Information processing device, information processing device operation method, radiographic system, and program

Abstract

To provide an information processing device capable of generating a material discrimination image from a radiation image even when a tube voltage of the radiation image changes while suppressing a calculation cost of machine learning.SOLUTION: An information processing device includes: an acquisition unit for acquiring a plurality of first radiation images of different radiation energies; and a generation unit for generating a plurality of first material discrimination images using the plurality of first radiation images as input data of a learned model. The learned model is a learned model obtained by using learning data with a set of a plurality of third radiation images of different radiation energies different from a radiation energy of a plurality of second radiation images, generated by using a plurality of second material discrimination images generated by energy subtraction processing using a plurality of second radiation images of different radiation energies as input data, and with the plurality of second material discrimination images as output data. The number of the plurality of third radiation images is greater than the number of the plurality of second radiation images.SELECTED DRAWING: Figure 10

Description

[Technical Field]

[0001] The present disclosure relates to an information processing device, an operation method of an information processing device, a radiation imaging system, and a program. [Background technology]

[0002] A commonly known imaging technique is time-multiplexed spectral imaging. In time-multiplexed spectral imaging, multiple radiation beams with different average energies are irradiated onto a subject over a short period of time, and the proportion of radiation beams with each average energy that pass through the subject and reach the radiation measurement surface is measured to distinguish the constituent materials of the subject. This type of time-multiplexed spectral imaging is also used to generate medical radiological images. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Publication No. 2023-120851 Summary of the Invention [Problem to be solved by the invention]

[0004] In the energy subtraction used in time-multiplexed spectral imaging, images are captured using multiple tube voltages, and a material-decomposed image is generated from the captured images. However, the combination of tube voltages may change. Therefore, there is a need for a process that can generate a material-decomposed image with high accuracy even when the tube voltage used to capture the radiation image changes.

[0005] Patent Document 1 proposes a method of creating a machine learning model for each tube voltage of a normal radiographic image and generating a material-decomposed image from a captured radiographic image using the learned model. However, creating a machine learning model for each tube voltage of a captured radiographic image increases the computational cost of machine learning.

[0006] An object of one embodiment of the present disclosure is to provide an information processing device that can generate a material decomposition image from a radiation image even when the tube voltage related to radiation imaging changes, while suppressing the computational cost of machine learning. [Means for solving the problem]

[0007] An information processing device according to an embodiment of the present disclosure includes an acquisition unit that acquires a plurality of first radiographic images relating to radiation energies that are different from each other, and a generation unit that generates a plurality of first material-decomposed images by using a set of the plurality of first radiographic images as input data for a trained model, wherein the trained model is a third radiographic image generated using a plurality of second material-decomposed images generated by energy subtraction processing using a plurality of second radiographic images relating to radiation energies that are different from each other from each other, and the trained model is obtained using training data in which a set of a plurality of third radiographic images relating to radiation energies that are different from and different from each other than the radiation energies of the plurality of second radiographic images is used as input data and the plurality of second material-decomposed images is used as output data, and the number of the plurality of third radiographic images is greater than the number of the plurality of second radiographic images. [Effects of the Invention]

[0008] According to one embodiment of the present disclosure, a material decomposition image can be generated from a radiation image even when the tube voltage for radiography changes, while suppressing the computational cost of machine learning. [Brief explanation of the drawings]

[0009] [Figure 1] 1 shows an example of the overall configuration of a radiation imaging system according to a first embodiment. [Figure 2] FIG. 2 is an equivalent circuit diagram of an example of a pixel of the radiation imaging apparatus according to the first embodiment. [Figure 3] 10 is a timing chart showing an example of a radiation imaging operation. [Figure 4] 10 is a timing chart showing an example of a radiation imaging operation. [Figure 5]FIG. 4 is a block diagram of a correction process according to the first embodiment. [Figure 6] FIG. 1 is a block diagram of signal processing for energy subtraction processing. [Figure 7] 1 shows an example of the configuration of a neural network related to a material decomposition image generation model. [Figure 8] 10 is a flowchart showing a process of generating input data for learning data of a material decomposition image generation model. [Figure 9A] 10 is a diagram illustrating an example of generation of input data for learning data of a material decomposition image generation model according to the first embodiment. [Figure 9B] 10 is a diagram illustrating an example of generation of input data for learning data of a material decomposition image generation model according to the first embodiment. [Figure 10] FIG. 10 is a block diagram showing an example of a generation process of a material decomposition image according to the first embodiment. [Figure 11] 4 is a flowchart showing a series of photographing processes according to the first embodiment. [Figure 12] 10 is a diagram illustrating an example of generation of input data for learning data of a material decomposition image generation model according to the second embodiment. [Figure 13] FIG. 10 is a block diagram illustrating an example of a process for generating a material decomposition image according to the second embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0010] Hereinafter, exemplary embodiments for carrying out the present disclosure will be described in detail with reference to the drawings. However, the dimensions, materials, shapes, and relative positions of components described in the following embodiments are arbitrary and can be changed depending on the configuration of an apparatus to which the present disclosure is applied or various conditions. In addition, the same reference numerals are used in the drawings to indicate identical or functionally similar elements.

[0011] In the following embodiments, the term "radiation" can include, in addition to X-rays, for example, alpha rays, beta rays, gamma rays, particle rays, and cosmic rays.

[0012] Energy subtraction processing refers to processing that calculates differences using images (energy images) related to different radiation energies to obtain, for example, material decomposition images of bone and soft tissue, contrast agent and water, etc., information on effective atomic number and surface density, etc. Note that energy subtraction processing may include, for example, correction processing such as offset correction processing as pre-processing, and image processing such as contrast adjustment processing as post-processing.

[0013] Furthermore, the term "energy subtraction image" can include, for example, a material decomposition image obtained using energy subtraction processing, an image showing effective atomic number and areal density, etc. Furthermore, the term "energy subtraction image" can also include a material decomposition image generated from virtual tube voltage images of different energies, etc. Furthermore, in the following embodiments, the term "energy subtraction image" can include an image inferred using a trained model obtained by training an image obtained using the above-described energy subtraction processing or an image generated from a virtual tube voltage image.

[0014] Furthermore, a virtual tube voltage image refers to a radiation image generated by inverse transformation using an energy subtraction image for an assumed tube voltage. Since the tube voltage applied to a radiation source during radiography corresponds to the radiation energy generated from the radiation source, any virtual tube voltage image relates to the radiation energy corresponding to the assumed tube voltage.

[0015] Furthermore, in the following, a machine learning model refers to a learning model based on a machine learning algorithm. Specific examples of machine learning algorithms include nearest neighbor algorithms, naive Bayes algorithms, decision trees, and support vector machines. Also included is deep learning, which uses a neural network to generate features and connection weighting coefficients for learning. Any available algorithm among the above can be used as appropriate and applied to the following embodiments. Furthermore, training data refers to training data, and is composed of a pair of input data and output data. Furthermore, supervised answer data refers to output data from training data (training data).

[0016] Furthermore, a trained model refers to a model that has been trained (learned) in advance using appropriate training data (learning data) for a machine learning model that follows an arbitrary machine learning algorithm, such as deep learning. However, although a trained model is obtained in advance using appropriate training data, it does not mean that no further learning is performed, and additional learning can also be performed. Additional learning can also be performed after the device is installed at the site of use. Note that obtaining output data from input data using a trained model is sometimes expressed as inference. Note that a material decomposition image generation model refers to a trained model that has been trained to output a material decomposition image when a radiological image is input.

[0017] [First embodiment] Currently, radiographic imaging devices using flat panel detectors (FPDs) made of semiconductor materials are widely used as imaging devices for radiological medical imaging diagnosis and non-destructive testing. For example, in medical imaging diagnosis, such radiographic imaging devices are used as digital imaging devices for capturing still images such as general radiography and video images such as fluoroscopy. Generally, FPDs use an integral sensor to detect radiation, which measures the total amount of charge generated by incident radiation quanta.

[0018] The following describes an information processing device used in a radiography system according to this embodiment and a method of operating the information processing device. Note that while this embodiment describes a medical radiography system in which the object to be examined is a human body, the technology according to this embodiment can also be applied to industrial radiography systems that use a subject as a base.

[0019] 1 shows an example of the overall configuration of a radiation imaging system according to this embodiment. The radiation imaging system of this embodiment includes a radiation generating device 101 including a radiation source, a radiation control device 102, a control device 103, a radiation imaging device 104, an input unit 150, and a display unit 120.

[0020] The radiation generating device 101 includes a radiation source such as a tube. The radiation generating device 101 generates radiation under the control of a radiation control device 102. The radiation control device 102 includes a control circuit, a processor, etc. Based on the control of the control device 103, the radiation control device 102 controls the radiation generating device 101 so that radiation is irradiated toward the subject Su and the radiation imaging device 104. More specifically, the radiation control device 102 can control imaging conditions such as the irradiation angle of radiation emitted by the radiation generating device 101, the radiation focal position, the tube voltage, and the tube current.

[0021] The radiation control device 102, the radiation imaging device 104, the input unit 150, and the display unit 120 are connected to the control device 103, and the control device 103 can control these. The control device 103 can perform, for example, various controls related to radiation imaging and image processing for spectral imaging.

[0022] The control device 103 is provided with an acquisition unit 131, a generation unit 132, a processing unit 133, a display control unit 134, and a storage unit 135. The acquisition unit 131 can acquire images captured by the radiation imaging device 104 and images generated by the generation unit 132. The acquisition unit 131 can also acquire various images from an external device (not shown) connected to the control device 103 via a network such as the Internet.

[0023] The generation unit 132 can generate a radiation image from an image (image information) captured by the radiation imaging device 104 and acquired by the acquisition unit 131. For example, the generation unit 132 can generate energy images (high-energy image and low-energy image) relating to different radiation energies from images captured by the radiation imaging device 104 by irradiating radiation of different energies. A method for generating the energy images will be described later.

[0024] The processing unit 133 generates an energy subtraction image based on energy images of different energies using a material decomposition image generation model. The material decomposition image generation model will be described later. The processing unit 133 can also perform image processing, analysis processing, and the like using the generated energy subtraction image, etc. Furthermore, the processing unit 133 can also function as an example of a learning unit that learns the material decomposition image generation model.

[0025] The display control unit 134 controls the display of the display unit 120 and can display, for example, information on the subject Su (examination object), information on radiography, various acquired and generated images, etc. on the display unit 120. The storage unit 135 can store, for example, information on the subject Su, information on radiography, and various acquired and generated images. The storage unit 135 can also include programs and the like for the control device 103 to perform various processes.

[0026] Here, the control device 103 can be configured by a computer provided with a processor and a memory. The control device 103 may be configured by a general computer or a computer dedicated to the radiation imaging system. The control device 103 may be, for example, a personal computer, such as a desktop PC, a notebook PC, or a tablet PC (portable information terminal). Furthermore, the control device 103 may be configured as a cloud-based computer in which some of the components are located in an external device.

[0027] Furthermore, each component of the control device 103 other than the storage unit 135 may be configured by a software module executed by a processor such as a CPU (Central Processing Unit) or an MPU (Micro Processing Unit). The processor may be, for example, a GPU (Graphical Processing Unit) or an FPGA (Field-Programmable Gate Array). Each component may be configured by a circuit that performs a specific function, such as an ASIC (Application Specific Integrated Circuit). The storage unit 135 may be configured by any storage medium, such as an optical disk such as a hard disk, or a memory.

[0028] The display unit 120 is configured with any monitor, and displays various information such as information about the subject Su, various images, and a mouse cursor in accordance with the operation of the input unit 150, under the control of the display control unit 134. The input unit 150 is an input device that issues instructions to the control device 103, and specifically includes a keyboard, a mouse, and the like. In this embodiment, the control device 103, the display unit 120, and the input unit 150 are configured separately, but these or some of them may be configured integrally. For example, the display unit 120 may be configured with a touch panel display, in which case the display unit 120 can also be used as the input unit 150.

[0029] The radiation imaging device 104 detects radiation that is irradiated from the radiation generation device 101 and passes through the subject Su, and captures a radiation image. The radiation imaging device 104 can be configured as, for example, an FPD. The radiation imaging device 104 is provided with a phosphor 141 that converts radiation into visible light, and a two-dimensional detector 142 that detects the visible light. The two-dimensional detector 142 includes a sensor in which pixels 20 that detect radiation quanta are arranged in an array of X columns and Y rows, and outputs image information according to the detected radiation dose.

[0030] An example of the pixel 20 will now be described with reference to Fig. 2. Fig. 2 shows an equivalent circuit diagram of an example of the pixel 20. The pixel 20 is provided with a photoelectric conversion element 201 and an output circuit unit 202. The photoelectric conversion element 201 may typically include a photodiode. The output circuit unit 202 is provided with an amplifier circuit unit 204, a clamp circuit unit 206, a sample-and-hold circuit unit 207, and a selection circuit unit 208.

[0031] The photoelectric conversion element 201 includes a charge accumulation section, which is connected to the gate of a MOS transistor 204a in the amplifier circuit section 204. The source of the MOS transistor 204a is connected to a current source 204c via a MOS transistor 204b. The MOS transistor 204a and the current source 204c form a source follower circuit. The MOS transistor 204b is an enable switch that turns on when an enable signal EN supplied to its gate becomes active level, putting the source follower circuit into an operating state.

[0032] In the example shown in FIG. 2, the charge storage section of the photoelectric conversion element 201 and the gate of the MOS transistor 204a form a common node. This node functions as a charge-voltage converter that converts the charge stored in the charge storage section of the photoelectric conversion element 201 into a voltage. A voltage V (=Q / C) appears in the charge-voltage converter, which is determined by the charge Q stored in the charge storage section and the capacitance C of the charge-voltage converter. The charge-voltage converter is connected to a reset potential Vres via a reset switch 203. When a reset signal PRES becomes active level, the reset switch 203 is turned on, and the potential of the charge-voltage converter is reset to the reset potential Vres.

[0033] The clamp circuit unit 206 clamps the noise output by the amplifier circuit unit 204 in response to the reset potential of the charge-voltage converter using a clamp capacitor 206a. In other words, the clamp circuit unit 206 is a circuit for canceling the noise described above from the signal output from the source follower circuit in response to the charge generated by photoelectric conversion in the photoelectric conversion element 201. This noise includes kTC noise generated during reset. Clamping is achieved by first setting the clamp signal PCL to an active level to turn on the MOS transistor 206b, and then setting the clamp signal PCL to an inactive level to turn off the MOS transistor 206b. The output side of the clamp capacitor 206a is connected to the gate of the MOS transistor 206c. The source of the MOS transistor 206c is connected to a current source 206e via a MOS transistor 206d. The MOS transistor 206c and the current source 206e form a source follower circuit. The MOS transistor 206d is an enable switch that turns on when an enable signal EN0 supplied to its gate becomes active, activating the source follower circuit.

[0034] A signal output from the clamp circuit unit 206 in response to charges generated by photoelectric conversion of the photoelectric conversion element 201 is written as an optical signal to the capacitor 207Sb via the switch 207Sa when the optical signal sampling signal TS becomes active. The signal output from the clamp circuit unit 206 when the MOS transistor 206b is turned on immediately after resetting the potential of the charge-voltage converter is a clamp voltage. The clamp voltage is written as noise to the capacitor 207Nb via the switch 207Na when the noise sampling signal TN becomes active. This noise includes an offset component of the clamp circuit unit 206. The switch 207Sa and the capacitor 207Sb form a signal sample-and-hold circuit 207S, and the switch 207Na and the capacitor 207Nb form a noise sample-and-hold circuit 207N. The sample-and-hold circuit unit 207 includes the signal sample-and-hold circuit 207S and the noise sample-and-hold circuit 207N.

[0035] When the driving circuit unit drives the row selection signal to an active level, the signal (optical signal) held in the capacitor 207Sb is output to the signal line 21S via the MOS transistor 208Sa and the row selection switch 208Sb. At the same time, the signal (noise) held in the capacitor 207Nb is output to the signal line 21N via the MOS transistor 208Na and the row selection switch 208Nb. The MOS transistor 208Sa forms a source follower circuit together with a constant current source (not shown) provided on the signal line 21S. Similarly, the MOS transistor 208Na forms a source follower circuit together with a constant current source (not shown) provided on the signal line 21N. The MOS transistor 208Sa and the row selection switch 208Sb form the signal selection circuit unit 208S, and the MOS transistor 208Na and the row selection switch 208Nb form the noise selection circuit unit 208N. The selection circuit section 208 includes a signal selection circuit section 208S and a noise selection circuit section 208N.

[0036] The pixel 20 may have an addition switch 209S that adds optical signals from multiple adjacent pixels 20. In addition mode, an addition mode signal ADD becomes active, and the addition switch 209S is turned on. As a result, the capacitances 207Sb of adjacent pixels 20 are connected to each other by the addition switch 209S, and the optical signals are averaged. Similarly, the pixel 20 may have an addition switch 209N that adds noises from multiple adjacent pixels 20. When the addition switch 209N is turned on, the capacitances 207Nb of adjacent pixels 20 are connected to each other by the addition switch 209N, and the noises are averaged. The adder 209 includes an addition switch 209S and an addition switch 209N.

[0037] The pixel 20 may also have a sensitivity change unit 205 for changing the sensitivity. The pixel 20 may include, for example, a first sensitivity change switch 205a, a second sensitivity change switch 205'a, and associated circuit elements. When the first change signal WIDE becomes active, the first sensitivity change switch 205a is turned on, and the capacitance value of the first additional capacitor 205b is added to the capacitance value of the charge-voltage conversion unit. This reduces the sensitivity of the pixel 20. When the second change signal WIDE2 becomes active, the second sensitivity change switch 205'a is turned on, and the capacitance value of the second additional capacitor 205'b is added to the capacitance value of the charge-voltage conversion unit. This further reduces the sensitivity of the pixel 20. By adding a function to reduce the sensitivity of the pixel 20 in this way, it becomes possible to receive a larger amount of light and widen the dynamic range. When the first change signal WIDE becomes active, the enable signal ENw may be turned on to cause the MOS transistor 204'a to operate as a source follower instead of the MOS transistor 204a.

[0038] The radiation imaging device 104 reads the output of the pixel circuit as described above and converts it into a digital value (image information) using an AD converter (not shown). The radiation imaging device 104 transfers the image information converted into a digital value to the control device 103. This allows the acquisition unit 131 of the control device 103 to acquire the radiation image.

[0039] Next, the radiographic imaging operation of the radiographic imaging system of this embodiment will be described with reference to FIGS. 3 and 4. FIGS. 3 and 4 show examples of various drive timings in imaging operations for performing energy subtraction processing in the radiographic imaging system of this embodiment. FIG. 3 shows an example of radiographic imaging operations when a relatively inexpensive tube that does not allow switching of tube voltage (energy) is used, while FIG. 4 shows an example of radiographic imaging operations when a tube that allows switching of tube voltage is used. The waveforms in the figures, with time on the horizontal axis, show the timing of X-ray exposure, synchronization signal, resetting of the photoelectric conversion element 201, driving of the sample-and-hold circuit unit 207, and readout of an image from the signal line 21. Note that in FIGS. 3 and 4, the waveform for X-rays represents the tube voltage. Also, while black and white areas are provided for X-rays, this is simply done to make it easier to distinguish the timing.

[0040] First, the example shown in FIG. 3 will be described. In this example, X-rays are emitted after the photoelectric conversion element 201 is reset. Ideally, the X-ray tube voltage is a rectangular wave, but it takes a finite amount of time for the tube voltage to rise and fall. In particular, when the exposure time is short with pulsed X-rays, the tube voltage can no longer be considered a rectangular wave, and takes on a waveform like that shown in the X-ray waveform in FIG. 3. For this reason, the energy of the X-rays differs during the rise, stable, and fall periods of the X-rays.

[0041] Therefore, after X-rays 301 in the rising phase are irradiated, sampling is performed by the noise sample-and-hold circuit 207N, and further, after X-rays 302 in the stable phase are irradiated, sampling is performed by the signal sample-and-hold circuit 207S. Thereafter, the difference between the signal on signal line 21N and the signal on signal line 21S is read out as an image. At this time, the signal (R1) of X-rays 301 in the rising phase is held in the noise sample-and-hold circuit 207N, and the sum of the signal (R1) of X-rays 301 in the rising phase and the signal (B) of X-rays 302 in the stable phase is held in the signal sample-and-hold circuit 207S. Therefore, an image 304 corresponding to the signal (B) of X-rays 302 in the stable phase is read out as the difference between the signal on signal line 21N and the signal on signal line 21S.

[0042] Next, after the irradiation of X-rays 303 in the falling phase and the readout of image 304 are completed, sampling is again performed by signal sample-and-hold circuit 207S. Thereafter, photoelectric conversion element 201 is reset, and sampling is again performed by noise sample-and-hold circuit 207N, and the difference between the signal on signal line 21N and the signal on signal line 21S is read out as an image. At this time, the noise sample-and-hold circuit 207N holds a signal in a state where no X-rays are irradiated. Furthermore, the signal sample-and-hold circuit 207S holds the sum of the signal (R1) of X-rays 301 in the rising phase, the signal (B) of X-rays 302 in the stable phase, and the signal (R2) of X-rays 303 in the falling phase. Therefore, image 306 corresponding to the sum of the signal (R1) of X-rays 301 in the rising phase, the signal (B) of X-rays 302 in the stable phase, and the signal (R2) of X-rays 303 in the falling phase is read out as the difference between the signal on signal line 21N and the signal on signal line 21S. Thereafter, by calculating the difference between image 306 and image 304, image 305 corresponding to the sum of the signal (R1) of X-ray 301 in the rising phase and the signal (R2) of X-ray 303 in the falling phase is obtained.

[0043] The timing for resetting the sample-and-hold circuit unit 207 and the photoelectric conversion element 201 is determined using a synchronization signal 307 indicating that X-ray exposure has started from the radiation generation device 101. A method for detecting the start of X-ray exposure may include measuring the tube current of the radiation generation device 101 and determining whether the current value exceeds a predetermined threshold. Alternatively, a configuration may be used in which, after resetting of the photoelectric conversion element 201 is completed, signals from the pixels 20 are repeatedly read out and whether the pixel values ​​exceed a predetermined threshold is determined. Furthermore, a configuration may be used in which an X-ray detector other than the two-dimensional detector 106 is built into the radiation imaging device 104 and whether the measured value exceeds a predetermined threshold is determined. In either case, sampling of the signal sample-and-hold circuit 207S, sampling of the noise sample-and-hold circuit 207N, and resetting of the photoelectric conversion element 201 are performed after a predetermined time has elapsed since the input of the synchronization signal 307.

[0044] In this way, it is possible to obtain an image 304 corresponding to the stable period of the pulsed X-rays and an image 305 corresponding to the sum of the rising and falling periods of the pulsed X-rays. Because the energy of the X-rays irradiated when generating these two images is different, energy subtraction processing can be performed by performing calculations between the two images.

[0045] Next, an example of radiographic operation when using a tube with switchable tube voltage will be described with reference to Fig. 4. This example differs from the example shown in Fig. 3 in that the X-ray tube voltage is actively switched.

[0046] In this example, first, the photoelectric conversion element 201 is reset, and then low-energy X-rays 401 are irradiated. Then, sampling is performed by the noise sample-and-hold circuit 207N, and the tube voltage is switched to irradiate high-energy X-rays 402. After the high-energy X-rays 402 are irradiated, sampling is performed by the signal sample-and-hold circuit 207S. Then, the tube voltage is switched to irradiate low-energy X-rays 403. Furthermore, the difference between the signals on the signal line 21N and the signal line 21S is read out as an image. At this time, the noise sample-and-hold circuit 207N holds the signal (R1) of the low-energy X-rays 401, and the signal sample-and-hold circuit 207S holds the sum of the signal (R1) of the low-energy X-rays 401 and the signal (B) of the high-energy X-rays 402. Therefore, an image 404 corresponding to the signal (B) of the high-energy X-ray 402 is read out as the difference between the signal on the signal line 21N and the signal on the signal line 21S.

[0047] Next, after the irradiation of low-energy X-rays 403 and the readout of image 404 are completed, sampling is again performed by signal sample-and-hold circuit 207S. Thereafter, photoelectric conversion element 201 is reset, and sampling is again performed by noise sample-and-hold circuit 207N, and the difference between the signal on signal line 21N and the signal on signal line 21S is read out as an image. At this time, the noise sample-and-hold circuit 207N holds a signal in a state where no X-rays are irradiated. Furthermore, the signal sample-and-hold circuit 207S holds the sum of the signal (R1) of low-energy X-rays 401, the signal (B) of high-energy X-rays 402, and the signal (R2) of low-energy X-rays 403. Therefore, an image 406 corresponding to the sum of the signal (R1) of low-energy X-rays 401, the signal (B) of high-energy X-rays 402, and the signal (R2) of low-energy X-rays 403 is read out as the difference between the signal on signal line 21N and the signal on signal line 21S. Thereafter, by calculating the difference between image 406 and image 404, image 405 corresponding to the sum of the signal (R1) of low-energy X-ray 401 and the signal (R2) of low-energy X-ray 403 is obtained.

[0048] The synchronization signal 407 is the same as in the example shown in Fig. 3. In this way, by acquiring images while actively switching the tube voltage, it is possible to make the energy difference between low-energy and high-energy images larger than in the method of Fig. 3. The method of acquiring low-energy and high-energy images is not limited to this. For example, an FPD including a stacked sensor formed of two types of materials (phosphors) with different X-ray absorption rates may be used as the radiation imaging device 104, and low-energy and high-energy images may be acquired based on the output from the sensor of each layer obtained by a single exposure.

[0049] Next, the energy subtraction processing method will be described. The energy subtraction processing in this embodiment includes, in addition to the signal processing of the energy subtraction processing, correction processing as pre-processing and image processing as post-processing.

[0050] First, the correction process, which is pre-processing, will be described with reference to Fig. 5. Fig. 5 is a block diagram of the correction process according to this embodiment. Note that in the following, this embodiment will be described assuming that a radiation imaging operation is performed according to the example shown in Fig. 3.

[0051] 3 without irradiating the radiation imaging device 104 with X-rays, and the acquisition unit 131 acquires the captured images. At this time, two images corresponding to images 304 and 306 are read out, with the first image (image 304) being image F_Odd and the second image (image 306) being image F_Even. Images F_Odd and F_Even are images corresponding to fixed pattern noise (FPN) of the radiation imaging device 104.

[0052] Next, X-rays are irradiated to the radiation imaging device 104 in the absence of a subject, and imaging is performed by the driving shown in Fig. 3, and the acquisition unit 131 acquires the captured images. At this time, two images corresponding to images 304 and 306 are read out, with the first image (image 304) being image W_Odd and the second image (image 306) being image W_Even. Images W_Odd and W_Even are images corresponding to the sum of signals due to the FPN of the radiation imaging device 104 and X-rays.

[0053] Therefore, by subtracting image F_Odd from image W_Odd and image F_Even from image W_Even, images WF_Odd and WF_Even from which the FPN of the radiation imaging apparatus 104 has been removed can be obtained. In this embodiment, such correction processing is called offset correction.

[0054] Image WF_Odd is an image corresponding to X-rays 302 in the stable period, and image WF_Even is an image corresponding to the sum of X-rays 301 in the rising period, X-rays 302 in the stable period, and X-rays 303 in the falling period. Therefore, by subtracting image WF_Odd from image WF_Even, an image corresponding to the sum of X-rays 301 in the rising period and X-rays 303 in the falling period is obtained. The energies of X-rays 301 in the rising period and X-rays 303 in the falling period are lower than the energy of X-rays 302 in the stable period. Therefore, by subtracting WF_Odd from WF_Even, a low-energy image W_Low in the absence of an object is obtained. Furthermore, a high-energy image W_High in the absence of an object is obtained from WF_Odd. In this embodiment, this correction process is called color correction.

[0055] Next, with a subject present, the radiographic imaging device 104 is irradiated with X-rays to perform imaging using the drive shown in Fig. 3, and the acquisition unit 131 acquires the captured images. At this time, two images corresponding to images 304 and 306 are read out, with the first image (image 304) designated as image X_Odd and the second image (image 306) designated as image X_Even. The generation unit 132 performs offset correction and color correction on these images in the same manner as when there is no subject, thereby obtaining a low-energy image X_Low when there is a subject and a high-energy image X_High when there is a subject.

[0056] Here, if the thickness of the subject is d, the linear attenuation coefficient of the subject is μ, the output of pixel 20 when there is no subject is I0, and the output of pixel 20 when there is a subject is I, the following equation (1) holds.

number

number

[0057] Therefore, by dividing the low-energy image X_Low when there is an object by the low-energy image W_Low when there is no object, the image of the attenuation rate at low energy (low-energy image Im L ) is obtained. Similarly, by dividing the high-energy image X_High when there is an object by the high-energy image W_High when there is no object, an image of the attenuation rate at high energy (high-energy image Im H In this embodiment, such a correction process is called gain correction. In this embodiment, the generation unit 132 performs correction processes including the offset correction, color correction, and gain correction described above to generate a low-energy image Im L and high-energy image Im H can be generated and obtained.

[0058] Next, before describing the energy subtraction process using the material decomposition image generation model according to this embodiment, the signal processing of the energy subtraction process will be described with reference to Fig. 6. Fig. 6 shows a block diagram of the signal processing of the energy subtraction process. In the signal processing of the energy subtraction process, the low-energy image Im obtained by the correction process described with reference to Fig. 5 is L and high-energy images H From the bone thickness image (bone image Im B ) and soft tissue thickness image (soft tissue image Im S ) is found.

[0059] First, let E be the energy of the X-ray photon, N(E) be the number of photons at energy E, B be the thickness of the bone, and S be the thickness of the soft tissue. Let μ be the linear attenuation coefficient of the bone at energy E. B (E), the linear attenuation coefficient of soft tissue at energy E is μ S (E), if the attenuation ratio is I / I0, the following equation (3) holds.

number

[0060] The number of photons N(E) at energy E is the X-ray spectrum. The X-ray spectrum can be obtained by simulation or actual measurement for the tube voltage. Also, the linear attenuation coefficient μ of bone at energy E is B (E) and the linear attenuation coefficient μ of soft tissue at energy E S (E) can be obtained from databases such as those of the National Institute of Standards and Technology (NIST). Therefore, by using equation (3), it is possible to calculate the attenuation ratio I / I0 for any bone thickness B, soft tissue thickness S, and X-ray spectrum N(E).

[0061] Here, the spectrum of low energy X-rays is N L (E) High-energy X-ray spectrum H If (E), the following equation (4) holds.

number

[0062] Here, we will explain the case where the Newton-Raphson method, which is a type of iterative solution, is used as a typical method for solving nonlinear simultaneous equations. First, let us define the number of iterations of the Newton-Raphson method as m, and the bone thickness after the mth iteration as B. m , the thickness of the soft tissue after the mth iteration is S m Then, the attenuation rate of high energy after the mth iteration, H m , and the low-energy attenuation rate L after the mth iteration m is expressed by the following equation (5).

number

number

[0063] In this case, the bone thickness B after the m+1th iterationm+1 and soft tissue thickness S m+1 is expressed by the following equation (7) using the attenuation rate of high energy H and the attenuation rate of low energy L.

number

number

number

[0064] By repeating this calculation, the attenuation rate of high energy after the mth iteration, H m The difference between the measured high-energy attenuation rate H and the measured high-energy attenuation rate L approaches 0. The same is true for the low-energy attenuation rate L. This results in the bone thickness B after the m-th iteration. m converges to the bone thickness B, and the m-th soft tissue thickness S m converges to the thickness S of the soft tissue. In this way, the nonlinear simultaneous equations shown in equation (4) can be solved. Therefore, by calculating equation (4) for all pixels, the low-energy image Im L and high-energy images H From the bone image B and soft tissue imaging S can be obtained.

[0065] For the sake of simplicity, the bone thickness B and the soft tissue thickness S are calculated by energy subtraction processing here, but this embodiment is not limited to this. For example, the thickness of water and the thickness of the contrast agent may be calculated by energy subtraction processing. In this case, the linear attenuation coefficient of water at energy E and the linear attenuation coefficient of the contrast agent at energy E may also be obtained from a database such as NIST. The thickness of any two types of material can be calculated by energy subtraction processing.

[0066] In the above description, the nonlinear simultaneous equations are solved using the Newton-Raphson method. However, the method for solving the nonlinear simultaneous equations is not limited to this. For example, iterative methods such as the least squares method and bisection method may also be used. Furthermore, the method for calculating the bone thickness B and the soft tissue thickness S is not limited to solving the nonlinear simultaneous equations using an iterative method. For example, the bone thickness B and the soft tissue thickness S for various combinations of high-energy attenuation rate H and low-energy attenuation rate L may be calculated in advance to generate a table, and the bone thickness B and the soft tissue thickness S may be calculated quickly by referring to the table.

[0067] In the above explanation, the high-energy image Im H and low-energy image Im L Energy subtraction processing was performed on the bone image Im B and soft tissue imaging S In contrast to this, the following describes the energy subtraction process using a material decomposition image generation model according to this embodiment. In this embodiment, the processing unit 133 generates a high-energy image Im using the material decomposition model. H and low-energy image Im L Based on the material decomposition image (bone image Im B and soft tissue imaging S )

[0068] (Material decomposition image generation model) The material decomposition image generation model according to this embodiment will be described below with reference to Fig. 7. In this embodiment, the material decomposition image generation model is stored in the storage unit 135 and used for processing by the processing unit 133, but the material decomposition image generation model may also be provided in an external device (not shown) connected to the control device 103.

[0069] The material decomposition image generation model according to this embodiment is a trained model obtained by performing training (learning) on ​​a machine learning algorithm. In this embodiment, training of the machine learning model according to the machine learning algorithm uses learning data configured as a group of pairs of input data, which is a tube voltage image expected to be processed, and output data, which is a material decomposition image corresponding to the input data.

[0070] Here, we will briefly explain a general trained model. A trained model is a machine learning model that has been trained (learned) in advance using appropriate training data for an arbitrary machine learning algorithm. The training data consists of one or more pairs of input data and output data (ground truth data). The format and combination of the input data and output data in the pairs that make up the training data may be suitable for a desired configuration, such as both being images, one being an image and the other being a numerical value, or one being composed of a group of multiple images and the other being a character string.

[0071] Specifically, for example, the training data may be a set of pairs of low-energy and high-energy images, and bone and soft tissue images generated by energy subtraction processing (hereinafter referred to as first training data).

[0072] In this case, when input data is input to the trained model, output data is output according to the design of the trained model. The trained model outputs output data that is likely to correspond to the input data, for example, according to trends trained using the training data. The trained model can also output the likelihood (reliability, probability) of each type of output data corresponding to the input data as a numerical value, for example, according to trends trained using the training data.

[0073] Specifically, for example, when a pair of a low-energy image and a high-energy image acquired by normal radiography is input to a machine learning model trained with first learning data, the machine learning model outputs material-decomposed images equivalent to bone images and soft tissue images generated by energy subtraction processing. Note that, from the viewpoint of maintaining quality, the machine learning model can be configured not to use its own output data as learning data.

[0074] Furthermore, the machine learning algorithm includes deep learning techniques such as convolutional neural networks (CNNs). In deep learning techniques, different parameter settings for layers and nodes constituting a neural network may affect the degree to which trends trained using training data can be reproduced in output data. For example, in a deep learning machine learning model using first training data, more appropriate parameter settings may result in more accurate material decomposition images being output.

[0075] Specifically, parameters in a CNN can include, for example, the filter kernel size, the number of filters, the stride value, and the dilation value set for the convolution layer, as well as the number of nodes output by the fully connected layer. The parameter set and the number of training epochs can be set to values ​​suitable for the usage of the trained model based on the training data. For example, the parameter set and the number of epochs can be set based on the training data to output a material decomposition image with higher accuracy.

[0076] Here is an example of a method for determining such parameter sets and the number of epochs. First, 70% of the pairs constituting the training data are used for training, and the remaining 30% are randomly set as those for evaluation. Next, the training pairs are used to train the machine learning model, and at the end of each training epoch, a training evaluation value is calculated using the evaluation pairs. The training evaluation value is, for example, the average value of a group of values ​​obtained by evaluating, using a loss function, the output when input data constituting each pair is input to the machine learning model being trained, and the output data corresponding to the input data. Finally, the parameter set and the number of epochs that result in the smallest training evaluation value are determined as the parameter set and the number of epochs for the machine learning model. Note that by dividing the pairs constituting the training data into those for training and those for evaluation and determining the number of epochs in this way, it is possible to prevent the machine learning model from overfitting the training pairs.

[0077] Here, the material decomposition image generation model according to this embodiment is configured as a module that outputs a pair of bone and soft tissue images based on an input pair of a low-energy image and a high-energy image.

[0078] In the image processing method constituting the energy subtraction processing executed by the processing unit 133 in this embodiment, processing using various machine learning algorithms such as deep learning is performed. Note that the image processing method may perform any existing processing such as various image filter processing, matching processing using a database of material decomposition images corresponding to similar images, and knowledge-based image processing in addition to processing using the machine learning algorithm.

[0079] An example of the configuration of a CNN related to a material decomposition image generation model according to this embodiment will be described below with reference to Fig. 7. Fig. 7 shows an example of the configuration of a material decomposition image generation model. The configuration shown in Fig. 7 is composed of a plurality of layer groups that are responsible for processing and outputting a group of input values. As shown in Fig. 7, the types of layers included in this configuration include a convolution layer, a downsampling layer, an upsampling layer, and a merging layer.

[0080] The convolution layer is a layer that performs convolution processing on a group of input values ​​according to parameters such as the set filter kernel size, the number of filters, the stride value, the dilation value, etc. Note that the number of dimensions of the filter kernel size may also be changed depending on the number of dimensions of the input image.

[0081] The downsampling layer is a layer that performs processing to reduce the number of output value groups to be less than the number of input value groups by thinning out or combining input value groups. Specifically, such processing includes, for example, max pooling processing.

[0082] An upsampling layer is a layer that performs processing to make the number of output values ​​greater than the number of input values ​​by duplicating input values ​​or adding values ​​interpolated from the input values, such as linear interpolation.

[0083] A synthesis layer is a layer that inputs a group of values, such as a group of output values ​​from a layer or a group of pixel values ​​that make up an image, from multiple sources and performs processing to combine them by concatenating or adding them.

[0084] In this configuration, a group of pixel values ​​constituting the input image Im710 is output through a convolution processing block and then synthesized in a synthesis layer with a group of pixel values ​​constituting the input image Im710. The synthesized group of pixel values ​​is then shaped into a material-decomposed image Im720 in the final convolution layer.

[0085] It should be noted that different parameter settings for the layers and nodes that make up the neural network may affect the degree to which trends trained from learning data can be reproduced during inference. In other words, in many cases, appropriate parameters differ depending on the implementation form, so parameters can be changed to preferred values ​​as needed.

[0086] In addition to changing the parameters as described above, changing the configuration of the CNN can sometimes result in better CNN characteristics, such as outputting images with more accurate material decomposition, shorter processing times, or shorter training times for machine learning models.

[0087] The CNN used in this embodiment is a U-net type machine learning model having an encoder function consisting of multiple layers including multiple downsampling layers and a decoder function consisting of multiple layers including multiple upsampling layers. That is, the CNN has a U-shaped structure having an encoder function and a decoder function. The U-net type machine learning model is configured (for example, using skip connections) so that position information (spatial information) obscured in multiple layers configured as encoders can be used in layers of the same dimension (layers corresponding to each other) in multiple layers configured as decoders.

[0088] Although not shown in the figure, as an example of a modification to the CNN configuration, for example, a batch normalization layer or an activation layer using a rectifier linear unit may be incorporated after the convolution layer.

[0089] Here, a GPU can perform efficient calculations by processing a larger amount of data in parallel. Therefore, when performing learning multiple times using a machine learning algorithm such as deep learning, it is effective to use a GPU for processing. Therefore, in this embodiment, a GPU is used in addition to a CPU for processing by the processing unit 133, which functions as an example of a learning unit. Specifically, when executing a learning program including a learning model, the CPU and GPU work together to perform calculations to perform learning. Note that calculations may be performed only by the CPU or the GPU in the processing of the learning unit. Furthermore, the energy subtraction processing according to this embodiment may also be implemented using a GPU, as with the learning unit. Note that if the trained model is provided in an external device, the processing unit 133 does not need to function as a learning unit.

[0090] The learning unit may also include an error detection unit and an update unit (not shown). The error detection unit obtains the error between correct data and output data output from the output layer of the neural network in response to input data input to the input layer. The error detection unit may use a loss function to calculate the error between the output data from the neural network and correct data. The update unit updates the connection weighting coefficients between the nodes of the neural network based on the error obtained by the error detection unit so as to reduce the error. This update unit updates the connection weighting coefficients using, for example, an error backpropagation method. The error backpropagation method is a technique for adjusting the connection weighting coefficients between the nodes of each neural network so as to reduce the error.

[0091] (Training data for material decomposition image generation model) Next, the learning data of the material decomposition image generation model according to this embodiment will be described. In this embodiment, input data for the learning data of the material decomposition image generation model is generated according to the flow shown in FIG.

[0092] In step S801, the acquisition unit 131 acquires a low-energy image Im L and high-energy image Im H The acquisition unit 131 acquires the high-energy image Im stored in the storage unit 135. H and low-energy image Im L Alternatively, the high-energy image Im may be acquired from an external device connected to the control device 103. H and low-energy image Im L The acquisition unit 131 may also acquire a low-energy image Im generated from an image captured by the radiation imaging device 104. L and high-energy image Im H may be obtained.

[0093] In step S802, the processing unit 133 generates a low-energy image Im by signal processing of the energy subtraction processing shown in FIG. L and high-energy images H From the bone thickness image (bone image Im B ) and soft tissue thickness image (soft tissue image Im S )

[0094] In step S803, the generation unit 132 generates an energy image (virtual tube voltage image Iv) at each tube voltage through simulation. The virtual tube voltage image Iv is obtained as shown in equation (10) by substituting the pixel value of the bone thickness image for the bone thickness B in equation (3) and the pixel value of the soft tissue thickness image for the soft tissue thickness S for each pixel. Note that, hereinafter, the process of generating a virtual tube voltage image from a material decomposition image is referred to as the inverse transformation of the energy subtraction process. FIG. 9A shows an example of the generated virtual tube voltage image Iv in this embodiment. In this embodiment, the generation unit 132 generates multiple pairs of high and low tube voltage images of the virtual tube voltage image Iv, as shown in FIG. 9A.

number

[0095] The tube voltage for the virtual tube voltage image Iv obtained by the simulation may be a plurality of predetermined voltage values. H _200 and 150kV Image Im L _150 pair 1 and 200kV image Im H _200 and 100kV Image Im L Pair 2 of 100 and 200kV image Im H _200 and 150kV Image Im L 9A shows pair 3 of 150 kV image Im as an example. H _150 and 100kV Image Im L _100 pair 4 and 150kV image Im H _150 and 65kV Image Im L This shows pair 5 of _65.

[0096] The tube voltage for the virtual tube voltage image Iv is calculated based on the low-energy image Im obtained in step S801. L and high-energy image Im H For example, the tube voltage for the virtual tube voltage image Iv may be determined using the tube voltage for the reference low-energy image Im obtained in step S801.L and high-energy image Im H For example, the tube voltage for the virtual tube voltage image Iv may be set to any voltage value within a range of ±5 kV from the tube voltage for the low-energy image Im obtained in step S801, which serves as a reference. L and high-energy image Im H The voltage may be increased or decreased by an arbitrary increment (increment), such as 5 kV or 1 kV, from the tube voltage related to the ion beam.

[0097] Here, FIG. 9B shows another example of a 102 kV image Im H _102 and 65kV Image Im L Pair 1 of _65 and 101kV Image Im H _101 and 65kV Image Im L Pair 2 of _65 and 100kV image Im H _100 and 65kV Image Im L 9B shows pair 3 of _65. H _99 and 65kV Image Im L Pair 4 of _65 and 100kV image Im H _100 and 64kV Image Im L This shows pair 5 of _64.

[0098] In the example of the pair of generated virtual tube voltage images Iv shown in FIGS. 9A and 9B, images of the same tube voltage, for example, 200 kV images Im H _200, etc. are included in multiple pairs. In contrast, only one virtual tube voltage image Iv of the same tube voltage may be generated. In this case, the processing unit 133 may generate a pair of virtual tube voltage images Iv to be used as learning data by appropriately combining the generated virtual tube voltage images Iv. Also, in the examples shown in FIGS. 9A and 9B, the processing unit 133 may generate a pair of virtual tube voltage images Iv to be used as learning data by appropriately combining already generated virtual tube voltage images Iv.

[0099] In step S804, the processing unit 133 uses the pair of virtual tube voltage images Iv generated in step S803 as input data, and calculates the bone image Im generated in step S802. B and soft tissue imaging Im S The machine learning model is trained using training data in which the low-energy image Im obtained in step S801 is used as output data. Through such training, the machine learning model can learn the relationship between various pairs of tube voltage images, which are input data, and material decomposition images, which are output data. Note that the input data used for training the machine learning model may be, for example, a virtual tube voltage image Iv with fine increments of about 1 kV as shown in FIG. 9B. In addition, the processing unit 133 trains the low-energy image Im obtained in step S801. L and high-energy image Im H The pair of bone images Im generated in step S802 may also be used as input data for the training data. B and soft tissue imaging Im S may be used.

[0100] By using the material decomposition image generation model that has been trained in this way, the processing unit 133 can acquire material decomposition images with high accuracy using pairs of low-energy images and high-energy images obtained at various tube voltages. Therefore, even if the tube voltage of the pair of low-energy images and high-energy images used as input data changes, the processing unit 133 can output material decomposition images with high accuracy.

[0101] Hereinafter, with reference to Fig. 10(a), the learning data of the material decomposition image generation model used by the processing unit 133 according to this embodiment will be described in more detail. In this embodiment, as shown in Fig. 10(a), the processing unit 133 uses a pair of virtual tube voltage images Iv of different tube voltages generated in step S803 as input data for the learning data. In addition, the processing unit 133 uses the bone image Im generated in step S802 as input data for the learning data. B and soft tissue imaging Sis set as the output data of the learning data. Note that even when virtual tube voltage images Iv with fine increments of about 1 kV tube voltage are generated in step S803, the processing unit 133 sets pairs of virtual tube voltage images Iv with different tube voltages generated in step S803 as input data of the learning data, as shown in FIG.

[0102] By using such training data, the burden of imaging to obtain input data for the training data can be reduced, while input data (high-energy images Im) corresponding to different tube voltages can be obtained. H and low energy image Im L ) combinations can be easily constructed. Furthermore, by constructing a trained model in this way, the many nonlinear calculation processes involved in the energy subtraction process can be included in inference using machine learning algorithms such as deep learning.

[0103] The material decomposition image generation model according to this embodiment can have two input channels and two output channels according to the input data and output data. However, the number of channels of the input data and output data of the material decomposition image generation model can be set appropriately. For example, the input data may be either a high-energy image or a low-energy image, and there may be one input channel. In addition, the output data may be a bone image Im B and soft tissue imaging S In these cases, the input data of the training data may be either a high-energy image or a low-energy image, and the output data may be a bone image Im B and soft tissue imaging S Furthermore, the number of channels may be three or more. For example, three or more radiographic images relating to different radiation energies may be used as input data, and the number of input channels may correspond to the number of the radiographic images.

[0104] Furthermore, the processing unit 133 according to this embodiment converts the bone image Im output from the material decomposition image generation model into a Band soft tissue imaging S Here, the image processing in this embodiment may be processing for performing any calculation on the energy subtraction image. B and soft tissue imaging S As image processing for the bone image Im, for example, adjustment processing such as contrast adjustment and gradation adjustment can be performed. Furthermore, as image processing, the processing unit 133 can apply, for example, a filter in the time direction such as a recursive filter or a filter in the space direction such as a Gaussian filter to the bone image Im. B and soft tissue imaging S In addition, the bone image Im may be used as the output data of the training data of the material decomposition image generation model. B and soft tissue imaging S Alternatively, a final image to be displayed after post-processing such as contrast correction may be used.

[0105] Furthermore, the processing unit 133 performs image processing on the bone image Im B and soft tissue imaging S In this case, the processing unit 133 uses the material decomposition image generation model to generate a low-energy image Im captured before the injection of the contrast agent. LM and high-energy images HM Bone thickness mask image from Im BM and the soft tissue thickness mask image Im SM Furthermore, the processing unit 133 uses the material decomposition image generation model to generate a low-energy image Im LL and high-energy images HL Live image of bone thickness from BL and live imaging of soft tissue thickness. SL Then, the processing unit 133 calculates the live image Im of the bone thickness. BL Bone thickness mask image from Im BM By subtracting the , a DSA image of the bone can be generated, and a live image of the soft tissue thickness can be obtained. SLSoft tissue thickness mask image Im SM By drawing the line, DSA images of soft tissue can be generated.

[0106] Next, a series of photographing processes according to this embodiment will be described with reference to Fig. 11. Fig. 11 is a flowchart showing a series of photographing processes according to this embodiment.

[0107] In step S1101, the operator inputs imaging conditions for radiography, such as tube voltage, tube current, and irradiation time, to the input unit 150. The control device 103 sets the imaging conditions in response to the operation of the operator.

[0108] In step S1102, radiation imaging is performed based on imaging conditions set in response to an operation by an operator. Specifically, the radiation control device 102 controls the radiation generation device 101 based on imaging conditions set by the control device 103. The radiation generation device 101 irradiates radiation toward the subject Su and the radiation imaging device 104 based on the control of the radiation control device 102. The radiation imaging device 104 detects radiation that has passed through the subject Su and transmits image information to the control device 103. The acquisition unit 131 of the control device 103 acquires the image information transmitted from the radiation imaging device 104.

[0109] In step S1103, the generation unit 132 performs correction processing including the offset correction, color correction, and gain correction described above based on the image information acquired by the acquisition unit 131, and generates a high-energy image Im H and low-energy image Im L The images W_odd and W_Even when no subject is placed, and the images F_Odd and F_Even when no X-rays are irradiated, which are used in the correction process, may be captured prior to capturing the image of the subject Su in step S1102. These images may also be captured under given imaging conditions and stored in the storage unit 135 in advance.

[0110] In step S1104, the processing unit 133 generates the high-energy image Im using the material decomposition image generation model, which is a trained model. H and low-energy image Im L Based on this, the bone image Im is an energy subtraction image. B and soft tissue imaging Im S Specifically, the processing unit 133 generates a high-energy image Im H and low-energy image Im L As the input data of the material decomposition image generation model, the bone image Im B and soft tissue imaging Im S Obtain and generate.

[0111] In step S1105, the processing unit 133 converts the bone image Im, which is the energy subtraction image generated in step S1104, into B and soft tissue imaging Im S The processing unit 133 performs image processing such as contrast adjustment and image size adjustment on the bone image Im. Note that any known method may be used as the adjustment method. In addition, the processing unit 133 applies a filter in the time direction such as a recursive filter or a filter in the space direction such as a Gaussian filter to the bone image Im. B and soft tissue imaging S The processing unit 133 may also apply the bone image Im B and soft tissue imaging S may be used to generate DSA images of bone and soft tissue.

[0112] In step S1106, the display control unit 134 displays the bone image Im B and soft tissue imaging Im S etc. are displayed on the display unit 120. Note that the display control unit 134 B and soft tissue imaging Im S may be displayed side by side on the display unit 120, or may be displayed by switching between them. Also, if a DSA image is generated in step S1105, the display control unit 134 can cause the display unit 120 to display the generated DSA image.

[0113] When the process in step S1106 is completed, a series of imaging processes according to this embodiment is completed. In this embodiment, the acquisition unit 131 acquires image information from the radiation imaging device 104, the generation unit 132 performs correction processing, and the acquisition unit 131 outputs the high-energy image Im generated by the generation unit 132. H and low-energy image Im L In response to this, the acquisition unit 131 acquires the high-energy image Im stored in the storage unit 135. H and low-energy image Im L Alternatively, the high-energy image Im may be acquired from an external device connected to the control device 103. H and low-energy image Im L The acquisition unit 131 may also acquire image information captured of the subject Su and image information used for correction processing from the storage unit 135 or an external device. In addition, although image processing was performed in step S1105, in step S1106, the bone image Im that has not been subjected to image processing may be acquired. B and soft tissue imaging Im S may be displayed as is.

[0114] As described above, the radiation imaging system according to this embodiment includes the radiation generating apparatus 101, the radiation imaging apparatus 104, and the control apparatus 103. The radiation generating apparatus 101 functions as an example of a radiation generating apparatus that generates radiation. The radiation imaging apparatus 104 functions as an example of a radiation imaging apparatus that detects radiation.

[0115] The control device 103 functions as an example of an information processing device including an acquisition unit 131 and a processing unit 133. The acquisition unit 131 functions as an example of an acquisition unit that acquires a plurality of first radiographic images related to different radiation energies. The processing unit 133 functions as an example of a generation unit that generates a plurality of first material-decomposed images by using a set of a plurality of first radiographic images as input data for the trained model. Here, the trained model is a third radiographic image generated using a plurality of second material-decomposed images generated by energy subtraction processing using a plurality of second radiographic images related to different radiation energies, and is a trained model obtained using training data in which a set of a plurality of third radiographic images related to radiation energies different from those of the plurality of second radiographic images is used as input data and a plurality of second material-decomposed images is used as output data. Here, the number of the plurality of third radiographic images is greater than the number of the plurality of second radiographic images.

[0116] More specifically, the processing unit 133 processes the acquired high-energy image Im H and low-energy image Im L as input data for the trained model, a plurality of energy subtraction images are acquired as output data from the material decomposition image generation model. Note that the energy subtraction images may include, for example, a plurality of material decomposition images obtained by decomposing different materials from each other. The plurality of material decomposition images may be, for example, an image showing the thickness of bone and an image showing the thickness of soft tissue, or an image showing the thickness of contrast agent and an image showing the thickness of water. Note that the material decomposition image generation model may have a plurality of input channels to which a plurality of images are respectively input.

[0117] According to the above configuration, the trained model according to this embodiment is trained using input data relating to different radiation energies, and is therefore able to output material-decomposed images based on radiographic images relating to various radiation energies. Here, the tube voltage relating to radiography corresponds to the radiation energy. Therefore, even when the tube voltage relating to radiography changes, the control device 103 according to this embodiment can input the radiographic image to the trained model and accurately generate a material-decomposed image from the radiographic image. Furthermore, the trained model according to this embodiment is a trained model obtained by training using, as input data for training data, multiple radiographic images relating to different radiation energies generated from acquired radiographic images. Therefore, the trained model according to this embodiment can reduce the computational costs of machine learning compared to generating a trained model for each tube voltage.

[0118] In this embodiment, the second radiographic image is obtained by radiography, and the plurality of third radiographic images can be obtained by calculation using the plurality of second material decomposition images. More specifically, the plurality of third radiographic images can be generated, for example, according to Equation (10). Therefore, the plurality of third radiographic images can be generated by inverse transformation of the energy subtraction process using linear attenuation coefficients related to the materials decomposed in the plurality of second material decomposition images, the plurality of second material decomposition images, and radiation spectra related to the radiation energies of the plurality of third radiographic images. In contrast, the energy subtraction process for generating the plurality of second material decomposition images can be performed, for example, according to Equations (3) to (9). Therefore, the energy subtraction process can be performed using linear attenuation coefficients related to the materials decomposed in the plurality of second material decomposition images, the plurality of second radiographic images, and radiation spectra related to the radiation energies of the plurality of second radiographic images.

[0119] The tube voltage corresponding to the radiation energies of the plurality of third radiographic images can be set to a tube voltage within ±5 kV of the tube voltage used when the plurality of second radiographic images were captured. In this case, by performing learning with a narrower range of change in the tube voltage of the input data, it is expected that the inference accuracy of the trained model can be improved and more accurate material decomposition images can be generated.

[0120] Furthermore, the plurality of third radiographic images may be generated at intervals of 5 kV or more for the tube voltages corresponding to the radiation energies of the plurality of third radiographic images, in other words, at tube voltage differences of 5 kV or more. In this case, extensive learning is performed on changes in the tube voltage of the input data, which is expected to improve the robustness of processing using the trained model against changes in the tube voltage of the radiographic images input during inference.

[0121] In this embodiment, the set of radiation images includes a high-energy image Im H and low-energy image Im L Although two radiographic images are used, the set of radiographic images may include two or more radiographic images, for example, three or more radiographic images. Therefore, for example, a set of multiple first radiographic images may include all of the multiple first radiographic images. Note that the number of radiographic images included in the set only needs to correspond to the number of input channels of the material decomposition image generation model. Therefore, the set of multiple first radiographic images and the set of multiple third radiographic images can correspond to the number of input channels of the trained model.

[0122] Furthermore, the control device 103 according to this embodiment can also function as an example of an information processing device (learning device) that causes a machine learning model to learn. In this case, the acquisition unit 131 can function as an example of an acquisition unit that acquires a plurality of first radiographic images associated with different radiation energies. Furthermore, the processing unit 133 can function as an example of a learning unit that trains the machine learning model. Here, the processing unit 133 can generate a plurality of material-decomposed images by energy subtraction processing using the plurality of first radiographic images. Furthermore, the processing unit 133 can use the plurality of material-decomposed images to generate a plurality of second radiographic images associated with radiation energies that are different from the radiation energies of the plurality of first radiographic images and that are different from each other. Furthermore, the processing unit 133 can train the machine learning model using training data in which a set of a plurality of second radiographic images is used as input data and a plurality of material-decomposed images is used as output data. Here, the number of the plurality of second radiographic images is greater than the number of the plurality of first radiographic images.

[0123] According to the above configuration, the control device 103 according to this embodiment generates a plurality of radiographic images representing different radiation energies from an acquired radiographic image, and uses the generated plurality of radiographic images as input data to train a machine learning model. This makes it possible to construct a material decomposition image generation model that can output a material decomposition image based on radiographic images representing various radiation energies. Therefore, the control device 103 can generate a material decomposition image from a radiographic image using the trained model, while suppressing the computational cost of machine learning, even when the tube voltage for radiography is changed.

[0124] In this case, the plurality of first radiographic images can be obtained by radiography, and the plurality of second radiographic images can be obtained by calculation using the plurality of material decomposition images. More specifically, the plurality of second radiographic images can be generated, for example, according to Equation (10). Therefore, the plurality of second radiographic images can be generated by inverse transformation of the energy subtraction process using linear attenuation coefficients related to the materials decomposed in the plurality of material decomposition images, the plurality of material decomposition images, and radiation spectra related to the radiation energies of the plurality of second radiographic images. In contrast, the energy subtraction process for generating the plurality of material decomposition images can be performed, for example, according to Equations (3) to (9). Therefore, the energy subtraction process can be performed using linear attenuation coefficients related to the materials decomposed in the plurality of material decomposition images, the plurality of first radiographic images, and radiation spectra related to the radiation energies of the plurality of first radiographic images. Note that the radiation spectrum indicates the number of photons at a radiation energy.

[0125] Furthermore, in this case as well, the tube voltage corresponding to the radiation energy of the plurality of second radiographic images can be set to a tube voltage within ±5 kV of the tube voltage used when the plurality of first radiographic images were captured. In this case, by performing learning with a narrower range of change in the tube voltage of the input data, it is expected that the inference accuracy of the trained model can be improved and more accurate material decomposition images can be generated.

[0126] Furthermore, the plurality of second radiographic images may be generated at intervals of 5 kV or more for the tube voltages corresponding to the radiation energies of the plurality of second radiographic images, in other words, at tube voltage differences of 5 kV or more. In this case, extensive learning is performed on changes in the tube voltage of the input data, which is expected to improve the robustness of processing using the trained model against changes in the tube voltage of the radiographic images input during inference.

[0127] [Second embodiment] Next, a radiography system including a material decomposition image generation model according to a second embodiment of the present disclosure will be described in detail. Note that the configuration of the radiography system according to this embodiment is similar to that of the radiography system according to the first embodiment, and therefore the same reference numerals will be used and the description thereof will be omitted. The following description of the radiography system according to this embodiment will focus on the differences from the radiography system according to the first embodiment.

[0128] In the first embodiment, a method for constructing a trained model for generating a material decomposition image from a set of a high-energy image and a low-energy image, and a method for generating a material decomposition image using the trained model were described. In contrast, in the present embodiment, a method for constructing a machine learning model for generating a material decomposition image from a single energy image, and a method for generating a material decomposition image using the trained model will be described.

[0129] As described above, when a set of high-energy and low-energy images is used as input data for a material decomposition image generation model, it is necessary to capture at least two images, one high-energy image and one low-energy image, during imaging. In contrast, for a machine learning model that generates a material decomposition image from a single energy image, it is sufficient to capture one energy image during imaging of the subject. This reduces the burden on the photographer during imaging and, if the subject is a human body, reduces the radiation dose.

[0130] The difference between this embodiment and the first embodiment is that in step S804 shown in Fig. 8, as shown in Fig. 12(a) and Fig. 13(a), each virtual tube voltage image Iv generated in step S803 is used as input data for the training data. Note that even when virtual tube voltage images Iv with fine increments of about 1 kV are generated in step S803, each virtual tube voltage image Iv generated in step S803 is used as input data for the training data, as shown in Fig. 12(b) and Fig. 13(b). Note that, as with the first embodiment, the output data for the training data is the bone image Im generated in step S802. B and soft tissue imaging Im Scan be used.

[0131] In step S1104 shown in FIG. 11, the processing unit 133 generates a bone image Im, which is an energy subtraction image, based on the energy image Im using a material separation image generation model, which is a trained model. B and soft tissue imaging Im S Specifically, the processing unit 133 receives the energy image Im as input data for the material decomposition image generation model, and generates a bone image Im as output data for the material decomposition image generation model. B and soft tissue imaging Im S In this embodiment, the high-energy image Im described in the first embodiment is acquired and generated as the energy image Im to be input to the material decomposition image generation model. H The low-energy image Im described in the first embodiment is used as the energy image Im to be input to the material decomposition image generation model. L The energy image Im input to the material decomposition image generation model may be an energy image at a certain tube voltage captured by a known method.

[0132] As described above, the acquisition unit 131 according to this embodiment functions as an example of an acquisition unit that acquires one first radiographic image. The trained model can be a trained model obtained using training data in which each of the plurality of third radiographic images is used as input data and the plurality of second material decomposition images is used as output data.

[0133] According to the above configuration, the control device 103 according to this embodiment can generate a material-decomposed image from a radiographic image by inputting one radiographic image into the trained model even when the tube voltage for radiography changes. Therefore, compared to the first embodiment, the burden on the operator during radiography can be reduced, and if the subject is a human body, the radiation dose can be reduced.

[0134] The control device 103 according to this embodiment can also function as an example of an information processing device (learning device) that causes a machine learning model to learn. In this case, the acquisition unit 131 acquires a plurality of first radiographic images associated with different radiation energies. The processing unit 133 uses, as input data, each of a plurality of second radiographic images generated using a plurality of material-decomposed images obtained by energy subtraction processing, and uses the plurality of material-decomposed images as output data to train the machine learning model. With this configuration, the control device 103 according to this embodiment can build a machine learning model that generates a material-decomposed image from a single radiographic image even when the tube voltage for radiography changes. Therefore, it is only necessary to capture one radiographic image when imaging a subject, which reduces the burden on the operator during imaging compared to the first embodiment and also reduces the radiation exposure dose when the subject is a human body.

[0135] In addition, in the material decomposition image generation model according to this embodiment, the output data is also the bone image Im B and soft tissue imaging S In this case, the output data of the training data may be the bone image Im B and soft tissue imaging S Either of the following may be used.

[0136] [Other embodiments] The present invention can also be realized by providing a program that implements one or more of the functions of the above-described embodiments to a system or device via a network or storage medium, and having one or more processors in the computer of the system or device read and execute the program. For example, acquired images can be transferred to another PC via a medical PACS for processing. The present invention can also be realized by a circuit (e.g., an ASIC) that implements one or more functions. A computer can have one or more processors or circuits, and can include multiple separate computers or a network of multiple separate processors or circuits to read and execute computer-executable instructions.

[0137] The processor or circuitry may include a central processing unit (CPU), a microprocessing unit (MPU), a graphics processing unit (GPU), an application specific integrated circuit (ASIC), or a field programmable gateway (FPGA). The processor or circuitry may also include a digital signal processor (DSP), a data flow processor (DFP), or a neural processing unit (NPU).

[0138] The above disclosure includes the following configurations, methods, and programs. (Configuration 1) an acquisition unit that acquires a plurality of first radiographic images associated with different radiation energies; a generation unit that generates a plurality of first material decomposition images by using the set of the plurality of first radiological images as input data for a trained model; Equipped with the trained model is a trained model obtained using training data in which a set of a plurality of third radiographic images, which are generated using a plurality of second material-decomposed images generated by an energy subtraction process using a plurality of second radiographic images associated with radiation energies different from each other, are used as input data, and the plurality of second material-decomposed images are used as output data, The number of the third radiographic images is greater than the number of the second radiographic images. (Configuration 2) an acquisition unit that acquires a first radiographic image; a generation unit that generates a plurality of first material decomposition images by using the first radiographic images as input data for a trained model; Equipped with the trained model is a trained model obtained using training data in which each of a plurality of third radiographic images, which are generated using a plurality of second material-decomposed images generated by an energy subtraction process using a plurality of second radiographic images associated with radiation energies different from each other, is used as input data, and the plurality of second material-decomposed images are used as output data, The number of the third radiographic images is greater than the number of the second radiographic images. (Configuration 3) an acquisition unit that acquires a plurality of first radiographic images associated with different radiation energies; a generation unit that generates a first material decomposition image by using the set of the plurality of first radiological images as input data for a trained model; Equipped with the trained model is a trained model obtained using training data in which a set of a plurality of third radiographic images, which are generated using a plurality of second material-decomposed images generated by an energy subtraction process using a plurality of second radiographic images associated with radiation energies different from each other, are used as input data, and one of the plurality of second material-decomposed images is used as output data; The number of the third radiographic images is greater than the number of the second radiographic images. (Configuration 4) an acquisition unit that acquires a first radiographic image; a generation unit that generates a first material decomposition image by using the first radiographic image as input data for a trained model; Equipped with the trained model is a trained model obtained using training data in which each of a plurality of third radiographic images, which are generated using a plurality of second material-decomposed images generated by an energy subtraction process using a plurality of second radiographic images associated with radiation energies different from each other, is used as input data, and one of the plurality of second material-decomposed images is used as output data, The number of the third radiographic images is greater than the number of the second radiographic images. (Configuration 5) 5. The information processing apparatus according to any one of configurations 1 to 4, wherein the plurality of third radiographic images are generated by inverse transformation of the energy subtraction processing using linear attenuation coefficients related to materials decomposed in the plurality of second material decomposition images, the plurality of second material decomposition images, and radiation spectra related to radiation energies of the plurality of third radiographic images. (Configuration 6) 6. The information processing device according to any one of configurations 1 to 5, wherein the energy subtraction processing is performed using linear attenuation coefficients related to materials decomposed in the plurality of second material decomposition images, the plurality of second radiological images, and radiation spectra related to radiation energies of the plurality of second radiological images. (Configuration 7) 6. The information processing device according to configuration 5 or 5, wherein the radiation spectrum indicates the number of photons in radiation energy. (Configuration 8) the plurality of second radiographic images are obtained by radiography; 8. The information processing apparatus according to any one of configurations 1 to 7, wherein the plurality of third radiological images are obtained by calculation using the plurality of second material decomposition images. (Configuration 9) 9. The information processing device according to any one of configurations 1 to 8, wherein a tube voltage corresponding to the radiation energy of the plurality of third radiographic images is a tube voltage within ±5 kV of a tube voltage when the plurality of second radiographic images were captured. (Configuration 10) 9. The information processing device according to any one of configurations 1 to 8, wherein the plurality of third radiographic images are generated at intervals of 5 kV or more for tube voltages corresponding to the radiation energies of the plurality of third radiographic images. (Configuration 11) 4. The information processing device according to configuration 1 or 3, wherein the plurality of sets of first radiographic images and the plurality of sets of third radiographic images correspond to the number of input-side channels of the trained model. (Configuration 12) an acquisition unit that acquires a plurality of first radiographic images associated with different radiation energies; generating a plurality of material decomposition images by energy subtraction processing using the plurality of first radiation images; generating a plurality of second radiographic images relating to radiation energies different from the radiation energies of the plurality of first radiographic images and different from each other using the plurality of material decomposition images; a learning unit that learns a machine learning model using training data in which each or a set of the plurality of second radiological images is input data and at least one of the plurality of material decomposition images is output data; and Equipped with The information processing device, wherein the number of the plurality of second radiographic images is greater than the number of the plurality of first radiographic images. (Configuration 13) 13. The information processing apparatus according to configuration 12, wherein the plurality of second radiographic images are generated by inverse transformation of the energy subtraction processing using linear attenuation coefficients related to materials decomposed in the plurality of material decomposition images, the plurality of material decomposition images, and radiation spectra related to radiation energies of the plurality of second radiographic images. (Configuration 14) 14. The information processing device according to configuration 12 or 13, wherein the energy subtraction processing is performed using linear attenuation coefficients related to materials decomposed in the plurality of material decomposition images, the plurality of first radiological images, and radiation spectra related to radiation energies of the plurality of first radiological images. (Configuration 15) the plurality of first radiographic images are obtained by radiography; 15. The information processing apparatus according to any one of configurations 12 to 14, wherein the plurality of second radiographic images are obtained by calculation using the plurality of material decomposition images. (Configuration 16) 16. The information processing device according to any one of configurations 12 to 15, wherein a tube voltage corresponding to the radiation energy of the plurality of second radiographic images is a tube voltage within ±5 kV of a tube voltage when the plurality of first radiographic images were captured. (Configuration 17) 16. The information processing device according to any one of configurations 12 to 15, wherein the plurality of second radiographic images are generated at intervals of 5 kV or more for tube voltages corresponding to the radiation energies of the plurality of second radiographic images. (Configuration 18) a radiography device for detecting radiation; An information processing device according to any one of configurations 1 to 17; A radiography system comprising: (Method 1) acquiring a plurality of first radiographic images at different radiation energies; generating a plurality of first material decomposition images by using the set of the plurality of first radiographic images as input data for a trained model; Including, the trained model is a trained model obtained using training data in which a set of a plurality of third radiographic images, which are generated using a plurality of second material-decomposed images generated by an energy subtraction process using a plurality of second radiographic images associated with radiation energies different from each other, are used as input data, and the plurality of second material-decomposed images are used as output data, A method for operating an information processing device, wherein the number of the plurality of third radiographic images is greater than the number of the plurality of second radiographic images. (Method 2) acquiring a first radiographic image; generating a plurality of first material decomposition images by using the first radiographic image as input data for a trained model; Including, the trained model is a trained model obtained using training data in which each of a plurality of third radiographic images, which are generated using a plurality of second material-decomposed images generated by an energy subtraction process using a plurality of second radiographic images associated with radiation energies different from each other, is used as input data, and the plurality of second material-decomposed images are used as output data, A method for operating an information processing device, wherein the number of the plurality of third radiographic images is greater than the number of the plurality of second radiographic images. (Method 3) acquiring a plurality of first radiographic images at different radiation energies; generating a first material decomposition image by using the set of the plurality of first radiographic images as input data for a trained model; Including, the trained model is a trained model obtained using training data in which a set of a plurality of third radiographic images, which are generated using a plurality of second material-decomposed images generated by an energy subtraction process using a plurality of second radiographic images associated with radiation energies different from each other, are used as input data, and one of the plurality of second material-decomposed images is used as output data; A method for operating an information processing device, wherein the number of the plurality of third radiographic images is greater than the number of the plurality of second radiographic images. (Method 4) acquiring a first radiographic image; generating a first material decomposition image by using the first radiographic image as input data for a trained model; Including, the trained model is a trained model obtained using training data in which each of a plurality of third radiographic images, which are generated using a plurality of second material-decomposed images generated by an energy subtraction process using a plurality of second radiographic images associated with radiation energies different from each other, is used as input data, and one of the plurality of second material-decomposed images is used as output data, A method for operating an information processing device, wherein the number of the plurality of third radiographic images is greater than the number of the plurality of second radiographic images. (Method 5) acquiring a plurality of first radiographic images at different radiation energies; generating a plurality of material decomposition images by energy subtraction processing using the plurality of first radiation images; generating a plurality of second radiographic images relating to radiation energies different from the radiation energies of the plurality of first radiographic images and different from each other, using the plurality of material decomposition images; training a machine learning model using training data in which each or a set of the plurality of second radiological images is used as input data and at least one of the plurality of material decomposition images is used as output data; Including, A method for operating an information processing device, wherein the number of the plurality of second radiographic images is greater than the number of the plurality of first radiographic images. (Program 1) A program that, when executed by a computer, causes the computer to execute each step of the method for operating an information processing device according to any one of methods 1 to 5.

[0139] Although the present invention has been described above with reference to the embodiments, the present invention is not limited to the above embodiments. The present invention also includes inventions that have been modified within the scope of the present invention and inventions equivalent to the present invention. Furthermore, the above-described embodiments can be combined as appropriate within the scope of the present invention. [Explanation of symbols]

[0140] 103: Control device (information processing device) 131: Acquisition Department 133: Processing unit (generation unit)

Claims

1. an acquisition unit that acquires a plurality of first radiographic images associated with different radiation energies; a generation unit that generates a plurality of first material decomposition images by using the set of the plurality of first radiological images as input data for a trained model; Equipped with the trained model is a trained model obtained using training data in which a set of a plurality of third radiographic images, which are generated using a plurality of second material-decomposed images generated by an energy subtraction process using a plurality of second radiographic images associated with radiation energies different from each other, are used as input data, and the plurality of second material-decomposed images are used as output data, The number of the third radiographic images is greater than the number of the second radiographic images.

2. an acquisition unit that acquires a first radiographic image; a generation unit that generates a plurality of first material decomposition images by using the first radiographic images as input data for a trained model; Equipped with the trained model is a trained model obtained using training data in which each of a plurality of third radiographic images, which are generated using a plurality of second material-decomposed images generated by an energy subtraction process using a plurality of second radiographic images associated with radiation energies different from each other, is used as input data, and the plurality of second material-decomposed images are used as output data, The number of the third radiographic images is greater than the number of the second radiographic images.

3. an acquisition unit that acquires a plurality of first radiographic images associated with different radiation energies; a generation unit that generates a first material decomposition image by using the set of the plurality of first radiological images as input data for a trained model; Equipped with the trained model is a trained model obtained using training data in which a set of a plurality of third radiographic images, which are generated using a plurality of second material-decomposed images generated by an energy subtraction process using a plurality of second radiographic images associated with radiation energies different from each other, are used as input data, and one of the plurality of second material-decomposed images is used as output data; The number of the third radiographic images is greater than the number of the second radiographic images.

4. an acquisition unit that acquires a first radiographic image; a generation unit that generates a first material decomposition image by using the first radiographic image as input data for a trained model; Equipped with the trained model is a trained model obtained using training data in which each of a plurality of third radiographic images, which are generated using a plurality of second material-decomposed images generated by an energy subtraction process using a plurality of second radiographic images associated with radiation energies different from each other, is used as input data, and one of the plurality of second material-decomposed images is used as output data, The number of the third radiographic images is greater than the number of the second radiographic images.

5. 5. The information processing apparatus according to claim 1, wherein the plurality of third radiographic images are generated by inverse transformation of the energy subtraction processing using linear attenuation coefficients related to materials decomposed in the plurality of second material decomposition images, the plurality of second material decomposition images, and radiation spectra related to radiation energies of the plurality of third radiographic images.

6. 5. The information processing apparatus according to claim 1, wherein the energy subtraction processing is performed using linear attenuation coefficients related to materials decomposed in the plurality of second material decomposition images, the plurality of second radiological images, and radiation spectra related to radiation energies of the plurality of second radiological images.

7. The information processing device according to claim 5 , wherein the radiation spectrum indicates the number of photons in radiation energy.

8. the plurality of second radiographic images are obtained by radiography; The information processing apparatus according to claim 1 , wherein the plurality of third radiographic images are obtained by a calculation using the plurality of second material decomposition images.

9. 5. The information processing device according to claim 1, wherein a tube voltage corresponding to the radiation energy of the plurality of third radiographic images is a tube voltage within ±5 kV of a tube voltage used when the plurality of second radiographic images were captured.

10. The information processing apparatus according to claim 1 , wherein the plurality of third radiographic images are generated at intervals of 5 kV or more for tube voltages corresponding to radiation energies of the plurality of third radiographic images.

11. The information processing device according to claim 1 , wherein the plurality of sets of first radiographic images and the plurality of sets of third radiographic images correspond to the number of input-side channels of the trained model.

12. an acquisition unit that acquires a plurality of first radiographic images associated with different radiation energies; generating a plurality of material decomposition images by energy subtraction processing using the plurality of first radiation images; generating a plurality of second radiographic images relating to radiation energies different from the radiation energies of the plurality of first radiographic images and different from each other using the plurality of material decomposition images; a learning unit that learns a machine learning model using training data in which each or a set of the plurality of second radiological images is input data and at least one of the plurality of material decomposition images is output data; and Equipped with The information processing device, wherein the number of the plurality of second radiographic images is greater than the number of the plurality of first radiographic images.

13. 13. The information processing apparatus according to claim 12, wherein the plurality of second radiographic images are generated by inverse transformation of the energy subtraction processing using linear attenuation coefficients related to materials decomposed in the plurality of material decomposition images, the plurality of material decomposition images, and radiation spectra related to radiation energies of the plurality of second radiographic images.

14. 13. The information processing apparatus according to claim 12, wherein the energy subtraction processing is performed using linear attenuation coefficients related to materials decomposed in the plurality of material decomposition images, the plurality of first radiographic images, and radiation spectra related to radiation energies of the plurality of first radiographic images.

15. the plurality of first radiographic images are obtained by radiography; The information processing apparatus according to claim 12 , wherein the plurality of second radiation images are obtained by a calculation using the plurality of material decomposition images.

16. The information processing apparatus according to claim 12 , wherein the tube voltages corresponding to the radiation energies of the second radiographic images are within ±5 kV of the tube voltage used when the first radiographic images were captured.

17. The information processing apparatus according to claim 12 , wherein the plurality of second radiographic images are generated at intervals of 5 kV or more for tube voltages corresponding to radiation energies of the plurality of second radiographic images.

18. a radiography device for detecting radiation; An information processing device according to any one of claims 1 to 4 and 12; A radiography system comprising:

19. acquiring a plurality of first radiographic images at different radiation energies; generating a plurality of first material decomposition images by using the set of the plurality of first radiographic images as input data for a trained model; Including, the trained model is a trained model obtained using training data in which a set of a plurality of third radiographic images, which are generated using a plurality of second material-decomposed images generated by an energy subtraction process using a plurality of second radiographic images associated with radiation energies different from each other, are used as input data, and the plurality of second material-decomposed images are used as output data, A method for operating an information processing device, wherein the number of the plurality of third radiographic images is greater than the number of the plurality of second radiographic images.

20. acquiring a first radiographic image; generating a plurality of first material decomposition images by using the first radiographic image as input data for a trained model; Including, the trained model is a trained model obtained using training data in which each of a plurality of third radiographic images, which are generated using a plurality of second material-decomposed images generated by an energy subtraction process using a plurality of second radiographic images associated with radiation energies different from each other, is used as input data, and the plurality of second material-decomposed images are used as output data, A method for operating an information processing device, wherein the number of the plurality of third radiographic images is greater than the number of the plurality of second radiographic images.

21. acquiring a plurality of first radiographic images at different radiation energies; generating a first material decomposition image by using the set of the plurality of first radiographic images as input data for a trained model; Including, the trained model is a trained model obtained using training data in which a set of a plurality of third radiographic images, which are generated using a plurality of second material-decomposed images generated by an energy subtraction process using a plurality of second radiographic images associated with radiation energies different from each other, are used as input data, and one of the plurality of second material-decomposed images is used as output data; A method for operating an information processing device, wherein the number of the plurality of third radiographic images is greater than the number of the plurality of second radiographic images.

22. acquiring a first radiographic image; generating a first material decomposition image by using the first radiographic image as input data for a trained model; Including, the trained model is a trained model obtained using training data in which each of a plurality of third radiographic images, which are generated using a plurality of second material-decomposed images generated by an energy subtraction process using a plurality of second radiographic images associated with radiation energies different from each other, is used as input data, and one of the plurality of second material-decomposed images is used as output data, A method for operating an information processing device, wherein the number of the plurality of third radiographic images is greater than the number of the plurality of second radiographic images.

23. acquiring a plurality of first radiographic images at different radiation energies; generating a plurality of material decomposition images by energy subtraction processing using the plurality of first radiation images; generating a plurality of second radiographic images relating to radiation energies different from the radiation energies of the plurality of first radiographic images and different from each other, using the plurality of material decomposition images; training a machine learning model using training data in which each or a set of the plurality of second radiological images is used as input data and at least one of the plurality of material decomposition images is used as output data; Including, A method for operating an information processing device, wherein the number of the plurality of second radiographic images is greater than the number of the plurality of first radiographic images.

24. A program that, when executed by a computer, causes the computer to execute each step of the method for operating an information processing device according to any one of claims 19 to 23.

Images