JP7742014B2 - Image providing device, image providing system, image providing method, and program - Google Patents

Description

The present invention relates to an image providing device that provides various information to subjects undergoing medical checkups, brain checkups, etc.

With health consciousness on the rise, the number of people undergoing comprehensive medical checkups aimed at early detection and treatment of illness is increasing year by year. Comprehensive medical checkups offer more detailed and tailored optional tests. One of these is the brain checkup. Cross-sectional images of the brain taken with an MRI device or similar device are presented as the results of the brain checkup.

Japanese Patent Application Laid-Open No. 2003-144222 discloses a technique for supporting setting of an imaging position in a medical imaging apparatus.
Patent Document 2 discloses a medical checkup report viewing system that allows a medical checkup report to be viewed from a remote location.

International Publication No. 2012/008296 Japanese Patent Application Laid-Open No. 2006-236061

However, the technology of Patent Document 1 is intended to reduce the burden on the operator of the medical imaging device, and is not intended to present a medical checkup report to a patient undergoing a medical checkup. Patent Document 2 presents a medical checkup report for a medical checkup, but does not describe the specific contents of the medical checkup report.
There is a demand for providing useful and satisfying medical checkup reports to patients undergoing brain checkups. In particular, there is a need to provide cross-sectional images of the brain suitable for brain checkups.

The present invention aims to provide an image providing device that can provide useful and highly satisfying image information to subjects undergoing medical checkups, brain checkups, etc.

In order to solve the above problem, a first embodiment of the image providing device of the present invention comprises a specific region extraction unit that extracts a specific region from a plurality of cross-sectional images of a subject's organs, a shape measurement unit that measures the shape of the specific region, an image division unit that sorts each of the cross-sectional images into a plurality of groups, an image selection unit that selects a representative image from each of the groups based on measurements taken by the shape measurement unit, and when the specific region is not extracted from the cross-sectional image in the group, selects the representative image based on positional information in the slice thickness direction of the cross-sectional image , and an image providing unit that provides the representative image.

A second embodiment of the image providing apparatus of the present invention is the first embodiment, wherein the organ is the brain and the specific region is a white matter hyperintensity area.
A third embodiment of the image providing device of the present invention is the first or second embodiment, wherein the organ is a brain and the specific region is a stroke region.

A fourth embodiment of the image providing device of the present invention is based on any one of the first to third embodiments, wherein the specific region extraction section extracts the specific region by pattern recognition.
A fifth embodiment of the image providing device of the present invention is based on any one of the first to fourth embodiments, wherein the shape measuring section measures the area, major axis, minor axis or perimeter of the specific region.
A sixth embodiment of the image providing device of the present invention is any of the first to seventh embodiments, in which the specific area extraction unit obtains a three-dimensional image of the specific area based on positional information in the slice thickness direction of the cross-sectional image and positional information on the slice plane of the specific area, and the shape measurement unit measures the volume of the three-dimensional image.

A seventh aspect of the image providing apparatus of the present invention is any one of the first to sixth aspects, wherein the image dividing section sets the groups based on position information of the cross-sectional images in the slice thickness direction.
An eighth aspect of the image providing device of the present invention is any one of the first to seventh aspects, wherein the image dividing section sets the groups based on the presence or absence of the specific area.
A ninth aspect of the image providing device of the present invention is any one of the first to eighth aspects, wherein the image dividing section sets the group based on position information on the slice plane of the specific region.
A tenth embodiment of the image providing device of the present invention is any of the first to ninth embodiments, wherein the image dividing section classifies the specific area into a plurality of areas and sets the groups based on the classification.
An eleventh embodiment of the image providing device of the present invention is any of the first to tenth embodiments, wherein the specific area extraction unit obtains a three-dimensional image of the specific area based on position information in the slice thickness direction of the cross-sectional image and position information on the slice surface of the specific area, and the image division unit sets the group based on the presence or absence of the three-dimensional image.
A twelfth embodiment of the image providing apparatus of the present invention is any one of the first to eleventh embodiments, wherein the image dividing section sets the groups based on position information in the slice thickness direction of the three-dimensional image.

A thirteenth embodiment of the image providing device of the present invention is any of the first to twelfth embodiments, in which the image selection unit calculates the similarity between the representative images, and if the similarity exceeds a predetermined reference value, reselects the representative image.
A fourteenth aspect of the image providing apparatus of the present invention is any one of the first to thirteenth aspects, wherein the similarity is calculated based on position information of the representative image in the slice thickness direction.
A fifteenth aspect of the image providing device of the present invention is any of the first to fourteenth aspects, wherein the image selecting section changes the type of the measurement value and reselects the representative image.
A sixteenth embodiment of the image providing device of the present invention is any of the first to fifteenth embodiments, wherein when the specific region is not extracted from the cross-sectional image in the group, the image selection unit selects the representative image based on position information in the slice thickness direction of the cross-sectional image.

A first embodiment of the image providing system of the present invention comprises any one of the first to sixteenth embodiments of the image providing device of the present invention, an image capturing device that captures the cross-sectional images, and a PACS server that receives the cross-sectional images from the image capturing device and transfers them to the image providing device, and the image providing device transmits the representative image to the PACS server.
A second embodiment of the image providing system of the present invention is the first embodiment, which is provided with a terminal device that communicates with the image providing device via a telecommunications network by operation of the subject, and the image providing device transmits the representative image to the terminal device based on instructions from the terminal device.

An embodiment of the image providing method of the present invention comprises: a specific region extraction step of extracting a specific region from a plurality of cross-sectional images of an organ of a subject using a specific region extraction unit ; a shape measurement step of measuring the shape of the specific region using a shape measurement unit ; an image division step of sorting each of the cross-sectional images into a plurality of groups using an image division unit; an image selection step of selecting a representative image from each of the groups using an image selection unit based on the measurement values obtained in the shape measurement step , and when the specific region is not extracted from the cross-sectional image in the group, selecting the representative image by the image selection unit based on positional information in the slice thickness direction of the cross-sectional image; and an image providing step of providing the representative image by the image providing unit .

An embodiment of the program of the present invention causes a CPU of an image providing device to function as a specific region extraction unit that extracts a specific region from a plurality of cross-sectional images of a subject's organs, a shape measurement unit that measures the shape of the specific region, an image division unit that sorts each of the cross-sectional images into a plurality of groups, an image selection unit that selects a representative image from each of the groups based on measurements taken by the shape measurement unit, and, when the specific region is not extracted from the cross-sectional image in the group, selects the representative image based on positional information in the slice thickness direction of the cross-sectional image , and an image providing unit that provides the representative image.

The image providing device of the present invention can provide useful and highly satisfying image information to subjects undergoing brain checkups and other similar procedures.

FIG. 1 is a diagram illustrating an example of the configuration of a brain image providing system 1. FIG. 2 is a diagram illustrating an example of the hardware configuration of a brain image providing device 30. FIG. 2 is a diagram illustrating an example of the software configuration of the brain image providing device 30. 10A and 10B are diagrams showing groups G1 to G3 of axial cross-sectional images P. FIG. 1 is a flowchart showing a brain image providing method. FIG. 10 is a diagram showing a presentation image Q. FIG. 10 is a diagram showing a medical checkup report R including a presentation image Q. FIG. 1 shows white matter hyperintensity areas W (hyperintensity areas around the lateral ventricles X, hyperintensity areas deep under the cortex Y).

The following describes an image provision device, image provision system, image provision method, and program according to embodiments of the present invention. As an example, we will explain the case of presenting a useful and satisfying medical checkup report to patients undergoing a brain checkup.

[Brain image provision system 1]
FIG. 1 is a diagram showing the overall configuration of a brain image providing system 1.
The brain image providing system (image providing system) 1 includes a brain image capturing device 10 , a PACS server 20 , and a brain image providing device 30 .

The brain imaging device (imaging device) 10 is, for example, an MRI (Magnetic Resonance Imaging) device or a CT (Computed Tomography) device.
The brain imaging device 10 may be a device such as a PET (positron emission tomography) device or a radioisotope imaging device (RI).

The brain imaging device 10 captures cross-sectional images of the brain of a subject (patient, examinee). These cross-sectional images (images P) are also called slices.
The brain imaging device 10 captures images P of a horizontal section perpendicular to the body axis direction (Axial), a frontal section perpendicular to the sagittal horizontal axis direction (Sagittal), and a sagittal section perpendicular to the frontal horizontal axis direction (Coronal).
The slice thickness (thickness of the cross section) and slice gap (the distance between the images P) are set to be constant, for example, 5.0 mm and 1.8 mm.
The brain imaging device 10 captures at least 10 to 20 horizontal slice images, coronal slice images, and sagittal slice images in one examination.
The multiple images P captured by the brain image capturing device 10 are transmitted to the PACS server 20 .

The PACS (Picture Archiving and Communication System) server 20 is a server that stores, views, manages, etc., a plurality of cross-sectional images captured by the brain image capturing device 10.
The PACS server 20 is also called a DICOM server. DICOM (Digital Imaging and Communication in Medicine) is a standard that defines the storage of images and their metadata, the method of transfer (data format) between devices, and so on.
The PACS server 20 is connected to the brain image capturing apparatus 10 via a LAN (Local Area Network). A plurality of brain image capturing apparatuses 10 may be connected to one PACS server 20.
The PACS server 20 stores the multiple images P received from the brain image capturing device 10 and transmits (transfers) them to the brain image providing device 30 .

Multiple client PCs 15 are connected to the PACS server 20 via a LAN. These client PCs 15 are terminals operated by doctors and others, and receive images P of the subject's brain from the PACS server 20. The client PCs 15 have a display on which the images P of the subject's brain are displayed.

The brain image providing device (image providing device) 30 is connected (interoperates) with the PACS server 20 via a network (telecommunications network) such as a LAN or the Internet. The brain image providing device 30 processes multiple images P received from the PACS server 20 and generates brain image information (presentation images Q) to be presented to the subject, as well as a medical checkup report R including the presentation images Q.

A portable information terminal 50 such as a smartphone carried by the subject can be connected to the brain image providing device 30 via the Internet.
The subject can operate a mobile information terminal (terminal device) 50 to communicate with the brain image providing device 30 and obtain (view, download, etc.) the subject's own brain image information (presented image Q, health check report R) from the brain image providing device 30.

[Hardware configuration of brain image providing device 30]
FIG. 2 is a diagram showing the hardware configuration of the brain image providing device 30.
The brain image providing device 30 includes a control unit 31. The control unit 31 includes a CPU 311, a ROM 312, a RAM 313, a HDD 314, and a network I/F 315.

The CPU 311 controls the overall operation by reading various programs (programs of the present invention) stored in the ROM 312 and executing various processes. The CPU 311 reads the programs from the ROM 312 and executes the brain image provision method (image provision method of the present invention). As a result, the CPU 311 causes the brain image provision device 30 to function as the specific region extraction unit, shape measurement unit, image division unit, image selection unit, and image provision unit of the present invention.

The RAM 313 is used as a temporary storage area such as a main memory or work area for the CPU 311 .
The HDD 314 is a large-capacity storage unit that stores image information and various programs. Cross-sectional images of the subject's brain are stored in the HDD 314.
The network I/F 315 is an interface that connects the brain image providing device 30 to a LAN or the Internet. The brain image providing device 30 receives a processing request (instruction) from the portable information terminal 50 or the like via the network I/F 315, and transmits various information to the portable information terminal 50 or the like.

[Software (program) configuration of brain image providing device 30]
FIG. 3 is a diagram showing the software configuration of the brain image providing device 30.
FIG. 4 is a diagram showing groups G1 to G3 of axial images P. As shown in FIG.
The brain image providing device 30 includes a specific region extracting unit 41 , a shape measuring unit 42 , an image dividing unit 43 , an image selecting unit 44 , and an image providing unit 45 .

The specific region extraction unit 41 extracts white matter hyperintensity regions (specific regions) W from all images P. The specific region extraction unit 41 uses existing technology to extract the white matter hyperintensity regions W. For example, the specific region extraction unit 41 extracts the white matter hyperintensity regions W using pattern recognition (including deep learning).
Information on the presence or absence of white matter hyperintensity areas W is linked to each image P.

The shape measurement unit 42 measures the shape of the white matter hyperintensity regions W extracted by the specific region extraction unit 41. Specifically, the shape measurement unit 42 measures the area of the white matter hyperintensity regions W. The shape measurement unit 42 also measures the position of the white matter hyperintensity regions W on the slice plane.
The shape measurement unit 42 measures the area of white matter hyperintensity regions W for all images P. If multiple white matter hyperintensity regions W exist in one image P, the area of each white matter hyperintensity region W is measured. The total area of the multiple white matter hyperintensity regions W is also measured.
Measurement values of the shape of the white matter hyperintensity areas W are associated with each image P.

The image division unit 43 divides (groups) a plurality of cross-sectional images of the brain of the subject into a plurality of groups. Specifically, the image division unit 43 divides horizontal cross-sectional images (a plurality of images P) perpendicular to the body axis direction among the cross-sectional images of the brain of the subject into a plurality of groups.
As shown in Figure 4, the brain is divided into three groups along the axial direction (upper brain layer G1, middle brain layer G2, and lower brain layer G3). The three groups G1 to G3 are layers along the axial direction of the brain.
The upper brain layer G1 is a layer (region) that mainly corresponds to the parietal lobe and also to the frontal lobe. The middle brain layer G2 is a layer that mainly corresponds to the frontal lobe and occipital lobe and also to the temporal lobe. The lower brain layer G3 is a layer that mainly corresponds to the temporal lobe and also to the frontal lobe.
The images P of the horizontal cross sections are divided into one of the groups of the upper brain layer G1, the middle brain layer G2, and the lower brain layer G3.

The image selection unit 44 selects, as a representative image D, the image P with the largest area of white matter hyperintensity region W in each of the groups G1 to G3.
In each of the groups G1 to G3, if white matter hyperintensity regions W cannot be extracted from multiple images P, a standard image P for that group is selected. The standard image P can be selected arbitrarily.

The image providing unit 45 presents the three representative images D selected by the image selecting unit 44, for example, arranged horizontally. The arrangement of the three representative images D can be set arbitrarily.
The arrangement of the representative images D may be set based on a hanging protocol, which can be set for each brain image capturing device 10, each subject (patient), each doctor, and each examination region.
The image providing unit 45 generates image information (presentation image Q) in which a plurality of representative images D are arranged, and then generates a medical checkup report R including the presentation image Q (see FIGS. 6 and 7).

The presentation image Q may be one piece of image information or multiple pieces of image information, as long as the presentation image Q is in an image format (image data) that allows the three representative images D to be displayed at the output (transmission) destination.
The health check report R includes, for example, the volume of white matter hyperintensity, P/D (near/deep lateral ventricle) score, previous diagnosis information, a graph showing the subject's position in data from the same age group, a graph showing changes over time, and standard text based on the analysis results.

Furthermore, the image providing unit 45 outputs (provides) the presented image Q and the medical checkup report R to the outside via a network such as a LAN or the Internet.
For example, the image providing unit 45 transmits the presentation image Q to the PACS server 20. As a result, the presentation image Q can be displayed on the display of the client PC 15 connected to the PACS server 20.
For example, the image providing unit 45 transmits the medical checkup report R to the portable information terminal 50 carried by the subject via the Internet. This allows the subject to view the medical checkup report R using the portable information terminal 50.

[Brain image provision method (image provision method)]
FIG. 5 is a flowchart showing a brain image providing method.
FIG. 6 is a diagram showing a presentation image Q. As shown in FIG.
FIG. 7 is a diagram showing a medical checkup report R including a presentation image Q.
The brain image providing method includes a specific region extraction step S1, a shape measurement step S2, an image division step S3, an image selection step S4, and an image providing step S5. It also includes an image display step S6.

Prior to the brain image providing method, the brain image providing device 30 obtains multiple images P from the brain image capturing device 10. Assume that the brain image providing device 30 obtains 30 images P (images P1 to P30). These 30 images P are assigned consecutive numbers from the top (top) to the bottom (bottom) in the body axis direction. In addition, each image P is assigned position information in the body axis direction (slice thickness direction). Specifically, image P1 is image data of the apex of the brain, and is assigned position information for this apex. Image P30 is image data of the bottom of the brain, and is assigned position information for this bottom.

(Specific area extraction step S1)
In the specific region extraction step S1, the specific region extraction unit 41 of the brain image providing device 30 extracts white matter hyperintensity regions (specific regions) W from each image P. The specific region extraction unit 41 extracts the white matter hyperintensity regions W by pattern recognition.

(Shape measurement process S2)
In the shape measurement step S2, the shape measurement unit 42 measures the area of the white matter hyperintensity regions W extracted from each image P. If multiple white matter hyperintensity regions W exist in a single image P, the area of each white matter hyperintensity region W is measured, or the total area is measured.
Furthermore, the shape measurement unit 42 also measures the position of the white matter hyperintensity area W on the slice surface.

(Image classification step S3)
In the image division step S3, the image division unit 43 divides the brain cross-sectional images of the subject into a plurality of groups. Specifically, the image division unit 43 divides the horizontal cross-sectional images (images P1 to P30) perpendicular to the body axis direction among the plurality of brain cross-sectional images into a plurality of groups.
As described above, the brain is divided into three groups (upper brain layer G1, middle brain layer G2, and lower brain layer G3) along the direction of the body axis.
For example, when there are 30 horizontal cross-sectional images P (images P1 to P30), 10 images are assigned to each of groups G1 to G3. Images P1 to P10 are assigned to the upper brain layer G1, images P11 to P20 to the middle brain layer G2, and images P21 to P30 to the lower brain layer G3.

(Image selection step S4)
In the image selection step S4, the image selection unit 44 selects a representative image D for each group.
First, the image selection unit 44 determines whether or not a white matter hyperintensity region W has been detected in each of the three groups G1 to G3 (S41).
The image selection unit 44 selects the image P (representative image D) in which the area of the white matter hyperintensity region W is largest among the groups in which the white matter hyperintensity region W is detected (S42). For example, image P4 from the upper brain layer G1, image P16 from the middle brain layer G2, and image P23 from the lower brain layer G3 are selected as representative images D.

On the other hand, if there is a group in which white matter hyperintensity regions W cannot be detected, the image selector 44 selects a standard image P for that group as the representative image D (S43). The image P located at the center (centre) of the group in the slice thickness direction is the standard image.
For example, if a white matter high signal intensity region W cannot be extracted from images P1 to P10 assigned to the upper brain layer G1, image P5 located at the center (centre) in the slice thickness direction is selected as the representative image D.
For example, if no white matter high-intensity signal areas W are extracted from the lower brain layer G3 only, image P25, for example, is selected from the lower brain layer G3, and images P4 and P16, for example, are selected from the upper brain layer G1 and middle brain layer G2, following the procedure described above.

(Image provision step S5)
In the image providing step S5, the image providing unit 45 presents the three representative images D (images P4, P16, and P23) selected by the image selecting unit 44.
The image providing unit 45 arranges the three representative images D, for example, in the order of the group, vertically or horizontally. When arranging them vertically, for example, image P4 is arranged on the top, image P16 is arranged in the center, and image P23 is arranged on the bottom. When arranging them horizontally, for example, image P4 is arranged on the right, image P16 is arranged in the center, and image P23 is arranged on the left.
The image providing unit 45 generates image information (presentation image Q) in which the representative images D are arranged, and further generates a medical checkup report R including this presentation image Q.
The presentation image Q is not limited to image information in which three representative images D are arranged, but may be any image format (image data) in which the three representative images D are displayed simultaneously or at different times.

(Image display step S6)
The image providing unit 45 outputs (provides) the presented image Q and the medical checkup report R to the outside via a network such as a LAN or the Internet.
The PACS server 20 receives the presentation image Q from the image provider 45. As a result, the client PC 15 connected to the PACS server 20 can download the presentation image Q and the medical checkup report R and print or display them.
This eliminates the need for the doctor (radiography interpreter) operating the client PC 15 to visually confirm the entire cross-sectional image (image P), and allows the doctor to confirm the extent and volume of the white matter high-intensity areas W at a glance of the presented image Q. Furthermore, the presented image Q and the medical checkup report R can be presented to the patient (subject) undergoing a brain checkup.

The mobile information terminal 50 owned by the person undergoing the brain checkup (subject) also receives the medical checkup report R including the presented image Q from the image providing unit 45 .
The examinee connects the mobile information terminal 50 to the brain image providing device 30 via the Internet, and then downloads the medical checkup report R from the image providing unit 45. This allows the examinee (subject) to view the medical checkup report R.

In this way, the brain image provision system 1, brain image provision device 30, and brain image provision method according to embodiments of the present invention can provide and present useful and highly satisfying image information (presented image Q, medical checkup report R) to subjects undergoing brain checkups. In other words, the extent and volume of white matter hyperintensity areas W are visualized, allowing subjects to easily confirm their own symptoms, etc. It also makes it possible to conduct a detailed investigation of cerebrovascular disorders that may be caused by aging, lifestyle habits, etc.

The above describes preferred embodiments of the present invention, but the present invention is not limited to these embodiments, and various modifications and variations are possible within the scope of the invention.

FIG. 8 shows white matter hyperintensity areas W (hyperintensity areas around the lateral ventricles X, hyperintensity areas deep under the cortex Y).
In the embodiment described above, in the image selection step S4, the specific region extraction unit 41 extracted white matter hyperintensity regions W from multiple images P, and the shape measurement unit 42 measured their areas. Furthermore, the white matter hyperintensity regions W may be further subdivided (classified) and subjected to similar processing. That is, the white matter hyperintensity regions W are subdivided into peri-lateral ventricle hyperintensity regions X and deep subcortical hyperintensity regions Y.
The specific region extraction unit 41 extracts a high-intensity area X around the lateral ventricle from each image P, and the shape measurement unit 42 measures its area. Similarly, the specific region extraction unit 41 extracts a high-intensity area Y in the deep subcortical region from each image P, and the shape measurement unit 42 measures its area.
Then, the image selection unit 44 selects as the representative image D the image with the largest area of the high signal intensity region X around the lateral ventricle or the largest area of the deep subcortical high signal intensity region Y.

Furthermore, the white matter high-intensity signal areas W may be subdivided into the following four areas for extraction and classification (grouping). Specifically, they may be subdivided into subcortical, deep area, periventricular, and corpus callosum for extraction and classification.

Instead of G1 to G3, the image division unit 43 may set (divide) groups based on the presence or absence of white matter hyperintensity areas W. Specifically, it may set group G4, which contains white matter hyperintensity areas W, and group G5, which does not contain white matter hyperintensity areas W.
Groups may also be set based on the presence or absence of a peri-lateral ventricle high signal area X or a deep subcortical high signal area Y. Specifically, group G6 is set up as a group containing a peri-lateral ventricle high signal area X, group G7 as a group containing a deep subcortical high signal area Y, and group G8 as a group containing neither a peri-lateral ventricle high signal area X nor a deep subcortical high signal area Y.
In this group setting, the same image P may be assigned to multiple groups (overlap is allowed). For example, if image P4 contains a peri-lateral ventricle high signal intensity area X and a deep subcortical high signal intensity area Y, image P4 will be assigned to both group G6 and group G7.

In the above-described embodiment, cross-sectional images of the brain are used, but cross-sectional images of other organs (internal organs such as the liver, teeth, and bones) may also be used.
Although the specific region extraction unit 41 extracted the white matter hyperintensity region W from the image P as the specific region, it may also extract other types of regions. For example, the specific region extraction unit 41 may extract a stroke region. Furthermore, the stroke region may be further subdivided (classified) into the following three categories for extraction and classification. Specifically, the stroke region may be subdivided into cerebral hemorrhage and cerebral hematoma (high absorption region), cerebral infarction (low absorption region), and early CT sign (unclear boundary), for extraction and classification.

In the above-described embodiment, the shape measurement unit 42 measures the area as the shape of a specific region (e.g., white matter hyperintensity region W), but other shapes may also be measured. The shape measurement unit 42 may measure the major axis, minor axis, perimeter, or the like as the shape of a specific region.
The image selection unit 44 may select, for example, the image (image P) with the longest perimeter of the white matter high-intensity area W as the representative image D.

The specific region extraction unit 41 may obtain a three-dimensional image of the specific region based on position information in the slice thickness direction of the image P and position information on the slice plane of the specific region. This allows the shape measurement unit 42 to measure the volume of this three-dimensional image.
The image selection unit 44 may select the representative image D from the image (the plurality of images P) having the largest volume of the three-dimensional image of the specific region.

The image segmentation unit 43 can set groups based on the positional information of specific regions (e.g., white matter hyperintensity regions W) on the slice plane. For example, it can distinguish between specific regions located in the right brain and specific regions located in the left brain (creating different groups and sorting them), or it can distinguish between specific regions located in the frontal lobe and specific regions located in the occipital lobe.

The specific region extraction unit 41 obtains a 3D image of the specific region based on position information in the slice thickness direction of the image and position information on the slice plane of the specific region. The image division unit 43 then sets groups based on the presence or absence of this 3D image.
The specific region does not exist as a flat body but as a three-dimensional object (a mass). Therefore, by finding and classifying (sorting) the three-dimensional objects (three-dimensional images) of the specific region, it is possible to efficiently select a noteworthy image P as a representative image D.
Specifically, when there are two three-dimensional images (image A, image B) of a white matter hyperintensity region W, the system sets up group G9 containing image A, group G10 containing image B, and group G11 containing neither image A nor image B. This allows for reliable selection of noteworthy images P as representative images D from each of the two white matter hyperintensity region clusters (groups).

The image division unit 43 may set the groups based on position information in the slice thickness direction or slice plane of the three-dimensional image.
For example, it distinguishes between the 3D images of specific regions in the right brain and those in the left brain (by creating different groups and sorting them), and it distinguishes between the 3D images of specific regions in the frontal lobe and those in the occipital lobe.
For example, if there are multiple 3D images of a specific region in the right brain, these may be grouped together, thereby drawing attention to a specific region in a specific location (such as the right brain).

A case where a plurality of representative images D are similar to each other will be described.
Images P that are adjacent or close to each other in the slice thickness direction may be selected from different groups as representative images D. For example, image P10 of group G1 and image P11 of group G2 are selected as representative images D.
Furthermore, when one image P is assigned to multiple groups in duplicate, the same image P may be selected from different groups as the representative image D. For example, image P4 is assigned to both group G6 and group G7 in duplicate, and image P4 is selected as the representative image D for group G6 and group G7.
In these cases, the representative images D are similar to one another, which reduces the usefulness of the presented image Q (health check report R).

Therefore, the representative image D is reselected as follows.
The image selection unit 44 calculates the similarity between the representative images D. If the similarity exceeds a predetermined reference value, the representative image D is reselected.
The similarity is calculated based on the position information in the slice thickness direction of the representative image D. For example, if the representative images D are located within 10 mm of each other in the slice thickness direction, the representative image D is reselected. In other words, "10 mm" is set as the "predetermined reference value" of the similarity.
The image selector 44 then changes the type of measurement value when reselecting the representative image D. For example, instead of selecting the area of the specific region (white matter hyperintensity region W), the image P (representative image D) with the largest perimeter of the specific region is selected.

In the above-described embodiment, the image division step S3 is performed after the specific region extraction step S1 and the shape measurement step S2, but the order is not limited to this.
If the grouping is not based on the characteristics of the specific region, the image division step S3 may be performed before the specific region extraction step S1. In this case, the image division step S3 only involves setting a plurality of groups.
On the other hand, when grouping is performed based on the characteristics of the specific region, the image division step S3 is performed after the specific region extraction step S1 and the shape measurement step S2.

In the above-described embodiment, horizontal cross-sectional images perpendicular to the body axis direction were used among the cross-sectional images of the subject's brain, but different cross-sectional images may also be used. Frontal cross-sectional images perpendicular to the sagittal horizontal axis direction or sagittal cross-sectional images perpendicular to the frontal horizontal axis direction may also be used. All of the horizontal cross-sectional images, frontal cross-sectional images, and sagittal cross-sectional images may be used, or any one or two of them may be used.

The brain image providing device 30 may be either a cloud-based or on-premise system. The connection (intercommunication) between the brain image providing device 30 and the PACS server 20 may be via either a LAN or the Internet.

1. Brain image provision system (image provision system)
10. Brain imaging device (imaging device)
20 PACS server
15 Client PC
30. Brain image providing device (image providing device)
31 control unit
311 CPU
41 Specific area extraction part
42 Shape measurement section
43 Image division unit
44 Image selection section
45 Image provider
50 Portable information terminal (terminal device)
G1 to G11 Groups
P, P1 to P20 images (cross-sectional images)
D Representative image
Q: Presentation image
R Health Check Report
W White matter hyperintensity area (specific area)
X High signal area around the lateral ventricle (specific area)
Y Deep subcortical high signal area (specific area)

Claims (19)

a specific region extraction unit that extracts a specific region from a plurality of cross-sectional images of an organ of a subject;
a shape measurement unit for measuring the shape of the specific region;
an image division unit that divides each of the cross-sectional images into a plurality of groups;
an image selection unit that selects a representative image from each of the groups based on the measurement values obtained by the shape measurement unit, and when the specific region is not extracted from the cross-sectional image in the group, selects the representative image based on position information of the cross-sectional image in the slice thickness direction;
an image providing unit that provides the representative image;
An image providing device comprising: the organ is the brain,
The image providing device according to claim 1 , wherein the specific region is a white matter hyperintensity region. the organ is the brain,
The image providing device according to claim 1 , wherein the specific region is a stroke region. The image providing device of claim 1, wherein the specific region extraction unit extracts the specific region through pattern recognition. The image providing device of claim 1, wherein the shape measuring unit measures the area, major axis, minor axis, or perimeter of the specific region. the specific region extraction unit obtains a three-dimensional image of the specific region based on position information of the cross-sectional image in a slice thickness direction and position information of the specific region on a slice plane;
The image providing device according to claim 1 , wherein the shape measuring unit measures a volume of the three-dimensional image. The image providing device of claim 1, wherein the image division unit sets the groups based on position information of the cross-sectional images in the slice thickness direction. The image providing device of claim 1, wherein the image division unit sets the group based on the presence or absence of the specific area. The image providing device of claim 8, wherein the image division unit sets the group based on position information of the specific region on the slice plane. The image providing device of claim 8, wherein the image division unit classifies the specific area into multiple categories and sets the groups based on these classifications. the specific region extraction unit obtains a three-dimensional image of the specific region based on position information of the cross-sectional image in a slice thickness direction and position information of the specific region on a slice plane;
The image providing device according to claim 1 , wherein the image dividing section sets the groups based on the presence or absence of the stereoscopic image. The image providing device of claim 11, wherein the image division unit sets the groups based on positional information in the slice thickness direction or slice plane of the three-dimensional image. The image providing device of claim 1, wherein the image selection unit calculates the similarity between the representative images, and if the similarity exceeds a predetermined reference value, reselects the representative image. The image providing device of claim 13, wherein the similarity is calculated based on position information of the representative image in the slice thickness direction. The image providing device of claim 13, wherein the image selection unit changes the type of the measurement value and reselects the representative image. The image providing device according to any one of claims 1 to 15 ;
an image capturing device for capturing the cross-sectional image;
a PACS server that receives the cross-sectional images from the image capturing device and transfers them to the image providing device;
Equipped with
An image providing system, wherein the image providing device transmits the representative image to the PACS server. a terminal device that communicates with the image providing device via a telecommunications network in response to an operation by the subject;
The image providing system according to claim 16 , wherein the image providing device transmits the representative image to the terminal device based on an instruction from the terminal device. a specific region extraction step of extracting specific regions from a plurality of cross-sectional images of an organ of a subject by a specific region extraction unit;
a shape measuring step of measuring the shape of the specific region by a shape measuring unit;
an image dividing step of dividing each of the cross-sectional images into a plurality of groups by an image dividing unit;
an image selection step of selecting a representative image from each of the groups by an image selection unit based on the measurement values obtained in the shape measurement step, and when the specific region is not extracted from the cross-sectional image in the group, selecting the representative image by the image selection unit based on position information in the slice thickness direction of the cross-sectional image;
an image providing step of providing the representative image by an image providing unit;
An image providing method comprising: The CPU of the image providing device
a specific region extraction unit that extracts a specific region from a plurality of cross-sectional images of an organ of a subject;
a shape measurement unit for measuring the shape of the specific region;
an image division unit that divides each of the cross-sectional images into a plurality of groups;
an image selection unit that selects a representative image from each of the groups based on the measurement values obtained by the shape measurement unit, and when the specific region is not extracted from the cross-sectional image in the group, selects the representative image based on position information of the cross-sectional image in the slice thickness direction;
an image providing unit that provides the representative image;
A program to function as a