Learning device, estimation device, learning method, estimation method, and program

Abstract

To reduce the burden required for acquiring an image that shows physiological activity.SOLUTION: A learning apparatus includes a model learning unit which updates a function image estimation model for estimating, based on image data of an appearance image which indicates a shape and composition of a biological tissue, image data of a function image which indicates physiological activity of each part of a biological tissue that satisfies an image condition for designating a biological tissue, until a predetermined end condition is satisfied, using a pair of image data of a function image of a biological tissue that satisfies the image condition and image data of an appearance image of the biological tissue.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a learning device, an estimation device, a learning method, an estimation method, and a program. [Background technology]

[0002] Pulmonary function images, which show the physiological activity of various parts of the lungs, such as images obtained by single photon emission computed tomography (SPECT) scans of the lungs, are sometimes used in disease treatments such as radiation therapy. Physiological activity is also called function. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Publication No. 2018-68814 [Patent Document 2] Japanese Patent Application Laid-Open No. 2013-192912 Summary of the Invention [Problem to be solved by the invention]

[0004] However, obtaining a pulmonary function image can be laborious, requiring the use of special drugs and at least two images, an exhalation image and an inhalation image, as described in the above-mentioned Patent Document 1. This issue is not limited to the lungs, but is a common problem for all living tissues.

[0005] In view of the above circumstances, an object of the present invention is to provide a technique for reducing the burden required to acquire an image showing physiological activity. [Means for solving the problem]

[0006] One aspect of the present invention is a learning device that includes a model learning unit that updates a functional image estimation model that estimates image data of a functional image, which is an image that shows the physiological activity of each part of biological tissue that satisfies image conditions specifying the biological tissue, based on image data of an appearance image, which is an image that shows the shape and composition of the biological tissue, using pairs of image data of the functional image of the biological tissue that satisfies the image conditions and image data of the appearance image of the biological tissue, until a predetermined termination condition is met.

[0007] One aspect of the present invention is an estimation device that includes: an appearance image data acquisition unit that acquires image data of an appearance image, which is an image showing the shape and composition of a target tissue, which is biological tissue that satisfies image conditions that specify the biological tissue; and an estimation unit that estimates image data of a functional image, which is an image showing the physiological activity of each part of the biological tissue that satisfies the image conditions, based on image data of the appearance image, which is an image showing the shape and composition of the biological tissue, using a trained functional image estimation model that is a mathematical model updated using pairs of image data of the functional image of the biological tissue that satisfies the image conditions and image data of the appearance image of the biological tissue, until a predetermined termination condition is met, and estimates image data of the functional image of the biological tissue indicated by the image data obtained by the appearance image data acquisition unit.

[0008] One aspect of the present invention is a learning method including a model learning step of updating a functional image estimation model that estimates image data of a functional image, which is an image showing the physiological activity of each part of biological tissue that satisfies image conditions specifying the biological tissue, based on image data of an appearance image, which is an image showing the shape and composition of the biological tissue, using pairs of image data of the functional image of the biological tissue that satisfies the image conditions and image data of the appearance image of the biological tissue, until a predetermined termination condition is met.

[0009] One aspect of the present invention is an estimation method comprising: an appearance image data acquisition step of acquiring image data of an appearance image, which is an image showing the shape and composition of a target tissue, which is biological tissue that satisfies image conditions specifying the biological tissue; and an estimation step of estimating image data of a functional image, which is an image showing the physiological activity of each part of the biological tissue that satisfies the image conditions, based on the image data of the appearance image, which is an image showing the shape and composition of the biological tissue, using a trained functional image estimation model, which is a mathematical model updated using pairs of image data of the functional image of the biological tissue that satisfies the image conditions and image data of the appearance image of the biological tissue, until a predetermined termination condition is met.

[0010] One aspect of the present invention is a program for causing a computer to function as the learning device described above.

[0011] One aspect of the present invention is a program for causing a computer to function as the above-described estimation device. [Effects of the Invention]

[0012] The present invention makes it possible to provide a technique that reduces the burden required to acquire images that show physiological activity. [Brief explanation of the drawings]

[0013] [Figure 1] FIG. 1 is an explanatory diagram illustrating a functional image acquisition system according to an embodiment. [Figure 2] FIG. 2 is a diagram illustrating an example of a hardware configuration of a learning device according to an embodiment. [Figure 3] FIG. 2 is a diagram showing an example of the configuration of a control unit included in the learning device according to the embodiment. [Figure 4] 10 is a flowchart showing an example of a flow of processing executed by a learning device according to an embodiment. [Figure 5] FIG. 2 is a diagram illustrating an example of a hardware configuration of an estimation apparatus according to an embodiment. [Figure 6] FIG. 2 is a diagram showing an example of the configuration of a control unit included in the estimation device according to the embodiment. [Figure 7] 1 is a flowchart showing an example of a flow of processing executed by an estimation device according to an embodiment. [Figure 8] FIG. 10 is a first explanatory diagram illustrating an example of the results of an experiment using the functional image acquisition system according to the embodiment. [Figure 9] FIG. 2 is a second explanatory diagram illustrating an example of the results of an experiment using the functional image acquisition system according to the embodiment. [Figure 10] FIG. 3 is a third explanatory diagram illustrating an example of the results of an experiment using the functional image acquisition system according to the embodiment. [Figure 11] FIG. 4 is a fourth explanatory diagram illustrating an example of the results of an experiment using the functional image acquisition system according to the embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0014] (Embodiment) 1 is an explanatory diagram illustrating a functional image acquisition system 100 according to an embodiment. The functional image acquisition system 100 acquires image data of functional images, which are images showing the physiological activity of each part of biological tissue (hereinafter referred to as "target tissue") that meets image conditions, which are conditions for specifying the biological tissue. More specifically, the functional image acquisition system 100 acquires image data of functional images based on image data of appearance images, which are images showing the shape and composition of the target tissue, such as CT (Computed Tomography) images.

[0015] The imaging condition may be any predetermined condition that specifies a biological tissue. The biological tissue specified by the imaging condition may be, for example, the lung. The biological tissue specified by the imaging condition may be, for example, the brain. The biological tissue specified by the imaging condition may be, for example, the large intestine. The biological tissue specified by the imaging condition does not necessarily have to be one, and may be multiple.

[0016] The functional image is, for example, an image obtained by a SPECT (Single Photon Emission Computed Tomography) examination. The functional image may also be, for example, an image obtained by a PET (Positron Emission Tomography) examination. 3 He or 129 The image may be an image obtained by MRI (Magnetic Resonance Imaging) using Xe as a contrast agent.

[0017] The appearance image may be one or more of a plurality of two-dimensional cross-sectional images obtained from a three-dimensional appearance image. The three-dimensional appearance image may be obtained, for example, by a three-dimensional CT scan. The functional image may also be one or more of a plurality of two-dimensional cross-sectional images obtained from a three-dimensional functional image. The three-dimensional functional image may be obtained, for example, by a three-dimensional SPECT examination, a three-dimensional PET examination, or a three-dimensional MRI.

[0018] The functional image acquisition system 100 includes a learning device 1 and an estimation device 2. The learning device 1 obtains, by machine learning, a mathematical model for estimating image data of a functional image of a target tissue based on image data of an appearance image of the target tissue.

[0019] More specifically, the learning device 1 updates a mathematical model (hereinafter referred to as a "functional image estimation model") that estimates image data of a functional image of a target tissue based on image data of an appearance image of the target tissue through machine learning until a predetermined condition for terminating learning (hereinafter referred to as a "learning termination condition") is satisfied. The learning device 1 acquires the functional image estimation model at the time when the learning termination condition is satisfied as the trained functional image estimation model.

[0020] The learning termination condition is, for example, that a predetermined number of updates have been performed. The learning termination condition may also be, for example, that the change in the functional image estimation model due to the updates is smaller than a predetermined change.

[0021] An example of learning will be described. In learning, pairs of image data of an external image of a target tissue and image data of a functional image of the target tissue (hereinafter referred to as "learning data") are used for learning. A collection of learning data is a so-called learning dataset. Hereinafter, image data of functional images included in the learning data will be referred to as learning functional image data. Also, hereinafter, image data of an estimated image output by a functional image estimation model will be referred to as estimated image data. The image data of an estimated image output by a functional image estimation model is image data of an estimated image estimated by the functional image estimation model.

[0022] Information D100 in FIG. 1 is an example of a training data set. Images D101-1 to D101-N (N is an integer equal to or greater than 1) are each an example of an appearance image. Images D102-1 to D102-N are each an example of a functional image. Each pair of image data of image D101-n and image data of image D102-n (n is an integer equal to or greater than 1 and equal to or less than N) is each an example of training data. Neural network D103 is an example of a neural network that represents a functional image estimation model. Each training data is input to neural network D103, thereby updating neural network D103.

[0023] In the learning, a functional image estimation model is executed on image data of the appearance image included in the learning data. In the learning, the functional image estimation model is updated based on the estimated image data output as a result of the execution of the functional image estimation model and the learning functional image data so as to reduce the difference between the estimated image data and the learning functional image data. The update is continued until the learning termination condition is satisfied. The functional image estimation model at the time when the learning termination condition is satisfied is the trained functional image estimation model.

[0024] The mathematical model to be trained is represented by a neural network consisting of an encoder that extracts features from an image using a convolutional layer, and a decoder that receives the extracted features and outputs a probability map. During training, the mathematical model is updated by updating the parameter values ​​of the neural network.

[0025] The estimation device 2 uses the trained functional image estimation model to estimate image data of the functional image of the biological tissue shown in the external image based on the image data of the input external image.

[0026] 2 is a diagram showing an example of the hardware configuration of a learning device 1 according to an embodiment. The learning device 1 includes a control unit 11 having a processor 91, such as a CPU (Central Processing Unit), and a memory 92 connected via a bus, and executes a program. By executing the program, the learning device 1 functions as a device including the control unit 11, an input unit 12, a communication unit 13, a memory unit 14, and an output unit 15.

[0027] More specifically, the processor 91 reads out a program stored in the storage unit 14 and stores the read out program in the memory 92. When the processor 91 executes the program stored in the memory 92, the learning device 1 functions as a device including a control unit 11, an input unit 12, a communication unit 13, a storage unit 14, and an output unit 15.

[0028] The control unit 11 controls the operation of various functional units included in the learning device 1. The control unit 11, for example, updates the functional image estimation model through learning. The control unit 11, for example, controls the operation of the output unit 15. The control unit 11, for example, records various information generated by executing learning of the functional image estimation model in the storage unit 14. The control unit 11, for example, records the obtained trained functional image estimation model in the storage unit 14.

[0029] Input unit 12 includes input devices such as a mouse, keyboard, and touch panel. Input unit 12 may be configured as an interface that connects these input devices to learning device 1. Input unit 12 accepts input of various types of information to learning device 1.

[0030] The communication unit 13 includes a communication interface for connecting the learning device 1 to an external device. The communication unit 13 communicates with the external device via wired or wireless communication. The external device is, for example, a device that transmits learning data. The communication unit 13 acquires learning data by communicating with the source of the learning data. The external device is, for example, the estimation device 2. The communication unit 13 outputs the trained functional image estimation model to the estimation device 2 by communicating with the estimation device 2. Note that the learning data does not necessarily have to be input via the communication unit 13, and may be input to the input unit 12. Furthermore, some or all of the learning data may be stored in advance in the storage unit 14.

[0031] The storage unit 14 is configured using a computer-readable storage medium device such as a magnetic hard disk drive or a semiconductor storage device. The storage unit 14 stores various information related to the learning device 1. The storage unit 14 stores information input via, for example, the input unit 12 or the communication unit 13. The storage unit 14 stores various information generated by, for example, executing learning of a functional image estimation model. The storage unit 14 stores, in advance, for example, a functional image estimation model before learning is performed. The storage unit 14 may also store the obtained trained functional image estimation model.

[0032] The output unit 15 outputs various types of information. The output unit 15 includes a display device such as a CRT (Cathode Ray Tube) display, a liquid crystal display, or an organic EL (Electro-Luminescence) display. The output unit 15 may be configured as an interface that connects these display devices to the learning device 1. The output unit 15 outputs, for example, information input to the input unit 12. The output unit 15 may also display, for example, the execution results of a functional image estimation model.

[0033] 3 is a diagram showing an example of the configuration of the control unit 11 included in the learning device 1 according to the embodiment. The control unit 11 includes a learning data acquisition unit 110, a model learning unit 120, a communication control unit 130, a memory control unit 140, and an output control unit 150.

[0034] The learning data acquisition unit 110 acquires learning data. The learning data acquisition unit 110 acquires learning data input to the input unit 12 or the communication unit 13. The learning data acquisition unit 110 may acquire learning data by reading out learning data already stored in the storage unit 14.

[0035] The model learning unit 120 uses the learning data to update the functional image estimation model until the learning end condition is satisfied. That is, the model learning unit 120 executes learning of the functional image estimation model using the learning data, thereby obtaining a trained functional image estimation model.

[0036] The communication control unit 130 controls the operation of the communication unit 13. The storage control unit 140 records various information in the storage unit 14. The output control unit 150 controls the operation of the output unit 15.

[0037] 4 is a flowchart showing an example of the flow of processing executed by the learning device 1 in the embodiment. The learning data acquisition unit 110 acquires learning data (step S101). Next, the model learning unit 120 executes a functional image estimation model on image data of an appearance image included in the learning data (step S102). By executing the functional image estimation model on the image data of the appearance image included in the learning data, image data of an image estimated by the functional image estimation model is obtained.

[0038] Next, the model learning unit 120 updates the functional image estimation model so as to reduce the difference between the image data of the image estimated by the functional image estimation model and the image data of the functional image included in the learning data (step S103). Next, the model learning unit 120 determines whether the learning end condition is satisfied (step S104). If the learning end condition is not satisfied (step S104: NO), the process returns to step S101, and new learning data is acquired. On the other hand, if the learning end condition is satisfied (step S104: YES), the process ends. The functional image estimation model at the time when the learning end condition is satisfied is the trained functional image estimation model.

[0039] 5 is a diagram illustrating an example of a hardware configuration of the estimation device 2 according to an embodiment. The estimation device 2 includes a control unit 21 having a processor 93 such as a CPU and a memory 94 connected via a bus, and executes a program. By executing the program, the estimation device 2 functions as a device including the control unit 21, an input unit 22, a communication unit 23, a storage unit 24, and an output unit 25.

[0040] More specifically, the processor 93 reads out a program stored in the storage unit 24 and stores the read out program in the memory 94. When the processor 93 executes the program stored in the memory 94, the estimation device 2 functions as a device including a control unit 21, an input unit 22, a communication unit 23, a storage unit 24, and an output unit 25.

[0041] The control unit 21 controls the operation of various functional units included in the estimation device 2. The control unit 21, for example, executes a trained functional image estimation model. The control unit 21, for example, controls the operation of the output unit 25. The control unit 21, for example, records various information generated by the execution of the trained functional image estimation model in the storage unit 24. The control unit 21, for example, records the estimation result of the trained functional image estimation model in the storage unit 24.

[0042] The input unit 22 includes input devices such as a mouse, a keyboard, and a touch panel. The input unit 22 may be configured as an interface that connects these input devices to the estimation device 2. The input unit 22 accepts input of various types of information to the estimation device 2.

[0043] The communication unit 23 includes a communication interface for connecting the estimation device 2 to an external device. The communication unit 23 communicates with the external device via wired or wireless communication. The external device is, for example, a device that is a sender of image data of the appearance image. The communication unit 23 acquires image data of the appearance image by communicating with the device that is a sender of the image data of the appearance image.

[0044] The external device is, for example, the learning device 1. The communication unit 13 acquires the trained functional image estimation model by communicating with the learning device 1. Note that image data of the appearance image acquired by the communication unit 23 is the target of execution of the trained functional image estimation model. Note that the target of execution of the trained functional image estimation model does not necessarily have to be input to the communication unit 23, but may be input to the input unit 22.

[0045] The storage unit 24 is configured using a computer-readable storage medium device such as a magnetic hard disk device or a semiconductor storage device. The storage unit 24 stores various information related to the estimation device 2. The storage unit 24 stores information input via, for example, the input unit 22 or the communication unit 23. The storage unit 24 stores, for example, a trained functional image estimation model. The storage unit 24 stores, for example, various information generated by executing the trained functional image estimation model.

[0046] The output unit 25 outputs various types of information. The output unit 25 includes a display device such as a CRT display, a liquid crystal display, or an organic EL display. The output unit 25 may be configured as an interface that connects these display devices to the estimation device 2. The output unit 25 outputs, for example, information input to the input unit 22. The output unit 25 may also display, for example, the results of estimation of a trained functional image estimation model.

[0047] 6 is a diagram showing an example of the configuration of the control unit 21 included in the estimation device 2 according to the embodiment. The control unit 21 includes an appearance image data acquisition unit 210, an estimation unit 220, a communication control unit 230, a storage control unit 240, and an output control unit 250.

[0048] The appearance image data acquisition unit 210 acquires image data of the appearance image. The appearance image data acquisition unit 210 acquires image data of the appearance image input to the input unit 22 or the communication unit 23.

[0049] The estimation unit 220 executes the trained functional image estimation model on the image data of the appearance image acquired by the appearance image data acquisition unit 210. The estimation unit 220 executes the trained functional image estimation model to estimate image data of the functional image of the biological tissue indicated by the image data that is the target of execution.

[0050] The communication control unit 230 controls the operation of the communication unit 23. The storage control unit 240 records various information in the storage unit 24. The output control unit 250 controls the operation of the output unit 25.

[0051] 7 is a flowchart showing an example of the flow of processing executed by the estimation device 2 in an embodiment. The appearance image data acquisition unit 210 acquires image data of an appearance image (step S201). Next, the estimation unit 220 executes a trained functional image estimation model on the image data of the appearance image obtained in step S201 (step S202). By executing the trained functional image estimation model on the image data of the appearance image obtained in step S201, the estimation unit 220 estimates image data of a functional image of biological tissue indicated by the image data obtained in step S201. Next, the output control unit 250 causes the output unit 25 to output the estimation result obtained in step S202 (step S203).

[0052] <Experimental Results> This section describes an example of the results of an experiment using 71 cases in which SPECT / CT ventilation examinations were performed. In the experiment, image data of functional images of pulmonary ventilation was estimated. In the experiment, images obtained by SPECT examinations and CT images of each patient were used. In the experiment, 5-fold cross-validation was performed. Two indices were used in the validation. One of the two indices was Spearman's rank correlation coefficient (Spearman r s The other of the two indices was the Dice similarity coefficient (DSC).

[0053] Fig. 8 is a first explanatory diagram illustrating an example of the results of an experiment using the functional image acquisition system 100 according to the embodiment. More specifically, Fig. 8 shows an example of the results of estimating image data of a functional image of pulmonary ventilation function using a trained functional image estimation model obtained using the functional image acquisition system 100.

[0054] FIG. 9 is a second explanatory diagram illustrating an example of the results of an experiment using the functional image acquisition system 100 of the embodiment. More specifically, FIG. 9 is data for comparison with the estimation result of FIG. 8, and is a functional image of the provider of the appearance image used to estimate the result of FIG. 8. More specifically, FIG. 9 is an image obtained by a SPECT test. That is, the image of FIG. 9 is a correct image for the estimation result shown in FIG. 8.

[0055] Figures 8 and 9 show that the image in Figure 8 is substantially the same as the image in Figure 9. The trained functional image estimation model obtained by the learning device 1 in this way estimates image data of the functional image with high accuracy based on image data of the appearance image.

[0056] 10 is a third explanatory diagram illustrating an example of the results of an experiment using the functional image acquisition system 100 according to the embodiment. More specifically, FIG. 10 shows the Spearman r index obtained as a result of the experiment. s The distribution of values ​​for DSC High, DSC Middle, and DSC Low is shown in Figure 10. MCD U-Net " indicates the result of estimation using a trained functional image estimation model. Figure 10 shows that the accuracy of estimation is higher than that of conventional methods such as the method described in Patent Document 1. DSC High, Middle, and Low represent high, medium, and low pulmonary ventilation functions, respectively.

[0057] 11 is a fourth explanatory diagram illustrating an example of the results of an experiment using the functional image acquisition system 100 according to the embodiment. As a result of the experiment, the index Spearman r sFIG. 11 shows that the DSC High value was 0.69±0.07 as a result of the experiment. FIG. 11 shows that the DSC Middle value was 0.50±0.06 as a result of the experiment. FIG. 11 shows that the DSC Low value was 0.75±0.04 as a result of the experiment. Thus, the results in FIG. 11 show that the accuracy of estimation is higher than that of conventional methods such as the method described in Patent Document 1. Note that the index Spearman r s The value is the correlation coefficient between the predicted value and the actual value, the DSC High value is the degree of agreement for high-functioning lung areas, the DSC Middle value is the degree of agreement for medium-functioning lung areas, and the DSC Low value is the degree of agreement for low-functioning lung areas.

[0058] The learning device 1 in this embodiment configured as described above updates, through learning, a mathematical model that estimates image data of a functional image based on image data of an appearance image. Using the acquired trained mathematical model, it is possible to estimate image data of a functional image with high accuracy, without using special agents such as contrast agents, based on a single appearance image. Therefore, the learning device 1 can reduce the burden required to acquire images showing physiological activity.

[0059] Furthermore, the estimation device 2 in this embodiment configured as above estimates image data of a functional image using the learned mathematical model obtained by the learning device 1. Therefore, the estimation device 2 can reduce the burden required to obtain an image showing physiological activity.

[0060] Furthermore, the functional image acquisition system 100 in this embodiment configured as described above includes the learning device 1. Therefore, the functional image acquisition system 100 can reduce the burden required to acquire images showing physiological activity.

[0061] (Variation) When the target tissue is the lung, the external appearance image may be an image of the extracted lung field.

[0062] When the target tissue is the lung and the external image is a CT image, the external image may be a normalized image in which values ​​equal to or greater than the maximum CT value of normal lung parenchyma are set to 1 and values ​​equal to or less than the minimum CT value are set to 0. The CT values ​​of normal lung parenchyma in CT images are often outside the range of images of biological tissues other than the lung. Therefore, this scaling can enhance the information contained in the lung parenchyma. As a result, the efficiency of learning the functional image estimation model is improved, and the estimation accuracy of the trained functional image estimation model is further improved. The maximum CT value is, for example, -250 HU, and the minimum CT value is, for example, -999 HU.

[0063] If the target tissue is the lung, the functional image may be an image obtained by SPECT normalization using the median pulmonary ventilation value. If the target tissue is the lung, the functional image may be an image obtained by SPECT normalization using the maximum pulmonary ventilation value.

[0064] Each of the functional image acquisition system 100, the learning device 1, and the estimation device 2 may be implemented using a plurality of information processing devices communicably connected via a network. In this case, each functional unit included in each of the functional image acquisition system 100, the learning device 1, and the estimation device 2 may be distributed and implemented in a plurality of information processing devices.

[0065] All or part of the functions of the functional image acquisition system 100, the learning device 1, and the estimation device 2 may be realized using hardware such as an ASIC (Application Specific Integrated Circuit), a PLD (Programmable Logic Device), or an FPGA (Field Programmable Gate Array). The program may be recorded on a computer-readable recording medium. Examples of computer-readable recording media include portable media such as flexible disks, optical magnetic disks, ROMs, and CD-ROMs, and storage devices such as hard disks built into computer systems. The program may be transmitted via a telecommunications line.

[0066] Although an embodiment of the present invention has been described above in detail with reference to the drawings, the specific configuration is not limited to this embodiment, and includes designs within the scope of the gist of the present invention. [Explanation of symbols]

[0067] 100...functional image acquisition system, 1...learning device, 2...estimation device, 11...control unit, 12...input unit, 13...communication unit, 14...memory unit, 15...output unit, 110...learning data acquisition unit, 120...model learning unit, 130...communication control unit, 140...memory control unit, 150...output control unit, 91...processor, 92...memory, 21...control unit, 22...input unit, 23...communication unit, 24...memory unit, 25...output unit, 210...appearance image data acquisition unit, 220...estimation unit, 230...communication control unit, 240...memory control unit, 250...output control unit, 91...processor, 92...memory, 93...processor, 94...memory

Claims

1. a model learning unit that updates a functional image estimation model that estimates image data of a functional image, which is an image showing physiological activity of each part of a biological tissue that satisfies an image condition specifying the biological tissue, based on image data of an appearance image, which is an image showing the shape and composition of the biological tissue, using pairs of image data of the functional image of the biological tissue that satisfies the image condition and image data of the appearance image of the biological tissue until a predetermined termination condition is satisfied; Equipped with the biological tissue satisfying the image conditions is the lung; The appearance image is a CT (Computed Tomography) image of the lung, normalized such that a CT value of a normal lung parenchyma equal to or greater than the maximum value is 1 and a CT value of a normal lung parenchyma equal to or less than the minimum value is 0. Learning device.

2. The functional image is an image obtained by a SPECT examination and normalized using the median pulmonary ventilation value. The learning device according to claim 1 .

3. an appearance image data acquisition unit that acquires image data of an appearance image that is an image showing the shape and composition of a target tissue that is a biological tissue that satisfies an image condition that specifies the biological tissue; an estimation unit that estimates image data of a functional image, which is an image showing physiological activity of each part of the biological tissue that satisfies the image conditions, based on image data of an appearance image, which is an image showing the shape and composition of the biological tissue, using a trained functional image estimation model that is a mathematical model updated using pairs of image data of the functional image of the biological tissue that satisfies the image conditions and image data of the appearance image of the biological tissue, based on image data obtained by the appearance image data acquisition unit, until a predetermined termination condition is satisfied; Equipped with the biological tissue satisfying the image conditions is the lung; The appearance image is a CT (Computed Tomography) image of the lung, normalized such that a CT value of a normal lung parenchyma equal to or greater than the maximum value is 1 and a CT value of a normal lung parenchyma equal to or less than the minimum value is 0. Estimation device.

4. a model learning step of updating a functional image estimation model that estimates image data of functional images, which are images showing physiological activity of each part of biological tissue that satisfy an image condition specifying the biological tissue, based on image data of appearance images, which are images showing the shape and composition of the biological tissue, using pairs of image data of functional images of biological tissue that satisfy the image condition and image data of appearance images of the biological tissue, until a predetermined termination condition is satisfied; and the biological tissue satisfying the image conditions is the lung; The appearance image is a CT (Computed Tomography) image of the lung, normalized such that a CT value of a normal lung parenchyma equal to or greater than the maximum value is 1 and a CT value of a normal lung parenchyma equal to or less than the minimum value is 0. How to learn.

5. an appearance image data acquisition step of acquiring image data of an appearance image which is an image showing the shape and composition of a target tissue which is a biological tissue that satisfies an image condition specifying the biological tissue; an estimation step of estimating image data of the functional image of the biological tissue indicated by the image data obtained in the appearance image data acquisition step, using a trained functional image estimation model, which is a mathematical model that estimates image data of a functional image, which is an image showing the physiological activity of each part of the biological tissue that satisfies the image conditions, based on image data of an appearance image, which is an image showing the shape and composition of the biological tissue, until a predetermined termination condition is satisfied, using a trained functional image estimation model that is a mathematical model that is updated using pairs of image data of the functional image of the biological tissue that satisfies the image conditions and image data of the appearance image of the biological tissue; and the biological tissue satisfying the image conditions is the lung; The appearance image is a CT (Computed Tomography) image of the lung, normalized such that a CT value of a normal lung parenchyma equal to or greater than the maximum value is 1 and a CT value of a normal lung parenchyma equal to or less than the minimum value is 0. Estimation method.

6. A program for causing a computer to function as the learning device according to claim 1 or 2.

7. A program for causing a computer to function as the estimation device according to claim 3.

Images