JP7487683B2 - Radiation image generating method and radiation image capturing device - Google Patents

Description

The present invention relates to a radiation image generating method and a radiation image capturing device, and in particular to a radiation image generating method and a radiation image capturing device for capturing images of a subject at multiple shooting angles.

Conventionally, a radiographic imaging device that captures images of a subject from multiple imaging angles is known (see, for example, Patent Document 1).

The X-ray CT device (radiographic imaging device) described in the above-mentioned Patent Document 1 includes an X-ray tube, a collimator, an X-ray detector, and a rotation mechanism. The radiographic imaging device disclosed in the above-mentioned Patent Document 1 generates a tomographic image by capturing images while rotating the X-ray tube and the X-ray detector using the rotation mechanism.

JP 2005-237779 A

Here, in the X-ray CT device (radiographic image capturing device) described in the above Patent Document 1, the subject is captured at each rotation angle (imaging angle) while being rotated by a rotation mechanism. By increasing the imaging angles (reducing the imaging angle interval), the image quality of the resulting tomographic image (reconstructed image) can be improved. However, increasing the imaging angles increases the number of images captured, and increases the imaging time required to generate a reconstructed image. In this case, in order to suppress the increase in imaging time, it is possible to shorten the imaging time at each imaging angle. However, shortening the imaging time at each imaging angle reduces the image quality of the images captured at each imaging angle, and the image quality of the reconstructed image also reduces. Therefore, it is desirable to suppress the deterioration of the image quality of the reconstructed image while shortening the imaging time.

This invention has been made to solve the above problems, and one object of the invention is to provide a radiographic image generating method and radiographic image capturing device that can reduce the image capture time while suppressing degradation of the image quality of the reconstructed image.

In order to achieve the above object, the radiation image generating method according to the first aspect of the present invention includes the steps of: capturing a plurality of radiation images at a plurality of imaging angles at which the relative angle between the subject and the imaging unit is different; selecting from the plurality of radiation images a first reference image captured at a first imaging angle and a second reference image captured at a second imaging angle different from the first imaging angle; selecting a first radiation image group and a second radiation image group captured at different imaging angles within an angle range that covers the imaging angles of each selected reference image; generating a first blurred image and a second blurred image based on each selected radiation image group; generating a first high-quality image based on the first reference image and the first blurred image, and generating a second high-quality image based on the second reference image and the second blurred image; and generating a reconstructed image based on the first high-quality image and the second high-quality image so as to reflect the three-dimensional structure of the subject.

The radiographic imaging device according to the second aspect of the present invention includes an imaging unit that images a subject, and an image processing unit that performs image processing on a plurality of radiographic images captured from a plurality of angles with different relative angles between the subject and the imaging unit. The image processing unit is configured to generate a first blurred image based on the first radiographic image group and generate a second blurred image based on the second radiographic image group from the plurality of radiographic images, the first blurred image being based on the first reference image and the first blurred image, and the second high-quality image being based on the second reference image and the second blurred image, among a first radiographic image group and a second radiographic image group captured at different imaging angles within an angle range that covers the imaging angles of a first reference image captured at a first imaging angle and a second reference image captured at a second imaging angle different from the first imaging angle, generate a first high-quality image based on the first reference image and the first blurred image, generate a second high-quality image based on the second reference image and the second blurred image, and generate a reconstructed image based on the first high-quality image and the second high-quality image, which is reconstructed to reflect the three-dimensional structure of the subject.

The radiation image generating method in the first aspect includes the steps of generating a first high quality image based on the first reference image and the first blurred image, generating a second high quality image based on the second reference image and the second blurred image, and generating a reconstructed image based on the first high quality image and the second high quality image so as to reflect the three-dimensional structure of the subject. Here, if the shooting time for shooting each radiation image is shortened in order to shorten the total shooting time required to generate the reconstructed image, the first reference image and the second reference image will be images with a lot of noise. On the other hand, since the first blurred image, which is an image in which the subject is blurred, is generated based on the first radiation image group, a sufficient amount of radiation can be obtained even if the time to shoot one radiation image is short. Furthermore, since the second blurred image, which is an image in which the subject is blurred, is generated based on the second radiation image group, a sufficient amount of radiation can be obtained even if the time to shoot one radiation image is short. Therefore, the first blurred image and the second blurred image are images in which the subject is blurred but with little noise.

The first and second reference images are noisy but the subject is not blurred, so edge information of the subject can be obtained from the first and second reference images. The first and second blurred images are noisy but the subject is not blurred, so pixel value information of the subject can be obtained. That is, by obtaining edge information of the subject from the first reference image and obtaining pixel value information of the subject from the first blurred image, even if the shooting time for each radiation image is shortened, an image (first high-quality image) in which noise is reduced and blur of the subject is suppressed can be generated. By obtaining edge information of the subject from the second reference image and obtaining pixel value information of the subject from the second blurred image, even if the shooting time for each radiation image is shortened, an image (second high-quality image) in which noise is reduced and blur of the subject is suppressed can be generated. Therefore, as described above, by generating a first high-quality image from a first blurred image that provides a sufficient amount of radiation and a first reference image, and by generating a second high-quality image from a second blurred image that provides a sufficient amount of radiation and a second reference image, it is possible to generate images in which degradation of image quality is suppressed even when the imaging time for capturing each radiation image is shortened. As a result, it is possible to provide a radiation image generating method that can reduce degradation of image quality of the reconstructed image while shortening the imaging time.

The radiographic imaging device in the second aspect includes an image processing unit that generates a first blurred image based on the first radiographic image group and a second blurred image based on the second radiographic image group, generates a first high-quality image based on the first reference image and the first blurred image, generates a second high-quality image based on the second reference image and the second blurred image, and generates a reconstructed image based on the first high-quality image and the second high-quality image, the reconstructed image being reconstructed to reflect the three-dimensional structure of the subject. This makes it possible to provide a radiographic imaging device that can reduce the imaging time while suppressing deterioration in the image quality of the reconstructed image, similar to the radiographic imaging method in the first aspect.

1 is a block diagram showing a configuration of a radiation image generating apparatus according to an embodiment; 3 is a schematic diagram for explaining an angle range when a radiological image generating device according to an embodiment captures an image of a subject; FIG. FIG. 2 is a schematic diagram for explaining a configuration for generating a reconstructed image from a plurality of radiation images. 5A and 5B are schematic diagrams for explaining a first reference image and a first radiographic image group. 10 is a schematic diagram for explaining a configuration in which an image processing unit generates a first blurred image according to an embodiment. FIG. 4 is a schematic diagram for explaining a configuration in which an image processing unit according to an embodiment generates a first high-quality image. FIG. 11 is a flowchart for explaining a process in which an image processing unit generates a reconstructed image according to an embodiment. 11 is a flowchart for explaining a process for generating a first blurred image by an image processing unit according to an embodiment. 10 is a flowchart for explaining a process for generating a first high-quality image by an image processing unit according to an embodiment. FIG. 13 is a block diagram showing a configuration of a radiation image generating device according to a modified example. 10 is a flowchart for explaining a process for generating a first high-quality image by an image processing unit according to a modified example.

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

A radiographic imaging device 100 according to this embodiment will be described with reference to Figures 1 to 6. The radiographic imaging device 100 is used for non-destructive testing and medical applications. In this embodiment, an example in which the radiographic imaging device 100 is used for non-destructive testing will be described.

(Overall configuration of the radiation image capturing device)
As shown in FIG. 1, the radiation image capturing apparatus 100 includes a capturing unit 1, a computer 2, a subject rotating mechanism 3, a display unit 4, and an input receiving unit 5.

The imaging unit 1 is configured to image a subject 90. In this embodiment, the imaging unit 1 includes an X-ray source 1a that irradiates X-rays, and an X-ray detector 1b that detects the X-rays irradiated from the X-ray source 1a. In this embodiment, an example in which a circuit board is imaged as the subject 90 will be described.

In this specification, the vertical direction is defined as the Z direction, the upward direction as the Z1 direction, and the downward direction as the Z2 direction. The direction from the X-ray source 1a toward the X-ray detector 1b is defined as the X direction, with one side defined as the X1 direction and the other side defined as the X2 direction. The direction perpendicular to the Z direction and the X direction is defined as the Y direction, with one side defined as the Y1 direction and the other side defined as the Y2 direction.

The X-ray source 1a is configured to irradiate the subject 90 with X-rays. Specifically, the X-ray source 1a is configured to generate X-rays by applying a high voltage.

The X-ray detector 1b is configured to detect X-rays emitted from the X-ray source 1a. The X-ray detector 1b is also configured to convert the detected X-rays into an electrical signal. The X-ray detector 1b is, for example, a flat panel detector (FPD). The X-ray detector 1b is configured with a plurality of conversion elements (not shown) and pixel electrodes (not shown) arranged on the plurality of conversion elements. The plurality of conversion elements and pixel electrodes are arranged in the Y direction and the Z direction at a predetermined period (pixel pitch). In this embodiment, the X-ray detector 1b outputs a detection signal (image signal) based on the detected X-ray dose. The image signal output from the X-ray detector 1b is sent to an image processing unit 21, which will be described later.

The computer 2 includes a first processor 2a, such as a CPU (Central Processing Unit), a second processor 2b, such as a GPU (Graphics Processing Unit) or an FPGA (Field-Programmable Gate Array) configured for image processing, memory such as a ROM (Read Only Memory) and a RAM (Random Access Memory), a storage unit 2c, and an image signal acquisition unit 2d.

The first processor 2a includes a control unit 20. The control unit 20 is configured to control the X-ray source 1a, the subject rotation mechanism 3, and the like. The control unit 20 is configured in software as a functional block realized by the first processor 2a executing various programs. The control unit 20 may be configured in hardware as a dedicated processing circuit.

The second processor 2b includes an image processing unit 21. The image processing unit 21 is configured to perform image processing on a plurality of radiation images captured from a plurality of angles having different relative angles between the subject 90 and the imaging unit 1. In this embodiment, the image processing unit 21 is configured to acquire a plurality of radiation images. Specifically, the image processing unit 22 is configured to capture a plurality of X-ray images 6 as a plurality of radiation images. That is, the image processing unit 21 acquires the X-ray image 6 based on an image signal acquired from the X-ray detector 1b via the image signal acquisition unit 2d. The image processing unit 21 is also configured to generate a first blurred image 8a (see FIG. 3) and a second blurred image 8b (see FIG. 3) described later based on the acquired plurality of X-ray images 6. The image processing unit 21 is also configured to generate a reconstructed image 10 (see FIG. 3) described later based on the first blurred image 8a and the second blurred image 8b. The reconstructed image 10 is an image reconstructed to reflect the three-dimensional structure of the subject 90. In this embodiment, the image processing unit 21 generates a tomographic image as the reconstructed image 10. That is, the radiation image capturing device 100 is a so-called CT (Computed Tomography) capturing device. Note that the tomographic image may be a two-dimensional tomographic image or a three-dimensional tomographic image.

The image processing unit 21 is configured in software as a functional block realized by the second processor 2b executing various programs. The image processing unit 21 may be configured in hardware as a dedicated processing circuit. Details of the configuration in which the image processing unit 21 generates the first blurred image 8 and the reconstructed image 10 will be described later.

The storage unit 2c is configured to store the X-ray image 6 acquired by the image signal acquisition unit 2d, the reconstructed image 10 generated by the image processing unit 21, the trained model 11, and various programs executed by the first processor 2a and the second processor 2b. The storage unit 2c includes a non-volatile memory such as a hard disk drive (HDD) or a solid state drive (SSD).

The image signal acquisition unit 2d is configured to acquire a detection signal (image signal) from the X-ray detector 1b. The image signal acquisition unit 2d is also configured to output the acquired image signal to the image processing unit 21. The image signal acquisition unit 2d is configured to acquire a plurality of radiation images. The image signal acquisition unit 2d includes, for example, an input/output interface.

The subject rotation mechanism 3 is configured to change the relative angle between the subject 90 and the imaging unit 1. Specifically, the subject rotation mechanism 3 is configured to change the relative angle between the subject 90 and the imaging unit 1 by rotating the subject 90 in a rotation direction around the axis 80. The subject rotation mechanism 3 includes, for example, a subject placement unit (not shown) on which the subject 90 is placed, and a drive unit (not shown) that rotates the subject placement unit. The axis 80 is an axis extending along the Z direction. In the example shown in FIG. 1, the axis 80 is provided at the center of the subject rotation mechanism 3. The subject 90 is also provided at the center of the subject rotation mechanism 3. That is, the axis 80 passes through the center of the subject 90 and is set at a position where the subject 90 can be rotated within the irradiation range of the X-rays irradiated from the X-ray source 1a. The subject rotation mechanism 3 is an example of a "relative angle change mechanism" in the claims.

The display unit 4 displays the reconstructed image 10 generated by the image processing unit 21. The display unit 4 includes, for example, a liquid crystal monitor.

The input reception unit 5 is configured to receive operation input from an operator. The input reception unit 5 includes, for example, input devices such as a keyboard and a mouse.

The trained model 11 is configured to generate a first high-quality image 9a (see FIG. 3) based on a first reference image 6a (see FIG. 3) and a first blurred image 8a (see FIG. 3) described later. The trained model 11 is also configured to generate a second high-quality image 9b (see FIG. 3) based on a second reference image 6b (see FIG. 3) and a first blurred image 8a (see FIG. 3) described later. Details of the configuration in which the trained model 11 generates the first high-quality image 9a and the second high-quality image 9b will be described later.

(Shooting angle)
Next, a configuration for changing the imaging angle when the radiation image capturing apparatus 100 captures an image of the subject 90 will be described with reference to FIG.

The control unit 20 controls the subject rotation mechanism 3 to rotate the subject 90 at a predetermined angle interval while photographing the subject 90 at each shooting angle. In the example shown in FIG. 2, the control unit 20 controls the subject rotation mechanism 3 to rotate the subject 90 360 degrees at predetermined angle intervals from 1 to 18. That is, in the example shown in FIG. 2, photographs are taken a total of 18 times by photographing at shooting angles 1 to 18. Also, in the example shown in FIG. 2, the predetermined angle interval is 20 degrees.

As shown in FIG. 2, the image processing unit 21 generates a reconstructed image 10 based on the X-ray images 6 captured at each of the imaging angles 1 to 18. As the number of X-ray images 6 captured (imaging angles) increases, a more detailed reconstructed image 10 can be generated, but the total imaging time required to generate the reconstructed image 10 also increases. Therefore, the total imaging time required to generate the reconstructed image 10 can be reduced by shortening the imaging time required to capture each X-ray image 6. However, shortening the imaging time required for each image reduces the amount of X-rays detected by the X-ray detector 1b per image. As a result, the image quality of the X-ray image 6 is reduced. When the image quality of the X-ray image 6 is reduced, the image quality of the reconstructed image 10 is also reduced. Therefore, it is desirable to suppress the deterioration of the image quality of the reconstructed image 10 obtained even when the imaging time required for each image is shortened.

In this embodiment, the image processing unit 21 generates a first blurred image 8a (see FIG. 3) in which the subject 90 is captured in a blurred state based on a first radiographic image group 7a (see FIG. 3) captured at an imaging angle within a predetermined angle range 30 including the imaging angle of a first reference image 6a (see FIG. 3), which is one of the multiple radiographic images. The predetermined angle range 30 includes the imaging angle of the first reference image 6a, which is one of the multiple radiographic images. Specifically, the predetermined angle range 30 includes the imaging angle of the first reference image 6a and a plurality of different imaging angles that are consecutive at a predetermined angle interval. More specifically, the predetermined angle range 30 includes a plurality of different imaging angles that are consecutive including imaging angles before and after the imaging angle of the first reference image 6a at a predetermined angle interval.

2, the X-ray image 6 of the subject 90 taken at angle 1 is set as the first reference image 6a. Also, in the example shown in FIG. 2, the predetermined angle range 30 includes two shooting angles at a predetermined angle interval before and after the shooting angle 1, with the shooting angle 1 at the center. That is, in the example shown in FIG. 2, the predetermined angle range 30 includes the shooting angle 3 and the shooting angle 17 at the center of the shooting angle 1, as shown by the arrow 30a. Note that when the reference image is changed, the predetermined angle range 30 is also changed. That is, the predetermined angle range is set to be a range including shooting angles at a predetermined angle interval before and after the selected X-ray image 6, with the selected X-ray image 6 at the center, each time an X-ray image 6 is selected as the reference image. For example, when an X-ray image 6 taken at shooting angle 10 is selected as the second reference image 6b, the shooting angles 8 to 12 are set as the predetermined angle range.

The image processing unit 21 also generates a first high-quality image 9a (see FIG. 3) by reducing noise from the first reference image 6a and reducing blur of the subject 90 appearing in the image based on the first reference image 6a and the first blurred image 8a. The image processing unit 21 also generates a second high-quality image 9b (see FIG. 3) by reducing noise from the second reference image 6b and the second blurred image 8b based on the second reference image 6b and the second blurred image 8b. The image processing unit 21 is also configured to generate a reconstructed image 10 based on the first high-quality image 9a and the second high-quality image 9b, which is reconstructed to reflect the three-dimensional structure of the subject 90.

(Generation of Reconstructed Images)
The configuration in which the image processing unit 21 generates a reconstructed image 10 from a plurality of radiation images (X-ray images 6) will be described with reference to FIG.

The image processing unit 21 selects, from the multiple radiation images (X-ray images 6), a first reference image 6a captured at a first imaging angle and a second reference image 6b captured at a second imaging angle different from the first imaging angle. The image processing unit 21 also selects a first radiation image group 7a and a second radiation image group 7b captured at different imaging angles within an angle range that covers the imaging angles of each selected reference image. Note that in this embodiment, the angle range that covers the imaging angle of the first reference image 6a is angle range 30, and the angle range that covers the imaging angle of the second reference image 6b is an angle range different from angle range 30.

The image processing unit 21 then generates a first blurred image 8a and a second blurred image 8b based on each of the selected radiographic image groups. The image processing unit 21 then generates a first high-quality image 9a based on the first reference image 6a and the first blurred image 8a. The image processing unit 21 also generates a second high-quality image 9b based on the second reference image 6b and the second blurred image 8b. After that, the image processing unit 21 generates a reconstructed image 10 based on the first high-quality image 9a and the second high-quality image 9b, which is reconstructed to reflect the three-dimensional structure of the subject 90.

(First Reference Image and First Radiation Image Group)
Next, the first reference image 6a and the first radiographic image group 7a will be described with reference to FIG.

The first reference image 6a is one of the multiple X-ray images 6 (radiographic images). In this embodiment, the image processing unit 21 sequentially selects the radiographic images at each of the multiple shooting angles including the first shooting angle and the second shooting angle as the reference image. Specifically, the image processing unit 21 sequentially selects the X-ray images 6 captured at shooting angles other than the shooting angle of the first reference image 6a and the shooting angle of the second reference image 6b as the reference image. In the following, an example will be described in which the X-ray image 6 captured at the first shooting angle (see FIG. 2) out of the multiple X-ray images 6 is selected as the first reference image 6a.

The first radiographic image group 7a is the X-ray images 6 included in a predetermined angle range 30 (see FIG. 2). In this embodiment, the X-ray images 6 captured at the second imaging angle, the third imaging angle, the seventeenth imaging angle, and the eighteenth imaging angle are respectively set as the second angle X-ray image 17a, the third angle X-ray image 17b, the seventeenth angle X-ray image 17c, and the eighteenth angle X-ray image group 7d. Note that the first radiographic image group 7a also includes the first reference image 6a itself.

The first reference image 6a and the first radiographic image group 7a are images captured with a shorter capture time than when the images are captured until the noise is reduced. In other words, the first reference image 6a and the first radiographic image group 7a are images with a small X-ray dose and a low SNR (signal-noise ratio). In other words, the first reference image 6a and the first radiographic image group 7a are images with a lot of noise. In the example shown in FIG. 4, the amount of noise in the first reference image 6a and the first radiographic image group 7a is expressed by hatching.

The second reference image 6b and the second radiographic image group 7b are similar to the first reference image 6a and the first radiographic image group 7a, except for the shooting angle. Therefore, a detailed description of the second reference image 6b and the second radiographic image group 7b will be omitted.

(First blurred image)
Next, a configuration in which the image processing unit 21 generates the first blurred image 8a will be described with reference to FIG.

The image processing unit 21 generates a first blurred image 8a based on the first radiographic image group 7a out of the first radiographic image group 7a and the second radiographic image group 7b captured at different imaging angles. Specifically, the image processing unit 21 generates the first blurred image 8a based on the first reference image 6a, the X-ray image 17a at the second angle captured at different imaging angles within the predetermined angle range 30, the X-ray image 17b at the third angle, the X-ray image 17c at the 17th angle, and the X-ray image 17d at the 18th angle. As shown in FIG. 5, the first blurred image 8a is an image in which the subject 90 is blurred. The first blurred image 8a is an image generated based on a total of five images, the first reference image 6a and the four X-ray images 6 included in the first radiographic image group 7a. Therefore, the first blurred image 8a is an image in which the subject 90 is blurred, but has a high SNR and low noise.

In this embodiment, the image processing unit 21 generates a first blurred image 8a by adding the first reference image 6a, the X-ray image 17a at the first angle, the X-ray image 17b at the third angle, the X-ray image 17c at the 17th angle, and the X-ray image 17d at the 18th angle. Note that adding the first reference image 6a and multiple radiographic images means adding, to each pixel of the first reference image 6a, the pixel values of the pixels of the multiple radiographic images that correspond to the position of each pixel of the first reference image 6a.

The configuration in which the image processing unit 21 generates the second blurred image 8b based on the second radiographic image group 7b is similar to the configuration in which the image processing unit 21 generates the first blurred image 8a based on the first radiographic image group 7a. Therefore, a detailed description of the configuration in which the image processing unit 21 generates the second blurred image 8b will be omitted.

(First high quality image)
Next, a configuration in which the image processing section 21 generates the first high quality image 9a will be described with reference to FIG.

In this embodiment, the image processing unit 21 generates a first high-quality image 9a based on the first reference image 6a, the first blurred image 8a, and the trained model 11.

The trained model 11 is a model that has been trained to output a clear image in which noise and blur of the subject 90 have been removed from a noisy image and an image in which the subject 90 is blurred. The trained model 11 includes, for example, Conditional Generative Adversarial Nets (CGAN).

In this embodiment, the image processing unit 21 is configured to generate a first high-quality image 9a by inputting the first reference image 6a and the first blurred image 8a to the trained model 11. As shown in FIG. 6, the first high-quality image 9a is an image in which noise has been reduced from the first reference image 6a and blur of the subject 90 in the image has been reduced. Here, since the first blurred image 8a is an image generated based on the first reference image 6a and the first radiation image group 7a, it can be considered to be an image captured for a longer shooting time than the first reference image 6a. That is, although the subject 90 is blurred in the first blurred image 8a, it can be said that the image is valid in terms of the distribution of pixel values. On the other hand, although the first reference image 6a is a noisy image, it is an image in which the subject 90 is not blurred, so that it can be said that the image is valid in terms of the pixel boundaries. In addition, the trained model 11 has learned to output a clear image in which noise and blur of the subject 90 have been removed from a noisy image and an image in which the subject 90 is blurred. Therefore, by inputting the first reference image 6a, which has valid pixel boundaries, and the first blurred image 8a, which has valid pixel value distribution, to the trained model 11, blurring of the subject 90 is suppressed and a first high-quality image 9a, which is an image with reduced noise, can be generated.

The configuration in which the image processing unit 21 generates the second high-quality image 9b is similar to the configuration in which the image processing unit 21 generates the first high-quality image 9a. Therefore, a detailed description of the configuration in which the image processing unit 21 generates the second high-quality image 9b will be omitted.

(Reconstruction image generation process)
Next, the process of the radiation image generating method in which the image processing unit 21 generates the reconstructed image 10 will be described with reference to FIG.

As shown in FIG. 7, the radiation image generating method includes step 101 of capturing a plurality of radiation images, step 102 of generating a first blurred image 8a, step 103 of generating a first high-quality image 9a, and step 104 of generating a reconstructed image 10.

In step 101, the image processing unit 21 captures multiple radiation images at multiple imaging angles with different relative angles between the subject 90 and the imaging unit 1. Specifically, in step 101, the image processing unit 21 acquires multiple radiation images captured at a predetermined angle interval. Also in step 101, the image processing unit 21 captures multiple radiation images by changing the relative angle between the subject 90 and the imaging unit 1 and capturing images at each angle. Specifically, the image processing unit 21 captures multiple X-ray images 6 as the multiple radiation images. In this embodiment, the image processing unit 21 captures 18 X-ray images 6 as the multiple X-ray images 6.

In step 102, the image processing unit 21 generates a first blurred image 8a based on the first reference image 6a and the first radiation image group 7a. The image processing unit 21 also generates a second blurred image 8b based on the second reference image 6b and the second radiation image group 7b. In this embodiment, the image processing unit 21 sequentially selects X-ray images 6 other than the first reference image 6a and the second reference image 6b from the 18 X-ray images 6 captured at all imaging angles from the first imaging angle to the 18th imaging angle, and generates blurred images for the sequentially selected reference images. In other words, the image processing unit 21 generates 18 blurred images including the first reference image 6a and the second reference image 6b.

In step 103, the image processing unit 21 generates a first high-quality image 9a based on the first reference image 6a and the first blurred image 8a. The image processing unit 21 also generates a second high-quality image 9b based on the second reference image 6b and the second blurred image 8b. In this embodiment, the image processing unit 21 generates multiple high-quality images from multiple reference images other than the first reference image 6a and the second reference image 6b and multiple blurred images other than the first blurred image 8a and the second blurred image 8b. That is, the image processing unit 21 generates 18 high-quality images from 18 reference images including the first reference image 6a and the second reference image 6b and 18 blurred images including the first blurred image 8a and the second blurred image 8b.

In step 104, the image processing unit 21 generates a reconstructed image 10 based on the first high-quality image 9a and the second high-quality image 9b, reconstructed so as to reflect the three-dimensional structure of the subject 90. Note that in this embodiment, the image processing unit 21 generates one reconstructed image 10 based on 18 high-quality images including the first high-quality image 9a and the second high-quality image 9b. Also, in this embodiment, the image processing unit 21 generates a tomographic image as the reconstructed image 10.

(Blurred image generation process)
Next, the process of generating a blurred image by the image processing unit 21 will be described with reference to FIG.

In step 102a, the image processing unit 21 selects, from the multiple radiation images (X-ray images 6), a first reference image 6a captured at a first imaging angle and a second reference image 6b captured at a second imaging angle different from the first imaging angle.

In step 102b, the image processing unit 21 selects a first radiographic image group 7a captured within an angle range 30 that covers the imaging angle of the selected first reference image 6a. Specifically, in step 102b, the image processing unit 21 acquires a first radiographic image group 7a captured at a plurality of different imaging angles that are consecutive at a predetermined angle interval from the imaging angle of the first reference image 6a from among the first radiographic image group 7a. More specifically, in step 102b, the image processing unit 21 acquires a first radiographic image group 7a captured at a plurality of different imaging angles that are consecutive, including imaging angles before and after the imaging angle of the first reference image 6a at a predetermined angle interval from among the first radiographic image group 7a. In this embodiment, four X-ray images 6 are acquired from the first radiographic image group 7a captured at four different imaging angles, including two imaging angles before and after the imaging angle of the first reference image 6a at a predetermined angle interval.

In addition, in step 102b, the image processing unit 21 selects a second radiographic image group 7b captured within an angle range covering the imaging angle of the selected second reference image 6b. Specifically, in step 102b, the image processing unit 21 acquires the second radiographic image group 7b captured at a plurality of different imaging angles that are consecutive at a predetermined angle interval from the imaging angle of the second reference image 6b from the second radiographic image group 7b. More specifically, in step 102b, the image processing unit 21 acquires the second radiographic image group 7b captured at a plurality of different imaging angles that are consecutive, including imaging angles before and after the imaging angle of the second reference image 6b at a predetermined angle interval from the second radiographic image group 7b from the second radiographic image group 7b. In this embodiment, four X-ray images 6 are acquired from the second radiographic image group 7b captured at four different imaging angles that include two imaging angles before and after the imaging angle of the second reference image 6b at a predetermined angle interval.

In step 102c, the image processing unit 21 generates a first blurred image 8a based on the first radiographic image group 7a selected in step 102b. Specifically, the image processing unit 21 adds the first radiographic image group 7a selected in step 102b. That is, the image processing unit 21 generates the first blurred image 8a by adding the first radiographic image group 7a. In step 102c, the image processing unit 21 generates a first blurred image 8a, which is an image in which the subject 90 is blurred, based on the first radiographic image group 7a. Note that in the process of step 102c, the image processing unit 21 generates the first blurred image 8a by adding the first reference image 6a together with the first radiographic image group 7a. In addition, the image processing unit 21 stores the generated first blurred image 8a in the storage unit 2c.

In step 102d, the image processing unit 21 generates a second blurred image 8b based on the second radiographic image group 7b selected in step 102b. Specifically, the image processing unit 21 adds the second radiographic image group 7b selected in step 102b. That is, the image processing unit 21 generates the second blurred image 8b by adding the second radiographic image group 7b. In step 102d, the image processing unit 21 generates a second blurred image 8b, which is an image in which the subject 90 is blurred, based on the second radiographic image group 7b. Note that in the process of step 102d, the image processing unit 21 generates the second blurred image 8b by adding the second reference image 6b together with the second radiographic image group 7b. In addition, the image processing unit 21 stores the generated second blurred image 8b in the storage unit 2c.

In step 102e, the image processing unit 21 generates blurred images corresponding to the reference images at all selected shooting angles other than the first reference image 6a and the second reference image 6b. Note that the image processing unit 21 generates blurred images corresponding to the reference images at all selected shooting angles in the same manner as the configuration for generating the first blurred image 8a (second blurred image 8b).

(High-quality image generation processing)
Next, the process of generating a high quality image by the image processing unit 21 will be described with reference to FIG.

In step 103a, the image processing unit 21 selects a first reference image 6a from the multiple radiation images (X-ray images 6).

In step 103b, the image processing unit 21 acquires the first blurred image 8a corresponding to the first reference image 6a selected in step 103a from the memory unit 2c.

In step 103c, the image processing unit 21 generates a first high-quality image 9a by inputting the first reference image 6a and the first blurred image 8a to the trained model 11.

In step 103d, the image processing unit 21 selects a second reference image 6b from the multiple radiation images (X-ray images 6).

In step 103e, the image processing unit 21 acquires the second blurred image 8b corresponding to the second reference image 6b selected in step 103d from the memory unit 2c.

In step 103f, the image processing unit 21 generates a second high-quality image 9b by inputting the second reference image 6b and the second blurred image 8b to the trained model 11.

In step 103g, the image processing unit 21 sequentially sets the radiographic images at each of the multiple shooting angles as reference images, and generates high-quality images for each of the reference images. Specifically, the image processing unit 21 sequentially sets the X-ray images 6 captured at the shooting angles other than the first reference image 6a and the second reference image 6b among the shooting angles 1 to 18 shown in FIG. 2 as reference images. In addition, the image processing unit 21 sequentially changes the predetermined angle range as the reference images are sequentially changed. The image processing unit 21 creates blurred images based on the reference images and the group of radiographic images at each of the shooting angles 1 to 18. In addition, the image processing unit 21 creates high-quality images based on the blurred images and the reference images created at each of the shooting angles 1 to 18. That is, the image processing unit 21 generates 18 reference images including the first reference image 6a and the second reference image 6b. In addition, the image processing unit 21 generates 18 blurred images including the first blurred image 8a and the second blurred image 8b. In addition, the image processing unit 21 generates 18 high-quality images including the first high-quality image 9a and the second high-quality image 9b.

(Effects of this embodiment)
In this embodiment, the following effects can be obtained.

In this embodiment, as described above, the method includes the steps of capturing a plurality of radiation images at a plurality of imaging angles at which the relative angles between the subject 90 and the imaging unit 1 are different; selecting a first reference image 6a captured at a first imaging angle and a second reference image 6b captured at a second imaging angle different from the first imaging angle from the plurality of radiation images; selecting a first radiation image group 7a and a second radiation image group 7b captured at different imaging angles within an angle range 30 that covers the imaging angles of each selected reference image; generating a first blurred image 8a and a second blurred image 8b based on each selected radiation image group; generating a first high-quality image 9a based on the first reference image 6a and the first blurred image 8a, and generating a second high-quality image 9b based on the second reference image 6b and the second blurred image 8b; and generating a reconstructed image 10 that is reconstructed to reflect the three-dimensional structure of the subject 90 based on the first high-quality image 9a and the second high-quality image 9b.

Here, if the imaging time for capturing each radiation image (X-ray image 6) is reduced in order to reduce the total imaging time required to generate the reconstructed image 10, the first reference image 6a and the second reference image 6b will be images with a lot of noise. On the other hand, the first blurred image 8a, which is an image in which the subject 90 is blurred, is generated based on the first radiation image group 7a, so that even if the time to capture one radiation image is short, a sufficient amount of radiation can be obtained. Also, the second blurred image 8b, which is an image in which the subject 90 is blurred, is generated based on the second radiation image group 7b, so that even if the time to capture one radiation image is short, a sufficient amount of radiation can be obtained. Therefore, the first blurred image 8a and the second blurred image 8b are images in which the subject 90 is blurred, but with little noise.

The first reference image 6a and the second reference image 6b are noisy, but the subject 90 is not blurred, so edge information of the subject 90 can be obtained from the first reference image 6a and the second reference image 6b. Also, the first blurred image 8a and the second blurred image 8b are images in which the subject 90 is blurred, but have little noise, so pixel value information of the subject 90 can be obtained. That is, by obtaining edge information of the subject 90 from the first reference image 6a and obtaining pixel value information of the subject 90 from the first blurred image 8a, it is possible to generate an image (first high-quality image 9a) in which noise is reduced and blur of the subject 90 is suppressed, even if the shooting time for shooting one radiation image is shortened. In addition, by acquiring edge information of the subject 90 from the second reference image 6b and acquiring pixel value information of the subject 90 from the second blurred image 8b, it is possible to generate an image (second high-quality image 9b) in which noise is reduced and blur of the subject 90 is suppressed, even if the shooting time for capturing each radiation image is shortened. Therefore, as described above, by generating the first high-quality image 9a from the first blurred image 8a from which a sufficient amount of radiation can be obtained and the first reference image 6a, and generating the second high-quality image 9b from the second blurred image 8b from which a sufficient amount of radiation can be obtained and the second reference image 6b, it is possible to generate an image in which degradation of image quality is suppressed, even if the shooting time for capturing each radiation image is shortened. As a result, it is possible to provide a radiation image generating method that can suppress degradation of the image quality of the reconstructed image 10 while shortening the shooting time.

In addition, in this embodiment, as described above, the radiographic imaging device 100 includes an imaging unit 1 that images the subject 90, and an image processing unit 21 that performs image processing on a plurality of radiographic images captured from a plurality of angles with different relative angles between the subject 90 and the imaging unit 1. The image processing unit 21 selects, from the plurality of radiographic images, a first reference image 6a captured at a first imaging angle and a second reference image 6b captured at a second imaging angle different from the first imaging angle, and selects a first radiographic image captured at a different imaging angle from the plurality of radiographic images within an angle range 30 that covers the imaging angles of the first reference image 6a captured at a first imaging angle and the second reference image 6b captured at a second imaging angle different from the first imaging angle. Of the image group 7a and the second radiographic image group 7b, a first blurred image 8a is generated based on the first radiographic image group 7a, and a second blurred image 8b is generated based on the second radiographic image group 7b; a first high-quality image 9a is generated based on the first reference image 6a and the first blurred image 8a, and a second high-quality image 9b is generated based on the second reference image 6b and the second blurred image 8b; and a reconstructed image 10 reconstructed to reflect the three-dimensional structure of the subject 90 is generated based on the first high-quality image 9a and the second high-quality image 9b.

As a result, it is possible to provide a radiographic imaging device 100 that can reduce the imaging time while suppressing deterioration in the image quality of the reconstructed image 10, similar to the above-mentioned radiographic image generating method.

In addition, the above embodiment has the following additional advantages:

That is, in the present embodiment, as described above, in the step of capturing a plurality of radiographic images, a plurality of radiographic images captured at a predetermined angular interval are acquired, and in the step of generating a first blurred image 8a and a second blurred image 8b, the first blurred image 8a is generated based on the first radiographic image group 7a captured at a plurality of different shooting angles that are consecutive at a predetermined angular interval from the shooting angle of the first reference image 6a among the first radiographic image group 7a, and the second blurred image 8b is generated based on the second radiographic image group 7b captured at a plurality of different shooting angles that are consecutive at a predetermined angular interval from the shooting angle of the second reference image 6b among the second radiographic image group 7b. As a result, when the number of first radiographic image groups 7a is the same in the configuration in which the first blurred image 8a is generated based on the first radiographic image group 7a that are not consecutive within the predetermined angular range 30 and the configuration of the present embodiment, the size of the angular range 30 can be made smaller in the configuration of the present embodiment. In addition, when the size of the angle range 30 is the same between the configuration in which the first blurred image 8a is generated based on a plurality of first radiographic image groups 7 that are not consecutive within a predetermined angle range 30 and the configuration in this embodiment, the configuration in this embodiment can increase the number of first radiographic image groups 7a used to generate the first blurred image 8a. Therefore, compared to the configuration in which the first blurred image 8a is generated based on the first radiographic image groups 7a that are not consecutive within the predetermined angle range 30, it is possible to suppress the blur of the subject 90 appearing in the first blurred image 8a from increasing. As a result, it is possible to suppress an increase in the amount of blur removed when generating the first high-quality image 9a, and therefore the image quality of the first high-quality image 9a can be improved. In addition, the image quality of the second high-quality image 9b can also be improved in the same way as the first high-quality image 9a.

In addition, in the present embodiment, as described above, in the step of generating the first blurred image 8a and the second blurred image 8b, the first blurred image 8a is generated based on the first radiographic image group 7a captured at a plurality of different continuous shooting angles including shooting angles before and after a predetermined angle interval of the shooting angle of the first reference image 6a among the first radiographic image group 7a, and the second blurred image 8b is generated based on the second radiographic image group 7b captured at a plurality of different continuous shooting angles including shooting angles before and after a predetermined angle interval of the shooting angle of the second reference image 6b among the second radiographic image group 7b. This makes it possible to generate the first blurred image 8a based on the first radiographic image group 7a captured at a plurality of different continuous shooting angles including shooting angles before and after a predetermined angle interval of the shooting angle of the first reference image 6a, so that it is possible to further suppress the blur of the subject 90 appearing in the first blurred image 8a from increasing. As a result, it is possible to further suppress the amount of blur removed when generating the first high-quality image 9a from increasing, so that the image quality of the first high-quality image 9a can be further improved. Furthermore, the image quality of the second high-image-quality image 9b can be further improved, just like the first high-image-quality image 9a.

In addition, in this embodiment, as described above, in the step of generating the first high-quality image 9a and the second high-quality image 9b, the radiation images at each of the multiple imaging angles including the first imaging angle and the second imaging angle are sequentially used as reference images, and high-quality images are generated for each of the reference images. This makes it possible to generate a reconstructed image 10 based only on the first high-quality image 9a and the second high-quality image 9b generated for each of the multiple radiation images. As a result, the image quality of the reconstructed image 10 can be improved compared to a configuration in which the reconstructed image 10 is generated based on the first high-quality image 9a, the second high-quality image 9b, and the radiation images.

In addition, in this embodiment, as described above, in the step of generating the first high-quality image 9a and the second high-quality image 9b, the first reference image 6a and the first blurred image 8a are input to the trained model 11 to generate the first high-quality image 9a, and the second reference image 6b and the second blurred image 8b are input to the trained model 11 to generate the second high-quality image 9b. This makes it possible to easily generate the first high-quality image 9a. Also, the second high-quality image 9b can be easily generated.

In addition, in this embodiment, as described above, in the step of generating the first blurred image 8a and the second blurred image 8b, the first radiographic image group 7a is added to generate the first blurred image 8a, and the second radiographic image group 7b is added to generate the second blurred image 8b. This makes it possible to easily generate the first blurred image 8a. Also, the second blurred image 8b can be easily generated.

In addition, in this embodiment, as described above, in the step of capturing multiple radiographic images, the relative angle between the subject 90 and the imaging unit 1 is changed, and multiple radiographic images are captured by capturing images at each angle. This makes it possible to suppress a decrease in the image quality of the reconstructed image 10 even when the imaging time at each angle is shortened while changing the relative angle between the subject 90 and the imaging unit 1. Therefore, the present invention can be applied to a radiographic image generating method in which imaging is performed while moving the subject 90 or the imaging unit 1.

In addition, in this embodiment, as described above, in the step of capturing a plurality of radiation images, a plurality of X-ray images 6 are captured as the plurality of radiation images, and in the step of generating a reconstructed image 10, a tomographic image is generated as the reconstructed image 10. This makes it possible to generate a tomographic image in which the image capturing time is shortened while suppressing degradation in image quality. Therefore, it is particularly preferable to apply the present invention to an image generation method for generating a tomographic image.

In addition, in this embodiment, as described above, the subject rotation mechanism 3 (relative angle change mechanism) that changes the relative angle is further provided, the imaging unit 1 includes an X-ray source 1a that irradiates X-rays and an X-ray detector 1b that detects the X-rays irradiated from the X-ray source 1a, and the image processing unit 21 is configured to acquire a plurality of X-ray images 6 as a plurality of radiation images, generate a first blurred image 8a and a second blurred image 8b based on the acquired plurality of X-ray images 6, and generate a reconstructed image 10. As a result, it is preferable to apply the present invention to a radiation image capturing device 100 that generates a reconstructed image 10 by capturing images while changing the relative angle between the subject 90 and the imaging unit 1 using the subject rotation mechanism 3 (relative angle change mechanism).

[Modification]
It should be noted that the embodiments disclosed herein are illustrative and not restrictive in all respects. The scope of the present invention is indicated by the claims, not by the description of the embodiments above, and further includes all modifications (variations) within the meaning and scope of the claims.

For example, in the above embodiment, an example of a configuration in which the image processing unit 21 generates the first high-quality image 9a using the trained model 11 has been shown, but the present invention is not limited to this. For example, the first high-quality image 9a may be generated without using the trained model 11, as in the image processing unit 120 provided in the radiographic imaging device 200 according to the modified example shown in FIG. 10.

The modified radiographic imaging device 200 differs from the radiographic imaging device 100 of the above embodiment in that it includes a computer 12 instead of a computer 2.

Computer 12 differs from computer 2 in the above embodiment in that computer 12 has second processor 12a instead of second processor 2b.

The second processor 12a differs from the second processor 2b in the above embodiment in that it has an image processing unit 120 instead of an image processing unit 21.

The image processing unit 120 generates a first high-quality image 9a based on the first reference image 6a and the first blurred image 8a. Specifically, the image processing unit 120 obtains image blur from the first blurred image 8a instead of from the learned model 11, removes the obtained image blur from the first blurred image 8a, and obtains edge information of the subject 90 from the first reference image 6a to generate the first high-quality image 9a. That is, the image processing unit 120 according to the modified example is configured to generate the first high-quality image 9a by calculation processing.

Next, referring to FIG. 11, a process in which the image processing unit 120 according to the modified example generates the first high-quality image 9a will be described. Note that the same reference numerals are used for processes similar to those of the image processing unit 21 according to the above embodiment, and detailed descriptions thereof will be omitted.

In steps 103a and 103b, the image processing unit 120 selects a first reference image 6a and obtains a first blurred image 8a corresponding to the selected first reference image 6a.

In step 103h, the image processing unit 120 generates a first high-quality image 9a based on the first reference image 6a and the first blurred image 8a. In step 103g, the image processing unit 120 generates the first high-quality image 9a by computational processing without using the trained model 11.

Then, the process proceeds to step 103d and step 103e, where the selected second reference image 6b and the second blurred image 8b corresponding to the second reference image 6b are obtained.

In step 103i, the image processing unit 120 generates a second high-quality image 9b based on the second reference image 6b and the second blurred image 8b. In step 103h, the image processing unit 120 generates the second high-quality image 9b by computational processing without using the trained model 11.

The other configurations of the image processing unit 120 according to the modified example are similar to the configurations of the image processing unit 21 according to the above embodiment. Furthermore, the effects of the radiographic image capturing device 200 according to the modified example are similar to the effects of the radiographic image capturing device 100 according to the above embodiment.

In addition, in the above embodiment, an example of a configuration in which the image processing unit 21 acquires, as the first radiographic image group 7a, a plurality of radiographic images captured at a plurality of different continuous shooting angles including shooting angles before and after the shooting angle of the first reference image 6a at a predetermined angle interval within the predetermined angle range 30, but the present invention is not limited to this. For example, the image processing unit may be configured to acquire, as the first radiographic image group, a plurality of radiographic images captured at a plurality of different non-consecutive shooting angles at a predetermined angle interval within the predetermined angle range 30. In addition, the image processing unit may be configured to acquire, as the second radiographic image group, a plurality of radiographic images captured at a plurality of different non-consecutive shooting angles at a predetermined angle interval.

In addition, in the above embodiment, an example of a configuration in which the subject rotation mechanism 3 rotates the subject 90 at 20 degree intervals as the predetermined angular interval has been shown, but the present invention is not limited to this. The predetermined angular interval when the subject rotation mechanism 3 rotates the subject 90 may be an angle other than 20 degrees. For example, the subject rotation mechanism 3 may rotate the subject 90 at 1 degree intervals as the predetermined angular interval. The predetermined angular interval may be set to any value.

In the above embodiment, the image processing unit 21 uses a total of five images, including one first reference image 6a and four X-ray images 6 included in the first radiation image group 7a, when generating the first blurred image 8a. However, the present invention is not limited to this. The image processing unit 21 may use any number of images when generating the first blurred image 8a as long as it uses one first reference image 6a and the first radiation image group 7a. For example, the image processing unit 21 may generate the first blurred image 8a using a total of three images, including one first reference image 6a and two first radiation image groups 7a. Note that as the number of images used when generating the first blurred image 8a increases, the pixel value of the first blurred image 8a increases. On the other hand, as the number of images used when generating the first blurred image 8a increases, the blur of the subject 90 appearing in the first blurred image 8a increases. In other words, there is a trade-off between the pixel value of the first blurred image 8a and the blur of the subject 90. Furthermore, there is a trade-off between the pixel value of the second blurred image 8b and the blur of the subject 90.

In addition, in the above embodiment, an example of a configuration in which the computer 2 includes a first processor 2a and a second processor 2b is shown, but the present invention is not limited to this. For example, the computer 2 may be configured to include only one processor. When the computer 2 includes only one processor, the processor may be configured to function as the control unit 20 and the image processing unit 21 by executing a program stored in the memory unit 2c.

In the above embodiment, the radiographic imaging device 100 is a so-called CT imaging device that captures images while rotating the subject 90, but the present invention is not limited to this. For example, the radiographic imaging device 100 may be configured as a PET (Positron Emission Tomography) imaging device, a SPECT (single photon emission CT) imaging device, or a tomosynthesis imaging device.

In the above embodiment, an example of a configuration in which the X-ray detector 1b outputs an image signal based on the amount of irradiated X-rays is shown, but the present invention is not limited to this. For example, the X-ray detector 1b may be configured as a so-called photon-counting type detector that detects X-rays as photons.

In the above embodiment, an example of a configuration in which the image processing unit 21 generates the first blurred image 8a by adding the first reference image 6a and the first radiographic image group 7a has been shown, but the present invention is not limited to this. For example, the image processing unit 21 may be configured to generate the first blurred image 8a by taking an additive average of the first reference image 6a and the first radiographic image group 7a. Also, the image processing unit 21 may be configured to generate the second blurred image 8b by taking an additive average of the second reference image 6b and the second radiographic image group 7b.

In addition, in the above embodiment, an example of a configuration in which the image processing unit 21 generates the first blurred image 8a and then generates the second blurred image 8b has been shown, but the present invention is not limited to this. For example, the image processing unit 21 may be configured to generate the first blurred image 8a after generating the second blurred image 8b. Furthermore, the image processing unit 21 may be configured to generate the first blurred image 8a and the second blurred image 8b in parallel (by parallel processing).

In addition, in the above embodiment, an example of a configuration in which the image processing unit 21 generates the first high-quality image 9a and then generates the second high-quality image 9b is shown, but the present invention is not limited to this. For example, the image processing unit 21 may be configured to generate the first high-quality image 9a after generating the second high-quality image 9b. Furthermore, the image processing unit 21 may be configured to generate the first high-quality image 9a and the second high-quality image 9b in parallel (by parallel processing).

In addition, in the above embodiment, an example of a configuration in which the image processing unit 21 acquires an X-ray image 6 as a radiation image has been shown, but the present invention is not limited to this. For example, the image processing unit 21 may be configured to acquire an image captured by a PET imaging device, an image captured by a SPECT imaging device, or the like as a radiation image.

In addition, in the above embodiment, an example of a configuration in which the subject rotation mechanism 3 rotates the subject 90 in a rotational direction around the axis 80 has been shown, but the present invention is not limited to this. For example, the subject rotation mechanism 3 may be configured to rotate the imaging unit 1 in a rotational direction around the axis 80.

In the above embodiment, the radiographic imaging device 100 is used for non-destructive testing, but the present invention is not limited to this. For example, the radiographic imaging device 100 may be used for medical purposes. That is, the radiographic imaging device 100 may be used to capture an image of a patient (human) as the subject 90.

[Aspects]
It will be appreciated by those skilled in the art that the exemplary embodiments described above are examples of the following aspects.

(Item 1)
capturing a plurality of radiographic images at a plurality of imaging angles that differ in relative angle between the subject and the imaging unit;
selecting a first reference image captured at a first imaging angle and a second reference image captured at a second imaging angle different from the first imaging angle from the plurality of radiation images;
selecting a first group of radiographic images and a second group of radiographic images captured at different imaging angles within an angle range that covers the imaging angles of each of the selected reference images;
generating a first blurred image and a second blurred image based on each of the selected radiation image groups;
generating a first high-quality image based on the first reference image and the first blurred image, and generating a second high-quality image based on the second reference image and the second blurred image;
generating a reconstructed image based on the first high quality image and the second high quality image so as to reflect a three-dimensional structure of the subject.

(Item 2)
In the step of capturing a plurality of radiation images, the plurality of radiation images are captured by capturing images at a predetermined angular interval;
2. The radiological image generating method according to item 1, wherein in the step of generating the first blurred image and the second blurred image, the first blurred image is generated based on the first radiological image group captured at a plurality of different imaging angles that are consecutive at a predetermined angle interval from the imaging angle of the first reference image among the first radiological image group, and the second blurred image is generated based on the second radiological image group captured at a plurality of different imaging angles that are consecutive at a predetermined angle interval from the imaging angle of the second reference image among the second radiological image group.

(Item 3)
3. The radiological image generating method according to item 2, wherein in the step of generating the first blurred image and the second blurred image, the first blurred image is generated based on the first radiological image group captured at a plurality of different continuous imaging angles including imaging angles before and after a predetermined angle interval from an imaging angle of the first reference image among the first radiological image group, and the second blurred image is generated based on the second radiological image group captured at a plurality of different continuous imaging angles including imaging angles before and after a predetermined angle interval from an imaging angle of the second reference image among the second radiological image group.

(Item 4)
4. The radiation image generating method according to any one of items 1 to 3, wherein in the step of generating the first high-quality image and the second high-quality image, the radiation images at each of a plurality of imaging angles including the first imaging angle and the second imaging angle are sequentially set as the reference image, and a high-quality image is generated for each of the reference images.

(Item 5)
The radiation image generating method according to any one of claims 1 to 4, wherein in the step of generating the first high-quality image and the second high-quality image, the first high-quality image is generated by inputting the first reference image and the first blurred image to a trained model, and the second high-quality image is generated by inputting the second reference image and the second blurred image to the trained model.

(Item 6)
6. The method for generating a radiological image according to any one of items 1 to 5, wherein in the step of generating the first blurred image and the second blurred image, the first blurred image is generated by adding the first group of radiological images, and the second blurred image is generated by adding the second group of radiological images.

(Item 7)
7. The method for generating a radiographic image according to any one of items 1 to 6, wherein in the step of capturing a plurality of radiographic images, the relative angle between the subject and the imaging unit is changed and imaging is performed at each angle, thereby capturing a plurality of radiographic images.

(Item 8)
In the step of capturing a plurality of radiation images, a plurality of X-ray images are captured as the plurality of radiation images;
8. The radiation image generating method according to any one of items 1 to 7, wherein in the step of generating the reconstructed image, a tomographic image is generated as the reconstructed image.

(Item 9)
An imaging unit that captures an image of a subject;
an image processing unit that performs image processing on a plurality of radiation images captured from a plurality of angles having different relative angles between the subject and the imaging unit,
The image processing unit includes:
a first radiographic image group and a second radiographic image group captured at different imaging angles within an angle range covering imaging angles of a first reference image captured at a first imaging angle and a second reference image captured at a second imaging angle different from the first imaging angle from the plurality of radiographic images, the first blurred image being generated based on the first radiographic image group, and the second blurred image being generated based on the second radiographic image group;
generating the first high-quality image based on the first reference image and the first blurred image, and generating the second high-quality image based on the second reference image and the second blurred image;
a radiographic imaging device configured to generate a reconstructed image based on the first high-quality image and the second high-quality image so as to reflect a three-dimensional structure of the subject;

(Item 10)
A relative angle changing mechanism for changing the relative angle is further provided,
the imaging unit includes an X-ray source that irradiates X-rays and an X-ray detector that detects the X-rays irradiated from the X-ray source;
10. The radiation image capturing device according to item 9, wherein the image processing unit is configured to acquire a plurality of X-ray images as the plurality of radiation images, generate the first blurred image and the second blurred image based on the acquired plurality of X-ray images, and generate the reconstructed image.

1 Imaging unit 1a X-ray source 1b X-ray detector 3 Object rotation mechanism (relative angle change mechanism)
Reference Signs List 6 X-ray image 7a First radiographic image group 7b Second radiographic image group 8a First blurred image 8b Second blurred image 9a First high-quality image 9b Second high-quality image 10 Reconstructed image 11 Trained model 21, 120 Image processing unit 30 Predetermined angle range 90 Subject 100, 200 Radiographic image capturing device

Claims (10)

capturing a plurality of radiographic images at a plurality of imaging angles that differ in relative angle between the subject and the imaging unit;
selecting a first reference image captured at a first imaging angle and a second reference image captured at a second imaging angle different from the first imaging angle from the plurality of radiation images;
selecting a first group of radiographic images and a second group of radiographic images captured at different imaging angles within an angle range that covers the imaging angles of each of the selected reference images;
generating a first blurred image and a second blurred image based on each of the selected radiation image groups;
generating a first high-quality image based on the first reference image and the first blurred image, and generating a second high-quality image based on the second reference image and the second blurred image;
generating a reconstructed image based on the first high quality image and the second high quality image so as to reflect a three-dimensional structure of the subject. In the step of capturing a plurality of radiation images, the plurality of radiation images are captured by capturing images at a predetermined angular interval;
2. The radiological image generating method according to claim 1, wherein in the step of generating the first blurred image and the second blurred image, the first blurred image is generated based on the first radiological image group captured at a plurality of different imaging angles that are successive at a predetermined angle interval from the imaging angle of the first reference image among the first radiological image group, and the second blurred image is generated based on the second radiological image group captured at a plurality of different imaging angles that are successive at a predetermined angle interval from the imaging angle of the second reference image among the second radiological image group. The radiological image generating method according to claim 2, wherein in the step of generating the first blurred image and the second blurred image, the first blurred image is generated based on the first radiological image group captured at a plurality of different continuous shooting angles including shooting angles before and after a predetermined angle interval of the shooting angle of the first reference image from among the first radiological image group, and the second blurred image is generated based on the second radiological image group captured at a plurality of different continuous shooting angles including shooting angles before and after a predetermined angle interval of the shooting angle of the second reference image from among the second radiological image group. The radiological image generating method according to any one of claims 1 to 3, wherein in the step of generating the first high-quality image and the second high-quality image, the radiological images at each of a plurality of shooting angles including the first shooting angle and the second shooting angle are sequentially set as reference images, and a high-quality image is generated for each of the reference images. The radiological image generating method according to any one of claims 1 to 4, wherein in the step of generating the first high-quality image and the second high-quality image, the first high-quality image is generated by inputting the first reference image and the first blurred image to a trained model, and the second high-quality image is generated by inputting the second reference image and the second blurred image to the trained model. The radiological image generating method according to any one of claims 1 to 5, wherein in the step of generating the first blurred image and the second blurred image, the first group of radiological images is added to generate the first blurred image, and the second group of radiological images is added to generate the second blurred image. The radiation image generating method according to any one of claims 1 to 6, wherein in the step of capturing a plurality of radiation images, the relative angle between the subject and the imaging unit is changed while capturing images at each angle, thereby capturing a plurality of radiation images. In the step of capturing a plurality of radiation images, a plurality of X-ray images are captured as the plurality of radiation images;
8. The radiation image generating method according to claim 1, wherein in the step of generating a reconstructed image, a tomographic image is generated as the reconstructed image. An imaging unit that captures an image of a subject;
an image processing unit that performs image processing on a plurality of radiation images captured from a plurality of angles where the relative angle between the subject and the imaging unit is different,
The image processing unit includes:
a first radiographic image group and a second radiographic image group captured at different imaging angles within an angle range covering imaging angles of a first reference image captured at a first imaging angle and a second reference image captured at a second imaging angle different from the first imaging angle from the plurality of radiographic images, the first radiographic image group and the second radiographic image group being captured at different imaging angles from each other;
generating a first high-quality image based on the first reference image and the first blurred image, and generating a second high-quality image based on the second reference image and the second blurred image;
a radiographic imaging device configured to generate a reconstructed image based on the first high-quality image and the second high-quality image so as to reflect a three-dimensional structure of the subject; A relative angle changing mechanism for changing the relative angle is further provided,
the imaging unit includes an X-ray source that irradiates X-rays and an X-ray detector that detects the X-rays irradiated from the X-ray source;
10. The radiographic imaging device according to claim 9, wherein the image processing unit is configured to acquire a plurality of X-ray images as the plurality of radiographic images, generate the first blurred image and the second blurred image based on the acquired plurality of X-ray images, and generate the reconstructed image.

Images