Image processing apparatus, image processing method, and recording medium

Abstract

To generate an image having a voxel value equivalent to that of an image by a CT apparatus equipped with a conventional energy integration type detector, and improved in ease of distinguishing different substances having an equivalent HU value.SOLUTION: An image processing apparatus comprising: a material decomposition information obtaining unit configured to obtain material decomposition information obtained by decomposing one or more materials from detection data corresponding to a plurality of X-ray energy ranges obtained by using an X-ray CT apparatus; a first image obtaining unit configured to obtain a first image obtained by performing processing including energy integration on the detection data; and a second image generating unit configured to generate a second image by performing image processing on the first image.SELECTED DRAWING: Figure 3

Description

[Technical Field]

[0001] The disclosure of this specification relates to an image processing device, an image processing method, an image processing program, and a recording medium. [Background technology]

[0002] In recent years, devices capable of detecting the amount of X-ray transmission in each X-ray energy range, such as photon-counting X-ray CT devices (PCCT: Photon Counting Computed Tomography) and dual-energy X-ray CT devices (DECT: Dual Energy Computed Tomography), have been put into practical use. By integrating and reconstructing the detection data of PCCT or DECT by X-ray energy, an integral image can be obtained. This integral image is equivalent to an image reconstructed from the detection data of a CT device equipped with a conventional energy-integrating detector, and is used as a reference when interpreting images.

[0003] Patent Document 1 discloses a technology for discriminating materials by using the characteristics of materials whose X-ray attenuation coefficients differ depending on the X-ray energy region. Patent Document 2 also discloses a technology for generating images, such as iodine-enhanced images and iodine-suppressed images, that emphasize or suppress regions of specific materials from detection data from PCCT or DECT. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] JP 2014-14445 A [Patent Document 2] Patent Publication No. 2021-13489 Summary of the Invention [Problem to be solved by the invention]

[0005] Creating iodine-enhanced or iodine-suppressed images makes it easier to distinguish a specific substance from other substances with the same HU (Hounsfield Unit) value. However, because the color and voxel values ​​(pixel values) of the area of ​​the specific substance are changed, the images differ from those of conventional CT scanners. Therefore, these images are often referred to when interpreting images for a limited purpose.

[0006] Therefore, the present invention aims to generate images that have voxel values ​​equivalent to those of images obtained by a CT device equipped with a conventional energy-integrated detector, while improving the ease of distinguishing between different materials that have equivalent HU values. [Means for solving the problem]

[0007] According to the present invention, there is provided an image processing apparatus including: a material decomposition information acquisition unit that acquires material decomposition information obtained by decomposing one or more materials from detection data corresponding to a plurality of X-ray energy regions acquired using an X-ray CT apparatus; a first image acquisition unit that acquires a first image obtained by performing processing including energy integration on the detection data; and a second image generation unit that performs image processing on the first image to generate a second image, wherein the image processing includes performing a filtering process based on the material decomposition information on each of regions in the first image that correspond to the material decomposition information, and performing a voxel value adjustment process that adjusts statistical values ​​obtained based on voxel values ​​of the regions that correspond to the material decomposition information so that they approach statistical values ​​obtained from voxel values ​​before the filtering process.

[0008] Furthermore, according to the present invention, there is provided an image processing method comprising: a material decomposition information acquisition step of acquiring material decomposition information obtained by decomposing one or more materials from detection data corresponding to a plurality of X-ray energy ranges acquired using an X-ray CT apparatus; a first image acquisition step of acquiring a first image obtained by performing processing including energy integration on the detection data; and a second image generation step of generating a second image by performing image processing on the first image, wherein the image processing includes: performing a filtering process based on the material decomposition information on each of the regions in the first image corresponding to the material decomposition information; and performing a voxel value adjustment process of adjusting statistical values ​​obtained based on voxel values ​​in the regions corresponding to the material decomposition information so that they approach statistical values ​​obtained from voxel values ​​before the filtering process.

[0009] The present invention also provides an image processing program for causing a computer to execute the image processing method, and a recording medium on which the image processing program is stored in a computer-readable format. [Effects of the Invention]

[0010] Integral images generated from PCCT or DECT can provide images that have voxel values ​​equivalent to images obtained from a CT device equipped with a conventional energy-integrated detector, while improving the ease of distinguishing between different materials with equivalent HU values. [Brief explanation of the drawings]

[0011] [Figure 1] FIG. 1 is a block diagram showing an example of an image processing system according to the present invention. [Figure 2] FIG. 2 is a block diagram showing an example of the hardware configuration of an image processing apparatus according to the present invention. [Figure 3] FIG. 3 is a block diagram showing an example of the functional configuration of the image processing apparatus according to the present invention. [Figure 4] FIG. 4 is a flow chart showing an example of an image processing method according to the present invention. [Figure 5] FIG. 5 is a flowchart showing an example of image processing in the second image generating step according to the present invention. [Figure 6] FIG. 6 is a conceptual diagram showing an example of image processing in the second image generating step according to the present invention. [Figure 7] FIG. 7 is a block diagram illustrating an example of a functional configuration of a trained model generation unit according to the second embodiment. [Figure 8] FIG. 8 is a flow diagram illustrating an example of machine learning in generating a trained model according to the second embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0012] The present invention will be described in detail below based on preferred embodiments thereof with reference to the accompanying drawings. Items described in other embodiments are given the same numbers and their description will be omitted. The configurations shown in the following embodiments are merely examples, and the present invention is not limited to the illustrated configurations.

[0013] <Embodiment 1> The image processing device according to the present invention includes a material decomposition information acquisition unit that acquires material decomposition information obtained by decomposing one or more materials from detection data corresponding to a plurality of X-ray energy regions acquired using an X-ray CT device; a first image acquisition unit that acquires a first image obtained by performing processing including energy integration on the detection data; and a second image generation unit that performs image processing on the first image to generate a second image, wherein the image processing includes performing a filtering process based on the material decomposition information on each of the regions in the first image that correspond to the material decomposition information, and performing a voxel value adjustment process that adjusts statistical values ​​obtained based on voxel values ​​(pixel values) of the regions that correspond to the material decomposition information so that they approach statistical values ​​obtained from voxel values ​​(pixel values) before the filtering process.

[0014] When a conventional CT image and an image that makes it easier to distinguish specific substances, such as an iodine-enhanced image, are displayed side by side, the viewer must move their eyes significantly, making it difficult to identify the correspondence. On the other hand, when the images are overlaid, the conventional CT image becomes difficult to see. Therefore, it is necessary to generate an image that makes it easy to distinguish between each substance and to easily associate the distinguishable substances with other substances. Using the image processing device according to the present invention, it is possible to obtain the effect of filter processing that makes it easier to distinguish and recognize voxel regions corresponding to each substance, while also generating an image that is more suitable for simultaneously viewing regions other than the voxel region corresponding to a specific substance.

[0015] (System Configuration) Fig. 1 is a block diagram showing an example of an image processing system according to the present invention. The image processing system 100 shown in Fig. 1 comprises an image processing device 101, an X-ray CT device 102, and a LAN (Local Area Network) 103 connecting these devices.

[0016] In this embodiment, the X-ray CT device 102 is a photon-counting type X-ray CT device. A dual energy type X-ray CT device (DECT) can also be used as the X-ray CT device.

[0017] The image processing device 101 generates and displays an image based on detection data acquired from the X-ray CT device 102. The LAN 103 is a network made up of communication devices conforming to standards such as IEEE (Institute of Electrical and Electronics Engineers) 802.3ab. The image processing device 101 may store the generated image in a server (not shown), and an image viewer (not shown) may read and display the image from the server. The generated image may also be used for purposes other than display, such as analysis and processing.

[0018] (Hardware configuration) Fig. 2 is a block diagram showing an example of the hardware configuration of an image processing device 101 according to the present invention. The image processing device 101 shown in Fig. 2 is composed of a storage medium 201, a ROM (Read Only Memory) 202, a CPU (Central Processing Unit) 203, and a RAM (Random Access Memory) 204. It also includes a LAN interface 205, an input interface 208, a display interface 206, and an internal bus 211. A keyboard 209 and a mouse 210 are connected to the image processing device 101 via the input interface 208. A display 207 is connected to the image processing device 101 via the display interface 206.

[0019] The storage medium 201 is a storage medium such as an SSD (Solid State Drive) that stores an OS (Operating System), processing programs for performing various processes according to this embodiment, and various information according to this embodiment. The ROM 202 stores a program, such as a BIOS (Basic Input Output System), for initializing hardware, reading the OS stored in the storage medium 201, and starting it up. The CPU 203 performs arithmetic processing when executing the BIOS, OS, and processing programs. The RAM 204 temporarily stores the processing programs and various data when the CPU 203 executes the BIOS, OS, and processing programs. The LAN interface 205 is an interface that complies with standards such as IEEE802.3ab and is used for communication via the LAN 103 . The display 207 is a display device such as an LCD (Liquid Crystal Display) that displays a user interface screen. The display interface 206 converts screen information to be displayed on the display 207 into display control signals and outputs them to the display 207. The input interface 208 receives signals based on key presses, button clicks, coordinate movement, etc. by the user from a keyboard 209 and a mouse 210. An internal bus 211 transmits signals when communication is performed between the various blocks. The image processing device according to this embodiment may be connected to an input / output device integrally formed as, for example, a touch panel, instead of the display 207, keyboard 209, and mouse 210.

[0020] The hardware configuration of the image processing device 101 described above is an example and can be modified as appropriate. Devices other than those described above may be added, or some devices may not be provided. Also, some devices may be replaced with other devices having similar functions. Furthermore, some functions may be provided by other devices via a network, and the functions constituting this embodiment may be distributed and realized among multiple devices.

[0021] (Functional configuration) FIG. 3 is a block diagram showing an example of the functional configuration of the image processing device 101 according to the present invention. The image processing device 101 includes a material decomposition information acquisition unit 310, a first image acquisition unit 312, and a second image generation unit 314. The image processing device 101 may further include a detection data acquisition unit 300, a trained model generation unit 316, an input unit 317, a display unit 318, and a memory unit 319. Hereinafter, a method for implementing the image processing method according to this embodiment using the image processing device 101 according to this embodiment will be described.

[0022] (Image processing method) An image processing method according to the present invention includes a material decomposition information acquisition step of acquiring material decomposition information obtained by decomposing one or more materials from detection data corresponding to a plurality of X-ray energy regions acquired using an X-ray CT apparatus; a first image acquisition step of acquiring a first image obtained by performing processing including energy integration on the detection data; and a second image generation step of generating a second image by performing image processing on the first image, wherein the image processing includes performing a filtering process on each of the regions in the first image corresponding to the material decomposition information based on the material decomposition information, and performing a voxel value adjustment process of adjusting statistical values ​​obtained based on voxel values ​​in the regions corresponding to the material decomposition information so that they approach statistical values ​​obtained from voxel values ​​before filtering. An image processing program according to the present invention is an image processing program for causing a computer to execute the image processing method according to the present invention. A recording medium according to the present invention is a recording medium storing the image processing program according to the present invention in a computer-readable format.

[0023] 4 is a flow diagram showing an example of an image processing method according to this embodiment. The image processing method according to this embodiment includes a material decomposition information acquisition step as step S402, a first image acquisition step as step S403, and a second image generation step as step S404. The image processing method according to this embodiment can further include a detection data acquisition step as step S401, and an image display step as step S405.

[0024] Step S401 is a detection data acquisition step in which the detection data acquisition unit 300 acquires detection data corresponding to a plurality of X-ray energy regions, which is acquired using the X-ray CT device 102, from the X-ray CT device 102.

[0025] Step S402 is a material decomposition information acquisition step in which the material decomposition information acquisition unit 310 acquires material decomposition information obtained by decomposing one or more materials from the detection data acquired in step S401.

[0026] Step S403 is a first image acquisition step in which the first image acquisition unit 312 acquires a first image obtained by performing processing including energy integration on the detection data acquired in step S401.

[0027] Step S404 is a second image generation step in which the second image generation unit 314 generates a second image by performing image processing on the first image acquired in step S403 based on the material decomposition information acquired in step S402. The image processing performed by the second image generation unit 314 includes performing filtering based on the material decomposition information on each of the regions in the first image corresponding to the material decomposition information, and performing voxel value adjustment processing to adjust statistical values ​​obtained based on voxel values ​​of the regions corresponding to the material decomposition information so that they approach statistical values ​​obtained from voxel values ​​of the regions corresponding to the material decomposition information before filtering.

[0028] Step S405 is an image display step in which the display unit 318 displays on the display 207 the second image generated in step S404. Each step will be described in detail below.

[0029] (Step S401: Detection data acquisition process) The detection data acquisition unit 300 acquires detection data corresponding to a plurality of X-ray energy regions acquired by the X-ray CT device 102 from the X-ray CT device 102. Here, the detection data is a sinogram for each X-ray energy region. A sinogram is data in which signals based on photon detection or X-ray detection for each energy region acquired by a detector are arranged with the arrangement of detectors as the X axis and the projection position (rotation angle) as the Y axis. In this embodiment, the detection data is acquired in accordance with DICOM (Digital Imaging and Communications in Medicine), an international standard for medical data communication. Note that the detection data may be acquired in accordance with other communication standards such as HTTP (Hypertext Transfer Protocol). Furthermore, the detection data may be transmitted from a device other than the X-ray CT device 102, such as a PACS (Picture Archiving and Communication System).

[0030] (Step S402: Material decomposition information acquisition process) The material decomposition information acquisition unit 310 acquires material decomposition information obtained by decomposing one or more materials from the detection data acquired by the detection data acquisition unit 300. In this embodiment, the material decomposition information is information indicating voxel regions corresponding to one or more individual materials based on the detection data. That is, the material decomposition information is information indicating regions within an image corresponding to specific materials, and can be a binary mask image in which each voxel value indicates the presence or absence of a specific material. Note that the material decomposition information can also be a multi-value mask image in which each voxel value indicates the type of material. The material decomposition information can also be coordinate information corresponding to the contour of a region. Note that, for example, multiple materials that do not need to be distinguished but are similar in energy can be assigned the same material decomposition information. The material decomposition information acquisition unit 310 may acquire the material decomposition information from an external device such as the storage unit 319 or a server. In this case, for example, the detection data can be transmitted to the server, the server can generate material decomposition information from the detection data, and the material decomposition information acquisition unit 310 can receive the material decomposition information generated by the server.

[0031] Material decomposition information is generated, for example, using the following procedure. First, the detection data is reconstructed for each X-ray energy range to obtain monochromatic X-ray image data for each X-ray energy range. Next, the likelihood of each voxel in the image being a particular material is evaluated based on the degree of match between the distribution of voxel values ​​in the X-ray energy direction obtained from the monochromatic X-ray image data and the distribution of HU values ​​in the X-ray energy direction estimated from the material's attenuation coefficient for each X-ray energy range. For example, if the X-ray energy ranges are 80 kV and 140 kV, the ratio of voxel values ​​estimated from the attenuation coefficient is approximately 1.7 for iodine and approximately 1.3 for calcium. In this case, in the 80 kV and 140 kV monochromatic X-ray images, a voxel value ratio close to 1.7 is evaluated as highly similar to iodine, and a voxel value ratio close to 1.3 is evaluated as highly similar to calcium. Next, for each voxel in the image, materials whose degree of match meets a predetermined standard are identified. Finally, the value "1" is set at the position of each voxel in the mask image corresponding to the identified substance, indicating that the identified substance is present. Note that the number of X-ray energy regions, the ratio of voxel values ​​between images for each X-ray energy region, and the types and number of substances to be identified are merely examples and are not limited to these.

[0032] (Step S403: First image acquisition step) In this embodiment, the first image acquisition unit 312 generates a first image from the detection data. The first image can be, for example, an X-ray energy integrated image generated by integrating the detection data in the X-ray energy direction (energy integration) and using a known back projection method or iterative image reconstruction method from the energy-integrated sinogram. The first image acquisition unit 312 may acquire the first image from the storage unit 319 or an external device such as a server. Note that the material decomposition information acquisition step in step S402 and the first image acquisition step in step S403 may be performed in either order, or may be performed simultaneously.

[0033] (Step S404: Second image generation process) The second image generation unit 314 performs image processing including filtering and voxel value adjustment on the first image based on the material decomposition information acquired in step S402, to generate a second image.

[0034] The image processing can include generating an intermediate image consisting of only the region corresponding to the material decomposition information that has been subjected to the voxel value adjustment processing, and integrating the intermediate image with the first image.

[0035] In this embodiment, the second image generation unit 314 performs filtering and voxel value adjustment on each region corresponding to the material decomposition information, and obtains an intermediate image consisting of only the regions corresponding to the material decomposition information that have been subjected to voxel value adjustment. Then, the second image generation unit 314 generates a second image based on the intermediate image.

[0036] In this embodiment, the second image generation unit 314 generates an intermediate image and combines it with the first image to generate the second image, but it is also possible to generate the second image based on the result of image processing without generating an intermediate image. Details of the processing in the second image generation unit 314 will be described later using Figures 5 and 6.

[0037] (Step S405: Image display process) The display unit 318 displays the second image on the display 207. The display unit 318 may display the second image in a predetermined area of ​​a user interface screen (not shown). The first image acquisition unit 312 and the second image generation unit 314 may transmit input data to an external server and receive the results of processing by the external server.

[0038] (Details of Step S404 (Second Image Generation Process)) FIG. 5 is a flowchart showing an example of image processing that the second image generating unit 314 performs based on the material decomposition information in step S404.

[0039] In step S501, the second image generation unit 314 acquires image processing conditions (filter type, parameters, voxel value adjustment method) for each material based on the material decomposition information for each material acquired in step S402. In this embodiment, the image processing conditions are stored in advance in the storage unit 319 and are read out from the storage unit 319.

[0040] In step S502, as image processing, the second image generation unit 314 first performs filtering on the region for each material in the first image based on the material decomposition information acquired in step S402. Here, the filtering is based on the filter type and parameters of the image processing conditions acquired in step S501.

[0041] In step S503, the second image generation unit 314 performs voxel value adjustment processing as image processing on the image that has been filtered in step S502. Here, the voxel value adjustment processing complies with the voxel value adjustment method of the image processing conditions acquired in step S501.

[0042] In step S504, the intermediate image consisting of only the region corresponding to the material decomposition information that has been subjected to the voxel value adjustment process is integrated with the first image, as will be described in detail below with reference to FIG.

[0043] FIG. 6 is a diagram showing an example of an overview of image processing in the second image generating step according to this embodiment. The second image generation unit 314 performs image processing on the first image 313-1 based on the material decomposition information 311-1 to 311-3, thereby obtaining intermediate images 315-1 to 315-3 for each piece of material decomposition information. Here, image processing corresponding to the material is performed on each region of the first image 313-1 indicated by the material decomposition information 311-1 to 311-3. First, filtering is performed on the regions of the first image 313-1 corresponding to the material decomposition information based on the material decomposition information 311-1 to 311-3. Then, voxel value adjustment processing is performed to adjust the statistical values ​​obtained based on the voxel values ​​of the regions corresponding to the material decomposition information so that they approach the statistical values ​​obtained from the voxel values ​​of the regions corresponding to the material decomposition information before filtering. That is, processing is performed so that the statistical values ​​of the voxel values ​​of the regions for each material approach the statistical values ​​of the voxel values ​​of the regions corresponding to the same material before filtering. The image processing is image processing related to texture, and is performed based on each image processing condition.

[0044] The image processing conditions include information on the type (filter type) of filter (image processing filter) in the filter processing, parameters for the filter of that type, and a method for adjusting voxel values ​​in the voxel value adjustment processing (voxel value adjustment method).

[0045] Specifically, the filter type of the filter process can be at least one filter type selected from the group consisting of contrast change, gamma correction, noise removal, sharpening, and smoothing. The parameter for contrast change is the rate at which contrast is increased or decreased. For gamma correction, the parameter is the gamma value. For noise removal using a median filter, the parameter is the aperture size. For sharpening using an unsharp mask, the parameters are the kernel size and strength k. For smoothing using a Gaussian filter, the parameters are the kernel size and standard deviation in the X and Y directions.

[0046] The voxel value adjustment method is a method of adjusting the statistical value obtained from the voxel values ​​after image processing of a region corresponding to a predetermined substance so that it approximately matches the statistical value obtained from the voxel values ​​before image processing. Here, the statistical value can be, for example, the average value or the median value. Methods include multiplying by a constant value or adding or subtracting so that the average value of the voxel values ​​after image processing of a region corresponding to a predetermined substance is approximately equal to the average value of the voxel values ​​before image processing.

[0047] For example, the voxel value adjustment process may be a process of adjusting the ratio of the statistical value (e.g., average value) of voxel values ​​after image processing to the statistical value (e.g., average value) of voxel values ​​before image processing (the statistical value (e.g., average value) of voxel values ​​after image processing / the statistical value (e.g., average value) of voxel values ​​before image processing) for a region corresponding to material decomposition information so that it falls within a predetermined range. In this case, the adjustment may be performed so that the ratio of the statistical value (e.g., average value) of voxel values ​​after image processing / the statistical value (e.g., average value) of voxel values ​​before image processing falls within a predetermined range. In this case, the adjustment may be performed so that the difference between the statistical value (e.g., average value) of voxel values ​​after image processing and the statistical value (e.g., average value) of voxel values ​​before image processing falls within a predetermined range.

[0048] Here, the image processing conditions may be a combination of multiple filter types and parameters. At least some of the filter types and parameters have different values ​​for each corresponding material decomposition information. In this embodiment, the material decomposition information acquisition unit 310 acquires two or more pieces of material decomposition information, and the filtering process may include applying filtering with filters having different parameters to regions corresponding to the two or more pieces of material decomposition information to be filtered. Furthermore, in this embodiment, the material decomposition information acquisition unit 310 acquires two or more pieces of material decomposition information, and the filtering process may include applying filtering with filters of different types to regions corresponding to the two or more pieces of material decomposition information to be filtered. In particular, for materials with similar HU values, filter types and their parameters are predefined and stored to facilitate distinction, for example, to increase the difference in contrast or sharpness. For example, in the case of iodine and calcium, image processing conditions are set such that a low-contrast smoothing filter is applied to the iodine region, and a high-contrast sharpening filter is applied to the calcium region. The image processing conditions listed here are merely examples, and other conditions may be used as long as they maintain the difference between the statistical values ​​obtained from the voxel values ​​before image processing within a region and the statistical values ​​obtained from the voxel values ​​after image processing within a certain range and improve the ease of distinguishing between materials. The image processing conditions may also be configured to be changeable by the user. In this case, the image processing conditions can be specified via the input unit 317.

[0049] The second image generation unit 314 generates a second image 330 by combining the intermediate images 315-1 to 315-3 (regions for each decomposed material) with a region 313-2 outside the region of the decomposed material in the first image. In this embodiment, addition is performed as the operation for combining. Note that the operation for combining is not limited to addition, and may be an operation in which voxel values ​​of regions corresponding to the material decomposition information in the first image 313-1 are replaced with voxel values ​​of corresponding regions in the intermediate images 315-1 to 315-3. The second image generating step may be performed without generating an intermediate image. Specifically, the first image may be subjected to image processing for each region based on the material decomposition information, and the results may be written directly into the regions of the first image 313-1 corresponding to the material decomposition information (regions for each decomposed material).

[0050] As described above, according to this embodiment, an integral image generated from PCCT or DECT can be obtained that has voxel values ​​equivalent to those of an image obtained from a CT device equipped with a conventional energy-integrated detector, while improving the ease of distinguishing between different materials with equivalent HU values.

[0051] (Modification of the first embodiment) In the first embodiment, the case where the material decomposition information acquired by the image processing device 101 is information indicating the presence or absence of a predetermined substance has been described. However, the material decomposition information is not limited to this, and may be a likelihood map indicating the likelihood of a predetermined substance.

[0052] In this case, the strength of application of the image processing conditions to the first image may be changed based on the likelihood of the material decomposition information. Specifically, if the likelihood is 100%, the image processing conditions may be applied with 100% strength, and if the likelihood is 50%, the image processing conditions may be applied with 50% strength.

[0053] According to this modified example, by changing the strength of application of the image processing conditions based on the likelihood of a predetermined substance, it is possible to weaken the distinction between substances when the substance likelihood is low.

[0054] <Embodiment 2> The image processing device according to the present invention can be an image processing device in which image processing conditions are acquired using a trained model, and the trained model is trained to output image processing conditions for generating an image in which the degree of separation of regions in the image corresponding to the material decomposition information is equal to or greater than a predetermined value, using as input material decomposition information obtained by decomposing one or more materials in detection data corresponding to a plurality of X-ray energy ranges acquired using an X-ray CT device, and an image obtained by performing processing on the detection data including energy integration.

[0055] In this embodiment, the image processing device 101 performs image processing using image processing conditions acquired by the second image generation unit 314 using a trained model generated by machine learning. Note that in this embodiment, the configuration of the image processing system 100, the hardware configuration and functional configuration of the image processing device 101, the flow of the image processing method, etc. are the same as those in embodiment 1. In this embodiment, the detection data acquisition process in step S401, the material decomposition information acquisition process in step S402, the first image acquisition process in step S403, and the image display process in step S405 are the same as those in embodiment 1, and therefore descriptions thereof will be omitted.

[0056] (Step S404: Machine learning process flow for generating a trained model to be used in the second image generation process) In this embodiment, the image processing device 101 has a trained model generation unit 316, which generates a trained model. The second image generation unit 314 acquires image processing conditions using the trained model. The trained model receives as input material decomposition information obtained by decomposing one or more materials in detection data corresponding to a plurality of X-ray energy ranges acquired using an X-ray CT device, and an image obtained by performing processing including energy integration on the detection data, and has been trained to output image processing conditions for generating an image in which the degree of separation of regions in the image corresponding to the material decomposition information is equal to or greater than a predetermined value.

[0057] The image processing conditions output by the trained model are used for image processing in the second image generation unit 314. Similar to the image processing conditions in the first embodiment, the image processing conditions output by the trained model consist of the filter type of the filter processing, parameters for that type of filter processing, and information on the voxel value adjustment method.

[0058] FIG. 7 is a diagram showing the functional configuration of the trained model generation unit 316 according to this embodiment. As shown in Figure 7, the trained model generation unit 316 includes a training detection data acquisition unit 700, a training material decomposition information acquisition unit 710, a training image acquisition unit 712, a training model acquisition unit 714, a training separation degree judgment image generation unit 720, a separation degree calculation unit 721, and a separation degree judgment unit 722.

[0059] FIG. 8 is a flow diagram showing an example of machine learning by the trained model generation unit 316 in the second embodiment.

[0060] In step S801, the learning detection data acquisition unit 700 acquires one or more pieces of learning detection data corresponding to a plurality of X-ray energy regions from the X-ray CT apparatus. Here, the method for acquiring the learning detection data can be the same as the detection data acquisition step in the first embodiment.

[0061] In step S802, the learning material decomposition information acquisition unit 710 acquires learning material decomposition information obtained by decomposing one or more materials in the one or more learning detection data acquired in step S801. Here, the method for acquiring (generating) the learning material decomposition information can be the same as the material decomposition information acquisition step in embodiment 1.

[0062] In step S803, the learning image acquisition unit 712 acquires a learning image obtained by performing processing including energy integration on the learning detection data. Here, the method of acquiring (generating) the learning image can be the same as the first image acquisition step in embodiment 1. Note that step S802 and step S802 may be performed either first or simultaneously.

[0063] In step S804, the learning separation degree judgment image generation unit 720 acquires image processing conditions including initial values ​​of parameters for filtering based on the learning material decomposition information acquired in step S802. Here, the initial values ​​of the parameters are all set to zero. Note that the initial values ​​of the parameters may be any other values, such as random values.

[0064] In step S805, the learning separation degree judgment image generating unit 720 generates a learning separation degree judgment image for the learning image acquired in step S803 using the image processing conditions acquired in step S804.

[0065] In step S806, the separability calculation unit 721 obtains the separability in the learning separation degree judgment image generated in step S805, based on one or more pieces of learning material decomposition information obtained in step S802.

[0066] The separability calculation unit 721 calculates the degree of separation of each material region corresponding to one or more pieces of material decomposition information for learning in the learning separation degree judgment image. The separability is an index representing the ease of distinguishing between a region corresponding to the material decomposition information and its neighboring surrounding regions. For example, the separability can be calculated based on the difference in luminance distribution between the inside and outside of the region corresponding to the material decomposition information. Specifically, the luminance distribution between the inside and outside of the region corresponding to the material decomposition information is normalized, and the absolute value of the difference in luminance distribution is integrated in the luminance value direction. The separability may also be calculated based on the strength of edges at the boundaries of the region corresponding to the material decomposition information, detected by edge detection. A Sobel filter, a Laplacian filter, a Canny filter, or the like may be used for edge detection. The separability may also be calculated based on the difference in spatial frequency spectrum between the inside and outside of the region corresponding to the material decomposition information. Specifically, spatial frequency spectra may be obtained by Fourier transforming the images inside and outside the region corresponding to the material decomposition information, and the absolute value of the difference in spatial frequency spectrum may be integrated in the spatial frequency direction.

[0067] In step S807, the separability determination unit 722 determines whether the separability acquired in step S806 is equal to or greater than a predetermined value. If the separability is equal to or greater than the predetermined value, the process ends, and if the separability is less than the predetermined value, the process proceeds to step S808. As the value of the separability, the user can specify a predetermined value via the input unit 317 depending on the ease of distinguishing between regions corresponding to the material decomposition information.

[0068] In step S808, the learning separation degree judgment image generation unit 720 acquires the next image processing conditions based on the search algorithm acquired by the learning model acquisition unit 714, and returns to step S805. Then, image processing is performed using the new image processing conditions, and a learning separation degree judgment image is generated again. Thereafter, steps S805 to S808 are repeated until the separability reaches a predetermined value or greater.

[0069] The learning separation degree judgment image generation unit 720 generates image processing conditions through machine learning using, for example, learning material decomposition information and learning images generated from multiple pieces of learning detection data. In machine learning, the parameters of each filter process constituting the image processing conditions are used as variables, and the parameters are adjusted so as to increase the separability calculated by the separability calculation unit 721. For parameter adjustment, known search algorithms such as linear search, binary tree search, grid search, and random search are used. Note that for parameter adjustment, known optimization algorithms such as steepest descent, SGD (Stochastic Gradient Descent), and Adam (Adaptive Moment Estimation) may also be used. Here, optimization algorithms can be used when the formula for calculating the separability is differentiable with respect to the parameters of each filter process, which are variables. Furthermore, for parameter adjustment, constraints may be set in advance on the parameter adjustment range based on smoothness, sharpness, etc. The adjustment range may also be specified by the user via the input unit 317. The algorithm for adjusting the parameters may be acquired via the learning model acquisition unit 714 and may be specified by the user via the input unit.

[0070] The trained model generated by the trained model generation unit 316 as described above is acquired by the second image generation unit 314, and the second image generation unit acquires image processing conditions using the trained model. At this time, the trained model receives the material decomposition information acquired by the material decomposition information acquisition unit 310 and the first image acquired by the first image acquisition unit 312 as inputs, and outputs the image processing conditions. In addition, the image processing conditions may be obtained using a trained model that has been machine-learned by a device other than the image processing device 101, and the image processing conditions may be stored in the memory unit 319 or an external server, etc., and the second image generation unit 314 may obtain the image processing conditions from the server and perform image processing.

[0071] According to this embodiment, in an integral image generated by PCCT or DECT, an image can be obtained that has voxel values ​​equivalent to those of an image obtained by a CT device equipped with a conventional energy-integrated detector, while improving the ease of distinguishing between different materials with equivalent HU values. Furthermore, by introducing separability, it becomes possible to obtain image processing conditions generated by a trained model obtained by machine learning, making it possible to obtain appropriate image processing conditions without manual intervention.

[0072] The disclosure of this embodiment includes the following configurations and methods. (Configuration 1) a material decomposition information acquisition unit that acquires material decomposition information obtained by decomposing one or more materials from detection data corresponding to a plurality of X-ray energy ranges acquired using the X-ray CT device; a first image acquisition unit that acquires a first image obtained by performing processing including energy integration on the detection data; a second image generation unit that performs image processing on the first image to generate a second image, The image processing performing a filtering process based on the material decomposition information on each of the regions in the first image corresponding to the material decomposition information; performing a voxel value adjustment process to adjust a statistical value obtained based on voxel values ​​of a region corresponding to the material decomposition information so that the statistical value obtained from the voxel values ​​before the filtering process approaches a statistical value obtained from the voxel values ​​before the filtering process. (Configuration 2) 2. The image processing device according to configuration 1, wherein the image processing includes generating an intermediate image consisting of only a region corresponding to the material decomposition information that has been subjected to the voxel value adjustment processing, and integrating the intermediate image with the first image. (Configuration 3) the voxel value adjustment process, wherein the statistical value is an average value, 3. The image processing device according to configuration 1 or 2, characterized in that the image processing device performs processing to adjust a ratio of an average value of voxel values ​​after the image processing to an average value of voxel values ​​before the image processing (average value of voxel values ​​after the image processing / average value of voxel values ​​before the image processing) for a region corresponding to the material decomposition information so that the ratio falls within a predetermined range. (Configuration 4) the voxel value adjustment process, wherein the statistical value is an average value, 4. The image processing device according to any one of configurations 1 to 3, wherein the image processing is a process of adjusting a difference between an average value of voxel values ​​after the image processing and an average value of voxel values ​​before the image processing for a region corresponding to the material decomposition information so that the difference falls within a predetermined range. (Configuration 5) The image processing conditions are acquired using a trained model, 5. The image processing device according to any one of configurations 1 to 4, wherein the trained model is trained to output image processing conditions for generating an image in which the degree of separation of regions corresponding to the material decomposition information is equal to or greater than a predetermined value, using as input material decomposition information obtained by decomposing one or more materials in detection data corresponding to a plurality of X-ray energy regions acquired using an X-ray CT device, and an image obtained by performing processing including energy integration on the detection data. (Configuration 6) 6. The image processing device according to configuration 5, wherein the degree of separation is calculated based on a difference in luminance distribution between the inside and outside of a region corresponding to the material decomposition information. (Configuration 7) 7. The image processing device according to configuration 5 or 6, wherein the degree of separation is calculated based on a difference in spatial frequency spectrum between the inside and outside of the region corresponding to the material decomposition information. (Configuration 8) 8. The image processing device according to any one of configurations 5 to 7, wherein the degree of separation is calculated based on edge strength at the boundary of the region corresponding to the material decomposition information. (Configuration 9) the material decomposition information acquisition unit acquires two or more pieces of material decomposition information, 9. The image processing device according to any one of configurations 1 to 8, wherein the filtering process includes performing filtering processes using filters with different parameters on regions corresponding to two or more pieces of material decomposition information to be subjected to the filtering process. (Configuration 10) the material decomposition information acquisition unit acquires two or more pieces of material decomposition information, 10. The image processing device according to any one of configurations 1 to 9, wherein the filtering process includes performing filtering processes using different types of filters on regions corresponding to two or more pieces of material decomposition information to be subjected to the filtering process. (Configuration 11) 11. The image processing device according to any one of configurations 1 to 10, wherein the filtering process includes performing filtering on an area corresponding to the material decomposition information to be subjected to the filtering process using a filter related to at least one selected from the group consisting of contrast change, gamma correction, noise removal, sharpening, and smoothing. (Method 12) a material decomposition information acquisition step of acquiring material decomposition information obtained by decomposing one or more materials from detection data corresponding to a plurality of X-ray energy ranges acquired using the X-ray CT device; a first image acquisition step of acquiring a first image obtained by performing processing including energy integration on the detection data; a second image generating step of generating a second image by performing image processing on the first image, The image processing performing a filtering process based on the material decomposition information on each of the regions in the first image corresponding to the material decomposition information; performing a voxel value adjustment process to adjust a statistical value obtained based on voxel values ​​of a region corresponding to the material decomposition information so that the statistical value is closer to a statistical value obtained from voxel values ​​before the filtering process. (Configuration 13) An image processing program for causing a computer to execute the image processing method according to Method 12. (Configuration 14) A recording medium storing the image processing program according to method 13 in a computer-readable format. [Explanation of symbols]

[0073] 101 Image processing device 310 Material Decomposition Information Acquisition Unit 312 First image acquisition unit 314... Second image generation unit S402: Material decomposition information acquisition process S403: First image acquisition step S404: Second image generation step

Claims

1. a material decomposition information acquisition unit that acquires material decomposition information obtained by decomposing one or more materials from detection data corresponding to a plurality of X-ray energy ranges acquired using the X-ray CT device; a first image acquisition unit that acquires a first image obtained by performing processing including energy integration on the detection data; a second image generation unit that performs image processing on the first image to generate a second image, The image processing performing a filtering process based on the material decomposition information on each of the regions in the first image corresponding to the material decomposition information; performing a voxel value adjustment process to adjust a statistical value obtained based on voxel values ​​of a region corresponding to the material decomposition information so that the statistical value obtained from the voxel values ​​before the filtering process approaches a statistical value obtained from the voxel values ​​before the filtering process.

2. 2. The image processing device according to claim 1, wherein the image processing includes generating an intermediate image consisting of only a region corresponding to the material decomposition information that has been subjected to the voxel value adjustment processing, and integrating the intermediate image with the first image.

3. the voxel value adjustment process, wherein the statistical value is an average value, 2. The image processing device according to claim 1, wherein the adjustment is performed so that a ratio of an average value of voxel values ​​after the image processing to an average value of voxel values ​​before the image processing (average value of voxel values ​​after the image processing / average value of voxel values ​​before the image processing) falls within a predetermined range for a region corresponding to the material decomposition information.

4. the voxel value adjustment process, wherein the statistical value is an average value, 2. The image processing device according to claim 1, wherein the image processing is performed to adjust a difference between an average value of voxel values ​​after the image processing and an average value of voxel values ​​before the image processing for a region corresponding to the material decomposition information so that the difference falls within a predetermined range.

5. The image processing conditions are acquired using a trained model, 2. The image processing device according to claim 1, wherein the trained model is trained to output image processing conditions for generating an image in which a degree of separation of regions corresponding to the material decomposition information is equal to or greater than a predetermined value, using as input material decomposition information obtained by decomposing one or more materials in detection data corresponding to a plurality of X-ray energy ranges acquired using an X-ray CT device, and an image obtained by performing processing including energy integration on the detection data.

6. 6. The image processing apparatus according to claim 5, wherein the degree of separation is calculated based on a difference in luminance distribution between the inside and outside of the region corresponding to the material decomposition information.

7. 6. The image processing apparatus according to claim 5, wherein the degree of separation is calculated based on a difference in spatial frequency spectrum between the inside and outside of the region corresponding to the material decomposition information.

8. 6. The image processing apparatus according to claim 5, wherein the degree of separation is calculated based on edge strength at the boundary of the region corresponding to the material decomposition information.

9. the material decomposition information acquisition unit acquires two or more pieces of material decomposition information, 9. The image processing device according to claim 1, wherein the filtering process includes performing filtering processes using filters with different parameters on regions corresponding to two or more pieces of material decomposition information to be subjected to the filtering process.

10. the material decomposition information acquisition unit acquires two or more pieces of material decomposition information, 9. The image processing device according to claim 1, wherein the filtering process includes performing filtering processes using different types of filters on regions corresponding to two or more pieces of material decomposition information to be subjected to the filtering process.

11. 9. The image processing device according to claim 1, wherein the filtering process includes performing filtering on an area corresponding to the material decomposition information to be filtered using a filter related to at least one selected from the group consisting of contrast modification, gamma correction, noise removal, sharpening, and smoothing.

12. a material decomposition information acquisition step of acquiring material decomposition information obtained by decomposing one or more materials from detection data corresponding to a plurality of X-ray energy ranges acquired using the X-ray CT device; a first image acquisition step of acquiring a first image obtained by performing processing including energy integration on the detection data; a second image generating step of generating a second image by performing image processing on the first image, The image processing performing a filtering process based on the material decomposition information on each of the regions in the first image corresponding to the material decomposition information; performing a voxel value adjustment process to adjust a statistical value obtained based on voxel values ​​of a region corresponding to the material decomposition information so that the statistical value is closer to a statistical value obtained from voxel values ​​before the filtering process.

13. An image processing program for causing a computer to execute the image processing method according to claim 12.

14. A recording medium storing the image processing program according to claim 13 in a computer-readable format.

Images