Information processing system, information processing method, and program
Patent Information
- Application Number
- PCT/JP2025/005635
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2025-02-19
- Publication Date
- 2026-08-27
Smart Images
Figure JP2025005635_27082026_PF_FP_ABST
Abstract
Description
Information Processing System, Information Processing Method, and Program
[0001] The present invention relates to an information processing system, an information processing method, and a program.
[0002] Techniques for estimating the risk of cancer in the breast region based on ultrasonic images are known (Patent Document 1).
[0003] Japanese Patent Application Laid-Open No. 2023-077820
[0004] Construction of a learned model for estimating an affected part from a captured image that captures a part of the human body, such as an ultrasonic image, is desired. For example, if a learned model can be constructed using the judgment of an expert on the captured image as teacher data, the accuracy of specifying the position of the affected part based on the input captured image can be improved. Therefore, in the present embodiment, supervised learning is adopted in the learning process for creating a learned model. On the other hand, in order to create teacher data used for supervised learning, it is necessary to add the judgment of an expert such as a doctor. However, when the judgment of an expert on the captured image is made by marking using an operation such as a mouse or a touch pen, individual differences appear in the mode of marking indicating the position of the affected part. In addition, such experts are busy, and there are also problems such as labor costs. If the marking speed is increased for an expert, the marking accuracy may be poor. In other words, even if the captured image with marking is adopted as teacher data as it is, the position of the affected part cannot be specified accurately. Or, when performing accurately, the labor of the expert becomes great.
[0005] The present invention has been made in view of such a situation, and an object thereof is to facilitate the construction of a model for specifying the position of an affected part based on an input captured image.
[0006] To achieve the above objective, one aspect of the present invention is an information processing system comprising: an operation receiving means for receiving an operation to specify a region of a diseased area included in an image of a part of the human body captured using a predetermined imaging method; a generation means for generating information of a modified designated region obtained by changing the designated region specified by the received operation; and a learning means for learning the generated information of the modified designated region.
[0007] An information processing method and program corresponding to the above-mentioned information processing system according to one aspect of the present invention are also provided as an information processing method and program corresponding to the information processing system according to one aspect of the present invention.
[0008] According to the present invention, it is possible to easily construct a model for identifying the location of the affected area based on the input captured image.
[0009] This figure shows an example of the overall configuration of an information processing system according to one embodiment of the present invention. This is a block diagram showing an example of the hardware configuration of the management server constituting the information processing system of Figure 1. This is a functional block diagram showing an example of the functional configuration of the management server and physician terminal constituting the information processing system of Figure 1. This is a flowchart showing an example of the processing flow of the management server. This figure shows a specific example of an image captured with a specified area associated with it. This figure shows a specific example of an image captured with a modified specified area associated with it. This figure shows a specific example of an image captured with a modified specified area associated with it. This figure shows a specific example of an image captured with a modified specified area associated with it. This figure shows a specific example of an image captured with a modified specified area associated with it. This is a graph used when evaluating information on the modified specified area.
[0010] Hereinafter, this embodiment will be described with reference to the drawings. <Information Processing System S> Figure 1 is a diagram showing an example of the overall configuration of an information processing system S according to one embodiment of the present invention. The information processing system S is an information processing system configured such that a management server 1 and physician terminals 2-1 to 2-r (where r is an integer value of 1 or more) are connected via a network N. The network N is, for example, the Internet, LAN (Local Area Network), VPN (Virtual Private Network), etc. Hereinafter, when it is not necessary to describe each of the physician terminals 2-1 to 2-r individually, they will be collectively referred to as "physician terminal 2".
[0011] The information processing system S is a system that enables automatic estimation of the location of an affected area contained in an image of a part of the human body taken using a predetermined imaging method. Here, "an image of a part of the human body taken using a predetermined imaging method" refers to, for example, an MRI (Magnetic Resonance Image) image taken in an MRI (Magnetic Resonance Image) examination, a CT (Computed Tomography) image taken in an CT (Computed Tomography) examination, or an ultrasound image taken in an ultrasound examination. Furthermore, "the location of the affected area" refers to, for example, the location where cancer has developed. The location of the affected area may be represented as a region of any specified shape within the image, or it may be represented by two-dimensional coordinates.
[0012] [Management Server 1] Management Server 1, which constitutes the information processing system S, is an information processing device that acts as a server for managing the entire information processing system S. Management Server 1 is capable of executing predetermined application programs that make the information processing system S available for use. Management Server 1 is capable of transmitting various types of information to the physician terminal 2 and external sources and executing various types of processing. Furthermore, Management Server 1 is capable of acquiring various types of information transmitted from the physician terminal 2 and external sources and executing various types of processing.
[0013] When the management server 1 receives an operation to specify the area of the affected area as an operation on an image of a part of the human body captured using a predetermined imaging method, it acquires the content of the operation and manages it in association with the image on which the operation was performed. The management server 1 may store the image in advance, or it may acquire it along with the content of the operation. In this embodiment, such an operation is performed by a physician using the physician terminal 2, and specifically, it is an operation to specify a rectangular area that includes at least a part of the area of the affected area by dragging with a mouse or the like. Hereinafter, the area of the image that is specified as the area of the affected area by the physician's operation will be called the "specified area". The area specification by dragging may be accepted as a free curve or as a geometric shape (for example, a rectangle or an ellipse).
[0014] In other words, captured images and information about the designated area associated with those images can serve as training data. However, as mentioned above, even if captured images with handwritten markings are used directly as training data, the location of the affected area cannot be accurately identified. Alternatively, accurately identifying the location of the affected area would require considerable effort from experts. Therefore, the management server 1 can use the information about the designated area as training data in the following way.
[0015] Management server 1 manages information about areas that have been modified by a physician's operation, linking it to the captured image. For example, management server 1 manages information about areas that have been modified by a physician's operation, specifically those that have been enlarged, reduced, moved, or deformed, linking it to the captured image. The area obtained by modifying the captured image based on the area designated by the physician as the affected area is called the "modified area."
[0016] The management server 1 generates information on the designated modification areas included in the captured image as candidate training data based on the captured image, and stores and manages it in a database. The management server 1 can use information on two or more designated modification areas from among multiple designated modification areas corresponding to one designated area. This makes the candidate training data available as conventional augmented data. Here, "augmented data" is data that has been expanded (augmented) in order to secure the large amount of image data necessary for machine learning from the number of captured images that can be prepared. Conventional augmented data is generated by processing the acquired image data, such as by scaling or affine transformation. In contrast, the augmented data according to this embodiment differs from conventional augmented data in that the captured image itself is not processed, but is left as is, and is generated by a simple process that performs one or more changes from scaling, moving, or transforming the designated area.
[0017] Management Server 1 prioritizes selecting high-quality training data from among the candidate training data that can contribute to the efficiency of machine learning resources. To this end, Management Server 1 evaluates the candidate training data and selects those that meet predetermined criteria as training data. Specifically, for example, Management Server 1 evaluates the number of edge pixels (hereinafter referred to as "edge pixels") detected and mapped by edge detection processing on the entire captured image, based on the number of edge pixels included in the designated change area and the total number of pixels (area) in the designated change area. Management Server 1 then selects those evaluation results that meet predetermined criteria as training data.
[0018] In this case, the management server 1 can evaluate the captured image and the information of the designated area associated with that image based on the change in the number of edge pixels included in the designated area, with the total number of pixels (area) of the designated area being used as a parameter. For example, the management server 1 can identify areas where there is a sharp change in the number of edge pixels included in the designated area in response to a change in the total number of pixels (area) of the designated area as the edge of the affected area, and can highly evaluate the information of the designated area before and after that area. Furthermore, if there are multiple stages of sharp changes in the number of edge pixels included in the designated area, the management server 1 can also perform an evaluation that takes into account the presence of other organs unrelated to the affected area.
[0019] In other words, both the captured image and the information on the designated modification area associated with that image can serve as training data. However, as mentioned above, the designated modification area is one or more modifications made to the designated area from enlargement, reduction, movement, and deformation. Therefore, even if the accuracy of the physician's marking is high (when the physician makes the marking at the expense of ease of use), the information on the designated modification area may not be suitable as training data for the affected area. On the other hand, even if the accuracy of the physician's marking is low (when the physician makes the marking prioritizing ease of use), the information on the designated modification area may still be suitable as training data for the affected area. Therefore, the management server 1 treats the captured image and the information on the designated modification area associated with that image as candidates for training data, regardless of the accuracy of the physician's marking.
[0020] Furthermore, the management server 1 is enabled to evaluate candidate training data based on differences in the external dimensions of the designated area to be changed. In this case, the management server 1 may prioritize the evaluation of information about the designated area to be changed that is closer in area to the external dimensions of the designated area before the change.
[0021] Furthermore, the management server 1 is capable of evaluating candidate training data based on information about the physician who performed the handwritten operations on the captured image (hereinafter referred to as "physician information"). In this case, the management server 1 can perform an evaluation that takes into account the characteristics of the handwritten operations performed by the physician. These characteristics of handwritten operations include, for example, the position, size, shape, and range of the designated area, and the distance between the affected area and the designated area. The functional configuration and processing flow of the management server 1 will be described later.
[0022] [Physician Terminal 2] Physician Terminal 2 is an information processing device installed in a medical institution and operated by a physician who examines patients. Physician Terminal 2 consists of an information processing device capable of displaying captured images, such as a personal computer or tablet terminal. Physician Terminal 2 is capable of executing a predetermined application program that enables the use of the information processing system S. Physician Terminal 2 is capable of performing various processes based on various information transmitted from the management server 1 and external sources, as well as various information entered by the physician. Physician Terminal 2 is also capable of transmitting various information to the management server 1 and external sources.
[0023] For example, physician terminal 2 acquires images (MRI images, CT images, ultrasound images, etc.) from various examination equipment (not shown) and displays them on a screen in a manner that allows handwritten input to specify the area of the affected area. When a handwritten input is made to specify the area of the affected area on the image displayed on the screen, physician terminal 2 accepts the input. Then, physician terminal 2 transmits information linking the image to the specified area to management server 1.
[0024] The processing performed by the management server 1 and the physician terminal 2, which constitute the information processing system S, is merely one example. In other words, as long as the information processing system S as a whole has the functionality to implement the above-mentioned processing, some or all of the functions to implement the above-mentioned processing may be shared or performed collaboratively within the information processing system S.
[0025] For example, some or all of the functions of the management server 1 may be assigned to other information processing devices within the information processing system S (e.g., physician terminal 2). Alternatively, some or all of the functions of other information processing devices within the information processing system S may be assigned to the management server 1. Furthermore, some or all of the functions of the management server 1 may be transferred to other servers, etc. (not shown). This facilitates processing within the entire information processing system S and allows for complementary processing.
[0026] <Hardware Configuration> [Hardware Configuration of Management Server 1] Figure 2 is a block diagram showing an example of the hardware configuration of the management server 1 that constitutes the information processing system S shown in Figure 1. The management server 1 includes a CPU (Central Processing Unit) 11, a ROM (Read Only Memory) 12, a RAM (Random Access Memory) 13, a bus 14, an input / output interface 15, an output unit 16, an input unit 17, a storage unit 18, a communication unit 19, and a drive 20.
[0027] The CPU 11 executes various processes according to the program recorded in the ROM 12 or the program loaded from the storage unit 18 into the RAM 13. The RAM 13 also stores data necessary for the CPU 11 to execute various processes. The CPU 11, ROM 12, and RAM 13 are interconnected via a bus 14. An input / output interface 15 is also connected to this bus 14.
[0028] The input / output interface 15 is connected to an output unit 16, an input unit 17, a storage unit 18, a communication unit 19, and a drive 20. The output unit 16 consists of a display, speaker, etc., and outputs various information as images, sounds, etc. The input unit 17 consists of a keyboard, mouse, touch panel, etc., and accepts input of various information. The storage unit 18 consists of a hard disk, DRAM (Dynamic Random Access Memory), etc., and stores various data. The communication unit 19 communicates with other devices via the aforementioned network N, which is configured as the Internet, etc.
[0029] The drive 20 is appropriately equipped with removable media 21, such as a magnetic disk, optical disk, magneto-optical disk, or semiconductor memory. Programs read from the removable media 21 by the drive 20 are installed in the storage unit 18 as needed. The removable media 21 can also store various types of data stored in the storage unit 18, just like the storage unit 18.
[0030] [Hardware Configuration of Doctor Terminal 2] Doctor Terminal 2 has the same hardware configuration as Management Server 1 shown in Figure 2. That is, Doctor Terminal 2 has a CPU, ROM, RAM, bus, input / output interface, output unit, input unit, storage unit, communication unit, drive, and removable media, corresponding to the CPU 11, ROM 12, RAM 13, bus 14, input / output interface 15, output unit 16, input unit 17, storage unit 18, communication unit 19, drive 20, and removable media 21 in Figure 2.
[0031] <Functional Configuration of Management Server 1 and Doctor Terminal 2> Figure 3 is a functional block diagram showing an example of the functional configuration of the management server 1 and doctor terminal 2 that constitute the information processing system S of Figure 1. In the CPU 11 of the management server 1, the acquisition unit 31, the data management unit 32, the generation unit 33 as a generation means, the evaluation unit 34 as an evaluation means, the learning unit 35 as a learning means, and the transmission control unit 36 all function during operation.
[0032] Furthermore, the storage unit 18 of the management server 1 is equipped with various databases. For example, there is an image database 41 that stores and stores images by associating them with designated areas or areas to be modified; a training data database 42 that stores and stores images used as training data and information on the areas to be modified associated with those images; and a physician database 43 that stores and stores physician information.
[0033] The acquisition unit 31 acquires various types of information. For example, the acquisition unit 31 acquires various types of information transmitted from the physician terminal 2 to the management server 1. Examples of information transmitted to the management server 1 include image capture data associated with a specified region and physician information.
[0034] The data management unit 32 stores and manages various types of information in the database of the storage unit 18. For example, the data management unit 32 stores and manages captured images acquired by the acquisition unit 31, which are associated with a designated area or a modified designated area, in the captured image DB 41. The data management unit 32 also stores and manages captured images adopted as training data by the learning unit 35 (described later), and information on the modified designated area associated with those captured images, in the training data DB 42. Furthermore, the data management unit 32 stores and manages physician information acquired by the acquisition unit 31 in the physician DB 43.
[0035] The generation unit 33 generates information on the modified specified area by changing a specified area associated with the captured image. For example, the generation unit 33 generates information on the modified specified area by performing one or more changes from enlargement, reduction, movement, and transformation on the specified area associated with the captured image. The information on the modified specified area generated by the generation unit 33, and the captured image associated with the information on the modified specified area, are treated as candidates for training data and are also used as conventional augmented data.
[0036] The evaluation unit 34 evaluates the captured image and the information of the designated change area associated with that image. For example, the evaluation unit 34 evaluates the information of the designated change area based on the number of edge pixels included in the designated change area and the total number of pixels (area) of the designated change area, which are edge pixels detected and mapped by edge detection processing on the entire captured image. In this case, the evaluation unit 34 can evaluate the information of the designated change area based on the change in the number of edge pixels included in the designated change area, with the total number of pixels (area) of the designated change area as a parameter. Furthermore, the evaluation unit 34 can evaluate the information of the designated change area based on the difference in the external size of the designated change area. For example, the evaluation unit 34 can give a higher evaluation to the information of the designated change area that is closer to the external size of the designated area before the change.
[0037] Furthermore, for example, the evaluation unit 34 evaluates the captured image and the information of the modification designation area associated with the captured image based on physician information relating to the physician who performed the handwritten operation on the captured image. In this case, the evaluation unit 34 evaluates the captured image and the information of the modification designation area associated with the captured image based on the characteristics of the handwritten operation performed by the physician.
[0038] The learning unit 35 learns information about the generated modified designated area and the captured image containing that modified designated area. For example, the learning unit 35 learns information about a modified designated area that has been modified by enlarging or shrinking a designated area associated with an captured image, and the captured image containing that modified designated area. Alternatively, for example, the learning unit 35 learns information about m modified designated areas (m being an integer between 1 and n) from the information about n modified designated areas (n being an integer of 1 or more) corresponding to one designated area associated with one captured image. In this case, the learning unit 35 can adopt information about m modified designated areas based on the evaluation results of the modified designated area information by the evaluation unit 34.
[0039] The transmission control unit 36 controls the communication unit 19 (see Figure 2) to transmit various types of information to the physician terminal 2 and to external sources.
[0040] Furthermore, in the CPU of the physician terminal 2, the acquisition unit 51, the display control unit 52, the operation reception unit 53 as a means for receiving handwritten operations, and the transmission control unit 54 function during operation. The acquisition unit 51 acquires various types of information. For example, the acquisition unit 51 acquires captured images transmitted to the physician terminal 2 from various examination equipment (not shown). The display control unit 52 controls the display of various types of information on the output unit's display. For example, the display control unit 52 controls the display of captured images acquired by the acquisition unit 51. The operation reception unit 53 accepts an operation on the captured image to specify an area that includes at least a part of the area of the affected area. The transmission control unit 54 controls the transmission of various types of information to the management server 1. For example, the transmission control unit 54 controls the transmission of captured images associated with the specified area to the management server 1.
[0041] <Processing flow of Management Server 1> Figure 4 is a flowchart showing an example of the processing flow of Management Server 1. When Management Server 1 receives an image with a specified region associated with it from Doctor Terminal 2 (YES in step S1), it acquires the image (step S2). The acquired image is stored in the image database 41. On the other hand, if no image with a specified region associated with it has been received (NO in step S1), Management Server 1 repeats the decision process in step S1.
[0042] The management server 1 generates information on the modified specified region by changing the specified region associated with the captured image acquired in step S2 (step S3). Then, the management server 1 evaluates the captured image and the information on the modified specified region associated with that captured image (step S4), and selects information on the m modified specified region from the information on the n modified specified regions corresponding to the 1 specified region associated with the 1 captured image, and performs learning (step S5). With this, the processing of the management server 1 is completed (END).
[0043] <Specific Example> [Specific Example of an Imaging Image with a Specified Region Linked]FIG. 5 is a diagram showing a specific example of an imaging image with a specified region linked. In FIG. 5, an imaging image 100, which is a CT image of a part of a patient's body cut out using a predetermined imaging method, is shown. In the imaging image 100, a rectangular specified region 200 designated by a doctor's operation as the affected area region is superimposed and displayed. The position, size, shape, and range of the specified region 200 in the imaging image 100 are linked to the imaging image 100 and stored and managed. Note that the shape of the specified region shown in FIG. 5 is not limited to a rectangle and may be a circle, an ellipse, or the like.
[0044] FIGS. 6A to 6D are diagrams showing specific examples of imaging images with a modified specified region linked. The management server 1 generates information on the modified specified region by performing one or more of enlargement, reduction, movement, and deformation on the specified region 200 linked to the imaging image 100 of FIG. 5 described above. In FIGS. 6A to 6D, each of the imaging images 501 to 504 with information on each of the modified specified regions 201 to 204 generated by performing a modification to enlarge or reduce the specified region 200 of FIG. 5 is shown. The imaging images 501 to 504 shown in FIGS. 6A to 6D do not perform processing on the imaging image 100 itself of FIG. 5, but perform a process of enlarging or reducing the outer size of the specified region 200 linked to the imaging image 100 of FIG. 5.
[0045] Therefore, the imaging images 501 to 504 shown in FIGS. 6A to 6D are data common to the imaging image 100 of FIG. 5, and only the outer sizes of the specified region and the modified specified region are different. Specifically, for example, the imaging image 501 shown in FIG. 6A is obtained by enlarging only the outer size of the specified region 200 of the imaging image 100 of FIG. 5 by 1.1 times. Also, the imaging image 502 shown in FIG. 6B is obtained by enlarging only the outer size of the specified region 200 of the imaging image 100 of FIG. 5 by 1.4 times. Also, the imaging image 503 shown in FIG. 6C is obtained by reducing only the outer size of the specified region 200 of the imaging image 100 of FIG. 5 by 0.9 times. Also, the imaging image 504 shown in FIG. 6D is obtained by reducing only the outer size of the specified region 200 of the imaging image 100 of FIG. 5 by 0.6 times.
[0046] The management server 1 evaluates the captured images 501 to 504 shown in Figures 6A to 6D and adopts the captured images that meet predetermined criteria and the designated modification areas associated with those captured images as training data. In the example of Figures 6A to 6D, the evaluation of the captured images 501 to 504 is performed based on the ratio of the number of edge pixels included in the designated modification area to the total number of pixels (area) of the designated modification area. Such an evaluation may be output as a comparable numerical value (score), or it may not be output exclusively as an internal process of the management server 1. Here, the dashed area 300 in Figures 6A to 6D indicates the edge of the affected area estimated by the detected group of edge pixels, and has a common position, size, and range in the captured images 501 to 504. Also, although the shape of the dashed area 300 shown in Figures 6A to 6D is drawn as a circle for simplification, it may be a complex shape that indicates the estimated edge of the affected area.
[0047] Here, for example, suppose that each of the captured images 501 and 503 is evaluated as meeting a predetermined criterion (receiving a high score) because the ratio of the number of edge pixels included in the designated modification area to the total number of pixels in each of the designated modification area 201 and 203 is high. In this case, captured images 501 and 503 are adopted as training data. On the other hand, suppose that each of the captured images 502 and 504 is evaluated as not meeting a predetermined criterion (receiving a low score) because the ratio of the number of edge pixels included in the designated modification area to the total number of pixels in each of the designated modification area 202 and 204 is low. In this case, captured images 502 and 504 are not adopted as training data.
[0048] Thus, in the examples of FIGS. 6A to 6D where there are four captured images (captured images 501 to 504) associated with the information of the changed designated areas, two of the four changed designated areas 201 to 204 obtained by changing the designated area 200 (see FIG. 5) of one captured image 100, namely the changed designated areas 201 and 203, and the captured images 501 and 503 associated with each of the changed designated areas 201 and 203 will be adopted as teacher data. Note that by finely setting the variation of the magnification when changing the outer size of the designated area, it is also possible to adopt more captured images and the information of the changed designated areas associated with those captured images as teacher data.
[0049] FIG. 7 is a diagram showing a specific example of a graph used when evaluating the information of the changed designated area. The graph of FIG. 7 shows the total number of pixels (area) of the changed designated area, which is a parameter, on the horizontal axis, and shows the number of edge pixels included in the changed designated area on the vertical axis. Looking at the shape of the curve L, in the areas A1 and A2 indicated by the broken line, a sharp change in the number of edge pixels included in the changed designated area can be seen. Among these, the area A1 with a smaller total number of pixels of the changed designated area is regarded as the edge of the affected part. Therefore, the information of the changed designated areas before and after the area A1 is highly evaluated.
[0050] On the other hand, the area A2 with a larger total number of pixels of the changed designated area is regarded as noise rather than the edge of the affected part. This noise is often, for example, the edge of another organ that has nothing to do with the affected part.
[0051] <Advantageous Effects of the Present Embodiment> According to the above-described embodiment, since the changed designated area, which is the changed designated area, is learned, even if the labor of handwritten marking on the captured image is reduced, learning can be performed using suitable samples without depending on the accuracy of the handwritten marking. As a result, the construction of a machine learning model for specifying the position of the affected part based on the input captured image is facilitated, and thus the accuracy of estimating the position of the affected part included in the captured image input to the constructed machine learning model can be improved.
[0052] <Other> Although one embodiment of the present invention has been described above, the present invention is not limited to the above-described embodiment, and any modifications, improvements, etc. that can achieve the objectives of the present invention are included in the present invention. For example, the specific examples shown in Figures 5 to 7 are merely examples for achieving the objectives of the present invention and are not particularly limited. For example, Figures 5 and 6 show CT images as imaging images, but the present invention is not limited to these, and MRI images, ultrasound images, etc., may also be used.
[0053] Furthermore, the series of processes described above can be executed by hardware or by software. In other words, the functional configuration described above is merely illustrative and not particularly limiting. That is, it is sufficient for the information processing system to have the functionality to execute the series of processes described above as a whole, and the functional blocks used to realize this functionality are not particularly limited to the examples given above.
[0054] Furthermore, the location of the functional blocks is not particularly limited and can be arbitrary. For example, the functional blocks of management server 1 may be transferred to other devices, or the functional blocks of other devices may be transferred to servers. Also, a single functional block may consist of hardware alone, software alone, or a combination of both.
[0055] When a series of processes are executed by software, the programs that make up that software are installed on a computer or other device from a network or storage medium. The computer may be a computer built into dedicated hardware. Alternatively, the computer may be a computer capable of performing various functions by installing various programs, such as a server, a general-purpose smartphone, or a personal computer.
[0056] Such recording media containing programs may consist not only of removable media (not shown) distributed separately from the main device to provide programs to users, but also of recording media provided to users in a state where they are pre-installed in the main device. Since programs can be distributed via a network, the recording media may be installed on or accessible from a computer connected to or capable of connecting to a network.
[0057] In this specification, the step of describing a program to be recorded on a recording medium includes not only processes that are performed chronologically in that order, but also processes that are not necessarily performed chronologically, but are executed in parallel or individually. Furthermore, in this specification, the term "system" refers to an overall system composed of multiple devices, means, etc.
[0058] In other words, the information processing system to which the present invention is applied can take various forms having the following configurations: (1) That is, the information processing system S to which the present invention is applied is an information processing system having: a handwritten operation receiving means (for example, an operation receiving unit 53 in Figure 3) that receives an operation (for example, a handwritten operation by a doctor) to specify an area of a diseased part included in an image of a part of the human body (for example, an MRI image, a CT image, an ultrasound image, etc.) taken using a predetermined imaging method; a generation means (for example, a generation unit 33 in Figure 3) that generates information of a modified specified area by changing the specified area specified by the received operation; and a learning means (for example, a learning unit 35 in Figure 3) that learns the generated information of the modified specified area.
[0059] As a result, the modified designated region is learned, reducing the effort required to mark the captured images, and enabling learning using suitable samples without depending on the accuracy of the marking. Consequently, it becomes easier to build a machine learning model for identifying the location of the affected area based on the input captured images, and this machine learning model can then accurately estimate the location of the affected area based on the input captured images.
[0060] (2) The learning means can also learn information about the modified designated area by enlarging or shrinking the designated area. This makes it possible to learn the modified designated area generated by a simple process that involves enlarging, shrinking, moving, or transforming the designated area, without processing the captured image itself.
[0061] (3) The learning means can also perform the learning by selecting information of m (m is an integer between 1 and n) of the specified change region from the information of n (n is an integer of 1 or more) of the specified change region that corresponds to the 1 specified region associated with the 1 captured image. This makes it possible to perform learning using more suitable samples without depending on the accuracy of the marking, even if the effort of marking the captured image is reduced.
[0062] (4) The system further includes an evaluation means (for example, an evaluation unit 34 in Figure 3) for evaluating the information of the designated change area, wherein the evaluation means evaluates the information of the designated change area based on the number of edge pixels included in the designated change area among the edge pixels detected and mapped by edge detection processing on the entire captured image, and the learning means can adopt the information of the designated change area of m based on the result of the evaluation of the information of the designated change area. This makes it possible to construct a machine learning model based on high-quality training data that can contribute to the efficient use of machine learning resources.
[0063] (5) Furthermore, the evaluation means can evaluate the information of the modified area based on the change in the number of edge pixels included in the modified area, with the area of the modified area as a parameter. This makes it possible to construct a machine learning model based on high-quality training data that can contribute to the efficiency of machine learning resources.
[0064] (6) Furthermore, the evaluation means can evaluate the information of the designated area for modification based on differences in the external size of the designated area for modification. This makes it possible to perform a first evaluation that takes into account the presence of other organs unrelated to the affected area, for example.
[0065] (7) Furthermore, the evaluation means can evaluate the information of the modified designated area more highly if it is closer to the external dimensions of the designated area. This makes it possible to construct a machine learning model based on high-quality training data that can contribute to the efficiency of machine learning resources.
[0066] (8) The evaluation means can also evaluate the information of the designated change area based on information about the person who performed the operation. This makes it possible to construct a machine learning model based on high-quality training data that takes into account the characteristics of the person performing the operation.
[0067] (9) The evaluation means can also evaluate the information of the designated area to be changed based on the characteristics of the operation performed by the person performing the operation. This makes it possible to construct a machine learning model based on high-quality training data that takes into account differences in the characteristics of the person performing the operation, such as the position, size, shape, and extent of the designated area, and the distance between the affected area and the designated area.
[0068] (10) Furthermore, the information processing method to which the present invention is applied can take various forms having the following configurations. That is, the information processing method to which the present invention is applied is an information processing method that includes the steps of: receiving an operation to specify a region of a diseased area included in an image of a part of the human body captured using a predetermined imaging technique; generating information of a modified specified region by changing the specified region specified by the received operation; and learning the generated information of the modified specified region.
[0069] (11) Furthermore, the program to which the present invention is applied can take various forms having the following configuration. That is, the program to which the present invention is applied is a program that causes an information processing system (for example, the information processing system S in Figure 1) to execute a control process that includes: receiving an operation to specify an area of a diseased part included in an image of a part of the human body taken using a predetermined imaging method; generating information of a modified specified area by changing the specified area specified by the received operation; and learning the generated information of the modified specified area.
[0070] 1: Management server, 2: Doctor terminal, 11: CPU, 16: Output unit, 17: Input unit, 18: Storage unit, 19: Communication unit, 31: Acquisition unit, 32: Data management unit, 33: Generation unit, 34: Evaluation unit, 35: Learning unit, 36: Transmission control unit, 41: Image capture DB, 42: Training data DB, 43: Doctor DB, 51: Acquisition unit, 52: Display control unit, 53: Operation reception unit, 54: Transmission control unit, S: Information processing system, N: Network
Claims
1. An information processing system comprising: an operation receiving means for receiving an operation to specify a region of a diseased area included in an image of a part of the human body captured using a predetermined imaging method; a generation means for generating information of a modified specified region obtained by changing the specified region specified by the received operation; and a learning means for learning the generated information of the modified specified region.
2. The information processing system according to claim 1, wherein the learning means learns information of the modified designated area that has been modified by expanding or shrinking the designated area.
3. The information processing system according to claim 2, wherein the learning means performs the learning by selecting information of m (m is an integer value between 1 and n) of the specified change region from information of n (n is an integer value of 1 or more) of the specified change region that corresponds to 1 of the captured image.
4. An information processing system according to claim 3, further comprising an evaluation means for evaluating information of the designated change region, wherein the evaluation means evaluates information of the designated change region based on the number of edge pixels included in the designated change region among the edge pixels which are pixels detected and mapped by edge detection processing on the entire captured image, and the learning means adopts the information of the designated change region m based on the result of the evaluation of information of the designated change region.
5. The information processing system according to claim 4, wherein the evaluation means evaluates information about the designated change area based on the change in the number of edge pixels included in the designated change area, with the area of the designated change area as a parameter.
6. The information processing system according to claim 4, wherein the evaluation means further evaluates the information of the designated change area based on differences in the external dimensions of the designated change area.
7. The information processing system according to claim 6, wherein the evaluation means evaluates the information of the modified designated area more highly, the information of which is closer to the external dimensions of the designated area.
8. The information processing system according to claim 6, wherein the evaluation means further evaluates the information of the designated change area based on information about the person who performed the operation.
9. The information processing system according to claim 8, wherein the evaluation means evaluates the information in the designated change area based on the characteristics of the operation performed by the person who performed the operation.
10. An information processing method comprising: receiving an operation to specify a region of a diseased area included in an image of a part of a human body captured using a predetermined imaging technique; generating information of a modified specified region obtained by changing the specified region specified by the received operation; and learning the generated information of the modified specified region.
11. A program that causes an information processing system to execute a control process including: receiving an operation to specify a region of a diseased area included in an image of a part of the human body captured using a predetermined imaging method; generating information of a modified specified region obtained by changing the specified region specified by the received operation; and learning the generated information of the modified specified region.