Next, a consistency index of the obtained intensity distribution is calculated (step S7), and it is determined whether the repetition condition is satisfied (step S8). If the repetition condition is not satisfied, the process returns to step S5. On the other hand, if the repetition condition is satisfied, the process proceeds to step S9. In this way, by changing the parameter a and repeating steps S5 to S7, consistency indexes for a plurality of parameters a are calculated. The calculated consistency index is preferably displayed to the user. As the repetition condition, for example, it is determined whether the consistency index seems to converge, and a threshold value is used for the determination. Different convergence threshold values may be set for each conversion function. Note that in the example shown in FIG. 4, an iterative process is used, but a solution may be obtained at once by parallel processing.

[0068] The obtained consistency index is evaluated to determine the optimal parameter a (step S9), and the intensity distribution for the optimal parameter a is obtained as a corrected projection image (step S10). It is preferable to use a particle swarm optimization method for this optimization, where multiple sets of parameters are temporarily defined and updated during iterative calculations.

[0069] Once the optimal parameter a is determined by the consistency index, reconstruction is performed using the obtained projection image (step S11). If the consistency index has been calculated for only a portion of the detector region, after iterative processing, a pseudo-dual energy resolution is performed for the entire region.

[0070] [GUI] The correction device 400 can communicate with the user via the input device 510 and the output device 520. The user can recognize information and perform input operations based on the GUI displayed on the output device 520. Figure 5 is a schematic diagram showing an example of the GUI.

[0071] The basic GUI screen 600 shown in Figure 5 displays the projection image drawing area, along with buttons for data loading, distribution settings, transformation function settings, correction execution, and reconstruction execution. The projection image drawing area displays the projection image on the detection surface within a frame, and after correction, the corrected projection image is displayed.

[0072] When the data loading button is clicked, for example, multiple accumulated measurement data will be presented, and candidate data for correction will be displayed for selection. The user can determine which measurement data to correct by selecting the appropriate data.

[0073] When the distribution setting button is clicked, a window 610 pops up displaying the incident X-ray distribution used for correction processing. The user can review the displayed incident X-ray distribution, adjust it as needed, and then click OK to finalize the incident X-ray distribution. For example, in the example shown in Figure 5, the box is designed to allow input of tube voltage and filter information. By default, the system may display a voltage higher than the measured tube voltage and the incident X-ray distribution determined by the filter conditions. Instead of inputting numerical values, the system may be designed to allow the user to move points on a graph.

[0074] When the button for setting the conversion function is clicked, window 620 pops up, displaying the function form and parameters. The user can review the displayed function form and parameters, adjust them as needed, and then click OK to confirm the function form and parameters. Alternatively, the user can use pre-defined parameters instead of entering them manually.

[0075] When the "Execute Correction" button is clicked, the correction process is executed, and window 630 pops up showing a graph of consistency metrics for the number of iterations or parameters. This graph represents the progress of the correction. The user can check whether the iteration process is converging and whether the optimal parameters are valid. Once the iteration process has converged, the corrected projection image is displayed in the projection image drawing area. If the "Execute Reconstruction" button is clicked again, a reconstructed image is generated using the corrected series of projection images and displayed on the screen.

[0076] [Examples] In fact, the projection image of composite sample C1, in which a total of four metals (two each of Al and Fe) were inserted into wood, was corrected. Figures 6(a) and (b) are reconstructed images using the projection images of composite sample C1 before and after correction, respectively.

[0077] As shown in Figure 6(a), the uncorrected projection image clearly shows severe black streaks between the metals. On the other hand, as shown in Figure 6(b), the reconstructed image using the corrected projection image shows reduced streaks, and multiple holes near the center of the wood are easily identifiable.

[0078] [Comparative Example] For composite sample C1, the projection image was corrected using only the Helgason-Ludwig consistency condition. Figure 7 shows the reconstructed image using the projection image corrected using only the Helgason-Ludwig consistency condition. In the projection image shown in Figure 7, there are regions in the Al portion where the linear absorption coefficient is smaller than that of the wood. Also, the hole in the center of the wood is no longer visible. Thus, it can be seen that the correction is insufficient when compared to the reconstructed image using the corrected projection image of the present invention shown in Figure 6(b). [Explanation of symbols]

[0079] 100 Systems 200 CT equipment 210 Rotation Control Unit 250 sample stage 260 X-ray source 270 detectors 280 Drive unit 300 Processing Units 310 Measurement data storage unit 320 Device information storage unit 330 Reconstruction part 340 Output Control Unit 400 Correction device 410 Projection image acquisition unit 420 Virtual Energy Setting Unit 430 Incident X-ray distribution setting section 440 Line Absorption Coefficient Model Setting Section 450 Transformation Function Model Setting Section 460 Dual Energy Splitter 470 Consistency Indicator Evaluation Department 480 Corrected Image Generation Unit L control bus 510 Input device 520 Output device 600 Basic screen 610-630 Presentation screen (window) C1 Composite Sample a parameter Db Incident intensity distribution of each monochromatic X-ray G transformation function NRMSD indicator

Claims

1. A correction device for correcting artifacts in CT images, A projection image acquisition unit that acquires the measured projection image, A virtual energy setting unit sets a virtual energy that is different from the measured energy, which is the energy of the X-rays irradiated when the measured projection image was acquired. A dual energy decomposition unit calculates a projection image under assumed conditions when X-rays of the virtual energy are irradiated, and performs dual energy decomposition using the measured projection image and the projection image under assumed conditions when X-rays of the virtual energy are irradiated. A consistency index evaluation unit that identifies the optimal condition among the assumed conditions by evaluating the provisionally corrected image obtained by the dual energy resolution using a consistency index, The system includes a correction image generation unit that generates a corrected image of the projection image obtained using the aforementioned optimal conditions, The correction device is characterized in that the consistency index is an index indicating the degree to which the Helgason-Ludwig condition is satisfied.

2. A correction device for correcting artifacts in CT images, A projection image acquisition unit that acquires the measured projection image, A virtual energy setting unit sets a virtual energy that is different from the measured energy, which is the energy of the X-rays irradiated when the measured projection image was acquired. A dual energy decomposition unit calculates a projection image under assumed conditions when X-rays of the virtual energy are irradiated, and performs dual energy decomposition using the measured projection image and the projection image under assumed conditions when X-rays of the virtual energy are irradiated. A consistency index evaluation unit that identifies the optimal condition among the assumed conditions by evaluating the provisionally corrected image obtained by the dual energy resolution using a consistency index, The system includes a correction image generation unit that generates a corrected image of the projection image obtained using the aforementioned optimal conditions, The consistency index evaluation unit is a correction device characterized by using a particle swarm optimization method to set the parameters of the transformation function among the assumed conditions.

3. The correction device according to claim 1 or 2, characterized in that the virtual energy setting unit sets the virtual energy higher than the measured energy.

4. The correction device according to claim 3, characterized in that the virtual energy setting unit sets the virtual energy such that the difference from the measured energy is greater than or equal to a predetermined value.

5. The correction device according to claim 1 or 2, characterized in that the dual energy decomposition unit uses a polynomial as the functional form of the transformation function used when calculating the projection image due to X-ray irradiation of the virtual energy.

6. The correction device according to claim 1 or 2, characterized in that the dual energy decomposition unit uses a linear absorption coefficient model corresponding to the measurement conditions when acquiring the measured projection image.

7. The correction device according to claim 1 or 2, characterized in that the virtual energy setting unit uses the setting of the incident X-ray distribution received from the user.

8. A correction method for correcting artifacts in CT images, Steps include obtaining the measured projection image, The steps include setting a virtual energy different from the measured energy, which is the energy of the X-rays irradiated when the measured projection image was acquired, The steps include: calculating the projection image under assumed conditions when X-rays of the virtual energy are irradiated, and performing dual energy decomposition using the measured projection image and the projection image under assumed conditions when X-rays of the virtual energy are irradiated; The steps include: identifying the optimal condition among the assumed conditions by evaluating the provisionally corrected image obtained by the dual energy resolution using a consistency index; The step includes generating a corrected image of the projection image when the above-mentioned optimal conditions are used, The correction method is characterized in that the consistency index is an index indicating the degree to which the Helgason-Ludwig condition is satisfied.

9. A correction program for correcting artifacts in CT images, The process of obtaining the measured projection image, A process to set a virtual energy different from the measured energy, which is the energy of the X-rays irradiated when the measured projection image was acquired, The process involves calculating the projection image under assumed conditions when X-rays of the virtual energy are irradiated, and performing dual energy decomposition using the measured projection image and the projection image under assumed conditions when X-rays of the virtual energy are irradiated. A process to identify the optimal condition among the assumed conditions by evaluating the provisionally corrected image obtained by the dual energy resolution using a consistency index, The computer is instructed to perform a process that generates a corrected image of the projection using the aforementioned optimal conditions. The correction program is characterized in that the consistency index is an index indicating the degree to which the Helgason-Ludwig condition is satisfied.

Images