Presumption device

Abstract

To provide a processing device and an estimation system.SOLUTION: The processing device comprises a receiving unit which receives estimated values of at least one of a bone density and a bone quantity estimated on the basis of a learned parameter from an estimation simple x-ray image obtained by imaging a target part of a first person.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] This disclosure relates to Estimation technology.

Background Art

[0002] Patent Document 1 describes a technique for diagnosing osteoporosis. Patent Document 2 describes a technique for estimating bone strength.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Patent Document 2

Summary of the Invention

Means for Solving the Problems

[0004] Estimation device is disclosed. In one embodiment, The estimation device includes an estimation unit, a detection unit, and a determination unit. The estimation unit estimates at least one estimated value of the bone density and bone mass of the first person based on the first learned parameters learned based on the simple X-ray image for estimation of the target part of the first person and the information regarding the health condition of the first person, the first simple X-ray image for learning of the target part of the second person, and the first learning data including the information regarding the health condition of the second person. The detection unit detects the fracture location of the first person based on the second learned parameters for fracture detection from the simple X-ray image for estimation. The determination unit determines whether the first person has osteoporosis based on the estimated value, the fracture location, and a predetermined criterion for whether it is osteoporosis.

Brief Description of the Drawings

[0006]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Figure 13

Figure 14

Figure 15

Figure 16

Figure 17

Figure 18

Embodiments for Carrying Out the Invention

[0007] Embodiment 1. FIG. 1 is a block diagram showing an example of the configuration of the computer device 1 according to Embodiment 1 of the present invention. The computer device 1 functions as an estimation device for estimating bone density. Hereinafter, the computer device 1 may be referred to as the "estimation device 1".

[0008] As shown in FIG. 1, the estimation device 1 includes, for example, a control unit 10, a storage unit 20, a communication unit 30, a display unit 40, and an input unit 50. The control unit 10, the storage unit 20, the communication unit 30, the display unit 40, and the input unit 50 are electrically connected to each other by, for example, a bus 60.

[0009] The control unit 10 can comprehensively manage the operation of the estimation device 1 by controlling other components of the estimation device 1. The control unit 10 can also be referred to as a control device or a control circuit. The control unit 10 includes at least one processor to provide control and processing capabilities for executing various functions, as will be described in more detail below.

[0010] According to various embodiments, at least one processor may be implemented as a single integrated circuit (IC), or as a plurality of communicatively connected integrated circuits (ICs) and / or discrete circuits. At least one processor can be executed according to various known techniques.

[0011] In one embodiment, the processor includes one or more circuits or units configured to execute one or more data calculation procedures or processes, for example, by executing instructions stored in a related memory. In other embodiments, the processor may be firmware (e.g., discrete logic components) configured to execute one or more data calculation procedures or processes.

[0012] According to various embodiments, the processor includes one or more processors, controllers, microprocessors, microcontrollers, application-specific integrated circuits (ASICs), digital signal processing devices, programmable logic devices, field-programmable gate arrays, or any combination of these devices or configurations, or combinations of other known devices and configurations, and may execute the functions described below. In this example, the control unit 10 includes, for example, a CPU (Central Processing Unit).

[0013] The storage unit 20 includes a non-temporary recording medium that can be read by the CPU of the control unit 10, such as a ROM (Read Only Memory) and a RAM (Random Access Memory). The storage unit 20 stores a control program 100 for controlling the estimation device 1. Various functions of the control unit 10 are realized by the CPU of the control unit 10 executing the control program 100 in the storage unit 20. The control program 100 can also be said to be a bone density estimation program for causing the computer device 1 to function as an estimation device. In this example, by the control unit 10 executing the control program 100 in the storage unit 20, an approximator 280 capable of outputting an estimated value 300 of bone density is formed in the control unit 10 as shown in FIG. 2. The approximator 280 includes, for example, a neural network 200. The control program 100 can also be said to be a program for causing the computer device 1 to function as the neural network 200. Hereinafter, the estimated value of bone density may be referred to as the "bone density estimated value". The configuration example of the neural network 200 will be described in detail later.

[0014] In addition to the control program 100, the storage unit 20 stores learned parameters 110 related to the neural network 200, estimation data 120 (hereinafter also referred to as "input information"), learning data 130, and teacher data 140. The learning data 130 and the teacher data 140 are data used when training the neural network 200. The learned parameters 110 and the estimation data 120 are data used when the learned neural network 200 estimates bone density.

[0015] The learning data 130 is the data input to the input layer 210 of the neural network 200 during the learning of the neural network 200. The learning data 130 is also called training data. The teacher data 140 is the data indicating the correct value of bone density. The teacher data 140 is compared with the output data output from the output layer 230 of the neural network 200 during the learning of the neural network 200. The learning data 130 and the teacher data 140 may be collectively called supervised learning data.

[0016] The estimation data 120 is the data input to the input layer 210 when the learned neural network 200 estimates bone density. The learned parameters 110 are the learned parameters within the neural network 200. It can be said that the learned parameters 110 are the parameters adjusted by the learning of the neural network 200. The learned parameters 110 include the weighting coefficients indicating the weights of the connections between artificial neurons. As shown in FIG. 2, the learned neural network 200 performs an operation based on the learned parameters 110 on the estimation data 120 input to the input layer 210, and outputs a bone density estimation value 300 from the output layer 230.

[0017] Note that the data input to the input layer 210 may be input to the input layer 210 via the input unit 50, or may be directly input to the input layer 210. When directly input to the input layer 210, the input layer 210 may be part or all of the input unit 50. Hereinafter, the bone density estimation value 300 may be referred to as the estimation result 300.

[0018] The communication unit 30 is connected to a communication network including the Internet or the like, either wired or wirelessly. The communication unit 30 can communicate with other devices such as a cloud server and a web server through the communication network. The communication unit 30 can input the information received from the communication network to the control unit 10. Also, the communication unit 30 can output the information received from the control unit 10 to the communication network.

[0019] The display unit 40 is, for example, a liquid crystal display or an organic EL display. By being controlled by the control unit 10, the display unit 40 can display various kinds of information such as characters, symbols, and figures.

[0020] The input unit 50 can receive an input from the user to the estimation device 1. The input unit 50 includes, for example, a keyboard and a mouse. The input unit 50 may include a touch panel capable of detecting a user's operation on the display surface of the display unit 40.

[0021] Note that the configuration of the estimation device 1 is not limited to the above example. For example, the control unit 10 may include a plurality of CPUs. Also, the control unit 10 may include at least one DSP. Further, all or some of the functions of the control unit 10 may be realized by a hardware circuit that does not require software for realizing the function. Also, the storage unit 20 may include a non-transitory computer-readable recording medium other than a ROM and a RAM. The storage unit 20 may include, for example, a small hard disk drive and an SSD (Solid State Drive). Also, the storage unit 20 may include a memory such as a USB (Universal Serial Bus) memory that is detachable from the estimation device 1. Hereinafter, a memory that is detachable from the estimation device 1 may be referred to as a "removable memory".

[0022] <Configuration Example of Neural Network> FIG. 3 is a diagram showing an example of the configuration of the neural network 200. In this example, the neural network 200 is, for example, a convolutional neural network (CNN). As shown in FIG. 3, the neural network 200 includes, for example, an input layer 210, a hidden layer 220, and an output layer 230. The hidden layer 220 is also called an intermediate layer. The hidden layer 220 includes, for example, a plurality of convolutional layers 240, a plurality of pooling layers 250, and a fully connected layer 260. In the neural network 200, the fully connected layer 260 exists before the output layer 230. And in the neural network 200, between the input layer 210 and the fully connected layer 260, the convolutional layers 240 and the pooling layers 250 are alternately arranged.

[0023] Note that the configuration of the neural network 200 is not limited to the example of FIG. 3. For example, the neural network 200 may include one convolutional layer 240 and one pooling layer 250 between the input layer 210 and the fully connected layer 260. Also, the neural network 200 may be a neural network other than a convolutional neural network.

[0024] <An Example of Estimation Data, Learning Data, and Teacher Data> The estimation data 120 includes image data of a simple X-ray image in which the bone to be estimated for bone density is imaged. The object to be estimated for bone density is, for example, a human. Therefore, it can be said that the estimation data 120 includes image data of a simple X-ray image in which a human bone is imaged. The learning data 130 includes image data of a plurality of simple X-ray images in which human bones are imaged. A simple X-ray image is a two-dimensional image and is also called a general X-ray image or a roentgenogram. Note that the object to be estimated for bone density may be other than a human. For example, the object to be estimated for bone density may be an animal such as a dog, a cat, or a horse. Also, the target bone is mainly cortical bone and cancellous bone derived from organisms, but the target bone may include an artificial bone mainly composed of calcium phosphate or a regenerated bone artificially manufactured by regenerative medicine or the like.

[0025] Hereinafter, the image data included in the estimation data 120 may be referred to as "estimation image data". Also, the simple X-ray image shown by the image data included in the estimation data 120 may be referred to as "estimation simple X-ray image". Also, the image data included in the learning data 130 may be referred to as "learning image data". Also, the simple X-ray image shown by the image data included in the learning data 130 may be referred to as "learning simple X-ray image". The learning data 130 includes a plurality of learning X-ray image data respectively showing a plurality of learning simple X-ray images.

[0026] As the imaging site of the estimation simple X-ray image, for example, the head, neck, chest, waist, hip joint, knee joint, ankle joint, foot, toe, shoulder joint, elbow joint, wrist joint, hand, finger or jaw joint is adopted. In other words, as the estimation data 120, there are image data of a simple X-ray image obtained by irradiating the head with X-rays, image data of a simple X-ray image obtained by irradiating the neck with X-rays, image data of a simple X-ray image obtained by irradiating the chest with X-rays, image data of a simple X-ray image obtained by irradiating the waist with X-rays, image data of a simple X-ray image obtained by irradiating the hip joint with X-rays, image data of a simple X-ray image obtained by irradiating the knee joint with X-rays, image data of a simple X-ray image obtained by irradiating the ankle joint with X-rays, image data of a simple X-ray image obtained by irradiating the foot with X-rays, image data of a simple X-ray image obtained by irradiating the toe with X-rays, image data of a simple X-ray image obtained by irradiating the shoulder joint with X-rays, image data of a simple X-ray image obtained by irradiating the elbow joint with X-rays, image data of a simple X-ray image obtained by irradiating the wrist joint with X-rays, image data of a simple X-ray image obtained by irradiating the hand with X-rays, image data of a simple X-ray image obtained by irradiating the finger with X-rays or image data of a simple X-ray image obtained by irradiating the jaw joint with X-rays. The simple X-ray image obtained by irradiating the chest with X-rays includes a simple X-ray image showing the lungs and a simple X-ray image showing the thoracic vertebrae. Note that the types of imaging sites of the estimation simple X-ray image are not limited to this. Also, the estimation simple X-ray image may be a front view in which the target site is imaged from the front or a side view in which the target site is imaged from the side.

[0027] Among the imaging regions of the plurality of learning X-ray images respectively indicated by the plurality of learning image data included in the learning data 130, for example, at least one of the head, neck, chest, waist, hip joint, knee joint, ankle joint, foot, toes, shoulder joint, elbow joint, wrist joint, hand, fingers, and jaw joint is included. In other words, the learning data 130 includes at least one of the image data of the simple X-ray images obtained by irradiating X-rays on the head, the image data of the simple X-ray images obtained by irradiating X-rays on the neck, the image data of the simple X-ray images obtained by irradiating X-rays on the chest, the image data of the simple X-ray images obtained by irradiating X-rays on the waist, the image data of the simple X-ray images obtained by irradiating X-rays on the hip joint, the image data of the simple X-ray images obtained by irradiating X-rays on the knee joint, the image data of the simple X-ray images obtained by irradiating X-rays on the ankle joint, the image data of the simple X-ray images obtained by irradiating X-rays on the foot, the image data of the simple X-ray images obtained by irradiating X-rays on the toes, the image data of the simple X-ray images obtained by irradiating X-rays on the shoulder joint, the image data of the simple X-ray images obtained by irradiating X-rays on the elbow joint, the image data of the simple X-ray images obtained by irradiating X-rays on the wrist joint, the image data of the simple X-ray images obtained by irradiating X-rays on the hand, the image data of the simple X-ray images obtained by irradiating X-rays on the fingers, and the image data of the simple X-ray images obtained by irradiating X-rays on the jaw joint, out of the 15 types of image data. The learning data 130 may include some types of the 15 types of image data, or may include all types of image data. Note that the types of the imaging regions of the learning simple X-ray images are not limited to this. Also, the plurality of learning simple X-ray images may include a frontal image or a lateral image. Further, the plurality of learning simple X-ray images may include both a frontal image and a lateral image of the same imaging region.

[0028] The teacher data 140 includes measurement values of bone density of a person having bones shown in the learning simple X-ray images included in the learning data 130, for each of the plurality of learning image data included in the learning data 130. The plurality of bone density measurement values included in the teacher data 140 include, for example, measurement values of bone density measured by irradiating X-rays to the lumbar spine, bone density measured by irradiating X-rays to the proximal femur, bone density measured by irradiating X-rays to the radius, bone density measured by irradiating X-rays to the middle phalanx, bone density measured by applying ultrasonic waves to the wrist, and at least one type of bone density measured by applying ultrasonic waves to the heel. Hereinafter, the bone density measurement value included in the teacher data 140 may be referred to as "reference bone density".

[0029] Here, as a method for measuring bone density, the DEXA (dual-energy X-ray absorptiometry) method is known. In a DEXA device that measures bone density using the DEXA method, when measuring the bone density of the lumbar spine, X-rays (specifically, two types of X-rays) are irradiated to the lumbar spine from the front. Also, in a DEXA device, when measuring the bone density of the proximal femur, X-rays are irradiated to the proximal femur from the front.

[0030] The teacher data 140 may include the bone density of the lumbar spine measured by a DEXA device, or may include the bone density of the proximal femur measured by a DEXA device. Note that the teacher data 140 may include bone density measured by irradiating X-rays to the target site from the side. For example, the teacher data 140 may include bone density measured by irradiating X-rays to the lumbar spine from the side.

[0031] Also, as another method for measuring bone density, the ultrasonic wave method is known. In a device that measures bone density using the ultrasonic wave method, for example, ultrasonic waves are applied to the wrist to measure the bone density of the wrist, or ultrasonic waves are applied to the heel to measure the bone density of the heel. The teacher data 140 may include bone density measured by the ultrasonic wave method.

[0032] Among the plurality of learning X-ray images shown by the plurality of learning image data included in the learning data 130, bones of a plurality of different people are respectively depicted. And, as shown in FIG. 4, in the storage unit 20, for each of the plurality of learning image data included in the learning data 130, the reference bone density of the person having the bone depicted in the learning X-ray image shown by the learning image data is associated. It can also be said that for each of the plurality of learning X-ray images used in the learning of the neural network 200, the reference bone density of the person having the bone depicted in the learning X-ray image is associated. The reference bone density associated with the learning image data is the bone density measured for the same person as the person having the bone depicted in the learning X-ray image at approximately the same time when the learning X-ray image shown by the learning image data was taken.

[0033] The site depicted in the learning X-ray image shown by the learning image data (in other words, the imaging site of the learning X-ray image) may or may not include the site (that is, the bone) where the reference bone density corresponding to the learning image data was measured. In other words, the site depicted in the learning X-ray image may or may not include the site where the reference bone density corresponding to the learning X-ray image was measured. As an example of the former, a case where learning image data showing a learning X-ray image depicting the lumbar region is associated with the reference bone density of the lumbar vertebra can be considered. As another example, a case where learning image data showing a hip joint is associated with the reference bone density of the proximal part of the femur can be considered. On the other hand, as an example of the latter, a case where learning image data showing the chest is associated with the reference bone density of the lumbar vertebra can be considered. As another example, a case where learning image data showing the knee joint is associated with the reference bone density of the heel can be considered.

[0034] In addition, the orientation of the part shown in the plain X-ray image indicated by the learning image data and the orientation of the X-ray irradiation on the target part in the measurement of the reference bone density corresponding to the learning image data may be the same or different. In other words, the orientation of the part shown in the learning plain X-ray image and the orientation of the X-ray irradiation on the target part in the measurement of the reference bone density corresponding to the learning plain X-ray image may be the same or different. As an example of the former, there may be a case where the learning image data showing a plain X-ray image of the chest taken from the front (hereinafter sometimes referred to as a "front chest plain X-ray image") is associated with the reference bone density measured by irradiating the lumbar spine with X-rays from the front. As another example, there may be a case where the learning image data showing a plain X-ray image of the waist taken from the front (hereinafter sometimes referred to as a "front waist plain X-ray image") is associated with the reference bone density measured by irradiating the proximal part of the femur with X-rays from the front. On the other hand, as an example of the latter, there may be a case where the learning image data showing a plain X-ray image of the waist taken from the side (hereinafter sometimes referred to as a "side waist plain X-ray image") is associated with the reference bone density measured by irradiating the lumbar spine with X-rays from the front. As another example, there may be a case where the learning image data showing a plain X-ray image of the knee joint taken from the side (hereinafter sometimes referred to as a "side knee plain X-ray image") is associated with the reference bone density measured by irradiating the proximal part of the femur with X-rays from the front.

[0035] In addition, among the plurality of learning image data included in the learning data 130, the plurality of simple X-ray images each represented may include a simple X-ray image in which the same type of part as the estimation simple X-ray image appears, or may include a simple X-ray image in which a part of a different type from the estimation simple X-ray image appears. As an example of the former, when the estimation simple X-ray image is a frontal chest simple X-ray image, it is conceivable that the plurality of learning simple X-ray images include a frontal chest simple X-ray image. As another example, when the estimation simple X-ray image is a simple X-ray image in which the knee joint is imaged from the front (hereinafter sometimes referred to as a "front knee simple X-ray image"), it is conceivable that the plurality of learning simple X-ray images include a lateral knee simple X-ray image. On the other hand, as an example of the latter, when the estimation simple X-ray image is a frontal lumbar simple X-ray image, it is conceivable that the plurality of learning simple X-ray images include a frontal chest simple X-ray image. As another example, when the estimation simple X-ray image is a lateral lumbar simple X-ray image, it is conceivable that the plurality of learning simple X-ray images include a front knee simple X-ray image.

[0036] In addition, the plurality of learning simple X-ray images may include a simple X-ray image in which a part in the same orientation as the estimation simple X-ray image appears, or may include a simple X-ray image in which a part in a different orientation from the estimation simple X-ray image appears. As an example of the former, when the estimation simple X-ray image is a frontal lumbar simple X-ray image, it is conceivable that the plurality of learning simple X-ray images include a frontal lumbar simple X-ray image. As another example, when the estimation simple X-ray image is a front knee simple X-ray image, it is conceivable that the plurality of learning simple X-ray images include a frontal chest simple X-ray image. On the other hand, as an example of the latter, when the estimation simple X-ray image is a lateral knee simple X-ray image, it is conceivable that the plurality of learning simple X-ray images include a front knee simple X-ray image. As another example, when the estimation simple X-ray image is a lateral lumbar simple X-ray image, it is conceivable that the plurality of learning simple X-ray images include a frontal chest simple X-ray image.

[0037] In addition, the teacher data 140 may include a reference bone density measured from a site (bone) included in the site shown in the simple X-ray image for estimation, or may include a reference bone density measured from a site (bone) not included in the site shown in the simple X-ray image for estimation. As an example of the former, when the simple X-ray image for estimation is a lumbar spine anteroposterior simple X-ray image, it is conceivable that the teacher data 140 includes the reference bone density of the lumbar vertebrae. On the other hand, as an example of the latter, when the simple X-ray image for estimation is a chest anteroposterior simple X-ray image, it is conceivable that the teacher data 140 includes the reference bone density of the middle phalanx.

[0038] In addition, the teacher data 140 may include a reference bone density measured by irradiating the target site with X-rays from the same direction as the direction of the site shown in the simple X-ray image for estimation, or may include a reference bone density measured by irradiating the target site with X-rays from a direction different from the direction of the site shown in the simple X-ray image for estimation. As an example of the former, when the simple X-ray image for estimation is a lumbar spine anteroposterior simple X-ray image, it is conceivable that the teacher data 140 includes the reference bone density measured by irradiating the lumbar vertebrae with X-rays from the front. On the other hand, as an example of the latter, when the simple X-ray image for estimation is a lumbar spine lateral simple X-ray image, it is conceivable that the teacher data 140 includes the reference bone density measured by irradiating the proximal part of the femur with X-rays from the front.

[0039] In this example, grayscale image data showing a simple X-ray image obtained by a simple X-ray imaging device (in other words, a general X-ray imaging device or an X-ray imaging device) is reduced in size and the number of its gradation levels is decreased, and this is used as learning image data and estimation image data. For example, consider a case where the number of a plurality of pixel data constituting the image data obtained by the simple X-ray imaging device is more than (1024×640) and the bit number of the pixel data is 16 bits. In this case, the number of a plurality of pixel data constituting the image data obtained by the simple X-ray imaging device is reduced to, for example, (256×256), (1024×512) or (1024×640), and the bit number of the pixel data is reduced to 8 bits, and this is used as learning image data and estimation image data. In this case, each of the learning simple X-ray image and the estimation simple X-ray image is composed of (256×256), (1024×512) or (1024×640) pixels, and the value of the pixel is represented by 8 bits.

[0040] Regarding the learning image data and the estimation image data, the control unit 10 of the estimation device 1 may generate them from the image data obtained by the simple X-ray imaging device, or a device other than the estimation device 1 may generate them from the image data obtained by the simple X-ray imaging device. In the former case, regarding the image data obtained by the simple X-ray imaging device, the communication unit 30 may receive it through the communication network, or it may be stored in the removable memory included in the storage unit 20. In the latter case, the communication unit 30 receives the learning image data and the estimation image data from another device through the communication network, and the control unit 10 may store the learning image data and the estimation image data received by the communication unit 30 in the storage unit 20. Alternatively, the learning image data and the estimation image data generated by another device may be stored in the removable memory included in the storage unit 20. Also, regarding the teacher data 140, the communication unit 30 may receive it through the communication network, and the control unit 10 may store the teacher data 140 received by the communication unit 30 in the storage unit 20. Alternatively, the teacher data 140 may be stored in the removable memory included in the storage unit 20. Note that the number and bit number of the pixel data of the learning image data and the estimation image data are not limited to the above.

[0041] <Example of Neural Network Learning> FIG. 5 is a diagram for explaining an example of the learning of the neural network 200. When the control unit 10 performs the learning of the neural network 200, as shown in FIG. 5, the learning data 130 is input to the input layer 210 of the neural network 200. Then, the control unit 10 adjusts the variable parameter 110a in the neural network 200 so that the error with respect to the teacher data 140 for the output data 400 output from the output layer 230 of the neural network 200 becomes small. More specifically, the control unit 10 inputs each learning image data in the storage unit 20 to the input layer 210. When the control unit 10 inputs the learning image data to the input layer 210, the control unit 10 inputs a plurality of pixel data constituting the learning image data to a plurality of artificial neurons constituting the input layer 210, respectively. Then, the control unit 10 adjusts the parameter 110a so that the error with respect to the reference bone density corresponding to the learning image data for the output data 400 output from the output layer 230 when the learning image data is input to the input layer 210 becomes small. As a method for adjusting the parameter 110a, for example, the error backpropagation method is adopted. The adjusted parameter 110a becomes the learned parameter 110 and is stored in the storage unit 20. The parameter 110a includes, for example, parameters used in the hidden layer 220. Specifically, the parameter 110a includes filter coefficients used in the convolutional layer 240 and weighting coefficients used in the fully connected layer 260. Note that the method for adjusting the parameter 110a, in other words, the learning method of the parameter 110a, is not limited to this.

[0042] In this way, the storage unit 20 stores the learned parameter 110 obtained by learning the relationship between the learning data 130 including the image data of a plurality of learning simple X-ray images and the measured value of the bone density as the teacher data 140 using the neural network 200.

[0043] In the above example, the estimation device 1 is performing the learning of the neural network 200, but other devices may perform the learning of the neural network 200. In this case, the learned parameters 110 generated by other devices are stored in the storage unit 20 of the estimation device 1. Also, it becomes unnecessary for the storage unit 20 to store the learning data 130 and the teacher data 140. Regarding the learned parameters 110 generated by other devices, the communication unit 30 may receive them through the communication network, and the control unit 10 may store the learned parameters 110 received by the communication unit 30 in the storage unit 20. Alternatively, the learned parameters 110 generated by other devices may be stored in the removable memory included in the storage unit 20.

[0044] In the neural network 200 learned as described above, a plurality of image data of learning simple X-ray images are input to the input layer 210 as the learning data 130, and the learned parameters 110a learned using the reference bone density as the teacher data 140 are included. As shown in FIG. 2 described above, the neural network 200 performs an operation based on the learned parameters 110a on the estimation data 120 input to the input layer 210, and outputs a bone density estimation value 300 from the output layer 230. When the estimation image data as the estimation data 120 is input to the input layer 210, a plurality of pixel data constituting the estimation image data are respectively input to a plurality of artificial neurons constituting the input layer 210. Then, the convolutional layer 240 performs an operation using the filter coefficients included in the learned parameters 110a, and the fully connected layer 260 performs an operation using the weighting coefficients included in the learned parameters 110a.

[0045] For example, when the estimation image data showing a frontal chest radiograph is input to the input layer 210, an estimated value 300 of the bone density of a person having bones in the chest shown in the frontal chest radiograph shown by the estimation image data is output from the output layer 230. Also, when the estimation image data showing a frontal lumbar radiograph is input to the input layer 210, an estimated value 300 of the bone density of a person having lumbar vertebrae included in the lumbar region shown in the frontal lumbar radiograph shown by the estimation image data is output from the output layer 230. Also, when the estimation image data showing a lateral lumbar radiograph is input to the input layer 210, an estimated value 300 of the bone density of a person having lumbar vertebrae included in the lumbar region shown in the lateral lumbar radiograph shown by the estimation image data is output from the output layer 230. Also, when the estimation image data showing a frontal knee radiograph is input to the input layer 210, an estimated value 300 of the bone density of a person having bones of the knee joint shown in the frontal knee radiograph shown by the estimation image data is output from the output layer 230. Also, when the estimation image data showing a lateral knee radiograph is input to the input layer 210, an estimated value 300 of the bone density of a person having bones of the knee joint shown in the lateral knee radiograph shown by the estimation image data is output from the output layer 230.

[0046] The estimated value 300 output from the output layer 230 may be represented by at least one of bone mineral density per unit area (g / cm 2 ), bone mineral density per unit volume (g / cm 3 ), YAM, T-score, and Z-score. YAM is an abbreviation for "Young Adult Mean" and may be called young adult mean percent. For example, from the output layer 230, an estimated value 300 represented by bone mineral density per unit area (g / cm 2 ) and an estimated value 300 represented by YAM may be output, or an estimated value 300 represented by YAM, an estimated value 300 represented by T-score, and an estimated value 300 represented by Z-score may be output.

[0047] Incidentally, the storage unit 20 may store a plurality of estimation data 120. In this case, the plurality of simple X-ray images respectively indicated by the plurality of estimation data 120 in the storage unit 20 may include a plurality of simple X-ray images in which the same type of part is imaged, or may include a plurality of simple X-ray images in which different types of parts are imaged. Further, the plurality of simple X-ray images for estimation may include a plurality of simple X-ray images in which the part is imaged from the same direction, or may include a plurality of simple X-ray images in which the part is imaged from different directions. In other words, the plurality of simple X-ray images for estimation may include a plurality of simple X-ray images in which the direction of the part imaged thereon is the same, or may include a plurality of simple X-ray images in which the direction of the part imaged thereon is different. The control unit 10 inputs each of the plurality of estimation data 120 in the storage unit 20 to the input layer 210 of the neural network 200, and a bone density estimation value 300 corresponding to each estimation data 120 is output from the output layer 230 of the neural network 200.

[0048] As described above, in this example, the image data of the simple X-ray image is used to perform the learning of the neural network 200 and the estimation of the bone density by the neural network 200. Since the image data of the simple X-ray image, in other words, the image data of the radiograph, is used in various examinations and the like in many hospitals, it can be easily obtained. Therefore, the bone density can be easily estimated without using an expensive device such as a DEXA device.

[0049] Also, by using the image data of the simple X-ray image taken for examination or the like as the estimation image data, it is possible to easily estimate the bone density by utilizing the opportunity of such examination or the like. Therefore, by using the estimation device 1, the service for hospital users can be improved.

[0050] In addition, in a frontal plain X-ray image of the chest or the like, bones may be less likely to be imaged due to the influence of organs. On the other hand, frontal plain X-ray images are likely to be taken in many hospitals. In this example, even when a frontal plain X-ray image in which bones are less likely to be imaged is used as a simple X-ray image for estimation or a simple X-ray image for learning, bone density can be estimated. Therefore, bone density can be easily estimated using the image data of the easily obtainable frontal plain X-ray image. In addition, frontal chest plain X-ray images are often taken in health check-ups and the like, and it can be said that they are particularly easily obtainable simple X-ray images. By using the frontal chest plain X-ray image as a simple X-ray image for estimation or a simple X-ray image for learning, bone density can be estimated even more easily.

[0051] In addition, in this example, even when a plurality of learning simple X-ray images include a simple X-ray image in which a part of a type different from that of the estimation simple X-ray image is imaged, bone density can be estimated from the image data of the estimation simple X-ray image. Therefore, the convenience of the estimation device 1 (in other words, the computer device 1) can be improved.

[0052] In addition, in this example, even when a plurality of learning simple X-ray images include a simple X-ray image in which a part in a direction different from that of the estimation simple X-ray image is imaged, bone density can be estimated from the image data of the estimation simple X-ray image. Therefore, the convenience of the estimation device 1 can be improved.

[0053] In addition, in this example, even when the part imaged in the learning simple X-ray image does not include the part (bone) at which the reference bone density corresponding to the learning simple X-ray image is measured, the neural network 200 can estimate the bone density based on the learned parameter 110. Therefore, the convenience of the estimation device 1 can be improved.

[0054] In addition, in this example, even when the orientation of the part shown in the learning simple X-ray image and the orientation of the X-ray irradiation of the target part in the measurement of the reference bone density corresponding to the learning simple X-ray image are different from each other, the neural network 200 can estimate the bone density based on the learned parameters 110. Therefore, the convenience of the estimation device 1 can be improved.

[0055] Also, in this example, even when the reference bone density measured from a part not included in the part shown in the estimation simple X-ray image is included in the teacher data 140, the bone density can be estimated from the image data of the estimation simple X-ray image. Therefore, the convenience of the estimation device 1 can be improved.

[0056] Note that the bone density estimated value 300 obtained by the estimation device 1 may be displayed on the display unit 40. Also, the bone density estimated value 300 obtained by the estimation device 1 may be used in other devices.

[0057] FIG. 6 is a diagram showing an example of a bone density estimation system 600 including the estimation device 1 and a processing device 500 that performs processing using the bone density estimated value 300 obtained by the estimation device 1. In the example of FIG. 6, the estimation device 1 and the processing device 500 can communicate with each other through the communication network 700. The communication network 700 includes, for example, at least one of a wireless network and a wired network. The communication network 700 includes, for example, a wireless LAN (Local Area Network) and the Internet.

[0058] In the estimation device 1, a communication network 700 is connected to the communication unit 30. The control unit 10 causes the communication unit 30 to transmit the bone density estimated value 300 to the processing device 500. The processing device 500 performs processing using the bone density estimated value 300 received from the estimation device 1 through the communication network 700. For example, the processing device 500 is a display device such as a liquid crystal display device, and displays the bone density estimated value 300. At this time, the processing device 500 may display the bone density estimated value 300 in a table or as a graph. Further, when a plurality of estimation devices 1 are connected to the communication network 700, the processing device 500 may display the bone density estimated values 300 obtained by the plurality of estimation devices 1. The configuration of the processing device 500 may be the same as the configuration of the estimation device 1 shown in FIG. 1, or may be different from the configuration of the estimation device 1.

[0059] Note that the processing using the bone density estimated value 300 executed by the processing device 500 is not limited to the above example. Further, the processing device 500 may communicate directly wirelessly or by wire with the estimation device 1 without going through the communication network 700.

[0060] <Other Examples of Estimation Data and Learning Data> <First Other Example> In this example, the learning data 130 includes, for each learning image data, information regarding the health state of a person having a bone shown in the learning simple X-ray image indicated by the learning image data. In other words, the learning data 130 includes, for each learning image data, information regarding the health state of the subject (the object to be examined) of the learning simple X-ray image indicated by the learning image data. Hereinafter, the information regarding the health state of the subject of the learning simple X-ray image may be referred to as "learning health-related information". Further, the information regarding the health state of the subject of the learning simple X-ray image indicated by the learning image data may be referred to as the learning health-related information corresponding to the learning image data.

[0061] The learning-related health information includes at least one type of information such as age information, gender information, height information, weight information, drinking habit information, smoking habit information, and information on the presence or absence of a history of fractures. The learning-related health information is database-ized for each person and generated as a CSV (Comma-Separated Value) format file or a text format file. Each of the age information, height information, and weight information is represented, for example, as numerical data of multiple bits. In the gender information, for example, "male" or "female" is represented by 1-bit data. In the drinking habit information, "with drinking habit" or "without drinking habit" is represented by 1-bit data. In the smoking habit information, "with smoking habit" or "without smoking habit" is represented by 1-bit data. In the information on the presence or absence of a history of fractures, "with fracture" or "without fracture" is represented by 1-bit data. Further, the learning-related health information may include the body fat percentage or subcutaneous fat percentage of the subject.

[0062] When the learning data 130 includes learning image data and the corresponding learning-related health information, the reference bone density (see FIG. 4) corresponding to the learning image data is also associated with the learning-related health information corresponding to the learning image data. That is, for the learning image data showing a simple X-ray image of a person's bone and the information on the health status of the person (learning-related health information), the measured value of the bone density of the person (reference bone density) is associated. In the learning of the neural network 200, the learning image data and the corresponding learning-related health information are simultaneously input to the input layer 210. Specifically, the learning image data is input to a part of the plurality of artificial neurons constituting the input layer 210, and the learning-related health information is input to another part of the plurality of artificial neurons. Then, when the learning image data and the corresponding learning-related health information are input to the input layer 210, the output data 400 output from the output layer 230 is compared with the reference bone density corresponding to the learning image data and the learning-related health information.

[0063] In this example, the estimation data 120 includes estimation image data and information regarding the health condition of a person having bones shown in the estimated simple X-ray image represented by the estimation image data. In other words, the estimation data 120 includes estimation image data and information regarding the health condition of the subject of the estimated simple X-ray image represented by the estimation image data. Hereinafter, the information regarding the health condition of the subject of the estimated simple X-ray image may be referred to as "estimation-related health information (hereinafter, also referred to as "individual data" in other embodiments)". Further, the information regarding the health condition of the subject of the estimated simple X-ray image represented by the estimation image data may be referred to as estimation-related health information corresponding to the estimation image data.

[0064] Similar to the learning-related health information, the estimation-related health information includes, for example, at least one type of information such as age information, gender information, height information, weight information, drinking habit information, smoking habit information, and information on the presence or absence of a fracture history. The estimation-related health information includes the same types of information as the learning-related health information. Further, similar to the learning-related health information, the estimation-related health information may include the body fat percentage or subcutaneous fat percentage of the subject.

[0065] In this example, when estimating bone density, the estimation image data and the corresponding estimation-related health information are simultaneously input to the input layer 210. Specifically, the estimation image data is input to a part of the plurality of artificial neurons constituting the input layer 210, and the estimation-related health information is input to another part of the plurality of artificial neurons. When the estimation image data and the estimation-related health information about a certain person are input to the input layer 210, an estimated value of the bone density of the certain person is output from the output layer 230.

[0066] In this way, by using not only the image data of the simple X-ray image but also the information regarding the health condition of the subject of the simple X-ray image, the estimation accuracy of bone density can be improved.

[0067] <Another second example> In this example, the learning data 130 includes image data of N (N≥2) learning simple X-ray images in which parts of the same person are depicted and the orientations of the parts depicted therein are different from each other. Hereinafter, the N learning simple X-ray images may be collectively referred to as a "learning simple X-ray image set".

[0068] The learning simple X-ray image set includes, for example, a front view and a side view of the same person. The learning simple X-ray image set includes, for example, a chest front simple X-ray image and a lumbar side simple X-ray image of a certain person. The image sizes of the front view and the side view included in the learning simple X-ray image set may be different from each other. For example, the horizontal width of the image size of the side view may be smaller than the horizontal width of the image size of the front view. Hereinafter, the image data of each learning simple X-ray image in the learning simple X-ray image set may be collectively referred to as a "learning image data set".

[0069] The learning data 130 includes learning image data sets for a plurality of different persons. As a result, the learning data 130 includes a plurality of learning image data sets. And, for one learning image data set, one reference bone density is associated. That is, for a learning image data set of a certain person, the measured value (reference bone density) of the bone density of the certain person is associated.

[0070] In the learning of the neural network 200 in this example, each training image dataset is input to the input layer 210. When one training image dataset is input to the input layer 210, the N pieces of training image data that make up the one training image dataset are simultaneously input to the input layer 210. For example, assume that the training image dataset is composed of the first training image data and the second training image data. In this case, the first training image data (for example, the image data of a frontal plain chest X-ray) is input to a part of the plurality of artificial neurons that make up the input layer 210, and the second training image data (for example, the image data of a lateral plain lumbar X-ray) is input to another part of the plurality of artificial neurons. Then, the output data 400 output from the output layer 230 when the training image dataset is input to the input layer 210 is compared with the reference bone density corresponding to the training image dataset.

[0071] Also, in this example, the estimation data 120 includes the image data of N simple X-rays in which the same person's parts are imaged and the orientations of the imaged parts are different from each other. Hereinafter, the N simple X-rays may be collectively referred to as an "estimation simple X-ray set".

[0072] The estimation simple X-ray set includes, for example, a frontal image and a lateral image of the same person. The estimation simple X-ray set includes, for example, a frontal plain lumbar X-ray and a lateral plain knee X-ray of a certain person. The image sizes of the frontal image and the lateral image included in the estimation simple X-ray set may be different from each other. For example, the horizontal width of the image size of the lateral image may be smaller than the horizontal width of the image size of the frontal image. Hereinafter, the image data of each estimation simple X-ray in the estimation simple X-ray set may be collectively referred to as an "estimation image dataset".

[0073] In this example, when the estimation data 120 is used to estimate bone density, N pieces of estimation image data constituting the estimation image data set are simultaneously input to the input layer 210. For example, assume that the estimation image data set is composed of the first estimation image data and the second estimation image data. In this case, the first estimation image data is input to a part of the plurality of artificial neurons constituting the input layer 210, and the second estimation image data is input to another part of the plurality of artificial neurons. When the estimation image data set of a certain person is input to the input layer 210, the estimated value of the bone density of the certain person is estimated from the output layer 230.

[0074] In this way, by using the image data of a plurality of simple X-ray images in which the parts of the same subject are imaged and the orientations of the imaged parts are different from each other, the estimation accuracy of bone density can be improved.

[0075] Note that the learning data 130 may include a learning image data set and learning health-related information. In this case, in the learning of the neural network 200, the learning image data set and the learning health-related information of the same person are simultaneously input to the input layer 210. Similarly, the estimation data 120 may include an estimation image data set and estimation health-related information. In this case, the estimation image data set and the estimation health-related information are simultaneously input to the input layer 210.

[0076] In each of the above examples, the same learned parameter 110 is used regardless of the type of bone shown in the X-ray image indicated by the image data for estimation. However, learned parameters 110 corresponding to the type of bone shown in the X-ray image indicated by the image data for estimation may be used. In this case, the neural network 200 has a plurality of learned parameters 110 corresponding to a plurality of types of bones respectively. The neural network 200 estimates the bone density using the learned parameter 110 corresponding to the type of bone shown in the X-ray image indicated by the input image data for estimation. For example, when the lumbar vertebra appears in the X-ray image indicated by the input image data for estimation, the neural network 200 uses the learned parameter 110 for estimating the bone density of the lumbar vertebra to estimate the bone density. Also, when the proximal part of the femur appears in the X-ray image indicated by the input image data for estimation, the neural network 200 uses the learned parameter 110 for estimating the bone density of the proximal part of the femur to estimate the bone density. The neural network 200 uses, for example, the learned parameter 110 instructed by the user through the input unit 50 among the plurality of learned parameters 110. In this case, the user instructs the learned parameter 110 to be used by the neural network 200 according to the type of bone shown in the X-ray image indicated by the image data for estimation input to the neural network 200.

[0077] In the learning of the neural network 200, a plurality of learning image data respectively indicating a plurality of X-ray images in which the same type of bone appears are used, and the learned parameter 110 corresponding to the type of bone is generated.

[0078] As described above, the estimation device 1 and the bone density estimation system 600 have been described in detail. However, the above description is illustrative in all aspects and the present disclosure is not limited thereto. Also, the various examples described above can be applied in combination as long as they do not conflict with each other. And it is understood that countless examples not illustrated can be assumed without departing from the scope of this disclosure.

[0079] Embodiment 2. FIG. 7 is a diagram showing an example of the configuration of the estimation device 1A according to the present embodiment. In the estimation device 1A, the approximator 280 further has a second neural network 900. The second neural network 900 can detect a fracture based on the learned parameters 910. Note that the estimation device 1A according to the present embodiment has the same configuration as the estimation device 1 according to the first embodiment, and the description of the same configuration will be omitted. For convenience of explanation, the neural network 200 described in the above example is referred to as the first neural network 200. The second neural network 900 has, for example, the same configuration as the first neural network 200.

[0080] The second neural network 900 can detect a fracture based on the same estimation image data as the estimation image data included in the estimation data 120 input to the first neural network 200. That is, from one piece of estimation image data, the first neural network 200 can estimate the bone density, and the second neural network 900 can detect a fracture. Note that the detection result 920 of the second neural network 900 may be output from the output layer 230 of the second neural network 900 as in the above example.

[0081] In the learning of the second neural network 900, learning image data showing a non-fractured bone and learning image data showing a fractured bone are used to learn the parameters. In the teacher data, for each learning image data, information indicating the current presence or absence of a fracture and information indicating the fracture location for the bone shown in the learning image data are associated. The teacher data may also include information indicating the past fracture history and information indicating the past fracture location. As a result, the second neural network 900 can detect the presence or absence and location of a fracture for the bone shown in the estimation image data based on the estimation image data, and output the detection result 920.

[0082] In addition, as shown in FIG. 8, the estimation device 1A according to the present embodiment may include a determination unit 930 that determines whether or not the subject has osteoporosis. The determination unit 930 can compare and consider the estimation result 300 of the first neural network 200 and the detection result 920 of the second neural network 900 to determine whether or not the subject has osteoporosis.

[0083] The determination unit 930 may determine osteoporosis based on, for example, an independent criterion or a well-known guideline. Specifically, when the detection result 920 indicates a fracture in the vertebral body or the proximal femur, the determination unit 930 may determine that the subject has osteoporosis. Further, when the bone density estimated value 300 output by the first neural network 200 indicates YAM, if YAM is less than 80% and the detection result 920 indicates a fracture other than the vertebral body and the proximal femur, the determination unit 930 may determine that the subject has osteoporosis. Also, when YAM indicated by the bone density estimated value 300 is 70% or less, the determination unit 930 may determine that the subject has osteoporosis.

[0084] In addition, in the estimation device 1A according to the present embodiment, as shown in FIG. 9, the approximator 28 may further include a third neural network 950. The third neural network 950 can classify the bones of the subject from the estimation image data included in the estimation data 120 based on the learned parameters 960.

[0085] For each pixel data of the input estimation image data, the third neural network 950 outputs part information indicating the part of the bone indicated by the pixel data. Thereby, the bones shown in the X-ray image indicated by the estimation image data can be classified. The part information may be called segmentation data.

[0086] For example, when the lumbar spine appears in the X-ray image shown by the input estimation image data, the third neural network 950 outputs, for each pixel data of the estimation image data, site information indicating which site among L1 to L5 of the lumbar spine the pixel data represents. For example, when a certain pixel data of the estimation image data indicates L1 of the lumbar spine, the third neural network 950 outputs site information indicating L1 as the site information corresponding to the pixel data.

[0087] The third neural network 950 uses the learned parameter 960 corresponding to the type of bone appearing in the X-ray image shown by the input estimation image data. The third neural network 950 has a plurality of learned parameters 960 corresponding to a plurality of types of bones respectively. The third neural network 950 uses the learned parameter 960 corresponding to the type of bone appearing in the X-ray image shown by the input estimation image data to classify the bones appearing in the X-ray image shown by the estimation image data. For example, when the lumbar spine appears in the X-ray image shown by the input estimation image data, the third neural network 950 uses the learned parameter 960 corresponding to the lumbar spine to classify the lumbar spine into L1 to L5. The third neural network 950 uses, among the plurality of learned parameters 960, for example, the learned parameter 960 instructed by the user through the input unit 50. In this case, the user instructs the learned parameter 960 used by the third neural network 950 according to the type of bone appearing in the X-ray image shown by the estimation image data input to the third neural network 950.

[0088] The third neural network 950 may classify the bones shown in the X-ray image represented by the input estimated image data into a first site where the implant is implanted, a second site where there is a tumor, and a third site where there is a fracture. In this case, for each pixel data of the estimated image data, the third neural network 950 outputs site information indicating which of the first site, the second site, and the third site the pixel data represents. When the pixel data represents a site other than the first site, the second site, and the third site, the third neural network 950 outputs site information indicating that the pixel data represents a site other than the first site, the second site, and the third site. When the third neural network 950 classifies the bones shown in the X-ray image represented by the estimated image data into the first site, the second site, and the third site, it can be said that the third neural network 950 detects the implant embedded in the bone shown in the X-ray image represented by the estimated image data, the fracture of the bone, and the tumor of the bone.

[0089] In the learning of the third neural network 950, a plurality of learning image data each representing a plurality of X-ray images showing the same type of bone are used to generate learned parameters 960 according to the type of the bone. When the third neural network 950 classifies the bones shown in the X-ray image represented by the input estimated image data into the first site, the second site, and the third site, the plurality of learning image data includes learning image data showing an X-ray image of a case where an implant is implanted, learning image data showing an X-ray image of a case where there is a tumor in the bone, and learning image data showing an X-ray image of a case where there is a fracture. Further, the teacher data includes, for each learning image data, annotation information for classifying the bone represented by the learning image data. The annotation information includes site information indicating the site of the bone represented by each pixel data of the corresponding learning image data.

[0090] Note that the first neural network 200 may estimate bone density for each part classified in the third neural network 950. In this case, as shown in FIG. 10, the first neural network 200 receives the estimation image data 121 and the part information 965 corresponding to each pixel data of the estimation image data 121, which is output by the third neural network 950 based on the estimation image data 121. The first neural network 200 outputs the bone density estimation value 300 for each part classified in the third neural network based on the learned parameter 110 corresponding to the type of bone shown in the X-ray image indicated by the estimation image data 121. For example, when the third neural network 950 divides the cervical vertebra shown in the X-ray image indicated by the estimation image data 121 into L1 to L5, the first neural network 200 individually outputs the bone density estimation value 300 for L1, the bone density estimation value 300 for L2, the bone density estimation value 300 for L3, the bone density estimation value 300 for L4, and the bone density estimation value 300 for L5.

[0091] In the learning of the first neural network 200, a plurality of learning image data respectively indicating a plurality of X-ray images in which the same type of bone appears are used, and the learned parameter 110 corresponding to the type of the bone is generated. Further, the teacher data includes the reference bone density of each part of the bone indicated by each learning image data.

[0092] The first neural network 200 uses, for example, the learned parameter 110 instructed by the user through the input unit 50 among the plurality of learned parameters 110. In this case, the user instructs the learned parameter 110 used by the first neural network 200 according to the type of bone shown in the X-ray image indicated by the estimation image data input to the first neural network 200.

[0093] When the third neural network 950 classifies the bones shown in the X-ray image represented by the estimation target image data into a first site where the implant is implanted, a second site where there is a tumor, and a third site where there is a fracture, the brightness of the first partial image data indicating the first site, the second partial image data indicating the second site where there is a tumor, and the third partial image data indicating the third site among the estimation target image data may be adjusted. FIG. 11 is a diagram showing a configuration example in this case.

[0094] As shown in FIG. 11, the adjustment unit 968 receives the estimation target image data 121 and the site information 965 corresponding to each pixel data of the estimation target image data 121 output by the third neural network 950 based on the estimation target image data 121. The adjustment unit 968 identifies the first partial image data, the second partial image data, and the third partial image data included in the estimation target image data based on the site information 965. Then, the adjustment unit 968 adjusts the brightness of the identified first partial image data, second partial image data, and third partial image data.

[0095] The adjustment unit 968 stores, for example, the brightness of the first site shown in a general X-ray image as the first reference brightness. The adjustment unit 968 also stores the brightness of the second site shown in a general X-ray image as the second reference brightness. Then, the adjustment unit 968 stores the brightness of the third site shown in a general X-ray image as the third reference brightness. The adjustment unit 968 subtracts the first reference brightness from the brightness of the first partial image data to adjust the brightness of the first partial image data. The adjustment unit 976 also subtracts the second reference brightness from the brightness of the second partial image data to adjust the brightness of the second partial image data. Then, the adjustment unit 978 subtracts the third reference brightness from the brightness of the third partial image data to adjust the brightness of the third partial image data. The adjustment unit 968 inputs, as the estimation target image data after brightness adjustment, the estimation target image data in which the brightness of the first partial image data, the second partial image data, and the third partial image data has been adjusted, to the first neural network 200. The first neural network 200 estimates the bone density of the bone shown in the X-ray image represented by the estimation target image data after brightness adjustment.

[0096] Here, it is not easy to correctly estimate the bone density from the first site where the implant is implanted, the second site where the tumor is present, and the third site where the fracture is present. As described above, by adjusting and reducing the brightness of the first partial image data indicating the first site, the second partial image data indicating the second site, and the third partial image data indicating the third site, it is possible to more correctly estimate the bone density of the bone shown in the X-ray image indicated by the estimation image data.

[0097] Note that the adjustment unit 968 may input the estimation image data in which the brightness of the first partial image data, the second partial image data, and the third partial image data is forcibly set to zero into the first neural network 200 as the estimation image data after brightness adjustment.

[0098] Also, the third neural network 950 may detect only one of the implant, the fracture, and the tumor. Also, the third neural network 950 may detect only two of the implant, the fracture, and the tumor. That is, the third neural network 950 may detect at least one of the implant, the fracture, and the tumor.

[0099] Also, the estimation device 1A may include the first neural network 200 and the third neural network 950 without including the second neural network 900. Also, the estimation device 1A may include at least one of the second neural network 900 and the third neural network 950 without including the first neural network 200.

[0100] As described above, the estimation device 1A has been described in detail, but the above description is illustrative in all aspects and this disclosure is not limited thereto. Also, the various examples described above can be applied in combination as long as they do not conflict with each other. And it is understood that countless examples not illustrated can be assumed without departing from the scope of this disclosure.

[0101] Embodiment 3. FIG. 12 is a diagram showing an example of the configuration of the estimation device 1B according to the present embodiment. The estimation device 1B includes a fracture prediction unit 980. The fracture prediction unit 980 can predict the probability of fracture, for example, based on the estimation result 300 of the neural network 200 of the estimation device 1 according to Embodiment 1. Specifically, for example, an arithmetic expression 990 showing the relationship between the estimation result related to bone density (such as bone density) and the probability of fracture is obtained from past literature or the like. The fracture prediction unit 980 stores the arithmetic expression 990. The fracture prediction unit 980 can predict the probability of fracture based on the input estimation result 300 and the stored arithmetic expression 990.

[0102] Further, the arithmetic expression 990 may be an arithmetic expression showing the relationship between the estimation result related to bone density and the probability of fracture after bone screw implantation. As a result, it becomes possible to consider a treatment plan including whether or not to implant a bone screw and drug administration.

[0103] Note that the estimation device 1B may include a second neural network 900. Further, the estimation device 1B may include a third neural network 950.

[0104] As described above, the estimation device 1B has been described in detail, but the above description is illustrative in all aspects and the present disclosure is not limited thereto. Further, the various examples described above can be applied in combination as long as they do not conflict with each other. And an infinite number of examples not illustrated can be considered to be within the scope of this disclosure without departing therefrom.

[0105] Embodiment 4. FIG. 13 shows the concept of the configuration of the estimation system 801 of the present embodiment.

[0106] The estimation system 801 of the present disclosure can estimate the future bone mass of a subject from an image showing the bone of the subject such as an X-ray image. The estimation system 801 of the present disclosure includes a terminal device 802 and an estimation device 803. Note that the bone mass is an index related to bone density and is a concept including bone density.

[0107] The terminal device 802 can acquire input information I for input to the estimation device 803. The input information I may be, for example, an X-ray image or the like. In this case, the terminal device 802 may be a device for a doctor or the like to take an X-ray image of a subject. For example, the terminal device 802 may be a simple X-ray imaging device (in other words, a general X-ray imaging device or a roentgen imaging device).

[0108] Note that the terminal device 802 is not limited to a simple X-ray imaging device. The terminal device 802 may be, for example, an X-ray fluoroscopy imaging device, a CT (Computed Tomography), an MRI (Magnetic Resonance Imaging), a SPECT (Single Photon Emission Computed Tomography)-CT, or tomography. In this case, the input information I may be, for example, an X-ray fluoroscopy image, a CT (Computed Tomography) image, an MRI (Magnetic Resonance Imaging) image, a bone scintigraphy image, or a tomography image.

[0109] The estimation system 801 is used, for example, for diagnosing osteoporosis or the like of a patient visiting a hospital. The estimation system 801 of the present disclosure takes a roentgenogram of a patient using, for example, a terminal device 802 installed in an X-ray room. Then, the image data is transferred from the terminal device 802 to the estimation device 803, and through the estimation device 803, not only the bone mass or bone density of the patient at the current time but also the bone mass or bone density of the patient in the future from the time of imaging can be estimated.

[0110] Note that the terminal device 802 does not necessarily transfer the input information I directly to the estimation device 803. In this case, for example, the input information I acquired by the terminal device 802 may be stored in a storage medium, and the input information I may be input to the estimation device 803 via the storage medium.

[0111] FIG. 14 shows a concept of the configuration of the estimation device 803 according to the present embodiment.

[0112] Based on the input information I acquired by the terminal device 802, the estimation device 803 can estimate the future bone mass or bone density of the subject. In the estimation device 803, the future bone mass or bone density of the subject can be estimated from the image data acquired by the terminal device 802, and the estimation result O can be output.

[0113] The estimation device 803 includes an input unit 831, an approximator 832, and an output unit 833. The input unit 831 is for inputting the input information I from the terminal device 802. The approximator 832 can estimate the future bone mass or bone density based on the input information I. The output unit 833 can output the estimation result O predicted by the approximator 832.

[0114] The estimation device 803 has various electronic components and circuits. As a result, the estimation device 803 can form each component. For example, the estimation device 803 can form at least one integrated circuit (e.g., IC: Integrated Circuit or LSI: Large Scale Integration) by integrating a plurality of semiconductor elements, or can form at least one unit by further integrating a plurality of integrated circuits, etc., to configure each functional part of the estimation device 803.

[0115] The plurality of electronic components may be, for example, active elements such as transistors or diodes, or passive elements such as capacitors. Note that the plurality of electronic components and the integrated circuits formed by integrating them can be formed by conventionally well-known methods.

[0116] The input unit 831 is for inputting information used by the estimation device 803. For example, input information I having an X-ray image acquired by the terminal device 802 is input to the input unit 831. The input unit 831 has a communication unit, and the input information I acquired by the terminal device 802 is directly input from the terminal device 802. Further, the input unit 831 may be provided with an input device capable of inputting the input information I or other information. The input device may be, for example, a keyboard, a touch panel, or a mouse.

[0117] The approximator 832 estimates the future bone mass or bone density of the subject based on the information input to the input unit 831. The approximator 832 has AI (Artificial Intelligence). The approximator 832 has a program that functions as AI and various electronic components and circuits for executing the program. The approximator 832 has a neural network.

[0118] The approximator 832 has been pre-trained on the relationship between inputs and outputs. That is, by applying machine learning to the approximator 832 using training data and teacher data, the approximator 832 can calculate an estimation result O from the input information I. Note that the training data or teacher data may be data corresponding to the input information I input to the estimation device 803 and the estimation result O output from the estimation device 803.

[0119] FIG. 15 shows the concept of the configuration of the approximator 832 of the present disclosure.

[0120] The approximator 832 has a first neural network 8321 and a second neural network 8322. The first neural network 8321 may be any neural network suitable for handling time-series information. For example, the first neural network 8321 may be a ConvLSTM network that combines LSTM (Long short-term memory) and CNN (Convolutional Neural Network). The second neural network 8322 may be, for example, a convolutional neural network composed of CNNs.

[0121] The first neural network 8321 has an encoding section E and a decoding section D. The encoding section E can extract the temporal changes and feature amounts of the positional information of the input information I. The decoding section D can calculate new feature amounts based on the feature amounts extracted by the encoding section E, the temporal changes of the input information I, and the initial values.

[0122] FIG. 16 shows the concept of the configuration of the first neural network 8321 of the present disclosure.

[0123] The encoding section E has a plurality of ConvLSTM layers (Convolutional Long short-term memory) E1. The decoding section D has a plurality of ConvLSTM layers (Convolutional Long short-term memory) D1. Each of the encoding section E and the decoding section D may have three or more ConvLSTM layers E1, D1. Also, the number of the plurality of ConvLSTM layers E1 and the number of the plurality of ConvLSTM layers D1 may be the same.

[0124] Note that the plurality of ConvLSTM layers E1 may each have different learning contents. The plurality of ConvLSTM layers D1 may each have different learning contents. For example, one ConvLTSM layer learns fine details such as changes in each pixel, while another ConvLSTM layer learns rough details such as changes in the overall image.

[0125] FIG. 17 shows the concept of the configuration of the second neural network 8322.

[0126] The second neural network 8322 has a conversion unit C. The conversion unit C can convert the feature amount calculated by the decoding unit D of the first neural network 8321 into bone mass or bone density. The conversion unit C has a plurality of convolutional layers C1, a plurality of pooling layers C2, and a fully connected layer C3. The fully connected layer C3 is located in front of the output unit 33. In the conversion unit C, the convolutional layer C1 and the pooling layer C2 are alternately arranged between the first neural network 8311 and the fully connected layer C3.

[0127] Note that the learning data is input to the encoding unit E of the approximator 832 during the learning of the approximator 832. The teacher data is compared with the output data output from the conversion unit C of the approximator 832 during the learning of the approximator 832. The teacher data is data indicating values measured using a conventional bone density measuring device.

[0128] The output unit 833 can display the estimation result O. The output unit 833 is, for example, a liquid crystal display or an organic EL display. The output unit 833 can display various types of information such as characters, symbols, and figures. The output unit 833 can display, for example, numbers or images.

[0129] The estimation device 803 of the present disclosure further has a control unit 834 and a storage unit 835. The control unit 834 can comprehensively manage the operation of the estimation device 803 by controlling other components of the estimation device 803.

[0130] The control unit 834 has, for example, a processor. The processor may include, for example, one or more processors, controllers, microprocessors, microcontrollers, application specific integrated circuits (ASICs), digital signal processing devices, programmable logic devices, or combinations of these devices or any calibration, or combinations of devices or calibration of other bases. The control unit 834 includes, for example, a CPU.

[0131] The storage unit 835 includes, for example, a non-transitory recording medium readable by the CPU of the control unit 834 such as a RAM (Random Access Memory) or a ROM (Read-Only Memory). A control program for controlling the estimation device 803 such as firmware is stored in the storage unit 835. Also, input information I to be input, learning data to be learned, and teacher data may be stored in the storage unit 835.

[0132] The processor of the control unit 834 can execute one or more data calculation procedures or processes according to the control program of the storage unit 835. Various functions of the control unit 834 are realized by the CPU of the control unit 834 executing the control program in the storage unit 11.

[0133] Note that the control unit 834 may perform other processes as preprocessing of the calculation process as necessary.

[0134] <An example of input information, learning data, and teacher data> The input information (hereinafter also referred to as the first input information I1) has image data in which the bone to be estimated for bone mass or bone density is imaged. The image data may be, for example, a simple X-ray image. The object to be estimated for bone mass or bone density is, for example, a person. In this case, it can be said that the first input information I1 is image data of a simple X-ray image in which a person's bone is imaged. A simple X-ray image is a two-dimensional image and is also called a general X-ray image or a roentgen image.

[0135] The first input information I1 is preferably a simple X-ray image that is relatively easy to obtain, but is not limited thereto. For example, by using an X-ray fluoroscopy image, a CT (Computed Tomography) image, an MRI (Magnetic Resonance Imaging) image, a bone scintigraphy image, or a tomosynthesis image as the input information, the bone mass or bone density may be estimated more accurately in some cases.

[0136] Note that the object for estimating the bone mass or bone density may be other than a human. For example, the object for estimating the bone mass or bone density may be an animal such as a dog, a cat, or a horse. Also, the target bone is mainly cortical bone and cancellous bone derived from a living organism, but the target bone may include an artificial bone mainly composed of calcium phosphate or a regenerated bone artificially manufactured by regenerative medicine or the like.

[0137] As the imaging site of the X-ray image, for example, the neck, chest, waist, proximal femur, knee joint, ankle joint, shoulder joint, elbow joint, wrist joint, finger joint, or jaw joint may be used. Note that the X-ray image may show a site other than bone. For example, in the case of a chest simple X-ray image, it may include an image of the lungs and an image of the thoracic vertebrae. The X-ray image may be a front view in which the target site is imaged from the front, or a side view in which the target site is imaged from the side.

[0138] The learning data or the teacher data may be any data corresponding to the input information I input to the estimation device 3 and the estimation result O output from the estimation device 3.

[0139] The learning data has the same type of information as the first input information I1. For example, if the first input information I1 is a simple X-ray image, the learning data may also have a simple X-ray image. Furthermore, when the first input information I1 is a chest simple X-ray image, the learning data may also have a chest simple X-ray image.

[0140] The training data includes training image data of a plurality of simple X-ray images in which bones are imaged. The imaging sites of the plurality of training image data include, for example, at least one of the neck, chest, waist, proximal femur, knee joint, ankle joint, shoulder joint, elbow joint, wrist joint, finger joint, and jaw joint. The training data may include some types of the 11 types of image data or may include all types of image data. Also, the plurality of training image data may include a front view or a side view.

[0141] The training data has bones of a plurality of different people imaged respectively. For each of the plurality of training image data, as teacher data, the actually measured value of the bone mass or bone density of the subject of each training image data is associated. The actually measured value of the bone mass or bone density is measured at approximately the same time as the time when the training image data was taken.

[0142] Also, the training image data of the training data may be a series of data with different time axes of the same person being imaged. That is, the training image data may include first training data having an X-ray image of a bone and second training data that is an image of the same person as the first training data and has an X-ray image taken after the first training data.

[0143] Also, the training image data of the training data may be a group of data with different ages, etc., by imaging the same part of other people. Also, the training image data of the training data may be a series of data with different time axes of imaging the same person and the same part.

[0144] The learning data and the first input information I1 may use the grayscale image data obtained by reducing the simple X-ray image taken by a simple X-ray imaging device (in other words, a general X-ray imaging device or an X-ray imaging device) and reducing the number of gradations thereof. For example, consider a case where the number of pixel data of the image data is more than (1024 × 640) and the number of bits of the pixel data is 16 bits. In this case, the number of pixel data is reduced to, for example, (256 × 256), (1024 × 512), or (1024 × 640), and the number of bits of the pixel data is reduced to 8 bits, which is used as the first input information I1 and the learning data.

[0145] The teacher data includes, for each of a plurality of learning image data included in the learning data, a measured value of bone mass or bone density of a bone included in the learning simple X-ray image shown by the learning image data. The bone mass or bone density may be measured, for example, by the DEXA (dual-energy X-ray absorptiometry) method or the ultrasonic method.

[0146] <Example of Neural Network Learning> The control unit 834 performs machine learning using the learning data and the teacher data on the approximator 832 so that the approximator 832 can calculate an estimation result O regarding bone mass or bone density from the input information I. The approximator 832 is optimized by known machine learning using the teacher data. The approximator 832 adjusts the variable parameters in the approximator 832 so that the difference between the pseudo-estimation result calculated from the learning data input to the encoding unit E and output from the conversion unit C and the teacher data becomes small.

[0147] Specifically, the control unit 834 inputs the learning data in the storage unit 835 to the encoding unit E. When the control unit 834 inputs the learning data to the encoding unit E, it inputs a plurality of pixel data constituting the learning image data to a plurality of artificial neurons constituting the encoding unit E respectively. Then, the control unit 834 adjusts the parameters so that the error with respect to the actually measured value of bone mass or bone density corresponding to the learning image data is reduced for the estimation result O output from the conversion unit C when the learning image data is input to the encoding unit E. The adjusted parameters become the learned parameters and are stored in the storage unit 835.

[0148] As a method for adjusting the parameters, for example, the error backpropagation method is adopted. The parameters include, for example, the parameters used in the encoding unit E, the decoding unit D, and the conversion unit C. Specifically, the parameters include the weighting coefficients used in the ConvLSTM layers of the encoding unit E and the decoding unit D, and the convolutional layer and the fully connected layer of the conversion unit C.

[0149] As a result, the approximator 832 performs an operation based on the learned parameters on the input information I input to the encoding unit E and outputs the estimation result O from the conversion unit C. When the X-ray image data as the input information I is input to the encoding unit E, a plurality of pixel data constituting this image data are respectively input to a plurality of artificial neurons constituting the input unit 831. Then, the ConvLSTM layer, the convolutional layer, and the fully connected layer can perform an operation using the weighting coefficients included in the learned parameters and output the estimation result O.

[0150] As described above, in the estimation system 801, the image data of the simple X-ray image is used to perform the learning of the approximator 832 and the estimation of the bone mass or bone density by the approximator 832. Therefore, by inputting the input information I to the estimation system 801, the future bone mass or bone density can be output as the estimation result O.

[0151] The estimation result O of the estimation system 801 may be an estimation result on a date in the future from the acquisition date of the input information I. For example, the estimation system 801 can estimate the bone mass or bone density from 3 months to 50 years after the time of imaging, more preferably from 6 months to 10 years after.

[0152] The estimation result O may be output as a value. For example, it may be represented by at least one of YAM (Young Adult Mean), T-score, and Z-score. For example, the output unit 833 may output an estimated value represented by YAM, or may output an estimated value represented by YAM, an estimated value represented by T-score, and an estimated value represented by Z-score.

[0153] Also, the estimation result O may be output as an image. When the estimation result O is an image, for example, an X-ray image-like image may be displayed. Note that an X-ray image-like image is an image imitating an X-ray image. Also, when a series of data with different time axes of the same person and the same part is imaged and learned using ConvLSTM, the temporal change of the image can be predicted. Thereby, a future image can be generated from an X-ray image at one point in time of another patient.

[0154] In addition to bones, internal organs, muscles, fat, or blood vessels may be imaged in the learning data and the input information I. Even in that case, highly accurate estimation can be performed.

[0155] The first input information I1 may have individual data (first individual data) of the subject. The first individual data may be, for example, age information, gender information, height information, weight information, or fracture history. As a result, highly accurate estimation can be performed.

[0156] The first input information I1 may have second individual data of the subject. The second individual data may include, for example, information on blood pressure, lipids, cholesterol, triglycerides, and blood glucose levels. As a result, highly accurate estimation can be performed.

[0157] The first input information I1 may include the lifestyle information of the subject. The lifestyle information may be information such as drinking habits, smoking habits, exercise habits, or eating habits. As a result, highly accurate estimation can be performed.

[0158] The first input information I1 may include the bone metabolism information of the subject. The bone metabolism information may be, for example, bone resorption ability or bone formation ability. These can be measured, for example, by at least one of the bone resorption markers type I collagen cross-linked N-telopeptide (NTX), type I collagen cross-linked C-telopeptide (CTX), tartrate-resistant acid phosphatase (TRACP-5b), deoxypyridinoline (DPD), the bone formation marker bone alkaline phosphatase (BAP), type I collagen cross-linked N-propeptide (P1NP), and the bone-related matrix marker undercarboxylated osteocalcin (ucOC). The bone resorption markers may be measured using serum or urine as a specimen.

[0159] In the estimation system 801, as the input information I, second input information I2 regarding the future planned actions of the subject may be further input. The second input information I2 may be, for example, information regarding improvement plans or post-improvement individual data, or information regarding improvement plans or post-improvement lifestyle, exercise habits, or eating habits. Specifically, the second input information I2 may be information such as weight data, drinking habits, smoking habits, sunlight exposure time, number of steps or walking distance per day, dairy product intake, or intake of foods rich in vitamin D such as fish and mushrooms after improvement. As a result, the estimation system 801 can show an estimated result O with improved future bone mass or bone density.

[0160] Also, the second input information I2 may be, for example, information regarding lifestyle habits planned to deteriorate. As a result, the estimation system can show an estimated result O with deteriorated future bone mass or bone density.

[0161] In the estimation system 801, as the input information I, third input information I3 regarding the therapy for the subject may be further input. The third input information I3 is, for example, information regarding physical therapy or drug therapy. Specifically, the third input information I3 may be at least one of a calcium drug, a female hormone drug, a vitamin drug, a bisphosphonate drug, a SERM (Selective Estrogen Receptor Modulator) drug, a calcitonin drug, a thyroid hormone drug, and a denosumab drug.

[0162] In the estimation system 801, the estimation result O may output a first result O1 based only on the first input information I1 and a second result O2 based on the first input information I1 and at least one of the second and third input information I2 and I3. As a result, the effects on future planned actions can be compared.

[0163] In the estimation system 801, the estimation result O may output not only the future bone mass or bone density but also the current result. As a result, the change in bone mass or bone density over time can be compared.

[0164] FIG. 18 shows the concept of the configuration of the approximator 832 of another embodiment of the estimation system 801.

[0165] The estimation device 803 of the estimation system 801 may have a first approximator 832a and a second approximator 832b. That is, in addition to the above approximator 832 (first approximator 832a), a second approximator 832b may be provided. The second approximator 832b may be, for example, a CNN.

[0166] In this case, in the estimation system 801, the first approximator 832a outputs the first image and the first value to the first output unit 833a as the first estimation result O1. Further, the second approximator 832b outputs the second value to the second output unit 833b as the second estimation result O2 from the first image from the first output unit 833a. As a result, the first value and the second value can be compared as the estimation result O of the future bone mass or bone density.

[0167] The estimation system 801 may output, as an estimation result O, a third value based on a first value and a second value. As a result, for example, a result of correcting the first value based on the second value (the third value) can be used as the estimation result O.

[0168] As described above, the estimation system 801 has been described in detail. However, the above description is illustrative in all aspects and this disclosure is not limited thereto. Also, the various examples described above are applicable in combination as long as they do not conflict with each other. And countless examples that are not illustrated can be assumed without departing from the scope of this disclosure.

Explanation of Signs

[0169] 1 Computer device (estimation device) 20 Storage unit 100 Control program 110, 910, 960 Learned parameters 120 Estimation data 130 Learning data 140 Teacher data 200 Neural network 210 Input layer 230 Output layer 280, 832 Approximator 500 Processing device 600 Bone density estimation system 801 Estimation system 802 Terminal device 803 Estimation device 831 Input unit 833 Output unit 834 Control unit 835 Storage unit 900, 8322 Second neural network 930 Judgment unit 950 Third neural network 980 Fracture prediction unit 8321 First neural network O Estimation result I Input information E Encoding section D Decoding section C Conversion section

Claims

1. An estimation unit that estimates at least one estimated value of the bone density and bone mass of the first person based on a simple X-ray image for estimation obtained by photographing the target part of the first person and the first learned parameters learned based on first learning data including a simple X-ray image for first learning obtained by photographing the target part of the second person and information regarding the health state of the second person; A detection unit that detects the fracture location of the first person based on the second learned parameters for fracture detection from the simple X-ray image for estimation; A determination unit that determines whether or not the first person has osteoporosis based on the estimated value, the fracture location, and a predetermined criterion for whether or not it is osteoporosis An estimation device comprising the above.

2. The estimation device according to Claim 1, wherein the target part of the first person includes at least one of the head, neck, chest, waist, hip joint, knee joint, ankle joint, foot, toe, shoulder joint, elbow joint, wrist joint, hand, finger, and jaw joint.

3. The estimation device according to Claim 1 or 2, wherein the simple X-ray image for estimation is an image obtained by photographing the chest of the first person, and at least one of the lungs and thoracic vertebrae of the first person is shown.

4. The estimation device according to any one of Claims 1 to 3, wherein the simple X-ray image for estimation is taken for purposes other than estimating the estimated value of the first person.

5. The estimation device according to Claim 4, wherein the purpose other than estimating the estimated value of the first person is an image taken for diagnosis or health check.

6. The estimation device according to any one of Claims 1 to 5, wherein the first learned parameters are based on the first learning data and first teacher data including measured values of at least one of the bone density and bone mass measured by measuring the second person.

7. The estimation device according to Claim 6, wherein the simple X-ray image for estimation and the simple X-ray image for first learning are front views.

8. The estimation device according to Claim 6 or 7, wherein the measured value is measured by irradiating the second person with X-rays from the front.

9. The estimation device according to Claim 8, wherein the measured value is a value measured by the dual-energy X-ray absorptiometry method.

10. The estimation device according to any one of claims 6 to 9, wherein the measured value is a value measured at at least one of the lumbar vertebra, proximal femur, radius, and middle phalanx of the second person, the estimation device.

11. The estimation device according to claim 6, wherein the measured value of the first teacher data is measured based on at least one of the arm and heel of the second person by an ultrasonic method, the estimation device.

12. The estimation device according to any one of claims 1 to 11, The estimated value is the bone mineral density per unit area (g / cm ２ ), the bone mineral density per unit volume (g / cm ３ ), at least one of the young adult mean percentage (YAM), T-score, and Z-score, an estimation device.

13. The estimation device according to any one of claims 1 to 12, wherein the simple X-ray image for estimation is obtained from a simple X-ray inspection device that has photographed the first person or a terminal connected to the simple X-ray inspection device, the estimation device.

14. The estimation device according to any one of claims 1 to 13, wherein the detection unit is a neural network that detects a fracture of the first person based on the second learned parameter from the simple X-ray image for estimation, the estimation device.

15. The estimation device according to any one of claims 1 to 14, wherein the simple X-ray image for estimation is a simple X-ray frontal image obtained by photographing the target part of the first person from the front, and the detection unit detects a fracture of the first person based on the second learned parameter from the simple X-ray frontal image, the estimation device.

16. The estimation device according to any one of claims 1 to 15, wherein the second learned parameter is second learning data including a second learning simple X-ray image showing a fractured bone of a third person and a third learning simple X-ray image showing a non-fractured bone of a fourth person, information indicating the presence or absence of a fracture and the fracture location of the bone shown in the second learning simple X-ray image, information indicating the presence or absence of a fracture of the bone shown in the third learning simple X-ray image, and information indicating the past fracture history and past fracture location of at least one of the third person and the fourth person, and is based on the second teacher data, the estimation device.

17. The estimation device according to any one of claims 1 to 16, wherein the simple X-ray image for estimation is a frontal image or a side image of a simple X-ray image of the skeleton of the first person, the estimation device.

Images