Image processing device, image processing system, and image processing method

Abstract

The present invention employs a technique different from conventional techniques to improve the quality of an image obtained by nuclear medicine imaging. In an image processing device (1), a first reconstruction processing unit (11) generates an initial image representing a concentration distribution of a radioactive substance by means of reconstruction processing with respect to list-mode data acquired by a PET device (90). A scatter component deriving unit (12) employs a scatter estimation simulation to derive, from the initial image, scatter component projection data representing a scatter component of the radiation. A second reconstruction processing unit (13) subjects the scatter component projection data to reconstruction processing to generate a low-resolution scatter image which represents the scatter component and which has a lower resolution than the initial image. An upscale processing unit (14) upscales the low-resolution scatter image to generate an upscaled scatter image having the same resolution as the initial image. An image correcting unit (15) subtracts the upscaled scatter image from the initial image to generate a scatter corrected image.

Description

[Technical Field]

[0001] One aspect of the present invention relates to an image processing device that processes images obtained in nuclear medicine imaging. [Background technology]

[0002] In recent years, various technologies related to nuclear medicine imaging have been proposed. For example, Patent Document 1 discloses a method for improving the quality of images obtained in PET (Positron Emission Tomography), which is an example of a nuclear medicine imaging method (the specific method in Patent Document 1 will be described later). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] US Patent Publication “US 7,397,035 B2” Summary of the Invention [Problem to be solved by the invention]

[0004] It is an object of one aspect of the present invention to improve the quality of images obtained in nuclear medicine imaging by using an unconventional approach. [Means for solving the problem]

[0005] In order to solve the above problem, an image processing device according to one aspect of the present invention includes: a first reconstruction processing unit that generates an initial image indicating the concentration distribution of a radioactive substance in the subject by performing reconstruction processing on list mode data of the subject acquired by a nuclear medicine imaging device; a scattered component derivation unit that derives scattered component projection data indicating the scattered components of the radiation emitted from the radioactive substance from the initial image by a scatter estimation simulation; a second reconstruction processing unit that generates a low-resolution scattered image indicating the scattered components and having a lower resolution than the initial image by performing reconstruction processing on the scattered component projection data; an upscaling processing unit that generates an upscaled scattered image having the same resolution as the initial image by upscaling the low-resolution scattered image; and an image correction unit that generates a scatter-corrected image by subtracting the upscaled scattered image from the initial image.

[0006] Furthermore, an image processing method according to one aspect of the present invention is an image processing method in which each step is executed by a computer, the image processing method including: a first reconstruction processing step of generating an initial image indicating the concentration distribution of a radioactive substance in the subject by performing reconstruction processing on list mode data of the subject acquired by a nuclear medicine imaging device; a scatter component derivation step of deriving scatter component projection data indicating the scatter components of the radiation emitted from the radioactive substance from the initial image by a scatter estimation simulation; a second reconstruction processing step of generating a low-resolution scatter image indicating the scatter components and having a lower resolution than the initial image by performing reconstruction processing on the scatter component projection data; an upscaling processing step of upscaling the low-resolution scatter image to generate an upscaled scatter image having the same resolution as the initial image; and an image correction step of generating a scatter-corrected image by subtracting the upscaled scatter image from the initial image. [Effects of the Invention]

[0007] According to one aspect of the present invention, the quality of images obtained in nuclear medicine imaging can be improved by a non-conventional technique. [Brief explanation of the drawings]

[0008] [Figure 1] FIG. 1 is a diagram showing the configuration of a main part of an image processing system according to a first embodiment. [Figure 2] FIG. 2 is a diagram showing an example of a processing flow of the image processing device according to the first embodiment. [Figure 3] 3A to 3C are diagrams showing examples of images generated by the image processing apparatus of the first embodiment. [Figure 4] FIG. 10 is a diagram showing the results of verification experiment 1. [Figure 5] FIG. 10 is a diagram showing the results of verification experiment 2. [Figure 6] FIG. 10 is a diagram showing an example of the flow of processing by the image processing device according to the second embodiment. [Figure 7] FIG. 1 is a diagram showing an example of a reconstructed image composed of a plurality of slice images. [Figure 8] FIG. 10 is a diagram illustrating the flow of tail fitting processing for one slice image. DETAILED DESCRIPTION OF THE INVENTION

[0009] [Embodiment 1] The image processing system 100 of the first embodiment will be described in detail below. For convenience of explanation, components having the same functions as those described in the first embodiment will be denoted by the same reference numerals in the following embodiments, and their descriptions will not be repeated. For simplicity, descriptions of matters similar to those in known technologies will also be omitted as appropriate.

[0010] Please note that each configuration and each numerical value described in this specification is merely an example unless otherwise specified. Therefore, unless otherwise specified, the positional relationship of each component is not limited to the example in each drawing. Also, please note that each drawing is intended to roughly explain the shape, structure, and positional relationship of each component, and is not necessarily drawn exactly as it actually is. In this specification, the expression "A to B" regarding two numbers A and B means "greater than or equal to A and less than or equal to B" unless otherwise specified.

[0011] (Brief explanation of PET) In PET, a drug labeled with a positron-emitting nuclide is injected (administered) into a subject (image target), and the distribution of the drug within the subject is imaged. In PET, a pair of radiation rays (annihilation radiation rays) emitted from the administered drug in directions approximately 180° apart are measured by coincidence counting using detectors arranged to surround the subject. In PET, the time-series data obtained by coincidence counting (coincidence counting data) is also called list-mode data. In PET, a reconstruction process is performed on the list-mode data to generate (derive) an image (reconstructed image) showing the concentration distribution of radioactive materials in the subject (hereinafter simply referred to as "concentration distribution"). The reconstructed image is typically a three-dimensional image.

[0012] However, the coincidence data includes "scattered coincidence data" in addition to "true coincidence data." Scattered coincidence data is data measured when one or both of a pair of emitted radiation rays are scattered within the subject. Scattered coincidence data is noise data, and therefore degrades the quality of the reconstructed image.

[0013] For this reason, various methods have been proposed to improve the quality of reconstructed images. For example, Patent Document 1 proposes a scatter correction method using the SSS (Single Scatter Simulation) method, which is one of the scatter estimation simulation methods. The SSS method is a method for estimating the scattered components of radiation in projection data space based on the concentration distribution and the attenuation coefficient distribution, assuming that the radiation scatters only once.

[0014] In the technology of Patent Document 1, scattered component projection data (data indicating the scattered component of radiation) is derived from a reconstructed image using the SSS method. The scattered component projection data can also be expressed as data obtained by coarse sampling. The scattered component projection data is then converted into interpolated list mode data by interpolation processing. Subsequently, the initial list mode data (original list mode data) is corrected using the interpolated list mode data. Then, reconstruction processing is performed on the corrected list mode data (corrected list mode data), thereby obtaining a reconstructed image in which the influence of the scattered component is reduced. In this way, Patent Document 1 proposes a method of improving the quality of a reconstructed image by correcting the list mode data (i.e., by performing correction in data space).

[0015] In response to this, the inventors of the present application (hereinafter simply referred to as "the inventors") have, as a result of extensive research, newly discovered that it is possible to improve the quality of reconstructed images using a method different from conventional techniques. Based on this new knowledge, the inventors have newly created an image processing system 100 (particularly, an image processing device 1).

[0016] (Configuration of image processing system 100) 1 is a diagram showing the configuration of the main parts of an image processing system 100. The image processing system 100 includes an image processing device 1 and a PET device 90. The image processing device 1 and the PET device 90 are communicatively connected by a known communication interface (not shown).

[0017] The image processing device 1 includes a control unit 10, a display unit 70, and a storage unit 80. The control unit 10 comprehensively controls each unit of the image processing device 1. The display unit 70 displays various images (for example, various images generated by the image processing device 1). The storage unit 80 stores various data and programs used in processing by the control unit 10.

[0018] In the first embodiment, a PET device 90 is illustrated as an example of a nuclear medicine imaging device according to one aspect of the present invention. FIG. 1 illustrates a helmet-type PET device 90. The PET device 90 has a helmet unit (semispherical gantry) that can be brought close to the head of a patient PT. The patient PT is an example of a subject according to one aspect of the present invention. The helmet unit has multiple detectors (not shown) arranged along the housing of the helmet unit. The PET device 90 further has a coincidence circuit (not shown) that acquires list mode data based on the detection results of the detectors. The PET device 90 transmits the list mode data to the image processing device 1 (more specifically, the control unit 10).

[0019] 1 is an example of a time-of-flight (TOF)-PET device and also an example of a proximity-type PET device. However, as will be apparent to those skilled in the art, the PET device according to one embodiment of the present invention is not limited to a helmet-type PET device. Therefore, it should be noted that the PET device according to one embodiment of the present invention is not limited to a TOF-PET device or a proximity-type PET device.

[0020] (An example of a processing flow in the image processing device 1) The image processing device 1 includes a first reconstruction processing unit 11, a scattered component derivation unit 12, a second reconstruction processing unit 13, an upscaling processing unit 14, and an image correction unit 15. Fig. 2 is a diagram showing an example of the processing flow in the image processing device 1. Fig. 3 is a diagram showing an example of each image generated by the image processing device 1.

[0021] Imax in FIG. 2 is a predetermined number indicating the maximum number of times the loop process in FIG. 2 is executed. Imax may be expressed as the upper limit number of times the scatter-corrected image is updated. Imax may be set as appropriate by the user of the image processing device 1. As an example, Imax may be set to 10. In the following description, a case where Imax>2 is illustrated, but it should be noted that Imax may also be set to 1. Imax may be any integer equal to or greater than 1.

[0022] First, the first reconstruction processor 11 acquires list mode data related to the subject from the PET device 90. Next, the first reconstruction processor 11 initializes the current image (S1). As an example, the image processing device 1 initializes all pixel values ​​of the current image to 1. However, the method for initializing the pixel values ​​of the current image is not limited to the above example. As an example, the pixel values ​​of the current image may be initialized using the number of pixels in the current image, the number of list mode data, and the sensitivity associated with the position of each pixel in the initial image.

[0023] Next, the image processing device 1 sets i to 1 (initial value) (S2). i is a counter value (loop counter) that indicates the number of times the loop processing in FIG. 2 is executed. Next, the first reconstruction processing unit 11 generates an updated image based on the current image and the list mode data (S3). For example, the first reconstruction processing unit 11 may generate an updated image at i=1 by performing reconstruction processing on the list mode data using a method similar to that described in Patent Document 1. Then, the image processing device 1 sets the updated image as the current image (S4).

[0024] In the example of processing in FIG. 2, when i=1, the current image set in S4 is also referred to as the initial image. As described above, the first reconstruction processing unit 11 generates the initial image based on the list mode data. As is clear from the above explanation, the initial image is an image that shows the concentration distribution of radioactive materials in the subject. As an example, the resolution (number of pixels) of the initial image is 140 (horizontal resolution) × 140 (vertical resolution) × 112 (number of slices). The initial image may also be referred to as a pre-scatter correction image. IMG1 in FIG. 3 is an example of an initial image generated by the first reconstruction processing unit 11.

[0025] In the following description, the current image corresponding to the counter value i set in S4 will also be referred to as IMGP(i). As described above, when i=1, IMGP(1) is set as IMG1.

[0026] Next, the image processing device 1 determines whether i is greater than 1 (S5). If i=1 (No in S5), the image processing device 1 sets the current image (i.e., the initial image) as a provisional scatter-corrected image (S6). The provisional scatter-corrected image may also be referred to as a current scatter-corrected image. Note that if i>1 (Yes in S5), the process proceeds to S7. The method for generating the provisional scatter-corrected image in S7 will be described later.

[0027] Next, the scattered component derivation unit 12 derives scattered component projection data (data indicating the scattered components of the radiation emitted from the radioactive material) from the provisional scatter-corrected image using, for example, a method similar to that described in Patent Document 1 (using the SSS method) (S8). When i=1, the scattered component derivation unit 12 derives scattered component projection data from the initial image.

[0028] Next, the second reconstruction processor 13 generates a low-resolution scatter image based on the scattered component projection data (S9). A low-resolution scatter image is an image that represents the scattered components of radiation emitted from radioactive materials and has a lower resolution than the current image (in other words, a lower resolution than the initial image). The low-resolution scatter image may also be referred to as a few-pixel scatter image. Specifically, the second reconstruction processor 13 generates the low-resolution scatter image by performing reconstruction processing on the scattered component projection data. The reconstruction processing method in S9 may be the same as in S3.

[0029] As an example, the second reconstruction processor 13 generates a low-resolution scatter image as an image with a resolution of 32 × 32 × 32. IMG2 in FIG. 3 is an example of a low-resolution scatter image generated by the second reconstruction processor 13. In this way, in the first embodiment, the scatter component projection data is imaged instead of being converted into list mode data after interpolation. In this respect, the processing of the first embodiment differs from the prior art.

[0030] Next, upscaling processing unit 14 generates an upscaled scattering image based on the low-resolution scattering image (S10). The upscaled scattering image in embodiment 1 has the same resolution as the current image (in other words, the same resolution as the initial image). As an example, upscaling processing unit 14 generates the upscaled scattering image by upscaling the resolution of the low-resolution scattering image to the same resolution as the current image using a known method. IMG3 in Figure 3 is an example of an upscaled scattering image generated by upscaling processing unit 14.

[0031] Next, the image correction unit 15 generates a scatter-corrected image by subtracting the upscaled scatter image from the current image (S11). That is, the image correction unit 15 generates IMGSC(i) by calculating IMGSC(i) = IMGP(i) - IMGUP(i). IMGSC(i) is the scatter-corrected image corresponding to the counter value i. Furthermore, IMGUP(i) is the upscaled scatter image corresponding to the counter value i.

[0032] By generating an upscaled scattered image in advance in S10, subtraction processing between two images having the same resolution becomes possible in S11. This subtraction processing allows the scattered component to be removed from the current image. That is, a scattered-corrected image is generated as an image in which the scattered component has been removed from the current image. When i=1, the image correction unit 15 generates IMGSC(1) by subtracting IMGUP(1) from the initial image IMGP(1).

[0033] Next, the image processing device 1 determines whether i has reached Imax (S12). If i has not reached Imax (No in S12), the image processing device 1 sets the i-th upscaled scattering image as the (i+1)-th interim upscaled scattering image (S13). The interim upscaled scattering image may also be referred to as the current upscaled scattering image. As an example, if i=1, the image processing device 1 sets IMGUP(i) as the second interim upscaled scattering image. Then, the image processing device 1 counts up i by 1 (S14). Then, the process returns to S3.

[0034] If i>1, the first reconstruction processing unit 11 generates an ith (e.g., second) updated image based on the current image (i-1th current image) generated in the previous S4 and the list mode data. Then, in S4, the ith updated image is set as the ith current image. In this way, in the process of FIG. 2, the current image is sequentially updated based on the list mode data. As described above, the first reconstruction processing unit 11 updates the current image related to the initial image based on the list mode data.

[0035] If i>1 (Yes in S5), the image corrector 15 generates a tentative scatter-corrected image by subtracting the tentative upscaled scattered image from the current image (S7). In other words, if i>1, the image corrector 15 generates the i-th tentative scatter-corrected image by subtracting the (i-1)-th upscaled scattered image from the i-th current image.

[0036] In this way, until i reaches Imax, as i is incremented, the current image is sequentially updated. And until i reaches Imax (until it becomes Yes in S12), the processes of S3 to S14 are repeatedly iterated. According to such loop processing, as i increases, the scattered components are sequentially removed from the current image.

[0037] Regarding the main processing of the current loop (the i-th loop) when 1 < i < Imax, it is as follows. First, in S8, the scattered component derivation unit 12 newly derives scattered component projection data by scattered estimation simulation from the provisional scattered correction image in the current loop (the i-th provisional scattered correction image) (that is, updates the scattered component projection data derived in the previous loop).

[0038] Subsequently, in S9, the second reconstruction processing unit 13 performs a reconstruction process on the scattered component projection data updated in S8 to newly generate a low-resolution scattered image (that is, updates the low-resolution scattered image generated in the previous loop).

[0039] Subsequently, in S10, the upscaling processing unit 14 generates a newly upscaled scattered image by upscaling the low-resolution scattered image updated in S9 (that is, updates the upscaled scattered image generated in the previous loop).

[0040] Subsequently, in S11, the image correction unit 15 subtracts the upscaled scattered image updated in S10 from the current image in the current loop to newly generate a scattered correction image (that is, updates the scattered correction image generated in the previous loop).

[0041] Thereafter, when i reaches Imax (YES in S12), the image corrector 15 outputs the Imax-th scatter-corrected image (i.e., the final scatter-corrected image) as the final image (S15). For example, the image corrector 15 may output the final image to the display unit 70. IMG4 in FIG. 4 is an example of the final image output by the image corrector 15. As is clear from the above description, the resolution of the final image is equal to the resolution of the initial image.

[0042] According to the process shown in FIG. 2, a final image is obtained as an image from which the scattered components contained in the initial image have been removed. In particular, by setting Imax to a large value, a final image from which the scattered components have been sufficiently removed can be obtained. As described above, in the first embodiment, unlike the conventional technology, the quality of the reconstructed image can be improved by performing correction to remove the scattered components in the image region.

[0043] Unlike the example in FIG. 2, in the image processing device 1, the upper limit number of updates for the upscaled scattering image (for convenience, referred to as Jmax) may be set to be less than the upper limit number of updates for the current image (i.e., Imax). In this way, in the processing in the image processing device 1, the timing to stop updating the upscaled scattering image may precede the timing to stop updating the current image. In this case, the image processing device 1 may use the last updated upscaled scattering image (the upscaled scattering image updated the Jmaxth time) to generate a scatter-corrected image in each subsequent loop.

[0044] (Verification experiment 1) The inventors conducted an experiment (for convenience, referred to as "Verification Experiment 1") to verify the effectiveness of the image processing method (scatter correction method) of Embodiment 1. In the following description, the scatter correction method of Patent Document 1 and the scatter correction method of Embodiment 1 will be referred to as the conventional method and the present method, respectively.

[0045] In Verification Experiment 1, an image quality evaluation phantom was used to evaluate the uniformity of images (PET images) obtained based on list mode data (more specifically, the uniformity of signal intensity in the images). The image quality evaluation phantom used in Verification Experiment 1 was a cylindrical phantom with a diameter of 16.5 cm and a height of 14.1 cm. The image quality evaluation phantom was designed to accommodate six spheres of different diameters (referred to as "Spheres 1 to 6" for convenience). The diameters of Spheres 1 to 6 were 10 mm, 13 mm, 17 mm, 22 mm, 28 mm, and 37 mm, respectively.

[0046] In verification experiment 1, the following radioactivity concentration ratio was set to 4:1 between spheres 1 to 4 and the background area: 18 Spheres 1 to 4 and Spheres 5 to 6 may be referred to as hot spheres and cold spheres, respectively (see also FIG. 4, which will be described below).

[0047] Starting from the timing when the signal intensity (radioactivity concentration) of the background region reached 5.3 kBq / mL, the image quality evaluation phantom was imaged for 30 minutes using the PET device 90. Images were then generated by applying the conventional method and the present method to the obtained list mode data.

[0048] Fig. 4 shows the results of Verification Experiment 1. Reference numerals 4000A, 4000B, and 4000C in Fig. 4 respectively indicate (i) an image generated by the conventional technology (Comparative Example 1), (ii) an image generated by the method of the present invention (Example 1), and (iii) a correct image (Correct Answer 1).

[0049] As shown in Figure 4, non-uniformity of signal intensity was observed in the background region in Comparative Example 1. Specifically, (i) the presence of a low-value region in the upper left of the background region and (ii) the presence of a high-value region in the upper left of the background region were observed in Comparative Example 1. The inventors confirmed that the background variability in Comparative Example 1 was 6.7%.

[0050] In contrast, it was confirmed that the non-uniformity of signal intensity in the background region was improved in Example 1 compared to Comparative Example 1. Specifically, the inventors confirmed that the background variability in Example 1 was 4.3%. As described above, Verification Experiment 1 confirmed that the method of the present invention makes it possible to obtain higher quality images than the conventional technology.

[0051] (Verification experiment 2) Subsequently, the inventors conducted a further verification experiment (for convenience, referred to as "Verification Experiment 2"). In Verification Experiment 2, the contrast of PET images was evaluated using a Hoffman brain phantom (an example of a phantom simulating a human brain). The Hoffman brain phantom was designed so that the signal intensity ratio (radioactivity concentration ratio) between the spinal fluid region (simulated cerebrospinal fluid region), the white matter region (simulated white matter region), and the gray matter region (simulated gray matter region) was 0:1:4.

[0052] In verification experiment 2, 20MBq 18 A F radioactive solution was sealed in a Hoffman brain phantom. The Hoffman brain phantom was then imaged for 30 minutes using PET scanner 90. Images were then generated from the list-mode data obtained by applying the conventional method and the present method, as in Verification Experiment 1.

[0053] Fig. 5 shows the results of Verification Experiment 2. Reference numerals 5000A, 5000B, and 5000C in Fig. 5 respectively indicate (i) an image generated by the conventional technology (Comparative Example 2), (ii) an image generated by the method of the present invention (Example 2), and (iii) a correct image (Correct Answer 2).

[0054] As shown in Figure 5, in Comparative Example 2, it was confirmed that the spinal fluid region (which should have been displayed in white) where the signal intensity should have been zero was tinted gray (i.e., noise was present).

[0055] Next, the inventors set ROIs (Regions of Interest) in the gray and white matter regions and calculated the gray-to-white matter contrast in the ROIs. As a result, the gray-to-white matter contrast was confirmed to be 45% in Comparative Example 2. The gray-to-white matter contrast is one of the index values ​​used to evaluate the influence of noise. In the absence of noise, the gray-to-white matter contrast is 100%. Therefore, the higher the gray-to-white matter contrast, the more noise is reduced.

[0056] In contrast, it was confirmed that the noise was reduced in Example 2 compared to Comparative Example 2. Specifically, in Example 2, the spinal fluid region was displayed in a color closer to white compared to Comparative Example 2. Furthermore, it was confirmed that the gray matter to white matter contrast was 51% in Example 2. As described above, Verification Experiment 2 also confirmed that the method of the present invention makes it possible to obtain higher quality images than conventional techniques.

[0057] The ratio of the gray-to-white matter contrast (51%) in Experimental Example 2 to the gray-to-white matter contrast (45%) in Comparative Example 2 is approximately 1.13. Therefore, Verification Experiment 2 confirmed that the present method improved image contrast by 13% compared to the conventional method.

[0058] (effect) As described above, the image processing system 100 (particularly, the image processing device 1) can improve the quality of images obtained in PET (more specifically, reconstructed images based on list mode data obtained from the PET device) using a method different from conventional methods.

[0059] In addition, the inventors have found through Verification Experiments 1 and 2 that the present method (a scattering correction method in image space) can further improve image quality compared to the conventional technique (a scattering correction method in data space). The inventors speculate as follows about the reason why the present method can improve image quality compared to the conventional method (more specifically, why scattering components, which are noise, can be removed more effectively).

[0060] Scatter component projection data is considered to be susceptible to the influence of spatial variations among multiple detectors arranged in a PET device. Therefore, in the interpolated list mode data generated by the conventional method, the influence of the variations in the scatter component projection data is considered to be exaggerated by the interpolation process. Therefore, in the conventional method, errors (deterioration in accuracy) are considered to occur when generating the scatter component projection data. As a result, the errors are considered to be insufficiently reduced even by correcting the list mode data.

[0061] On the other hand, because an image is defined by multiple regularly arranged pixel data, it is considered to be spatially organized data compared to general numerical data (e.g., list-mode data). Therefore, when performing interpolation processing in image space (upscaling a low-resolution scatter image in the present method), it is considered to be less susceptible to the spatial variations of multiple detectors compared to interpolation processing in data space (generating scatter component projection data in conventional methods). As a result, it is considered that the present method will generate relatively few errors when generating an upscaled scatter image. It is considered that the present method can remove scatter components more effectively than conventional methods by correcting the current image using such an upscaled scatter image.

[0062] Incidentally, it is known that the scattered coincidence counts in a PET device depend on the size of the device. Specifically, it is known that the scattered coincidence counts tend to increase as the size of the PET device decreases. For this reason, for example, in a helmet-type PET device, the influence of scattered components is greater than in other general PET devices.

[0063] However, as described above, the present technique can effectively remove scattered components from the reconstructed image. Therefore, high-quality images can be obtained even when a helmet-type PET device is used. For this reason, the present technique is particularly suitable for helmet-type PET devices (e.g., PET device 90). Similarly, the present technique is suitable for proximity-type PET devices.

[0064] [Embodiment 2] The processing flow of the image processing device 1 is not limited to the example shown in Fig. 2. Fig. 6 is a diagram showing an example of the processing flow of the image processing device 1 in embodiment 2. S22 in Fig. 6 is the same processing as S2 in Fig. 2, and therefore description thereof will be omitted.

[0065] First, the first reconstruction processor 11 generates an initial image based on the list mode data (S21). Then, after S22, the scattered component derivation unit 12 derives scattered component projection data based on the initial image (S23). S23 corresponds to S8 when i=1.

[0066] Next, second reconstruction processor 13 reconstructs a low-resolution scattering image based on the scattered component projection data (S24), and upscale processor 14 upscales the low-resolution scattering image to generate an upscaled scattering image (S25).

[0067] Subsequently, the image correction unit 15 generates a scattered light correction image by subtracting the upscaled scattered light image from the initial image (S26). That is, in the second embodiment, the image correction unit 15 generates IMGSC(i) as IMGSC(i) = IMG1 - IMGUP(i). Thus, in the second embodiment, unlike the first embodiment, subtraction from the initial image (IMG1) is performed regardless of the counter value i.

[0068] Subsequently, the image processing apparatus 1 determines whether i has reached Imax (S27). If i has not reached Imax (No in S27), the scattered light component derivation unit 12 derives scattered light component projection data based on the scattered light correction image (S28). Then, the image processing apparatus 1 increments i by 1 (S29).

[0069] Thus, until i reaches Imax, the scattered light component projection data is sequentially updated as i is incremented. Then, the process returns to S24. That is, until i reaches Imax (until Yes in S27), the processes of S24 to S29 are repeatedly executed. According to such loop processing, as i increases, the scattered light components are sequentially removed from the scattered light correction image.

[0070] When 1 < i < Imax, in the current loop (the i-th loop), in S24 and S25, the same processes as in S9 and S10 are executed. Subsequently, in S26, the image correction unit 15 newly generates a scattered light correction image by subtracting the upscaled scattered light image updated in S25 from the initial image (that is, updates the scattered light correction image generated in the previous loop).

[0071] Thereafter, when i reaches Imax (YES in S27), the image correction unit 15 outputs the Imax-th scattered light correction image (that is, the final scattered light correction image) as the final image. Also by the process of FIG. 6, the final image is obtained as an image from which the scattered light components included in the initial image have been removed. Therefore, the second embodiment also has the same effect as the first embodiment.

[0072] <00002,63>In addition, Imax may also be set to 1 in the second embodiment. That is, S24 to S29 do not have to be repeated repeatedly in the second embodiment. As is clear from Fig. 6, when Imax = 1, S28 and S29 are not executed.

[0073] [Embodiment 3] In conventional techniques, tail fitting is performed in data space. On the other hand, the image processing device 1 can perform tail fitting in image space. In particular, the image processing device 1 can perform tail fitting for the entire three-dimensional image, or for each slice of the three-dimensional image. Therefore, the image processing device 1 is expected to reduce errors caused by tail fitting. The concept of tail fitting in the image processing device 1 will be described below with reference to FIGS. 7 and 8.

[0074] FIG. 7 is a diagram showing an example of a reconstructed image IMG composed of multiple slice images (two-dimensional images). The IMG in FIG. 7 is a three-dimensional image that represents an image of an object OBJ. OBJ is another example of a subject according to one embodiment of the present invention. Slice n in FIG. 7 refers to the nth slice image (e.g., the nth from the top) in the stack direction of multiple slice images. Below, tail fitting for slice 1 (the topmost slice image) will be described, but the description also applies to any slice n.

[0075] 7, slice 1 includes an object region 711 (a region where the image of OBJ is expressed) and a tail region 712 (a peripheral region of OBJ). The object region 711 can be extracted by a known method (for example, by an extraction method based on an attenuation coefficient image).

[0076] 8 is a diagram illustrating the flow of tail fitting processing for a certain slice image (e.g., slice 1). First, (i) a scattered image IMG11 of OBJ and (ii) an uncorrected image IMG12 of OBJ are generated based on slice 1. Then, tail fitting based on a known method is performed based on IMG11 and IMG12 to calculate the scale factor c.

[0077] Next, IMG11 is multiplied by c to generate a tail-fitted scatter image IMG13. IMG13 may also be referred to as a tail-fitted scatter image. Then, IMG13 is subtracted from IMG12 to generate a scatter-corrected image IMG14. In this way, the image processing device 1 can remove the scatter component of IMG12 using IMG13.

[0078] [Embodiment 4] As will be apparent to those skilled in the art, the nuclear medicine imaging technique according to one aspect of the present invention is not limited to PET, but may be any technique that (i) is capable of reconstructing an image based on list-mode data, and (ii) is capable of deriving scatter component projection data using scatter estimation simulation.

[0079] Another example of a nuclear medicine imaging technique according to an embodiment of the present invention is SPECT (Single Photon Emission-Computed Tomography). List-mode data in SPECT is time-series data of single counts. As described above, the nuclear medicine imaging device according to an embodiment of the present invention is not limited to a PET device, and a SPECT device, for example, is also included in the category of the nuclear medicine imaging device.

[0080] [Software implementation example] The functions of the image processing system 100 (hereinafter simply referred to as the "system") are realized by a program for causing a computer to function as the system, and by a program for causing a computer to function as each control block of the system (particularly each part included in the control unit 10).

[0081] In this case, the system includes a computer having at least one control device (e.g., a processor) and at least one storage device (e.g., a memory) as hardware for executing the program. The control device and storage device execute the program, thereby realizing the functions described in each of the above embodiments.

[0082] The program may be recorded on one or more non-transitory computer-readable recording media. The recording media may or may not be included in the system. In the latter case, the program may be supplied to the system via any wired or wireless transmission medium.

[0083] Furthermore, some or all of the functions of the control blocks can be realized by logic circuits. For example, an integrated circuit in which a logic circuit that functions as each of the control blocks is formed is also included in the scope of the present invention. In addition, the functions of the control blocks can also be realized by, for example, a quantum computer.

[0084] Furthermore, each process described in each of the above embodiments may be executed by AI (Artificial Intelligence). In this case, the AI ​​may run on the control device or on another device (for example, an edge computer or a cloud server).

[0085] [Additional Notes] One aspect of the present invention is not limited to the above-described embodiments, and various modifications are possible within the scope of the claims. Embodiments obtained by appropriately combining the technical means disclosed in different embodiments are also included in the technical scope of one aspect of the present invention. [Explanation of symbols]

[0086] 1. Image processing device 10 Control Unit 11 First reconstruction processing unit 12 Scattered component derivation section 13 Second reconstruction processing unit 14 Upscale processing section 15 Image correction section 90 PET device (nuclear medicine imaging device) 100 Image Processing System IMG1 Initial image IMG2 low-resolution scattering image IMG3 upscaled scattering image IMG4 Final image (final scatter-corrected image) PT Patient (Subject)

Claims

1. a first reconstruction processing unit that generates an initial image showing a concentration distribution of a radioactive substance in a subject by performing reconstruction processing on list mode data related to the subject acquired by a nuclear medicine imaging apparatus; a scattered component derivation unit that derives scattered component projection data indicating scattered components of the radiation emitted from the radioactive material from the initial image by a scattering estimation simulation; a second reconstruction processing unit that generates a low-resolution scattered image that indicates the scattered component and has a lower resolution than the initial image by performing reconstruction processing on the scattered component projection data; an upscaling processor that upscales the low-resolution scattering image to generate an upscaled scattering image having the same resolution as the initial image; an image corrector that generates a scatter-corrected image by subtracting the upscaled scatter image from the initial image.

2. the first reconstruction processor updates a current image relative to the initial image based on the list mode data; the image corrector generates a temporary scatter-corrected image by subtracting the previously generated upscaled scatter image from the updated current image; the scatter component derivation unit derives the scatter component projection data from the tentative scatter-corrected image through the scatter estimation simulation, thereby updating the previously derived scatter component projection data; the second reconstruction processing unit generates a new low-resolution scattered image by performing reconstruction processing on the updated scattered component projection data, thereby updating the previously generated low-resolution scattered image; the upscaling processor upscales the updated low-resolution scattering image to generate a new upscaled scattering image, thereby updating the previously generated upscaled scattering image; The image processing device according to claim 1 , wherein the image correction unit updates the previously generated scatter-corrected image by subtracting the updated upscaled scatter image from the updated current image to generate a new scatter-corrected image.

3. the scatter component derivation unit derives the scatter component projection data from the scatter-corrected image previously generated by the scatter estimation simulation, thereby updating the scatter component projection data previously derived; the second reconstruction processing unit generates a new low-resolution scattered image by performing reconstruction processing on the updated scattered component projection data, thereby updating the previously generated low-resolution scattered image; the upscaling processor upscales the updated low-resolution scattering image to generate a new upscaled scattering image, thereby updating the previously generated upscaled scattering image; The image processing device according to claim 1 , wherein the image correction unit updates the previously generated scatter-corrected image by subtracting the updated upscaled scatter image from the initial image to generate a new scatter-corrected image.

4. The image processing device according to claim 1 , wherein the image correction unit outputs the final scattering-corrected image obtained as a result of repeated updates up to a predetermined upper limit number of updates.

5. The image processing device according to claim 1 , wherein the scattered component deriving unit executes the scattering estimation simulation using an SSS (Single Scatter Simulation) method.

6. The image processing device according to any one of claims 1 to 5, An image processing system comprising the above nuclear medicine imaging device.

7. 7. The image processing system according to claim 6, wherein the nuclear medicine imaging device is a PET (Positron Emission Tomography) device.

8. 8. The image processing system according to claim 7, wherein the PET device is a helmet-type PET device.

9. An image processing method in which each step is executed by a computer, a first reconstruction processing step of generating an initial image showing a concentration distribution of a radioactive substance in a subject by performing reconstruction processing on list mode data related to the subject acquired by a nuclear medicine imaging device; a scattered component deriving step of deriving scattered component projection data indicating scattered components of the radiation emitted from the radioactive material from the initial image by a scatter estimation simulation; a second reconstruction processing step of generating a low-resolution scatter image representing the scatter component and having a resolution lower than that of the initial image by performing reconstruction processing on the scatter component projection data; an upscaling step of upscaling the low-resolution scattering image to generate an upscaled scattering image having the same resolution as the initial image; and an image correction step of subtracting the upscaled scatter image from the initial image to generate a scatter-corrected image.

Images