JP2022035432A - Information processing program, information processing method and information processing device - Google Patents

Abstract

To provide an information processing program, an information processing method and an information processing device for generating data suitable for learning medical images.SOLUTION: An image generation system includes steps of: repeatedly inputting a first input value to a generated model learned from a medical image to generate a first image while changing the first input value; repeating the process until a distance between the generated first image and an existing medical image is equal to or less than a threshold value, so as to search the first input value that generates a first image similar to the existing medical image; and inputting a second input value to the generated model in which the searched first input value is weighted and averaged to generate the second input value, so as to generate a second image similar to a medical image.SELECTED DRAWING: Figure 4

Description

The present invention relates to an information processing program, an information processing method, and an information processing apparatus.

In the medical field, various diagnoses using X-ray images are performed. For example, the presence of an internal device such as a pacemaker is confirmed based on a chest X-ray image.
The confirmation of the internal device using the X-ray image is performed by a person visually confirming the X-ray image.

In addition, based on a large amount of X-ray images, data is analyzed using AI (Artificial Intelligence) to predict internal devices. However, the chest X-ray image of a child is smaller than that of an adult X-ray image, and sufficient learning data cannot be prepared. Therefore, data analysis using AI becomes insufficient and accurate prediction cannot be made.

Therefore, the image data is expanded by performing image processing on a small amount of image data. For example, a new image is created by performing image processing such as left-right inversion and rotation on a chest X-ray image, or a new X-ray image is synthesized by weighted averaging of a plurality of X-ray images. Is being done.

Japanese Unexamined Patent Publication No. 2000-175896

However, in the conventional image data expansion method as described above, a chest X-ray image in which the internal positions are reversed on the left and right is generated by inverting the left and right, or is usually taken by rotating the image. A chest x-ray image with a never-before-seen angle is generated. Further, by synthesizing a new image by weighted averaging a plurality of chest X-ray images, unnatural data may be generated as a chest X-ray image.
In one aspect, the invention aims to generate data suitable for learning.

Therefore, this information processing program is an information processing program that generates an image similar to an existing medical image and executes a data expansion process for expanding the medical image data by a computer, and is a generation model in which the medical image is learned. 1 The input value is repeatedly input while changing the input value to search for an input value that can generate an image similar to the existing medical image, and the second input value generated by weighted averaging the searched input values is used as the generation model. Input to cause the computer to perform a process of generating an image similar to the medical image.

According to one embodiment, data suitable for learning can be generated.

It is a figure which shows typically the structure of the image generation system as an example of an embodiment. It is a figure for demonstrating the processing of the 1st image generation part in the image generation system as an example of an embodiment. It is a figure for demonstrating the processing of the 2nd input value generation part in the image generation system as an example of embodiment. It is a flowchart for demonstrating the process in the image generation system as an example of an embodiment. It is a figure which illustrates the hardware composition of the image processing apparatus of the image generation system as an example of an embodiment.

Hereinafter, embodiments of the information processing program, information processing method, and information processing apparatus will be described with reference to the drawings. However, the embodiments shown below are merely examples, and there is no intention of excluding the application of various modifications and techniques not specified in the embodiments. That is, the present embodiment can be variously modified and implemented within a range that does not deviate from the purpose. Further, each figure does not have the purpose of having only the components shown in the figure, but may include other functions and the like.

(A) Configuration FIG. 1 is a diagram schematically showing a configuration of an image generation system 1 as an example of an embodiment.
This image generation system 1 realizes a data expansion process for expanding medical image data by generating a medical image, and in the following examples, an example using generation of a child's chest X-ray image will be shown. ..

As shown in FIG. 1, the image generation system 1 includes an image processing device 10 and a known image DB 201. The image processing device 10 includes a first input value generation unit 101, a first image generation unit 102, an input value correction unit 103, a comparison unit 104, a second input value generation unit 105, and a second image generation unit 106. The image processing device 10 is communicably connected to the known image DB 201.
The known image DB 201 is, for example, a known medical image database and stores chest X-ray images of a plurality of children. The known image DB 201 outputs a chest X-ray image of a child in response to a request from the image generation system 1. Hereinafter, a child's chest X-ray image output from the known image DB 201 may be referred to as a sample image.

The first image generation unit 102 uses an image generation model to generate an image according to an input value (input value, first input value). For the image generation model, for example, GAN (Generative Adversarial Networks) may be used.
FIG. 2 is a diagram for explaining the processing of the first image generation unit 102 in the image generation system 1 as an example of the embodiment.

The first image generation unit 102 inputs an image having a predetermined format, which is generated by the first input value generation unit 101 described later, into the image generation model, and the image corresponds to the first input value. Is output. Hereinafter, the image generated by the first image generation unit 102 may be referred to as a first generated image.

Further, the first image generation unit 102 outputs an image corresponding to the corrected first input value by inputting the first input value corrected by the input value correction unit 103 described later into the image generation model. ..

The first input value generation unit 101 generates a first input value to be input to the first image generation unit 102. The first input value is configured as, for example, a combination of a plurality of elements (for example, about 100 dimensions) configured as a variable having a value from 0 to 1.
The first input value generation unit 101 generates a first input value, for example, by combining a plurality of randomly generated elements.

The comparison unit 104 compares the image generated by the first image generation unit 102 (first generated image) with the chest X-ray image (sample image) of the child read from the known image DB 201.

The comparison unit 104 calculates the distance between the first generated image and the sample image. The calculation of the distance between the two images by the comparison unit 104 can be realized by using various known methods. For example, the comparison unit 104 extracts feature quantity vectors from the first generated image and the sample image, respectively, and measures the distance between these feature quantity vectors. The distance of the feature amount vector can be calculated by using various known methods, and for example, the Euclidean distance, the Manhattan distance, the Mahalanobis distance, and the like may be used.
It may be said that the comparison unit 104 calculates the degree of similarity between the first generated image and the sample image.

Further, the comparison unit 104 stores a first input value corresponding to the first generated image whose distance from the sample image is equal to or less than a predetermined threshold value in a predetermined storage area such as the memory unit 12 or the storage device 13 (see FIG. 5). To memorize.

That is, it can be said that the comparison unit 104 searches for a first input value for generating a first generated image in which the distance from the sample image is equal to or less than a predetermined threshold value. Further, the first input value corresponding to the first generated image in which the distance from the sample image is equal to or less than a predetermined threshold value may be referred to as a search result.

The input value correction unit 103 is set so that the difference (distance) from the sample image is smaller than the first input value which is the source of the first generated image compared with the sample image by the comparison unit 104. Make corrections. For example, the input value correction unit 103 corrects the first input value by appropriately increasing or decreasing the value of each element constituting the first input value.

The first input value corrected by the input value correction unit 103 is input to the first image generation unit 102, and the first image generation unit 102 inputs the first input value to the image generation model again to generate the first generated image. Generated.

The comparison unit 104 again calculates the distance from the sample image of the image generated based on the corrected first input value. Based on this calculation result, the input value correction unit 103 indicates that the first generated image generated based on the corrected first input value is larger than the first generated image generated based on the first input value before correction. Check if the difference (distance) from the sample image has become smaller.

The input value correction unit 103 repeatedly corrects the first input value, generates the first generated image by the first image generation unit 102, and compares the sample image with the comparison unit 104, thereby causing a difference (difference from the sample image). The first input value is corrected so that the distance) becomes smaller.
The input value correction unit 103 transfers the corrected first input value to the first image generation unit 102 by storing it in a predetermined storage area of the memory unit 12 or the storage device 13, for example. The first image generation unit 102 inputs the first input value corrected in this way into the image generation model, and outputs an image corresponding to the corrected first input value.

The second input value generation unit 105 uses the first input value (search result) corresponding to the first generated image in which the distance from the sample image is equal to or less than a predetermined threshold value as a result of comparison by the comparison unit 104. An input value (second input value) for generating an image similar to the image is generated.

The second input value generation unit 105 weighted averages the values of all the search results (the first input value corresponding to the first generated image whose distance to the sample image is equal to or less than a predetermined threshold value), thereby performing the second input. Generate a value.
Then, the second input value generation unit 105 randomly changes the weight of the weighted average to generate a plurality of second input values.

FIG. 3 is a diagram for explaining the processing of the second input value generation unit 105 in the image generation system 1 as an example of the embodiment.
In the example shown in FIG. 3, an example of generating three second input values based on n search results 1 to n (n is a natural number of 1 or more) is shown.
The second input value generation unit 105 generates the second input value by using the following equation (1).

上記式（１）において、第２入力値は、i=0からi=nまでのw_ir_iの値を足し合わせることにより求められる（nは探索結果の総数）。
なお、上記式（１）において、r_iは、探索結果iのベクトルを示す。また、w_iは、探索結果iの重みを示し、以下の式（２）で表される。

In the above equation (1), the second input value is obtained by adding the values of w _i r _i from i = 0 to i = n (n is the total number of search results).
In the above equation (1), r _i indicates the vector of the search result i. Further, w _i indicates the weight of the search result i and is expressed by the following equation (2).

また、上記式（２）において、x_iは0以上1以下の一様乱数である。

Further, in the above equation (2), x _i is a uniform random number of 0 or more and 1 or less.

Here, an example of calculating the second input value based on the search results 1 and 2 exemplified below is shown.
Search result 1: [0.14, 0.22,…, 0.42]
Search result 2: [0.20, 0.28,…, 0.36]
The second input value generation unit 105 calculates, for example, a weighted average value in which the weights of the first element "0.14" of the search result 1 and the first element "0.20" of the search result 2 are randomly changed. , The first element "0.18" of the second input value is calculated. In this way, the second input value generation unit 105 completes the second input value by calculating the weighted average value in which the weights are randomly changed between the corresponding elements for each element of the search results 1 and 2. ..
Second input value: [0.18, 0.28,…, 0.36]
It is desirable that the second input value generation unit 105 generate the second input value by using all of the obtained search results. However, the present invention is not limited to this, and the second input value generation unit 105 may generate a second input value by using a part of the obtained search results. This corresponds to the case of w _i = 0.

The first image generation unit 102, the comparison unit 104, the input value correction unit 103, and the second input value generation unit 105 described above input the first input value to the image generation model trained from the medical image (child's chest X-ray image). It functions as a search unit that searches for input values (search results) that can generate images similar to existing medical images by repeatedly inputting while changing.

The second image generation unit 106 uses the image generation model to generate an image (second generated image) corresponding to the second input value generated by the second input value generation unit 105. The second image generation unit 106 generates a child's chest X-ray image similar to the sample image by inputting the second input value generated by the second input value generation unit 105 into the image generation model.

The second image generation unit 106 functions as an image generation unit that generates an image similar to a medical image by inputting a second input value generated by weighted averaging the searched input values into the image generation model.
The second image generation unit 106 generates an image using the same image generation model as the first image generation unit 102, and detailed description thereof will be omitted.
The second generated image thus generated is used, for example, as training data for a learning model for diagnostic imaging.

(B) Operation The processing in the image generation system 1 as an example of the embodiment configured as described above will be described with reference to the flowcharts (steps S1 to S8) shown in FIG.
In step S1, the first input value is generated by combining a plurality of elements randomly generated by the first image generation unit 102.
In step S2, the first image generation unit 102 inputs the first input value generated by the first input value generation unit 101 into the image generation model to generate the first generated image.
In step S3, the comparison unit 104 calculates the distance between the first generated image and the chest X-ray image (sample image) of the child read from the known image DB 201.

In step S4, the comparison unit 104 confirms whether the distance between the first generated image and the sample image is equal to or less than a predetermined threshold value. As a result of the confirmation, if the distance between the first generated image and the sample image is larger than the threshold value (see the NO route in step S4), the process proceeds to step S8.

In step S8, the input value correction unit 103 has a difference (distance) from the sample image with respect to the first input value which is the source of the first generated image compared with the sample image by the comparison unit 104. Make corrections to make it smaller. After that, the process returns to step S2, and the first image generation unit 102 inputs the modified first input value to the image generation model to generate the first generated image.

On the other hand, if the distance between the first generated image and the sample image is equal to or less than the threshold value as a result of the confirmation in step S4 (see the YES route in step S4), the process proceeds to step S5. In step S5, the input value generation unit 105 determines as a search result a first input value for generating a first generated image in which the distance from the sample image is equal to or less than a predetermined threshold value.

In step S6, the second input value generation unit 105 resembles the sample image by using the first input value (search result) corresponding to the first generated image in which the distance from the sample image is equal to or less than a predetermined threshold value. Generate an input value (second input value) for generating an image.

In step S7, the second image generation unit 106 generates a child's chest X-ray image similar to the sample image by inputting the second input value generated by the second input value generation unit 105 into the image generation model. do. After that, the process ends.

(C) Effect As described above, according to the image generation system 1 as an embodiment of the present invention, the first image generation unit 102 produces an image of the first input value randomly generated by the first input value generation unit 101. The first generated image is generated by inputting to the generation model. Further, the comparison unit 104 calculates the distance by comparing the first generated image with the existing chest X-ray image of the child acquired from the known image DB 201.

Then, a plurality of first input values (search results) for generating the first generated image in which the distance from the sample image is equal to or less than the threshold value are stored, and the second input value generation unit 105 stores all the search results (sample). A second input value is generated by weighted averaging the values of the first input value) corresponding to the first generated image whose distance to the image is equal to or less than a predetermined threshold.

The second image generation unit 106 generates a similar image similar to the sample image by inputting the second input value thus generated into the image generation model. As a result, the data can be naturally expanded as a child's chest X-ray image, and effective analysis / prediction can be performed on the child's chest X-ray image.

When the second input value generation unit 105 generates the second input value, the weight of the weighted average is randomly changed to generate a plurality of second input values. Thereby, by appropriately changing the weighted average weight, a similar image similar to the sample image can be easily generated. Further, by changing the weight of the weighted average, a plurality of second input values can be easily generated, whereby a required number of second generated images similar to the sample image can be easily generated. can.

The search result used for calculating the second input value is the first input value for generating the first generated image whose distance is equal to or less than the threshold value as a result of comparison with the sample image by the comparison unit 104. That is, a plurality of second input values are generated by randomly changing the weight of the weighted average with respect to the first input value that generates the first generated image similar to the sample image. As a result, the second generated image generated based on the second input value also resembles the sample image.

(D) Others FIG. 5 is a diagram illustrating a hardware configuration of an image processing device 10 of an image generation system 1 as an example of an embodiment.

The image processing device 10 is an information processing device (computer), and is, for example, a processor 11, a memory unit 12, a storage device 13, a graphic processing device 14, an input interface 15, an optical drive device 16, a device connection interface 17, and a network interface. It has 18 as a component. These components 11 to 18 are configured to be communicable with each other via the bus 19.

The processor (processing unit) 11 controls the entire image processing device 10. The processor 11 may be a multiprocessor. The processor 11 is, for example, one of a CPU, an MPU (Micro Processing Unit), a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), a PLD (Programmable Logic Device), and an FPGA (Field Programmable Gate Array). You may. Further, the processor 11 may be a combination of two or more types of elements among the CPU, MPU, DSP, ASIC, PLD, and FPGA.

Then, the processor 11 executes a control program for the image processing device 10 (image generation program: not shown), whereby the first input value generation unit 101, the first image generation unit 102, and the input value illustrated in FIG. 1 are executed. Functions as a correction unit 103, a comparison unit 104, a second input value generation unit 105, and a second image generation unit 106 are realized. That is, when the processor 11 executes the control program (image generation program) for the image processing device 10, the functions as the search unit and the image generation unit are realized.

The image processing device 10 generates, for example, the first input value generation unit 101 and the first image by executing a program (image generation program, OS program) recorded on a computer-readable non-temporary recording medium. Functions as a unit 102, an input value correction unit 103, a comparison unit 104, a second input value generation unit 105, and a second image generation unit 106 are realized. OS is an abbreviation for Operating System.

The program describing the processing content to be executed by the image processing apparatus 10 can be recorded on various recording media. For example, a program to be executed by the image processing device 10 can be stored in the storage device 13. The processor 11 loads at least a part of the program in the storage device 13 into the memory unit 12 and executes the loaded program.

Further, the program to be executed by the image processing device 10 (processor 11) can be recorded on a non-temporary portable recording medium such as an optical disk 16a, a memory device 17a, and a memory card 17c. The program stored in the portable recording medium can be executed after being installed in the storage device 13 by control from the processor 11, for example. The processor 11 can also read and execute the program directly from the portable recording medium.

The memory unit 12 is a storage memory including a ROM (Read Only Memory) and a RAM (Random Access Memory). The RAM of the memory unit 12 is used as the main storage device of the image processing device 1. At least a part of the program to be executed by the processor 11 is temporarily stored in the RAM. Further, various data necessary for processing by the processor 11 are stored in the memory unit 12.

The storage device 13 is a storage device such as a hard disk drive (HDD), SSD (Solid State Drive), and storage class memory (SCM), and stores various data. The storage device 13 is used as an auxiliary storage device for the image processing device 10. The storage device 13 stores an OS program, a control program, and various data. The control program includes an image generation program.

As the auxiliary storage device, a semiconductor storage device such as an SCM or a flash memory can also be used. Further, RAID (Redundant Arrays of Inexpensive Disks) may be configured by using a plurality of storage devices 13.

The storage device 13 may function as the above-mentioned known image DB 201. Further, the storage device 13 has the first input value generated by the first input value generation unit 101, the second input value generated by the second image generation unit 106, and the first input value corrected by the input value correction unit 103. , Various data generated by each unit may be stored, such as a first generated image generated by the first input value generation unit 101 and a second generated image generated by the second image generation unit 106.

A monitor 14a is connected to the graphic processing device 14. The graphic processing device 14 displays an image on the screen of the monitor 14a according to a command from the processor 11. Examples of the monitor 14a include a display device using a CRT (Cathode Ray Tube), a liquid crystal display device, and the like.

A keyboard 15a and a mouse 15b are connected to the input interface 15. The input interface 15 transmits signals sent from the keyboard 15a and the mouse 15b to the processor 11. The mouse 15b is an example of a pointing device, and other pointing devices can also be used. Other pointing devices include touch panels, tablets, touchpads, trackballs and the like.

The optical drive device 16 reads the data recorded on the optical disk 16a by using a laser beam or the like. The optical disc 16a is a portable non-temporary recording medium in which data is readablely recorded by reflection of light. Examples of the optical disk 16a include a DVD (Digital Versatile Disc), a DVD-RAM, a CD-ROM (Compact Disc Read Only Memory), a CD-R (Recordable) / RW (ReWritable), and the like.

The device connection interface 17 is a communication interface for connecting peripheral devices to the image processing device 10. For example, a memory device 17a or a memory reader / writer 17b can be connected to the device connection interface 17. The memory device 17a is a non-temporary recording medium equipped with a communication function with the device connection interface 17, for example, a USB (Universal Serial Bus) memory. The memory reader / writer 17b writes data to the memory card 17c or reads data from the memory card 17c. The memory card 17c is a card-type non-temporary recording medium.

The network interface 18 is connected to the network. The network interface 18 transmits / receives data via the network. Other information processing devices, communication devices, and the like may be connected to the network. The image processing device 10 may be connected to the known image DB 201 via, for example, the network interface 18 and the network.

The disclosed technique is not limited to the above-described embodiment, and can be variously modified and implemented without departing from the spirit of the present embodiment. Each configuration and each process of the present embodiment can be selected as necessary, or may be combined as appropriate.

In the above-described embodiment, the second input value generation unit 105 generates the second input value using the search results (first input value) 1 to n searched by the comparison unit 104, and the second image generation unit 106. Generates a child's chest x-ray image similar to, but is not limited to, a sample image.

For example, the second image generation unit 106 inputs the search results (first input value) 1 to n searched by the comparison unit 104 into the image generation model as they are to generate an image similar to a child's chest X-ray image. May be good.
Further, in the above-described embodiment, an example of generating a chest X-ray image of a child has been shown, but the present invention is not limited thereto. For example, it may be used for the generation of an adult chest X-ray image, or may be applied to the generation of an X-ray image other than the chest. For example, it may be applied to the generation of an image other than an X-ray image, such as an image obtained from a CT (Computed Tomography) or MRI (Magnetic Resonance Imaging) inspection device, and can be modified in various ways. ..

Furthermore, it is not limited to medical images, and can be applied to the generation of various other images.
Further, according to the above-mentioned disclosure, it is possible to carry out and manufacture this embodiment by a person skilled in the art.

(E) Additional notes The following additional notes will be further disclosed with respect to the above embodiments.
(Appendix 1)
It is an information processing program that generates an image similar to an existing medical image and executes a data expansion process that expands the medical image data by a computer.
By repeatedly inputting the first input value to the generated model in which the medical image is learned while changing the first input value, an input value that can generate an image similar to the existing medical image is searched for.
An information processing program that inputs a second input value generated by weighted averaging of the searched input values to the generation model, and causes the computer to execute a process of generating an image similar to the medical image.

(Appendix 2)
The computer is made to execute a process of changing the first input value so that the distance between the image generated by inputting the first input value into the generation model and the existing medical image is equal to or less than the threshold value. , The information processing program according to Appendix 1.

(Appendix 3)
The information processing program according to Appendix 1 or 2, which causes the computer to execute a process of randomly generating the first input value and inputting the first input value into the generation model.

(Appendix 4)
The information processing program according to any one of Supplementary note 1 to 3, wherein the computer is made to execute a process of randomly changing the weight of the weighted average to generate the second input value.

(Appendix 5)
When a computer performs data expansion processing that expands medical image data by generating an image similar to an existing medical image.
A process of repeatedly inputting a first input value to a generated model learned from a medical image while changing the first input value to search for an input value capable of generating an image similar to the existing medical image.
An information processing method comprising a process of inputting a second input value generated by weighted averaging of searched input values into the generation model to generate an image similar to the medical image.

(Appendix 6)
It is characterized by comprising a process of changing the first input value so that the distance between the image generated by inputting the first input value into the generation model and the existing medical image is equal to or less than the threshold value. The information processing method according to Appendix 5.

(Appendix 7)
The information processing method according to Supplementary Note 5 or 6, further comprising a process of randomly generating the first input value and inputting the first input value into the generation model.

(Appendix 8)
The information processing method according to any one of Supplementary note 5 to 7, further comprising a process of randomly changing the weight of the weighted average to generate the second input value.

(Appendix 9)
An information processing device that performs data expansion processing that generates an image similar to an existing medical image and expands the medical image data.
A search unit that searches for an input value that can generate an image similar to the existing medical image by repeatedly inputting the first input value to the generated model that has learned the medical image while changing the first input value.
An information processing apparatus comprising: an image generation unit for inputting a second input value generated by weighted averaging of searched input values into the generation model to generate an image similar to the medical image.

(Appendix 10)
The search unit
The feature is that the process of changing the first input value is performed so that the distance between the image generated by inputting the first input value into the generation model and the existing medical image is equal to or less than the threshold value. The information processing apparatus according to Appendix 9.

(Appendix 11)
The search unit
The information processing apparatus according to Appendix 9 or 10, wherein the first input value is randomly generated and the first input value is input to the generation model.

(Appendix 12)
The search unit
The information processing apparatus according to any one of Supplementary note 9 to 11, wherein the weight of the weighted average is randomly changed to generate the second input value.

1 Image processing system 10 Image processing device (information processing device)
11 CPU
12 Memory section 13 HDD
14 Graphic processing device 14a Monitor 15 Input interface 15a Keyboard 15b Mouse 16 Optical drive device 16a Optical disk 17 Device connection interface 17a Memory device 17b Memory reader / writer 17c Memory card 18 Network interface 19 Bus 101 1st input value generator 102 1st image generation Part 103 Input value correction part 104 Comparison part 105 Second input value generation part 106 Second image generation part 201 Known image DB

Claims (6)

It is an information processing program that generates an image similar to an existing medical image and executes a data expansion process that expands the medical image data by a computer.
By repeatedly inputting the first input value to the generated model in which the medical image is learned while changing the first input value, an input value that can generate an image similar to the existing medical image is searched for.
An information processing program that inputs a second input value generated by weighted averaging of the searched input values to the generation model, and causes the computer to execute a process of generating an image similar to the medical image. The computer is made to execute a process of changing the first input value so that the distance between the image generated by inputting the first input value into the generation model and the existing medical image is equal to or less than the threshold value. , The information processing program according to claim 1. The information processing program according to claim 1 or 2, wherein the computer is made to execute a process of randomly generating the first input value and inputting the first input value into the generation model. The information processing program according to any one of claims 1 to 3, wherein the computer is made to execute a process of randomly changing the weight of the weighted average to generate the second input value. When a computer performs data expansion processing that expands medical image data by generating an image similar to an existing medical image.
A process of repeatedly inputting a first input value to a generated model learned from a medical image while changing the first input value to search for an input value capable of generating an image similar to the existing medical image.
An information processing method comprising a process of inputting a second input value generated by weighted averaging of searched input values into the generation model to generate an image similar to the medical image. An information processing device that performs data expansion processing that generates an image similar to an existing medical image and expands the medical image data.
A search unit that searches for an input value that can generate an image similar to the existing medical image by repeatedly inputting the first input value to the generated model that has learned the medical image while changing the first input value.
An information processing apparatus comprising: a generation unit for inputting a second input value generated by weighted averaging of searched input values into the generation model to generate an image similar to the medical image.

Images