X-ray image processing apparatus, X-ray image processing method, and program

Abstract

To provide an X-ray image processing device that can reduce the burden on the user when evaluating the shape of the vertebral bodies. [Solution] This X-ray image processing device 100 comprises a display unit 101 and an image processing unit 2. The image processing unit 2 includes a vertebral region estimation unit 20 that individually estimates multiple vertebral regions from an X-ray image 10, a vertebral image generation unit 21 that generates multiple vertebral images containing any of the multiple vertebral regions, a position estimation unit 22 that estimates evaluation positions for evaluating the shape of the vertebral body from the multiple vertebral images, and an evaluation parameter calculation unit 24 that calculates evaluation parameters for evaluating the shape of the vertebral body for each vertebral body based on the evaluation positions. The display unit 101 displays the evaluation parameters calculated for each vertebral body by the evaluation parameter calculation unit 24.

Description

Technical Field

[0001] This invention relates to an X-ray image processing apparatus, an X-ray image processing method, and a program, and particularly to an X-ray image processing apparatus, an X-ray image processing method, and a program for estimating an evaluation position for evaluating the shape of a vertebral body.

Background Art

[0002] Conventionally, an X-ray image processing apparatus for evaluating the shape of a vertebral body has been known. Such an X-ray image processing apparatus is disclosed, for example, in Japanese Patent Application Laid-Open No. 2009-219763.

[0003] Japanese Patent Application Laid-Open No. 2009-219763 discloses an image measuring apparatus that displays a side image of a vertebral body taken from the side of a human body on a display unit. In this image measuring apparatus, for the side image displayed on the display unit, when the user operates a mouse button, the positions of the upper edge of the anterior margin, the lower edge of the anterior margin, the upper edge of the center, the lower edge of the center, the upper edge of the posterior margin, and the lower edge of the posterior margin of the vertebral body are specified. Then, the image measuring apparatus measures the anterior edge height, the center height, and the posterior edge height of the vertebral body based on the specified positions. The user evaluates the shape of the vertebral body based on the measured anterior edge height, center height, and posterior edge height of the vertebral body.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] However, in Japanese Patent Application Laid-Open No. 2009-219763, it is necessary for the user to specify the positions of the upper edge of the anterior margin, the lower edge of the anterior margin, the upper edge of the center, the lower edge of the center, the upper edge of the posterior margin, and the lower edge of the posterior margin of the vertebral body. For this reason, there is a problem that the burden on the user for evaluating the shape of the vertebral body is large.

[0006] This invention was made to solve the above-mentioned problems, and one of its objectives is to provide an X-ray image processing device, an X-ray image processing method, and a program that can reduce the burden on the user in evaluating the shape of the vertebral body. [Means for solving the problem]

[0007] To achieve the above objective, the X-ray image processing apparatus according to the first aspect of this invention comprises an image acquisition unit that acquires an X-ray image showing multiple vertebral bodies, an image processing unit that processes the X-ray image acquired by the image acquisition unit, and a display unit. The image processing unit inputs the X-ray image into a first trained model that has been trained using a first training X-ray image showing multiple vertebral body regions as training data, thereby estimating multiple vertebral body regions individually from the X-ray image, and the vertebral body region estimation unit processes the X-ray image input into the first trained model and the multiple vertebral body regions estimated by the vertebral body region estimation unit. The system includes: a vertebral body image generation unit that generates multiple vertebral body images containing any of the estimated multiple vertebral body regions; a position estimation unit that estimates evaluation positions for evaluating the shape of the vertebral body from the multiple vertebral body images by inputting the multiple vertebral body images into a second pre-trained model trained using a second training X-ray image showing the anterior, central, and posterior edges of the vertebral body as training data; and an evaluation parameter calculation unit that calculates evaluation parameters for evaluating the shape of the vertebral body for each vertebral body based on the evaluation positions. The display unit displays the evaluation parameters calculated for each vertebral body by the evaluation parameter calculation unit.

[0008] Furthermore, in order to achieve the above objective, the X-ray image processing method according to the second aspect of this invention comprises the steps of: acquiring an X-ray image showing multiple vertebral bodies; individually estimating multiple vertebral regions from the X-ray image by inputting the X-ray image into a first trained model that has been trained using a first training X-ray image showing multiple vertebral regions as training data; generating multiple individual vertebral images that include any of the estimated multiple vertebral regions based on the X-ray image input into the first trained model and the estimated multiple vertebral regions; estimating evaluation positions for evaluating the shape of the vertebral bodies from the multiple individual vertebral images by inputting the multiple individual vertebral images into a second trained model that has been trained using a second training X-ray image showing the anterior, central, and posterior edges of the vertebral bodies as training data; calculating evaluation parameters for evaluating the shape of the vertebral bodies based on the evaluation positions for each vertebral body; and displaying the evaluation parameters calculated for each vertebral body on a display unit.

[0009] Furthermore, in order to achieve the above objective, a program according to the third aspect of this invention includes the steps of: acquiring an X-ray image showing multiple vertebral bodies; individually estimating multiple vertebral regions from the X-ray image by inputting the X-ray image into a first trained model that has been trained using a first training X-ray image showing multiple vertebral regions as training data; generating multiple individual vertebral images that include any of the estimated multiple vertebral regions based on the X-ray image input into the first trained model and the estimated multiple vertebral regions; estimating evaluation positions for evaluating the shape of the vertebral bodies from the multiple individual vertebral images by inputting the multiple individual vertebral images into a second trained model that has been trained using a second training X-ray image showing the anterior, central, and posterior edges of the vertebral bodies as training data; calculating evaluation parameters for evaluating the shape of the vertebral bodies based on the evaluation positions for each vertebral body; and displaying the evaluation parameters calculated for each vertebral body on a display unit. [Effects of the Invention]

[0010] In the X-ray imaging apparatus in the first phase, the X-ray image processing method in the second phase, and the program in the third phase, as described above, a first pre-trained model is used to individually estimate multiple vertebral regions from an X-ray image, generating individual vertebral images containing each of the estimated vertebral regions. A second pre-trained model is used to estimate evaluation positions for evaluating the shape of the vertebral body from the individual vertebral images, generating a vertebral shape evaluation image containing the vertebral body and the evaluation position. This allows the user to specify the evaluation position, using the evaluation position automatically estimated using the second pre-trained model as an aid. As a result, the user's burden in evaluating the shape of the vertebral body can be reduced. Therefore, the time required to evaluate the shape of the vertebral body can be shortened. Furthermore, because the evaluation position is automatically estimated using the second pre-trained model, the variability in the estimation of the evaluation position is suppressed compared to when the evaluation position is specified manually. In other words, the reproducibility of specifying the evaluation position is improved. This reduces the negative impact on the evaluation of the shape of the vertebral body caused by poor reproducibility of specifying the evaluation position. [Brief explanation of the drawing]

[0011] [Figure 1] This is a schematic diagram showing the overall configuration of an X-ray image processing device according to one embodiment. [Figure 2] This is a schematic diagram illustrating the configuration of an X-ray imaging device. [Figure 3] This diagram illustrates the evaluation positions used to assess the shape of the vertebral body. [Figure 4] This is a schematic diagram illustrating the learning method of a learning model according to one embodiment. [Figure 5] This figure illustrates the estimation of the vertebral body region according to one embodiment, and the estimation of evaluation positions for evaluating the shape of the vertebral body. [Figure 6] This diagram illustrates how to rotate individual vertebral images so that the centerline of the rectangular vertebral body is horizontal. [Figure 7] This diagram illustrates how to rotate individual vertebral images so that the centerline of the trapezoidal vertebral body becomes horizontal. [Figure 8] This figure illustrates an image used to evaluate the shape of a vertebral body according to one embodiment. [Figure 9] This is a diagram illustrating an overall X-ray image according to one embodiment. [Figure 10] This is a diagram illustrating a partial X-ray image according to one embodiment. [Figure 11] This figure illustrates the analysis result image according to one embodiment. [Figure 12] This figure illustrates an image created from analysis results according to one embodiment. [Figure 13] This is a flowchart illustrating an X-ray image processing method according to one embodiment. [Modes for carrying out the invention]

[0012] The following describes embodiments of the present invention based on the drawings.

[0013] Referring to Figures 1 to 13, the configuration of an X-ray image processing device 100 according to one embodiment will be described. The X-ray image processing device 100 is used to diagnose the shape of the vertebral body 40 caused by a fracture of the vertebral body 40 of a patient 80 (see Figure 2).

[0014] As shown in FIG. 3, there are a quantitative evaluation method and a semi - quantitative evaluation method for determining the shape of the vertebral body 40 caused by a fracture of the vertebral body 40 from the X - ray images 10 of the thoracic and lumbar vertebrae. As a quantitative evaluation method, the QM method (Quantitative Measurement) is known. In the QM method, in the X - ray images 10 of the thoracic and lumbar vertebrae, the anterior edge height (A), the central height (C), and the posterior edge height (P), which are the vertebral body heights of the anterior edge 41, the center 42, and the posterior edge 43 of the vertebral body 40, respectively, are measured. The presence or absence of a fracture of the vertebral body 40 is determined from the ratios of the anterior edge height (A), the central height (C), and the posterior edge height (P). For example, if either the ratio C / A, which is the ratio of the central height (C) to the anterior edge height (A), or the ratio C / P, which is the ratio of the central height (C) to the posterior edge height (P), is less than 0.8, it is determined that the vertebral body 40 is fractured. Also, if the ratio A / P, which is the ratio of the anterior edge height (A) to the posterior edge height (P), is less than 0.75, it is determined that the vertebral body 40 is fractured. Note that the determination of whether the vertebral body 40 is fractured or not is made by the user (doctor). The X - ray image processing apparatus 100 described below is used to assist the user (doctor) in determining the fracture of the vertebral body 40 and does not determine whether the vertebral body 40 is fractured or not.

[0015] (Configuration of the X - ray Image Processing Apparatus) As shown in FIG. 1, the X - ray image processing apparatus 100 includes an image acquisition unit 1, an image processing unit 2, and a storage unit 3.

[0016] The image acquisition unit 1 is configured to acquire an X - ray image 10 in which a plurality of vertebral bodies 40 (see FIG. 5) are imaged. In the present embodiment, the image acquisition unit 1 is configured to acquire, for example, the X - ray image 10 taken by the X - ray imaging apparatus 200 from an image server 210 such as a PACS (Picture Archiving and Communication System). The image acquisition unit 1 includes, for example, an input / output interface.

[0017] The image processing unit 2 is configured to process the X-ray image 10 acquired by the image acquisition unit 1. The image processing unit 2 is a computer composed of a CPU (Central Processing Unit), ROM (Read Only Memory), RAM (Random Access Memory), and a GPU (Graphics Processing Unit). The image processing unit 2, consisting of the CPU and other hardware components, includes the following functional blocks of software (program 2a): a vertebral region estimation unit 20, a vertebral individual image generation unit 21, a position estimation unit 22, an image generation unit 23, an evaluation parameter calculation unit 24, a judgment result input unit 25, a position change input unit 26, a vertebral addition unit 27, a selection unit 28, a display control unit 29, an analysis result output unit 30, and a threshold determination unit 31. Details of each functional block of the image processing unit 2 will be described later.

[0018] In this embodiment, the memory unit 3 is configured to store a first trained model 3a that has been trained using a first training X-ray image 12a (see Figure 4) showing multiple vertebral regions 40a as training data. Specifically, the first trained model 3a is trained using the first training X-ray image 12a showing multiple vertebral regions 40a and a first labeled image 13a in which labels have been assigned to multiple vertebral bodies 40 in the first training X-ray image 12a as training data. The memory unit 3 is also configured to store a second trained model 3b that has been trained using a second training X-ray image 12b (see Figure 4) showing the anterior edge 41, central part 42, and posterior edge 43 of the vertebral body 40 as training data. Specifically, the second training X-ray image 12b is trained using the following as training data: the second training X-ray image 12b showing the anterior edge 41, central 42, and posterior edge 43 of the vertebral body 40, and the second labeled image 13b in which labels are assigned to the anterior edge 41, central 42, and posterior edge 43 of the vertebral body 40 in the second training X-ray image 12b. The storage unit 3 includes, for example, an HDD (Hard Disk Drive) or non-volatile memory.

[0019] (Configuration of X-ray imaging equipment) As shown in Figure 2, the X-ray imaging device 200 comprises an X-ray source 50, an X-ray detection unit 51, an imaging device control unit 52, and an imaging device image processing unit 53. The imaging device control unit 52 is electrically connected to the X-ray source 50 and the imaging device image processing unit 53. The X-ray detection unit 51 is also electrically connected to the imaging device image processing unit 53. The X-ray imaging device 200 generates an X-ray image 10 showing multiple vertebrae 40 by imaging a patient 80 in a lateral recumbent position. The X-ray imaging device 200 then sends the generated X-ray image 10 to an image server 210. In the example shown in Figure 2, electrical connections are illustrated with dashed lines, and information input and output are illustrated with solid arrows.

[0020] The X-ray source 50 generates X-rays when a high voltage is applied. The X-rays generated by the X-ray source 50 are configured to irradiate the X-ray detection unit 51 in the direction in which it is located.

[0021] The X-ray detection unit 51 detects X-rays emitted from the X-ray source 50 and converts the detected X-rays into electrical signals. The X-ray detection unit 51 is, for example, an FPD (Flat Panel Detector). The detection signal (image signal) from the X-ray detection unit 51 is sent to the image processing unit 53 of the imaging device.

[0022] The imaging device control unit 52 is configured to control the X-ray imaging device 200. The imaging device control unit 52 includes, for example, a CPU, ROM, and RAM.

[0023] The imaging device image processing unit 53 is configured to generate an X-ray image 10 based on a detection signal sent from the X-ray detection unit 51. The imaging device image processing unit 53 includes, for example, a processor such as a GPU or an FPGA (Field-Programmable Gate Array) configured for image processing.

[0024] The X-ray image 10 generated in the imaging device image processing unit 53 is sent to the image server 210.

[0025] (Manufacturing a learning model) As shown in Figure 4, the training method for the first learning model 4a includes the steps of: acquiring a first training X-ray image 12a in step 110; acquiring a first label image 13a in step 111; and training the first learning model 4a to individually estimate multiple vertebral regions 40a (see Figure 5) from the first training X-ray image 12a using the first training X-ray image 12a and the first label image 13a as training data in step 112. The first learning model 4a is, for example, a convolutional neural network (CNN) or incorporates a convolutional neural network in part.

[0026] The training method for the second learning model 4b includes the steps of: acquiring a second training X-ray image 12b (step 110a); acquiring a second label image 13b (step 111a); and training the second learning model 4b to estimate evaluation positions for evaluating the shape of the vertebral body 40 from the second training X-ray image 12b (step 112a), using the second training X-ray image 12b and the second label image 13b as training data. The second learning model 4b is, for example, a convolutional neural network (CNN) or incorporates a convolutional neural network in part.

[0027] (Vertebral body area estimation department) In this embodiment, as shown in Figure 5, the vertebral region estimation unit 20 uses a first trained model 3a to individually estimate multiple vertebral regions 40a from the X-ray image 10. As a result, a vertebral label image 120 is obtained in which the positions of multiple vertebral bodies 40 are identified in the X-ray image 10.

[0028] (Vertebral body image generation unit) In this embodiment, the vertebral body image generation unit 21 generates a vertebral body image 130 that includes one of the multiple vertebral body regions 40a estimated by the vertebral body region estimation unit 20. Specifically, the vertebral body image generation unit 21 generates a vertebral body image 130 of a predetermined size based on the center of each of the multiple vertebral body regions 40a estimated by the vertebral body region estimation unit 20. That is, the vertebral body image 130 is an image with a predetermined length L1 in the vertical direction and a predetermined length L2 in the horizontal direction, centered on the center of each of the multiple vertebral body regions 40a estimated by the vertebral body region estimation unit 20.

[0029] Furthermore, as shown in Figures 6 and 7, the vertebral body image generation unit 21 may rotate the vertebral body image 130 so that the center line 40b of the vertebral body 40 depicted in the vertebral body image 130 becomes horizontal. That is, instead of the upper or lower side of the roughly rectangular vertebral body 40 shown in Figure 6 being aligned horizontally, the vertebral body image 130 may be rotated so that the center line 40b connecting the centers of the left and right sides of the vertebral body 40 in the vertical direction becomes horizontal. As a result, even if the vertebral body 40 has a trapezoidal shape as shown in Figure 7, for example, the vertebral body image 130 is rotated so that the vertebral body 40 is aligned horizontally. In this way, by estimating the evaluation position using the second trained model 3b on the vertebral body image 130 rotated so that the center line 40b of the vertebral body 40 is horizontal, it becomes possible to improve the accuracy of the evaluation position estimation.

[0030] (Position estimation part) In this embodiment, as shown in Figure 5, the position estimation unit 22 uses a second trained model 3b to estimate evaluation positions for evaluating the shape of the vertebral body 40 from the individual vertebral body images 130. Specifically, the position estimation unit 22 uses a second trained model 3b, which was trained using a second training X-ray image 12b showing the upper edge 41a of the anterior edge 41, the lower edge 41b of the anterior edge 41, the upper edge 42a of the central 42, the lower edge 42b of the central 42, the upper edge 43a of the posterior edge 43, and the lower edge 43b of the posterior edge 43 of the vertebral body 40 as training data, to estimate the positions of the upper edge 41a of the anterior edge 41, the lower edge 41b of the anterior edge 41, the upper edge 42a of the central 42, the lower edge 42b of the central 42, the upper edge 43a of the posterior edge 43, and the lower edge 43b of the posterior edge 43 from the individual vertebral body images 130. Note that these positions are examples of "evaluation positions" in the claims.

[0031] (Evaluation parameter calculation unit) In this embodiment, the evaluation parameter calculation unit 24 calculates evaluation parameters for evaluating the shape of the vertebral body 40 based on the evaluation position. Specifically, as shown in Figure 8, the evaluation parameter calculation unit 24 calculates the anterior edge height (A) based on the upper edge 41a and lower edge 41b of the anterior edge 41, which are estimated by the position estimation unit 22. The evaluation parameter calculation unit 24 also calculates the central height (C) based on the upper edge 42a and lower edge 42b of the central 42. The evaluation parameter calculation unit 24 also calculates the posterior edge height (P) based on the upper edge 43a and lower edge 43b of the posterior edge 43. The evaluation parameter calculation unit 24 then calculates the ratio of central height (C) to anterior edge height (A), C / A, the ratio of central height (C) to posterior edge height (P), and the ratio of anterior edge height (A) to posterior edge height (P), A / P, as evaluation parameters for evaluating the shape of the vertebral body 40. Note that C / A, C / P, and A / P are examples of the "first ratio," "second ratio," and "third ratio" in the claims, respectively.

[0032] (Image generation unit) In this embodiment, as shown in Figure 8, the image generation unit 23 generates a vertebral body shape evaluation image 140 that includes the vertebral body 40 and the evaluation position. Specifically, the image generation unit 23 generates a vertebral body shape evaluation image 140 that includes evaluation parameters in addition to the vertebral body 40 and the evaluation position. The vertebral body shape evaluation image 140 generated by the image generation unit 23 is displayed on a display unit 101 (see Figure 1) which is provided separately from the X-ray image processing device 100. Note that the display unit 101 may be provided in the X-ray image processing device 100.

[0033] In this embodiment, the image generation unit 23 generates a vertebral body shape evaluation image 140, which includes an overall X-ray image 141 showing multiple vertebral bodies 40. The image generation unit 23 also generates an overall X-ray image 141 on which position markers 141a indicating predetermined positions of each of the multiple vertebral body regions 40a estimated by the vertebral body region estimation unit 20 are superimposed. The position markers 141a are, for example, circular in shape. The predetermined positions are, for example, the centroid positions of the vertebral body regions 40a.

[0034] In this embodiment, the image generation unit 23 generates a whole X-ray image 141 in which multiple vertebral bodies 40 and identification markers 141b that identify the multiple vertebral bodies 40 are superimposed. In Figure 8, the letters T2 to T12 are displayed next to the vertebral bodies 40 as identification markers 141b.

[0035] In this embodiment, the image generation unit 23 generates a vertebral body shape evaluation image 140, which includes a partial X-ray image 142 in which evaluation positions are superimposed on an image of a predetermined vertebral body 40. The image generation unit 23 generates a partial X-ray image 142 in which the positions of the upper edge 41a of the anterior edge 41, the lower edge 41b of the anterior edge 41, the upper edge 42a of the central 42, the lower edge 42b of the central 42, the upper edge 43a of the posterior edge 43, and the lower edge 43b of the posterior edge 43 are superimposed on the partial X-ray image 142. For example, the above positions are represented by circular markers (black circles shown in Figure 8).

[0036] In this embodiment, the image generation unit 23 generates a partial X-ray image 142 in which the identification markers 141b (T2 to T12) are superimposed. For example, the identification marker 141b (T8) is displayed above the partial X-ray image 142.

[0037] Furthermore, the image generation unit 23 generates a vertebral body shape evaluation image 140 that includes C / A, C / P, and A / P as evaluation parameters. For example, next to the identification marker 141b (T2~T12) that identifies the vertebral body 40, the C / A, C / P, and A / P of the vertebral body 40 corresponding to the identification marker 141b are displayed. In other words, the image generation unit 23 generates a list image 143 in which the identification marker 141b and C / A, C / P, and A / P are displayed in a list format.

[0038] (Judgment result input section) The judgment result input unit 25 receives the user's judgment result regarding whether or not the vertebral body 40 is deformed (whether or not it is fractured). For example, the judgment result is entered on the display unit 101 by the user. In Figure 8, the judgment result (G0, G1, G2, or G3, etc.) is entered in the SQ column next to the evaluation parameter.

[0039] (Position change input section) In this embodiment, the position change input unit 26 changes the evaluation position of the vertebral body shape evaluation image 140. For example, the user drags and drops one of the circular markers of the evaluation position (upper edge 41a of the anterior edge 41, lower edge 41b of the anterior edge 41, upper edge 42a of the central 42, lower edge 42b of the central 42, upper edge 43a of the posterior edge 43, and lower edge 43b of the posterior edge 43) on the partial X-ray image 142 displayed on the display unit 101. This changes the evaluation position. In addition, the evaluation parameters are recalculated in accordance with the change in the evaluation position.

[0040] (Additional vertebral body) In this embodiment, the vertebral body addition unit 27 adds vertebral bodies 40 that are captured on the overall X-ray image 141. For example, the user clicks with the mouse on the overall X-ray image 141 displayed on the display unit 101 in an area that has not been estimated as a vertebral body region 40a by the vertebral body region estimation unit 20. As a result, a predetermined range centered on the clicked position is newly added as a vertebral body region 40a. The position estimation unit 22 then estimates the evaluation positions (upper edge 41a of the anterior edge 41, lower edge 41b of the anterior edge 41, upper edge 42a of the central 42, lower edge 42b of the central 42, upper edge 43a of the posterior edge 43, and lower edge 43b of the posterior edge 43) for the vertebral bodies 40 added to the vertebral body addition unit 27. The image generation unit 23 then generates a partial X-ray image 142 in which the evaluation positions are superimposed on the vertebral bodies 40 added to the vertebral body addition unit 27. Furthermore, the evaluation parameter calculation unit 24 calculates C / A, C / P, and A / P as evaluation parameters for the added vertebral body 40. The image generation unit 23 generates a list image 143 that includes the evaluation parameters of the added vertebral body 40. For example, if the vertebral body region 40a of the vertebral body 40 corresponding to T12 has not been estimated by the vertebral body region estimation unit 20, the user clicks the vertebral body 40 corresponding to T12 with the mouse. As a result, the evaluation parameters of the vertebral body 40 corresponding to T12 are calculated, and a vertebral body shape evaluation image 140 including the calculated evaluation parameters is generated.

[0041] (Selection unit, display control unit) In this embodiment, the selection unit 28 selects the position marker 141a on the overall X-ray image 141. Then, when the selection unit 28 selects the position marker 141a, the display control unit 29 displays the partial X-ray image 142 corresponding to the selected position marker 141a on the display unit 101.

[0042] The display control unit 29 either displays the overall X-ray image 141 and the partial X-ray image 142 corresponding to the selected vertebral body 40 side by side on the display unit 101, or switches between displaying the overall X-ray image 141 and the partial X-ray image 142 corresponding to the selected vertebral body 40 on the display unit 101. In Figure 8, the overall X-ray image 141 and the partial X-ray image 142 are displayed side by side on the display unit 101. A list image 143 is also displayed next to the partial X-ray image 142. Furthermore, as shown in Figure 9, if a position marker 141a is selected while only the overall X-ray image 141 is displayed, the display of the overall X-ray image 141 may be switched to display the partial X-ray image 142 and list image 143 corresponding to the selected position marker 141a, as shown in Figure 10. Furthermore, while the partial X-ray image 142 and evaluation parameters are displayed, clicking a toggle button (not shown) with the mouse switches the display to show only the overall X-ray image 141 shown in Figure 9.

[0043] For example, as shown in Figure 8, suppose a partial X-ray image 142 of T8 is displayed on the display unit 101, and the user clicks on a position marker 141a other than T8 on the overall X-ray image 141 with the mouse. As a result, the selection unit 28 selects the clicked position marker 141a, and the display control unit 29 displays the partial X-ray image 142 corresponding to the clicked position marker 141a on the display unit 101.

[0044] Furthermore, in this embodiment, as shown in Figure 8, the image generation unit 23 generates an overall X-ray image 141 in which a selection marker 141c indicating the vertebral body 40 selected by the selection unit 28 is superimposed on the overall X-ray image 141. The image generation unit 23 generates an overall X-ray image 141 in which the selection marker 141c is superimposed on the position marker 141a of the vertebral body 40 selected by the selection unit 28. For example, the selection marker 141c is a circle along the outer edge of the position marker 141a. Also, the color of the selection marker 141c and the color of the position marker 141a are different. Also, the color of the identification marker 141b of the vertebral body 40 selected by the selection unit 28 is different from the color of the identification marker 141b of the vertebral body 40 that has not been selected. In Figure 8, the vertebral body 40 T8 is selected, and the letters T8 are shown in bold to indicate that they are different from the others. Furthermore, the identification marker 141b, T8, for the selected vertebral body 40 is displayed above the partial X-ray image 142. In addition, the evaluation parameters for the selected vertebral body 40 are enclosed in a square frame in the list image 143.

[0045] Furthermore, the vertebral body shape evaluation image 140 includes a slider 144a for adjusting the sharpness of the overall X-ray image 141 (partial X-ray image 142), and a slider 144b for adjusting the contrast. The vertebral body shape evaluation image 140 also displays the patient ID and patient name of the patient 80.

[0046] (Analysis result output section) In this embodiment, as shown in Figure 8, the analysis result output unit 30 commands the output of the analysis results of the shape of the vertebral bodies 40. For example, a report button 145 is displayed in the upper left of the vertebral body shape evaluation image 140 displayed on the display unit 101. When the user clicks the report button 145 with the mouse, the output of the analysis results is commanded. When the analysis result output unit 30 commands the output of the analysis results, the image generation unit 23 generates an analysis result image 150, as shown in Figure 11, which includes an overall X-ray image 141 showing multiple vertebral bodies 40 and evaluation parameters (C / A, C / P, and A / P) for each of the multiple vertebral bodies 40. The analysis result image 150 is displayed on the display unit 101. The overall X-ray image 141 and the list image 143 are displayed side by side on the display unit 101. In addition to the evaluation parameters, the judgment results (G0, G1, G2, or G3, etc.) are also displayed.

[0047] Furthermore, in this embodiment, as shown in Figure 11, the image generation unit 23 generates an analysis result image 150 which includes a whole X-ray image 141 in which multiple vertebral bodies 40 and identification markers 141b (T2~T12) that identify the multiple vertebral bodies 40 are superimposed, and evaluation parameters associated with the identification markers 141b. The analysis result image 150 also displays the analysis result ID (Study ID), the date and time of analysis (Study Date), the patient ID (Patient ID), the patient name (Patient Name), gender, date of birth, etc.

[0048] In this embodiment, when the analysis result output unit 30 commands the image generation unit 23 to output analysis results, it generates an analysis result image 150 that includes a whole X-ray image 141 showing multiple vertebral bodies 40, evaluation parameters for each of the multiple vertebral bodies 40, and partial X-ray images 142 selected by the user. Specifically, when the analysis result output unit 30 commands the output of analysis results, the analysis result creation image 160 (user interface) shown in Figure 12 is displayed on the display unit 101. The analysis result creation image 160 displays the analysis result image 150 along with a partial X-ray image selection unit 161 for selecting the partial X-ray images 142 to be displayed in the analysis result image 150. For example, four partial X-ray image selection units 161 are displayed. In addition, identification markers 141b (T2~T12) are displayed in the lower right corner of the partial X-ray images 142.

[0049] Furthermore, the analysis result image 160 displays a comment input section 162 for entering comments, a username input section 163 for entering the user's name, and a destination selection section 164 for selecting the destination to send the analysis results. The analysis result image 160 also displays a send button 165. When the user clicks the send button 165 with the mouse on the display unit 101, the analysis result image 150 is sent to the image server 210.

[0050] (Threshold determination unit) In this embodiment, the threshold determination unit 31 determines whether the deformation of the vertebral body 40 exceeds a predetermined threshold based on the evaluation parameters. The image generation unit 23 then, based on the determination result of the threshold determination unit 31, makes the display of at least one of the identification marker 141b and the evaluation parameters different for vertebral bodies 40 that exceed the predetermined threshold and vertebral bodies 40 that do not exceed the predetermined threshold. For example, in the analysis result images 150 shown in Figures 11 and 12 (vertebral body shape evaluation image 140 shown in Figure 8), the color of the evaluation parameters of vertebral bodies 40 where either C / A or C / P is less than 0.8, or where A / P is less than 0.75, is different from the color of the evaluation parameters of other vertebral bodies 40. For example, since the identification marker 141b for T10 has a C / A of less than 0.8 (0.75), the color of the evaluation parameter (C / A) for vertebral body T9 40 is different from the color of the evaluation parameters of other vertebral bodies 40. In Figures 11 and 12 (Figure 8), the evaluation parameters for T9 are shown in bold. Furthermore, the color of the identification marker 141b for vertebrae 40 exceeding a predetermined threshold may be different from the color of the identification marker 141b for vertebrae 40 not exceeding the predetermined threshold.

[0051] (X-ray image processing method) Next, we will explain the X-ray image processing method.

[0052] As shown in Figure 13, in step 300, the image acquisition unit 1 acquires an X-ray image 10 showing multiple vertebral bodies 40 from an image server 210, such as a PACS.

[0053] In step 301, the vertebral region estimation unit 20 uses the first trained model 3a to individually estimate multiple vertebral regions 40a from the X-ray image 10.

[0054] In step 302, the vertebral body image generation unit 21 generates a vertebral body image 130 that includes one of the estimated multiple vertebral body regions 40a.

[0055] In step 303, the position estimation unit 22 uses the second trained model 3b to estimate the evaluation position for evaluating the shape of the vertebral body 40 from the individual vertebral body images 130. The evaluation parameter calculation unit 24 then calculates the evaluation parameters based on the evaluation position.

[0056] In step 304, the image generation unit 23 generates a vertebral body shape evaluation image 140 that includes the vertebral body 40 and the evaluation position. Specifically, the image generation unit 23 generates an overall X-ray image 141, a partial X-ray image 142 in which the evaluation position is superimposed, and a list image 143.

[0057] In step 305, if the evaluation position is changed by the position change input unit 26, in step 305a, the evaluation parameter calculation unit 24 calculates the evaluation parameters based on the changed evaluation position.

[0058] In step 306, if the vertebral body 40 captured in the overall X-ray image 141 is added by the vertebral body addition unit 27, in step 306a, the position estimation unit 22 estimates the evaluation position for the vertebral body 40 added to the vertebral body addition unit 27. The image generation unit 23 generates a partial X-ray image 142 in which the evaluation position is superimposed on the vertebral body 40 added to the vertebral body addition unit 27. The evaluation parameter calculation unit 24 calculates evaluation parameters based on the added evaluation position.

[0059] In step 307, if the analysis result output unit 30 is instructed to output the analysis result of the shape of the vertebral body 40, in step 307a, the image generation unit 23 generates an analysis result image 150 which includes a whole X-ray image 141, a partial X-ray image 142 selected by the user, and a list image 143.

[0060] (Effects of this embodiment) In this embodiment, the following effects can be obtained.

[0061] In this embodiment, as described above, a first pre-trained model 3a is used to individually estimate multiple vertebral regions 40a from the X-ray image 10, and individual vertebral images 130 containing each of the estimated multiple vertebral regions 40a are generated. A second pre-trained model 3b is used to estimate evaluation positions for evaluating the shape of the vertebral body 40 from the individual vertebral images 130, and a vertebral shape evaluation image 140 containing the X-ray image 10 of the vertebral body 40 and the evaluation positions is generated. This allows the user to specify evaluation positions with the evaluation positions automatically estimated using the second pre-trained model 3b as an aid. As a result, the burden on the user to evaluate the shape of the vertebral body 40 can be reduced. Therefore, the time required to evaluate the shape of the vertebral body 40 can be shortened. In addition, because the evaluation positions are automatically estimated using the second pre-trained model 3b, the variability in the estimation of evaluation positions is suppressed compared to when evaluation positions are specified manually. That is, the reproducibility of specifying evaluation positions is improved. This reduces the negative impact on the evaluation of the shape of the vertebral body 40 caused by poor reproducibility of specifying evaluation positions.

[0062] Furthermore, in the above embodiment, the following additional effects can be obtained by configuring it as follows.

[0063] In this embodiment, as described above, the image processing unit 2 further includes an evaluation parameter calculation unit 24 that calculates evaluation parameters for evaluating the shape of the vertebral body 40 based on the evaluation position, and the image generation unit 23 generates a vertebral body shape evaluation image 140 that includes the evaluation parameters in addition to the X-ray image 10 of the vertebral body 40 and the evaluation position. As a result, the evaluation parameters are calculated automatically, saving the user the trouble of calculating the evaluation parameters. Consequently, the burden on the user to evaluate the shape of the vertebral body 40 can be further reduced.

[0064] In this embodiment, as described above, the image generation unit 23 generates a vertebral body shape evaluation image 140 that includes C / A, C / P, and A / P as evaluation parameters. This makes it possible to appropriately evaluate the shape of the vertebral body 40 based on C / A, C / P, and A / P.

[0065] In this embodiment, as described above, the image processing unit 2 further includes a position change input unit 26 for changing the evaluation position of the vertebral body shape evaluation image 140. This allows the user to change the evaluation position if the estimated evaluation position is not appropriate, thereby enabling a more appropriate evaluation of the shape of the vertebral body 40.

[0066] In this embodiment, as described above, the vertebral body image generation unit 21 generates a vertebral body image 130 of a predetermined size based on the center of each of the multiple vertebral body regions 40a estimated by the vertebral body region estimation unit 20. This makes it easy to generate a vertebral body image 130 that includes the entire vertebral body region 40a.

[0067] In this embodiment, as described above, the image generation unit 23 generates a vertebral body shape evaluation image 140 which includes a whole X-ray image 141 showing multiple vertebral bodies 40 and a partial X-ray image 142 in which the evaluation position is superimposed on the image of a predetermined vertebral body 40. As a result, the evaluation position is superimposed on the partial X-ray image 142, making it easy to visually identify the evaluation position on the vertebral body 40.

[0068] In this embodiment, as described above, the image processing unit 2 further includes a vertebral body addition unit 27 that adds vertebral bodies 40 captured in the overall X-ray image 141. The position estimation unit 22 estimates the evaluation position for the vertebral bodies 40 added to the vertebral body addition unit 27, and the image generation unit 23 generates a partial X-ray image 142 in which the evaluation position is superimposed on the vertebral bodies 40 added to the vertebral body addition unit 27. This makes it possible to add vertebral body regions 40a that were not estimated by the vertebral body region estimation unit 20, so that the shape of all vertebral bodies 40 can be evaluated.

[0069] In this embodiment, as described above, the image processing unit 2 further includes a selection unit 28 that selects a position marker 141a on the overall X-ray image 141, and a display control unit 29 that, when a position marker 141a is selected by the selection unit 28, displays a partial X-ray image 142 corresponding to the position marker 141a on the display unit 101. As a result, by selecting a position marker 141a of a vertebral body 40 that the user wants to focus on, a partial X-ray image 142 of the selected vertebral body 40 is displayed, so the user can easily view the partial X-ray image 142 of the vertebral body 40 that the user wants to focus on.

[0070] In this embodiment, as described above, when the image generation unit 23 receives a command from the analysis result output unit 30 to output the analysis results, it generates an analysis result image 150 that includes an overall X-ray image 141 showing multiple vertebral bodies 40 and evaluation parameters for each of the multiple vertebral bodies 40. As a result, after the user determines that the evaluation position is appropriate, the user can easily view the analysis result image 150 by inputting a command to output the analysis results of the shape of the vertebral bodies 40.

[0071] In this embodiment, as described above, when the image generation unit 23 is instructed by the analysis result output unit 30 to output the analysis results, it generates an analysis result image 150 that includes a whole X-ray image 141 showing multiple vertebral bodies 40, evaluation parameters for each of the multiple vertebral bodies 40, and a partial X-ray image 142 selected by the user. This allows the user to view the whole X-ray image 141 and evaluation parameters, as well as the partial X-ray image 142 of the vertebral body 40 of interest.

[0072] In this embodiment, as described above, the image generation unit 23 generates an analysis result image 150 which includes an overall X-ray image 141 in which multiple vertebral bodies 40 and identification markers 141b that identify the multiple vertebral bodies 40 are superimposed, and evaluation parameters associated with the identification markers 141b. This allows the user to easily recognize the correspondence between the vertebral bodies 40 in the overall X-ray image 141 and the evaluation parameters.

[0073] In this embodiment, as described above, the image generation unit 23 makes the display mode of at least one of the identification marker 141b and the evaluation parameter different for vertebrae 40 that exceed a predetermined threshold and vertebrae 40 that do not exceed a predetermined threshold. This allows the user to easily distinguish between vertebrae 40 that exceed a predetermined threshold and vertebrae 40 that do not exceed a predetermined threshold.

[0074] [Differentiation] The embodiments disclosed herein should be considered in all respects to be illustrative and not restrictive. The scope of the present invention is indicated by the claims rather than the description of the embodiments above, and further includes all modifications (exceptions) within the meaning and scope of the claims.

[0075] For example, the above embodiment shows an example in which both the evaluation position and evaluation parameters are included in the vertebral body shape evaluation image 140, but the present invention is not limited thereto. For example, the vertebral body shape evaluation image 140 may not include evaluation parameters.

[0076] Furthermore, in the above embodiment, an example was shown in which the upper edge 41a of the anterior edge 41, the lower edge 41b of the anterior edge 41, the upper edge 42a of the central 42, the lower edge 42b of the central 42, the upper edge 43a of the posterior edge 43, and the lower edge 43b of the posterior edge 43 were estimated as evaluation positions from the vertebral body images 130, but the present invention is not limited thereto. For example, positions other than those described above may be estimated as evaluation positions.

[0077] Furthermore, in the above embodiment, the vertebral body image generation unit 21 is shown as generating a vertebral body image 130 of a predetermined size based on the center of each of the multiple vertebral body regions 40a estimated by the vertebral body region estimation unit 20, but the present invention is not limited thereto. For example, the vertebral body image 130 of a predetermined size may be generated based on a position other than the center of each of the multiple vertebral body regions 40a estimated by the vertebral body region estimation unit 20.

[0078] Furthermore, although the above embodiment shows an example in which the image processing unit 2 includes a vertebral body addition unit 27, the present invention is not limited thereto. For example, the image processing unit 2 does not have to include a vertebral body addition unit 27. In this case, evaluation parameters for only the vertebral bodies 40 estimated by the vertebral body region estimation unit 20 are calculated and displayed.

[0079] Furthermore, although the above embodiment shows an example in which the analysis result image 150 includes a partial X-ray image 142 selected by the user, the present invention is not limited thereto. For example, the analysis result image 150 may not include a partial X-ray image 142.

[0080] Furthermore, although the above embodiment shows an example in which the color of at least one of the identification marker 141b and the evaluation parameter differs between vertebrae 40 that exceed a predetermined threshold and vertebrae 40 that do not exceed a predetermined threshold, the present invention is not limited thereto. For example, at least one of the identification marker 141b and the evaluation parameter of vertebrae 40 that exceed a predetermined threshold may be enclosed in a border.

[0081] [Pattern] Those skilled in the art will understand that the exemplary embodiments described above are specific examples of the following embodiments.

[0082] (Item 1) An image acquisition unit that acquires X-ray images showing multiple vertebral bodies, The system includes an image processing unit that processes the X-ray image acquired by the image acquisition unit, The aforementioned image processing unit, A vertebral region estimation unit that individually estimates the multiple vertebral regions from the X-ray image using a first pre-trained model that has been trained using a first training X-ray image showing multiple vertebral regions as training data, A vertebral body image generation unit generates a vertebral body image that includes any of the plurality of vertebral body regions estimated by the vertebral body region estimation unit, A position estimation unit estimates evaluation positions for evaluating the shape of the vertebral body from the individual vertebral body images, using a second pre-trained model trained with a second training X-ray image showing the anterior, central, and posterior edges of the vertebral body as training data. An X-ray image processing apparatus comprising: an image generation unit that generates a vertebral body shape evaluation image including the vertebral body and the evaluation position.

[0083] (Item 2) The aforementioned image processing unit, The system further includes an evaluation parameter calculation unit that calculates evaluation parameters for evaluating the shape of the vertebral body based on the evaluation position, The X-ray image processing apparatus according to item 1, wherein the image generation unit generates a vertebral body shape evaluation image that includes the evaluation parameters in addition to the vertebral body and the evaluation position.

[0084] (Item 3) The position estimation unit uses the second trained model, which has been trained using the second training X-ray image showing the upper edge of the anterior edge, the lower edge of the anterior edge, the upper edge of the central part, the lower edge of the central part, the upper edge of the posterior edge, and the lower edge of the posterior edge of the vertebral body as training data, to estimate the positions of the upper edge of the anterior edge, the lower edge of the anterior edge, the upper edge of the central part, the lower edge of the central part, the upper edge of the posterior edge, and the lower edge of the posterior edge from the individual vertebral body images. The evaluation parameter calculation unit, Based on the upper edge and lower edge of the front edge estimated by the position estimation unit, the front edge height is calculated; based on the upper edge and lower edge of the center, the center height is calculated; and based on the upper edge and lower edge of the rear edge, the rear edge height is calculated. As evaluation parameters for evaluating the shape of the vertebral body, a first ratio is calculated, which is the ratio of the central height to the anterior edge height; a second ratio is calculated, which is the ratio of the central height to the posterior edge height; and a third ratio is calculated, which is the ratio of the anterior edge height to the posterior edge height. The X-ray image processing apparatus according to item 2, wherein the image generation unit generates a vertebral body shape evaluation image that includes the positions of the vertebral body, the upper edge of the anterior edge, the lower edge of the anterior edge, the upper edge of the center, the lower edge of the center, the upper edge of the posterior edge, and the lower edge of the posterior edge, and the first ratio, the second ratio, and the third ratio.

[0085] (Item 4) The X-ray image processing apparatus according to any one of items 1 to 3, further comprising a position change input unit for changing the evaluation position of the vertebral body shape evaluation image.

[0086] (Item 5) The X-ray image processing apparatus according to any one of items 1 to 4, wherein the vertebral body-specific image generation unit generates a vertebral body-specific image of a predetermined size based on the center of each of the plurality of vertebral body regions estimated by the vertebral body region estimation unit.

[0087] (Item 6) The X-ray image processing apparatus according to any one of items 1 to 5, wherein the image generation unit generates a vertebral body shape evaluation image that includes a whole X-ray image showing the plurality of vertebral bodies and a partial X-ray image in which the evaluation position is superimposed on a predetermined image of the vertebral body.

[0088] (Item 7) The image processing unit further includes a vertebral body addition unit that adds the vertebral bodies captured in the overall X-ray image, The position estimation unit estimates the evaluation position for the vertebral body added to the vertebral body addition unit. The X-ray image processing apparatus according to item 6, wherein the image generation unit generates a partial X-ray image in which the evaluation position is superimposed on the vertebral body added to the vertebral body addition unit.

[0089] (Item 8) The image generation unit generates the overall X-ray image in which position markers indicating predetermined positions for each of the plurality of vertebral regions estimated by the vertebral region estimation unit are superimposed. The aforementioned image processing unit, On the overall X-ray image, a selection unit for selecting the position marker, The X-ray image processing apparatus according to item 6 or 7, further comprising: a display control unit that, when the position marker is selected by the selection unit, displays the partial X-ray image corresponding to the position marker on the display unit.

[0090] (Item 9) The aforementioned image processing unit, An evaluation parameter calculation unit calculates evaluation parameters for evaluating the shape of the vertebral body based on the evaluation position estimated by the position estimation unit, The system further includes an analysis result output unit that commands the output of the analysis results of the vertebral body shape, The X-ray image processing apparatus according to any one of items 6 to 8, wherein the image generation unit generates an analysis result image including the overall X-ray image showing the plurality of vertebrae and the evaluation parameters for each of the plurality of vertebrae when the analysis result output unit is instructed to output the analysis result.

[0091] (Item 10) The X-ray image processing apparatus according to item 9, wherein the image generation unit generates an analysis result image that includes the overall X-ray image showing the plurality of vertebrae, the evaluation parameters for each of the plurality of vertebrae, and the partial X-ray image selected by the user, when the analysis result output unit is instructed to output the analysis result.

[0092] (Item 11) The X-ray image processing apparatus according to item 10, wherein the image generation unit generates the analysis result image, which includes the overall X-ray image in which the plurality of vertebrae and identification markers for identifying the plurality of vertebrae are superimposed, and the evaluation parameters associated with the identification markers.

[0093] (Item 12) The image processing unit further includes a threshold determination unit that determines whether the shape of the vertebral body exceeds a predetermined threshold based on the evaluation parameters, The X-ray image processing apparatus according to any one of items 9 to 11, wherein the image generation unit, based on the determination result of the threshold determination unit, makes the display mode of at least one of the identification marker and the evaluation parameter different for vertebrae that exceed a predetermined threshold and vertebrae that do not exceed a predetermined threshold.

[0094] (Item 13) The steps include obtaining an X-ray image showing multiple vertebral bodies, The process involves using a first trained model, which has been trained using a first training X-ray image showing multiple vertebral regions as training data, to individually estimate the multiple vertebral regions from the X-ray image, The steps include generating a vertebral body image that includes any of the estimated multiple vertebral body regions, Using a second trained model trained with a second training X-ray image showing the anterior, central, and posterior edges of the vertebral body as training data, the step of estimating the evaluation position for evaluating the shape of the vertebral body from the individual vertebral body images, An X-ray image processing method comprising the step of generating a vertebral body shape evaluation image including the vertebral body and the evaluation position.

[0095] (Item 14) The steps include obtaining an X-ray image showing multiple vertebral bodies, The process involves using a first trained model, which has been trained using a first training X-ray image showing multiple vertebral regions as training data, to individually estimate the multiple vertebral regions from the X-ray image, The steps include generating a vertebral body image that includes any of the estimated multiple vertebral body regions, Using a second trained model trained with a second training X-ray image showing the anterior, central, and posterior edges of the vertebral body as training data, the step of estimating the evaluation position for evaluating the shape of the vertebral body from the individual vertebral body images, A program comprising the step of generating a vertebral body shape evaluation image including the vertebral body and the evaluation position. [Explanation of Symbols]

[0096] 1 Image acquisition unit 2 Image Processing Unit 2a Program 3a First pre-trained model 3b Second pre-trained model 10 X-ray image 12a 1st teacher X-ray image 12b X-ray image for second teacher 20 Vertebral body area estimation section 21 Vertebral body-specific image generation unit 22 Position estimation part 23 Image generation unit 24 Evaluation parameter calculation unit 26 Position change input section 27 Additional vertebral body 28 Selection Section 29 Display Control Unit 30 Analysis Result Output Unit 31 Threshold determination unit 40 vertebral bodies 40a vertebral region 41 Leading edge 41a Upper edge of the anterior edge (evaluation position) 41b Lower edge of the anterior margin (evaluation position) 42 center 42a Upper center edge (evaluation position) 42b Lower center edge (evaluation point) 43 Trailing edge 43a Upper edge of the trailing edge (evaluation position) 43b Lower edge of the posterior margin (evaluation position) 100 X-ray imaging processing equipment 101 Display section 130 individual vertebral images 140 Vertebral Body Shape Evaluation Images 141 Overall X-ray image 141a Location indicator 141b Identification mark 142 Partial X-ray image 150 Analysis result images A Front edge height C center height P trailing edge height C / A ratio 1 C / P ratio (second ratio) A / P Third Ratio

Claims

1. An image acquisition unit that acquires X-ray images showing multiple vertebral bodies, An image processing unit that processes the X-ray image acquired by the image acquisition unit, It includes a display unit, The aforementioned image processing unit, A vertebral region estimation unit estimates the multiple vertebral regions individually from the X-ray image by inputting the aforementioned X-ray image into a first trained model that has been trained using a first training X-ray image showing multiple vertebral regions as training data. A vertebral body image generation unit generates multiple vertebral body images, each containing one of the multiple vertebral body regions estimated by the vertebral body region estimation unit, based on the X-ray image input to the first trained model and the multiple vertebral body regions estimated by the vertebral body region estimation unit. A position estimation unit estimates evaluation positions for evaluating the shape of a vertebra from multiple vertebral body images by inputting multiple vertebral body images into a second trained model that has been trained using a second training X-ray image showing the anterior, central, and posterior edges of the vertebral body as training data. Includes an evaluation parameter calculation unit that calculates evaluation parameters for each vertebra to evaluate the shape of the vertebra based on the evaluation position, The X-ray image processing apparatus is characterized in that the display unit displays the evaluation parameters calculated for each vertebral body by the evaluation parameter calculation unit.

2. The image processing unit further includes an image generation unit that generates a vertebral shape evaluation image that includes a vertebral body corresponding to the vertebral region included in the vertebral body-specific image input to the second trained model and information indicating the evaluation position, The X-ray image processing apparatus according to claim 1, wherein the image generation unit generates a vertebral body shape evaluation image that includes the evaluation parameters in addition to the vertebral body and the evaluation position.

3. The position estimation unit uses the second trained model, which has been trained using the second training X-ray image showing the upper edge of the anterior edge, the lower edge of the anterior edge, the upper edge of the center, the lower edge of the center, the upper edge of the posterior edge, and the lower edge of the posterior edge of the vertebral body as training data, to estimate the positions of the upper edge of the anterior edge, the lower edge of the anterior edge, the upper edge of the center, the lower edge of the center, the upper edge of the posterior edge, and the lower edge of the posterior edge from the individual vertebral body images. The evaluation parameter calculation unit, Based on the upper edge and lower edge of the front edge estimated by the position estimation unit, the front edge height is calculated; based on the upper edge and lower edge of the center, the center height is calculated; and based on the upper edge and lower edge of the rear edge, the rear edge height is calculated. As evaluation parameters for evaluating the shape of the vertebral body, a first ratio is calculated, which is the ratio of the central height to the anterior edge height; a second ratio is calculated, which is the ratio of the central height to the posterior edge height; and a third ratio is calculated, which is the ratio of the anterior edge height to the posterior edge height. The X-ray image processing apparatus according to claim 2, wherein the image generation unit generates a vertebral body shape evaluation image that includes the positions of the vertebral body, the upper edge of the anterior edge, the lower edge of the anterior edge, the upper edge of the center, the lower edge of the center, the upper edge of the posterior edge, and the lower edge of the posterior edge, and the first ratio, the second ratio, and the third ratio.

4. The X-ray image processing apparatus according to claim 2, wherein the image processing unit further includes a position change input unit for changing the evaluation position of the vertebral body shape evaluation image.

5. The X-ray image processing apparatus according to claim 1, wherein the vertebral body-specific image generation unit generates a vertebral body-specific image of a predetermined size based on the center of each of the plurality of vertebral body regions estimated by the vertebral body region estimation unit.

6. The X-ray image processing apparatus according to claim 2, wherein the image generation unit generates a vertebral body shape evaluation image that includes a whole X-ray image showing the plurality of vertebral bodies and a partial X-ray image in which the evaluation position is superimposed on an image of a predetermined vertebral body.

7. The image processing unit further includes a vertebral body addition unit that adds the vertebral bodies captured in the overall X-ray image, The position estimation unit estimates the evaluation position for the vertebral body added to the vertebral body addition unit. The X-ray image processing apparatus according to claim 6, wherein the image generation unit generates a partial X-ray image in which the evaluation position is superimposed on the vertebral body added to the vertebral body addition unit.

8. The image generation unit generates the overall X-ray image in which position markers indicating predetermined positions of each of the multiple vertebral regions estimated by the vertebral region estimation unit are superimposed. The aforementioned image processing unit, On the overall X-ray image, a selection unit for selecting the position marker, The X-ray image processing apparatus according to claim 6, further comprising a display control unit that, when the position marker is selected by the selection unit, displays the partial X-ray image corresponding to the position marker on the display unit.

9. The image processing unit further includes an analysis result output unit that commands the output of the analysis results of the shape of the vertebral body, The X-ray image processing apparatus according to claim 6, wherein the image generation unit generates an analysis result image including the overall X-ray image showing the plurality of vertebrae and the evaluation parameters for each of the plurality of vertebrae when the analysis result output unit commands the output of the analysis result.

10. The X-ray image processing apparatus according to claim 9, wherein the image generation unit generates an analysis result image that includes the overall X-ray image showing the plurality of vertebrae, the evaluation parameters for each of the plurality of vertebrae, and the partial X-ray image selected by the user, when the analysis result output unit is instructed to output the analysis result.

11. The X-ray image processing apparatus according to claim 10, wherein the image generation unit generates an analysis result image which includes the plurality of vertebrae and identification markers for identifying the plurality of vertebrae superimposed on the overall X-ray image and the evaluation parameters associated with the identification markers.

12. The image processing unit further includes a threshold determination unit that determines whether the shape of the vertebral body exceeds a predetermined threshold based on the evaluation parameters, The X-ray image processing apparatus according to claim 9, wherein the image generation unit, based on the determination result of the threshold determination unit, makes the display mode of at least one of the identification markers for identifying the plurality of vertebrae and the evaluation parameters different for vertebrae that exceed a predetermined threshold and vertebrae that do not exceed a predetermined threshold.

13. The vertebral body image generation unit rotates the multiple vertebral body images so that the centerlines of the vertebral bodies depicted in the multiple vertebral body images are horizontal. The X-ray image processing apparatus according to claim 1, wherein the position estimation unit estimates the evaluation position for evaluating the shape of the vertebrae from the plurality of vertebrae-specific images by inputting the plurality of vertebrae-specific images rotated by the vertebrae-specific image generation unit into a second trained model.

14. The steps include obtaining an X-ray image showing multiple vertebral bodies, The steps include: inputting the aforementioned X-ray image into a first trained model that has been trained using a first training X-ray image showing multiple vertebral regions as training data, thereby individually estimating the multiple vertebral regions from the X-ray image; The steps include generating multiple vertebral body images, each containing one of the estimated multiple vertebral body regions, based on the X-ray image input to the first trained model and the estimated multiple vertebral body regions, The process involves inputting multiple images of individual vertebrae into a second trained model, which has been trained using a second training X-ray image showing the anterior, central, and posterior edges of the vertebrae as training data, thereby estimating evaluation positions for evaluating the shape of the vertebrae from the multiple images of individual vertebrae. A step of calculating evaluation parameters for each vertebra to evaluate the shape of the vertebra based on the evaluation position, An X-ray image processing method comprising the step of displaying the evaluation parameters calculated for each vertebral body on a display unit.

15. The steps include obtaining an X-ray image showing multiple vertebral bodies, The steps include: inputting the aforementioned X-ray image into a first trained model that has been trained using a first training X-ray image showing multiple vertebral regions as training data, thereby individually estimating the multiple vertebral regions from the X-ray image; The steps include generating multiple vertebral body images, each containing one of the estimated multiple vertebral body regions, based on the X-ray image input to the first trained model and the estimated multiple vertebral body regions, The process involves inputting multiple images of individual vertebrae into a second trained model, which has been trained using a second training X-ray image showing the anterior, central, and posterior edges of the vertebrae as training data, thereby estimating evaluation positions for evaluating the shape of the vertebrae from the multiple images of individual vertebrae. A step of calculating evaluation parameters for each vertebra to evaluate the shape of the vertebra based on the evaluation position, A program comprising the step of displaying the evaluation parameters calculated for each vertebral body on a display unit.

Images