Information processing device, control method, and program

Abstract

This technology aims to reduce the likelihood of misinterpreting the temporal changes in injuries and illnesses due to differences in the detectors used for testing. [Solution] The information processing device detects the state of the object to be inspected from an image of the object to be inspected. The information processing device includes an acquisition means for acquiring information of a first detector used for detecting the state of the object to be inspected and information of a second detector used when state detection was performed on the object to be inspected in the past, and a control means for controlling the device to use the second detector to detect the state of the object to be inspected if the object to be inspected is subject to ongoing observation and the first detector and the second detector are different.

Description

Technical Field

[0004] , , , , , , , ,

[0006] , , , ,

[0005] , , , , , , , ,

[0007] , , ,

[0001] The present invention relates to a technique for detecting a specific injury state from an image of an inspection target using Artificial Intelligence (AI / artificial intelligence).

Background Art

[0002] In the medical field, a technique for detecting a specific injury state from an image of an inspection target by inference processing using an AI model (learning model) as a detector is known.

[0003] Patent Document 1 describes a method for searching for case data with a similar disease progression from a comparison between an inspection image and a medical image included in the case data. Patent Document 2 describes a method for detecting an injury from an image captured under imaging conditions according to the inspection purpose.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Patent Document 2

Summary of the Invention

Problems to be Solved by the Invention

[0005] The learning model can improve detection performance by being trained with learning data and can detect new injury states. However, due to changes in detection performance for the same injury due to training, it may become impossible to correctly observe the temporal changes of the injury.

[0006] Patent Document 1 is not suitable for observing the temporal changes of an injury, and in Patent Document 2, the detector may change from the previous inspection, and it may become impossible to correctly observe the temporal changes of the injury.

[0007] This invention has been made in view of the above problems, and its purpose is to realize a technology that reduces the inability to correctly observe the changes in injuries and illnesses over time due to different detectors used in the examination. [Means for solving the problem]

[0008] To solve the above problems and achieve the objective, the present invention provides an information processing device for detecting the state of an object to be inspected from an image of the object to be inspected, comprising: acquisition means for acquiring information of a first detector used for detecting the state of the object to be inspected and information of a second detector used when state detection was performed on the object to be inspected in the past; and control means for controlling the device to use the second detector to detect the state of the object to be inspected if the object to be inspected is subject to ongoing observation and the first detector and the second detector are different. [Effects of the Invention]

[0009] According to the present invention, it is possible to reduce the inability to correctly observe the changes in injuries and illnesses over time due to differences in the detectors used for the examination. [Brief explanation of the drawing]

[0010] [Figure 1] Hardware configuration diagram of the information processing device according to this embodiment. [Figure 2] A flowchart illustrating the control process of the information processing device according to Embodiment 1. [Figure 3] A diagram illustrating an example of a database subject to follow-up observation according to Embodiment 1. [Figure 4] A flowchart illustrating the control process of the information processing device according to Embodiment 2. [Figure 5] A diagram illustrating inspection information related to Embodiment 2. [Figure 6] A flowchart illustrating the control process of the information processing device according to Embodiment 3. [Modes for carrying out the invention]

[0011] The embodiments will be described in detail below with reference to the attached drawings. Note that the following embodiments do not limit the invention as defined in the claims. While the embodiments describe multiple features, not all of these features are essential to the invention, and the features may be combined in any way. Furthermore, in the attached drawings, identical or similar configurations are given the same reference numerals, and redundant descriptions are omitted.

[0012] [Definition of Terms] In this embodiment, the patients managed by the medical system have patient information such as name, age, gender, and medical history registered in the system's database, and each patient is pre-assigned a unique patient management number. The patient information can then be retrieved using the patient management number as the key.

[0013] Test information refers to information related to a test. Each test is uniquely identified by a pre-assigned test management number and registered in association with patient information. Test information includes the purpose of the test, detection results, test site, test location, and detector identification information. Past test information can be retrieved using the patient management number as the key.

[0014] The purpose of the examination is to identify the illness or injury being examined, such as atopic dermatitis, pressure ulcers (bedsores), injuries, gingivitis, and freckles. Any suspected illness or injury is acceptable, and it may differ from the final diagnosis.

[0015] The detection results include one or more test subjects for each injury or illness being tested, and status detection information for each test subject. The test subject is the state of the injury or illness being tested. The status detection information is detailed information about the state of the test subject as detected by the detector.

[0016] The areas examined include the head, chest, breasts, abdomen, waist, buttocks, back, shoulders, elbows, knees, hands, and legs.

[0017] The inspection locations are places such as examination rooms, laboratories, bathrooms, and indoor areas of inpatient wards within hospital facilities.

[0018] The detector used in the inspection of this embodiment is prepared by an AI model (learning model) for each inspection purpose or inspection target of a specific injury or illness, and detects the state of the injury or illness that is the inspection purpose from the inspection target image (hereinafter referred to as the inspection image) of the inspection target and additional information such as the shooting parameters of the inspection image. The inspection image is an image obtained by shooting the inspection target of the injury or illness that is the inspection purpose of the patient. The inspection target is the symptom of the injury or illness that is the inspection purpose. When the injury or illness that is the inspection purpose is pressure ulcer, it includes skin damage, granulation tissue, necrotic tissue, etc. The state of the injury or illness is the location, area, swelling degree, unevenness information, etc. of the injury or illness in the patient's body.

[0019] The method for detecting injuries or illnesses in this embodiment includes not only the inference process by machine learning using a detector, but also a rule-based detection method using luminance, color information, or edges, and combinations thereof may also be used. The inference process by machine learning includes not only the inference process using one model, but also the inference process using multiple models. Furthermore, when contradictions occur in the inference results of multiple models, it also includes a method of making the inference results consistent by using the detection results based on rules.

[0020] The detector can be uniquely identified based on the identification information unique to the detector. The identification information of the detector only needs to be able to identify the detector that can obtain the same detection result for the same inspection image. If the detector can be identified by information such as the inspection date and time and the inspection facility information, those information can also be used. When the state of the injury or illness that is the inspection purpose is detected by the detector from the inspection image and the additional information, the detection result and the identification information of the used detector are associated with the inspection information and the inspection image and stored.

[0021] In this embodiment, the recommended detector (hereinafter referred to as the recommended detector) is the detector recommended by the service provider as being most suitable for examining the injury or illness for the purpose of examination. The detector's detection performance improves by training a learning model with training data, enabling it to detect new injury or illness conditions. In other words, the learning model of the recommended detector is a learning model generated by training the learning model of the detector used in the previous examination (previous detector) with training data. Therefore, the detector using the latest learning model is likely to become the recommended detector. On the other hand, machine learning has the problem of overfitting due to bias in the training data, and the latest detector does not necessarily produce the desired results. For this reason, in this embodiment, the latest detector at the time of examination is designated as the recommended detector.

[0022] The learning model is composed of a neural network, and in this embodiment, it is composed of a Convolutional Neural Network (CNN). However, the learning model in this embodiment is not limited to a CNN and may be composed of other neural networks such as Recurrent Neural Networks (RNNs) or fully connected networks. The computation method of the neural network is well known, so its explanation is omitted. It is composed of a neural network, etc.

[0023] Furthermore, inference processing can be performed by an image processing processor such as a GPU (Graphics Processing Unit). A GPU is a processor capable of performing a large number of multiply-accumulate operations and has the computational processing power to perform matrix operations and other calculations for the neural network that constitutes the learning model in a short amount of time. In addition, the inference processing may be performed collaboratively by the processor (CPU) of the control unit 101 of the information processing device 100 (described later in Figure 1) and the GPU, or the calculation may be performed by either the processor (CPU) of the control unit 101 or the GPU.

[0024] <Device configuration> In this embodiment, a computer device operates as an information processing device 100. The processing of the information processing device in this embodiment may be implemented by a single computer device, or the functions may be distributed among multiple computer devices as needed. The multiple computer devices are connected to each other in a manner that allows them to communicate with one another.

[0025] Furthermore, computer devices can take various forms, including personal computers (desktop PCs, notebook PCs), tablet PCs, smartphones, and cloud computing.

[0026] The information processing device 100 includes a control unit 101, a non-volatile memory 102, a work memory 103, a storage device 104, an input device 105, an output device 106, a communication interface (I / F) 107, and a system bus 108.

[0027] The control unit 101 includes a CPU, MPU, or other arithmetic processing processor that comprehensively controls the entire information processing device 100. The non-volatile memory 102 is a ROM that stores programs and parameters executed by the processor of the control unit 101. Here, the program refers to a program for executing control processing, which will be described later in Figures 2, 4, and 5. The work memory 103 is a RAM that temporarily stores programs read from the non-volatile memory 102, as well as constants and variables for executing the program.

[0028] The storage device 104 is an internal device such as a solid-state drive (SSD) or hard disk drive (HDD) built into the information processing device 100, or an external device such as an SSD or HDD that is detachably connected to the information processing device 100. The storage device 104 stores applications that realize the control processing described later in Figures 2, 4, and 5, as well as detectors and inference parameters used in the control processing, examination images, and associated information. The detectors and database may be managed by an external electronic medical record system or the like and configured to be obtainable via the communication interface 107.

[0029] The input device 105 is an operating component such as a mouse, keyboard, or touch panel that accepts user input and outputs operation instructions to the control unit 101. The output device 106 is a display device such as a display or monitor made of an LCD or organic EL, which displays data such as the GUI (Graphical User Interface) of an application, inspection images, inspection information, and detection results.

[0030] The communication interface 107 connects to networks such as the Internet and a LAN (Local Area Network) for communication. The system bus 108 includes an address bus, a data bus, and a control bus that connect the components 101 to 107 of the information processing device 100 for data exchange.

[0031] The non-volatile memory 102 stores the operating system (OS), which is the basic software executed by the control unit 101, and control programs including applications that work with the OS to realize advanced functions. The non-volatile memory 102 also stores applications and detectors used in the control processing described later in Figures 2, 4, and 5, as well as the monitoring target database described later in Figure 3. Note that the detectors and database may be obtained from an external electronic medical record system via the communication interface 107.

[0032] The functions of the information processing device 100 in this embodiment are realized by software provided by an application. The application is assumed to have software for utilizing the basic functions of the OS installed on the information processing device 100. The OS of the information processing device 100 may also have software for realizing the processing in this embodiment.

[0033] <Control Processing> Next, with reference to Figure 2, the control processing performed by the information processing device 100 according to this embodiment will be described.

[0034] Figure 2 is a flowchart showing the process of detecting the state of the object being inspected from the inspection image according to this embodiment.

[0035] The process shown in Figure 2 is achieved by the control unit 101 executing an application stored in the non-volatile memory 102 or the storage device 104. In the process shown in Figure 2, the current examination image, patient information, and past examination information are associated and stored in the storage device 104. The same applies to Figures 4 and 6, which will be described later.

[0036] In step S201, the control unit 101 loads the examination image, patient information, and past examination information associated with the patient information from the storage device 104 into the work memory 103.

[0037] In step S202, the control unit 101 searches for past examination information for the same examination purpose and the same body part using the patient management number of the patient information from step S201 as a key, loads the matching examination information for the same examination purpose into the work memory 103, and determines whether or not there are detection results using a detector. If there are detection results using a detector, the control unit 101 proceeds to step S203, and if there are no detection results using a detector, it proceeds to step S208.

[0038] In step S203, the control unit 101 obtains the observation target ID and detector (previous detector) identification information from the latest inspection information (previous inspection information) among the past inspection information for the same inspection purpose obtained in step S202. The observation target ID is an identification number assigned to each inspection target for the inspection purpose, as will be described later in Figure 3.

[0039] In step S204, the control unit 101 determines whether the examination purpose is for an injury or illness that is subject to follow-up observation, based on the observation target ID of the previous examination information obtained in step S203. The determination of whether an injury or illness is subject to follow-up observation is made by checking whether there is an observation target ID that matches the examination purpose obtained in step S202 in the observation target DB, which will be described later in Figure 3. If no observation target ID is assigned, it is determined that the injury or illness is not subject to observation. If the injury or illness is subject to follow-up observation, the control unit 101 proceeds to step S205, and if it is not subject to follow-up observation, it proceeds to step S208.

[0040] In step S205, the control unit 101 obtains the identification information of the recommended detector corresponding to the observation target ID obtained in step S204 from the observation target DB in Figure 3, and compares it with the identification information of the previous detector obtained in step S202. If the identification information of the recommended detector and the identification information of the previous detector are the same, the control unit 101 determines that the recommended detector and the previous detector are a match and proceeds to step S208. If the identification information of the recommended detector and the identification information of the previous detector are different, the control unit 101 determines that the recommended detector and the previous detector are not a match and proceeds to step S206.

[0041] In step S206, the control unit 101 determines whether a predetermined period has elapsed since the last inspection. If the control unit 101 determines that a predetermined period has elapsed since the last inspection, it proceeds to step S208; otherwise, it proceeds to step S207. For example, the predetermined period may be six months.

[0042] In step S207, the control unit 101 performs state detection on the inspection image using the detector corresponding to the identification information of the previous detector acquired in step S203, and stores the detection result in the work memory 103 in association with the observation target ID.

[0043] In step S208, the control unit 101 performs state detection on the inspection image using the detector corresponding to the identification information of the recommended detector acquired in step S205, and stores the detection result in the work memory 103 in association with the observation target ID.

[0044] In step S209, the control unit 101 determines whether a detector and detection result to be used for the next inspection have been selected. In this embodiment, for example, the detection result from step S207 and the detection result from step S208 are selectively displayed on the output device 106, and the control unit 101 determines the detection result selected by the user. If the control unit 101 determines that a detector and detection result to be used for the next inspection have been selected, it proceeds to step S210; otherwise, it terminates the process.

[0045] In step S210, the control unit 101 associates the identification information of the detector selected in step S209, the detection result from the selected detector, the inspection information including the observation target ID, and the inspection image, saves them to the storage device 104, and terminates the process.

[0046] <Databases under observation> Figure 3 illustrates a database to be monitored according to this embodiment.

[0047] The database in Figure 3 is referenced by the control unit 101 when determining the injury or illness to be monitored in step S204 of Figure 2.

[0048] Column 301 is the observation target ID column, where an identification number assigned to each test subject is registered.

[0049] Column 302 is the "Examination Purpose" column, where the examination purposes for patients under observation are registered.

[0050] Column 303 is the column to be inspected, and it stores the status of the inspected object as detected by the detector.

[0051] Column 304 is a state detection information column, where detailed information about the state of the object being inspected, as detected by the detector, is registered.

[0052] Column 305 is the identification information column for recommended detectors, and it contains information to uniquely identify the recommended detector according to the disease or injury being examined and the condition of the person being examined.

[0053] Records 306-323 contain the inspection purpose, inspection target, state detection information, and recommended detector identification information for each observation target ID.

[0054] <Control procedures when the disease being examined is atopic dermatitis> Next, with reference to Figures 2 and 3, we will describe an example of control processing for examination images of a patient who visits the clinic monthly for treatment of atopic dermatitis.

[0055] In step S201, the control unit 101 loads the examination image, patient information, and examination information from the storage device 104 into the work memory 103. The examination information is assumed to have been specified by the physician using the GUI displayed on the output device 106, with the disease being examined being atopic dermatitis, the subject being erythema, and the examination site being the inside of the right elbow.

[0056] In step S202, the control unit 101 uses the patient management number of the patient information from step S201 as a key to search for past examination information for the same illness and the same body part, and determines whether or not there are any detection results using the detector from the past examination information. In this embodiment, since the patient visits the hospital every month, let's assume that a detection result from about one month ago is found.

[0057] In step S203, the control unit 101 obtains the observation target ID and detector identification information from the most recent inspection information among the past inspection information for the same inspection purpose obtained in step S202. In this embodiment, the observation target ID is 13, and the identification information of the previous detector is AD_0200.

[0058] In step S204, the control unit 101 determines whether the purpose of the examination is for an injury or illness that is subject to follow-up observation, based on the observation target ID of the most recent past examination information obtained in step S202. The control unit 101 loads the follow-up observation target DB shown in Figure 3 from the storage device 104 into the work memory 103 and searches for the corresponding observation target ID. If there are multiple observation target IDs in the examination information, the follow-up observation target DB is searched for each observation target ID. In this embodiment, the observation target ID is 13 and exists in the follow-up observation target DB, so it is determined that it is subject to follow-up observation, and the process proceeds to step S205.

[0059] In step S205, the control unit 101 determines whether the recommended detector and the previous detector obtained in step S202 match. If the purpose of the test is atopic dermatitis, the identification information 305 of the recommended detector is AD_0300, while the identification information of the previous detector is AD_0200. Since the recommended detector and the previous detector are different, the process proceeds to step S206.

[0060] In step S206, the control unit 101 determines whether a predetermined period has elapsed since the last inspection. For example, if the predetermined period is six months and the last inspection was approximately one month ago, it is determined that the predetermined period has not elapsed, and the process proceeds to step S207.

[0061] In step S207, the control unit 101 detects the state of injury or illness from the detector corresponding to the detector identification information AD_0200 acquired in step S203, as well as from the inspection image and associated information.

[0062] In step S209, the control unit 101 determines whether a detector to be used for the next inspection has been selected. In this embodiment, assuming that the detection result from step 207 has been selected, the process proceeds to step S210.

[0063] In step S210, the control unit 101 associates the identification information of the detector selected in step S209, the detection result from the selected detector, the inspection information including the observation target IDs 13 and 14, and the inspection image, and stores them in the storage device 104, then terminates the process.

[0064] <Explanation of effects> As explained above, according to this embodiment, if an examination image is specified, the purpose of the examination is for an injury or illness to be monitored over time, and the previous detector is different from the recommended detector, the previous detector is used for detection. This reduces the possibility of being unable to correctly observe the changes in the injury or illness over time due to changes in the detection performance of the detector.

[0065] For example, in the case of a patient with atopic dermatitis who visits the clinic monthly, the area and size of the erythema recorded during the previous examination are detected using the same detector as before. This allows for a comparison with past detection results, making it easier to observe whether the symptoms are improving or worsening. As a result, doctors can accurately observe the changes in the condition over time, and patients can receive appropriate treatment.

[0066] Furthermore, if either the previous detector or the recommended detector is selected as the detector to be used for the next examination, the identification information of the selected detector, the detection results from the selected detector, and the examination information and examination images including the observed subject ID are stored in association. This allows the condition of the injury or illness to be detected using the same detector as the previous examination during the next examination.

[0067] For first-time patients and those who have not visited the clinic for a considerable period and do not require strict follow-up, the recommended detector is used to detect the condition of the patient being examined, thereby reducing the frequency of use of older detectors. This reduces the need for service providers to continue maintaining older detectors when follow-up is unnecessary due to a long gap since the last visit.

[0068] In this embodiment, an example was described in which a single examination image is input to the detector to detect the state of injury or illness. However, a configuration in which multiple examination images are input to the detector to perform inference processing of the state of injury or illness and the inference results are consistent using a rule-based approach is also possible.

[0069] Furthermore, although this embodiment has described a detector for detecting the area of ​​erythema in atopic dermatitis as an example, it is not limited to this. In the case of periodontal disease examination, the oral cavity may be photographed from multiple viewpoints, such as the five-image method or the thirteen-image method, and the state of periodontal disease may be detected from each image using multiple detectors. A comprehensive determination may also be made from the inference results of the images of the occlusal surfaces of the maxilla and mandible, as well as the frontal, left-side, and right-side images.

[0070] [Embodiment 2] Embodiment 2 describes an example in which the state of injury or illness is detected using both the previously used detector and the recommended detector on the examination image, and the detection results of both are compared to determine which detector and detection result should be used for the next examination. Embodiment 2 also describes a case in which there are multiple examination targets for a single examination purpose.

[0071] The configuration of the information processing device and the database to be monitored according to this embodiment are the same as those in Figures 1 and 3 of Embodiment 1.

[0072] Figure 4 is a flowchart showing the process of detecting the state of the object being inspected from the inspection image according to this embodiment.

[0073] In step S400, the control unit 101 loads patient information and examination information associated with the examination image from the storage device 104 into the work memory 103.

[0074] In step S401, the control unit 101 searches for past examination information for the same illness and the same body part, using the patient management number of the patient information acquired in S400 as a key, and determines whether or not there are any detection results using a detector from the past examination information. If there are detection results using a detector, the control unit 101 proceeds to step S402, and if not, proceeds to step S415.

[0075] In step S402, the control unit 101 determines whether the purpose of the inspection information acquired in step S401 is for follow-up observation. If the control unit 101 determines that the purpose of the inspection is for follow-up observation, it proceeds to step S403; if it determines that it is not for follow-up observation, it proceeds to step S415. The determination of whether or not it is for follow-up observation is the same as in step S204 in Figure 2.

[0076] In step S403, the control unit 101 assigns 0 to counter i for each of the inspection targets included in the inspection objective of step S402. Counter i is a number assigned to each inspection target in order to detect the state of all inspection targets included in the inspection objective.

[0077] In step S404, the control unit 101 determines whether the counter i is less than or equal to the number of items to be inspected. If the counter i is less than or equal to the number of items to be inspected, the process proceeds to step S405. If the counter i is greater than the number of items to be inspected, the process proceeds to step S416. The number of items to be inspected is the number of items to be inspected for the same inspection purpose in the database under observation.

[0078] In step S405, the control unit 101 loads the i-th observation target ID of the inspection purpose that is the subject of the follow-up observation target DB into the work memory 103 and determines whether or not it is present in past inspection information. If the i-th observation target ID is present in past inspection information, the control unit 101 proceeds to step S406; otherwise, it proceeds to step S414.

[0079] In step S406, the control unit 101 obtains the identification information 305 of the recommended detector corresponding to the i-th observation target ID from the progress observation target DB in Figure 3, and compares it with the identification information of the previous detector corresponding to the same observation target ID in the past inspection information obtained in step S405. If the identification information of the recommended detector and the identification information of the previous detector are the same, the control unit 101 determines that the recommended detector and the previous detector are a match and proceeds to step S414. If the identification information of the recommended detector and the identification information of the previous detector are different, the control unit 101 determines that the recommended detector and the previous detector are not a match and proceeds to step S407.

[0080] In step S407, the control unit 101 uses the previously detected detector to detect the state of the i-th inspection target on the inspection image.

[0081] In step S408, the control unit 101 uses the recommended detector to detect the state of the i-th object to be inspected on the inspection image.

[0082] In step S409, the control unit 101 determines whether the difference between the detection result in step S407 and the detection result in step S408 is less than a threshold. If the difference is less than the threshold, the process proceeds to step S410; if the difference is greater than or equal to the threshold, the process proceeds to step S411. The threshold is uniquely determined for each object being inspected based on the purpose of the inspection, the area being inspected, etc.

[0083] In step S410, the control unit 101 determines whether to use the recommended detector and the detection results from the recommended detector for the next inspection. If it does, it proceeds to step S412; otherwise, it proceeds to step S411. In this embodiment, for example, the output device 106 displays a message such as "Do you want to use the detection results from the recommended detector?" along with "Yes" and "No" buttons. If the user presses the "Yes" button, it is determined that the detection results from the recommended detector will be used. If the user presses the "No" button, it is determined that the detection results from the previous detector will be used.

[0084] In step S411, the control unit 101 stores the detection result from the previous detector in step S407 in the work memory 103, associating it with the observation target ID.

[0085] In step S412, the control unit 101 stores the detection result from the recommended detector in step S408 or S414 in the work memory 103, associating it with the observation target ID.

[0086] In step S413, the control unit 101 increments the counter i and returns the process to step S404.

[0087] In step S414, the control unit 101 uses the recommended detector to detect the state of the i-th object to be inspected.

[0088] In step S415, the control unit 101 uses the recommended detector to detect the state of all inspection targets, associates the observation target ID with the detection result for each inspection target, and stores it in the work memory 103.

[0089] In step S416, the control unit 101 determines whether a detector and detection result to be used for the next inspection have been selected. In this embodiment, for example, the detection result from step S411 and the detection result from step S412 are displayed on the output device 106 in a selectable manner, and the control unit 101 determines the detection result selected by the user. If the control unit 101 determines that a detector and detection result to be used for the next inspection have been selected, it proceeds to step S417; otherwise, it terminates the process.

[0090] In step S417, the control unit 101 associates the identification information of the detector selected in step S416, the detection result from the selected detector, the inspection information including the observation target ID, and the inspection image, saves them to the storage device 104, and terminates the process.

[0091] <Control procedures when the injury / illness being examined is a pressure ulcer> Figure 5 illustrates the inspection information associated with the inspection image in a past inspection according to this embodiment.

[0092] Figure 5(a) illustrates the examination information associated with the image of the subject examined in the previous examination, and Figure 5(b) illustrates the examination information associated with the image of the subject examined in the current examination. Figure 5(a) is referenced in step S405 of Figure 4 when determining whether or not the subject ID of the subject in question is present in the past examination information. Items common to both Figure 5(a) and Figure 5(b) are given the same number and their explanations are omitted.

[0093] Column 501 is the observation target ID column, where an identification number assigned to each test subject is registered.

[0094] Column 502 is the "Examination Purpose" column, where the examination purposes for patients under observation are registered.

[0095] Column 503 is the column to be inspected, and it stores the status of the inspected object as detected by the detector.

[0096] Column 504 is a state detection information column, where detailed information about the state of the object being inspected, as detected by the detector, is registered.

[0097] Column 505 is a detector identification information column, containing information to uniquely identify the detector corresponding to the detection result associated with the inspection image.

[0098] Records 506-515 contain the inspection purpose, inspection target, state detection information, and detector identification information for each observation target ID.

[0099] Assuming that past examination information is shown in Figure 5(a) and the follow-up target database is shown in Figure 3, the control process when the purpose of the examination is pressure ulcers will be explained with reference to Figure 4. Then, Figure 5(b) shows the examination information in which the detection results from the control process in Figure 4 are associated with the examination images.

[0100] In step S400, the control unit 101 loads patient information and examination information associated with the image to be processed used for the examination from the storage device 104 into the work memory 103. The patient is one who is receiving treatment while saving images of the progress of a pressure ulcer on the buttocks once a week, and the condition of the injury is detected from the examination image, and the detection result is saved in association with the examination image.

[0101] In step S401, the control unit 101 searches for past examination information for the same injury and the same body part, using the patient management number of the patient information acquired in step S400 as the key. In this case, since the patient's pressure ulcer progress is saved as images once a week, the examination information from one week ago will be found.

[0102] In step S402, the control unit 101 determines whether the purpose of the examination information acquired in step S401 is for a patient subject to follow-up observation. The purpose of the examination is pressure ulcers, and it is present in the examination purpose column 302 of the patient subject database, so it is determined to be a patient subject to follow-up observation, and the process proceeds to step S403.

[0103] In step S403, the control unit 101 assigns 1 to the counter i to be inspected.

[0104] <If the recommended detector matches the previous detector (YES in step S406)> In step S404, the control unit 101 determines whether the counter i is less than or equal to the number of subjects to be examined. In the follow-up target DB, the subjects to be examined for pressure ulcers are records 306 to 311 in Figure 3, so the counter i is 6. The counter i is 1, which is less than or equal to 6, so the process proceeds to step S405.

[0105] In step S405, the control unit 101 loads the 0th observation target ID of the target examination purpose in the follow-up target DB into the work memory 103 and determines whether or not it exists in the past examination information. In the follow-up target DB, the 1st observation target ID for the examination purpose of pressure ulcers is 1, and since there is an observation target ID that matches in the observation target column 506 in Figure 5(a), the process proceeds to step 406.

[0106] In step S406, the control unit 101 determines whether the recommended detector corresponding to the first observation target ID1 matches the previous detector corresponding to the observation target ID1 in past inspection information. If it determines that they match, the process proceeds to step S414; otherwise, the process proceeds to step S407.

[0107] The identification information 305 for the recommended detector in the observation target column 306 of the database under observation in Figure 3 is BS_0003, and the detector identification information 505 in the observation target column 506 of Figure 5 is also BS_0003. Therefore, the recommended detector and the previous detector match.

[0108] In step S414, the control unit 101 detects the first object to be inspected with the recommended detector. In this embodiment, the control unit 101 detects the area and size of the pressure ulcer skin injury using the recommended detector BS_0003.

[0109] In step S412, the control unit 101 stores the detection result from the recommended detector in the work memory 103, associating it with the observation target ID. The observation target ID is 1, the inspection purpose is pressure ulcers, the inspection target is skin injury, and the detector identification information is BS_0003. The detected region and area are then stored in the work memory 103. The stored inspection information corresponds to the record in the observation target column 510 in Figure 5(b).

[0110] In step S413, the control unit 101 increments the counter i and returns to step S404. The counter i becomes 2.

[0111] <If the recommended detector and the previous detector do not match (NO in step S406), the difference in detection results is less than the threshold (YES in step S409), and the recommended detector is selected (YES in step S410)> In step S404, the control unit 101 determines whether the counter i is less than the number of items to be inspected. In this embodiment, the counter i is 2, which is less than or equal to 6, so the process proceeds to step S405.

[0112] In step S405, the control unit 101 loads the second observation target ID for the examination purpose in the follow-up target DB into the work memory 103 and determines whether or not it exists in the past examination information. In the follow-up target DB, the second observation target ID for the examination purpose of pressure ulcers is 2, and since there is an observation target ID that matches in the observation target column 507 in Figure 5, the process proceeds to step 406.

[0113] In step S406, the control unit 101 determines whether the recommended detector corresponding to the second observation target ID2 matches the previous detector corresponding to the observation target ID2 in the past inspection information.

[0114] In Figure 3, the identification information 305 for the recommended detector in the observation target column 307 of the database under observation is BS_0005, and in Figure 5, the detector identification information 505 in the observation target column 507 is BS_0003. Therefore, it is determined that the recommended detector and the previous detector do not match, and the process proceeds to step S407.

[0115] In step S407, the control unit 101 uses the previously detected detector BS_0003 to detect the state of the first object to be inspected. The area of ​​damage up to the dermis detected is 16 square centimeters.

[0116] In step S408, the control unit 101 uses the recommended detector BS_0005 to detect the condition of the first object to be inspected. The area of ​​the detected damage extending to the dermis is 15 square centimeters.

[0117] In step S409, the control unit 101 determines whether the difference between the result of step S407 and the result of step S408 is less than a threshold. The threshold is set to 2 square centimeters. The result of step S407 is 16 square centimeters, and the detection result of step S408 is 15 square centimeters, so the difference is 1 square centimeter, which is less than the threshold, and the process proceeds to step S410.

[0118] In step S410, the control unit 101 determines whether to use the recommended detector and the detection result from the recommended detector for the next inspection. In this embodiment, the difference between the detection result of the recommended detector and the detection result of the previous detector is only 1 square centimeter, and the user determines that the difference is acceptable, so the "Yes" button is pressed, and the process proceeds to step S412.

[0119] In step S412, the control unit 101 stores the detection result from the recommended detector in the work memory 103, associating it with the observation target ID. The observation target ID is 2, the inspection purpose is pressure ulcers, the inspection target is damage extending to the dermis, and the detector identification information is BS_0005. The detected region and area are stored in the work memory 103. The stored inspection information corresponds to the record in the observation target column 511 in Figure 5(b).

[0120] In step S413, the control unit 101 increments counter i and returns the process to step S404. Counter i becomes 3.

[0121] <If the recommended detector and the previous detector do not match (NO in step S406), or if the difference in detection results is greater than or equal to the threshold (NO in step S409)> In step S404, the control unit 101 determines whether the counter i is less than the number of items to be inspected. Since the counter i is 3 and is less than or equal to 6, the process proceeds to step S405.

[0122] In step S405, the control unit 101 loads the third observation target ID of the target examination in the follow-up target DB into the work memory 103 and determines whether or not it exists in past examination information. In the follow-up target DB, the third observation target ID for the examination target of pressure ulcers is 3, and since there is an observation target ID that matches in the observation target column 507 in Figure 5(a), the process proceeds to step 406.

[0123] In step S406, the control unit 101 determines whether the recommended detector corresponding to the third observation target ID 3 matches the previous detector corresponding to the observation target ID 3 in the past inspection information.

[0124] In Figure 3, the identification information 305 for the recommended detector in the observation target column 308 of the database under observation is BS_0005, and in Figure 5(a), the detector identification information 505 in the observation target column 508 is BS_0003. Therefore, it is determined that the recommended detector and the previous detector do not match, and the process proceeds to step S407.

[0125] In step S407, the control unit 101 uses the previously detected detector BS_0003 to detect the condition of the first object to be inspected. The area of ​​damage up to the subcutaneous tissue detected is 10 square centimeters.

[0126] In step S408, the control unit 101 uses the recommended detector BS_0005 to detect the condition of the first object to be inspected. The area of ​​damage up to the subcutaneous tissue detected is 7 square centimeters.

[0127] In step S409, the control unit 101 determines whether the difference between the detection result in step S407 and the detection result in step S408 is less than a threshold. The threshold is set to 1 square centimeter. The detection result in step S407 is 10 square centimeters, and the detection result in step S408 is 7 square centimeters, so the difference is 3 square centimeters, which is determined to be greater than or equal to the threshold, and the process proceeds to step S411.

[0128] In step S411, the control unit 101 stores the previously detected result from the detector in the work memory 103, associating it with the observation target ID. The observation target ID is 3, the inspection purpose is pressure ulcers, the inspection target is damage extending to the subcutaneous tissue, and the detector identification information is BS_0003. The detected region and area are then stored in the work memory 103. The stored inspection information corresponds to the record in the observation target column 512 in Figure 5(b).

[0129] In step S413, the control unit 101 increments counter i and returns the process to step S404. Counter i becomes 4.

[0130] <If the recommended detector and the previous detector do not match (NO in step S406), the difference in detection results is less than the threshold (YES in step S409), or the previous detector is selected (NO in step S410)> In step S404, the control unit 101 determines whether the counter i is less than the number of items to be inspected. Since the counter i is 4 and is less than or equal to 6, the process proceeds to step S405.

[0131] In step S405, the control unit 101 loads the fourth observation target ID of the target examination in the follow-up observation target DB into the work memory 103 and determines whether or not it exists in the past examination information. In the follow-up observation target DB, the fourth observation target ID for the examination target of pressure ulcers is 4, and since there is an observation target ID that matches in the observation target column 507 in Figure 5(a), the process proceeds to step S406.

[0132] In step S406, the control unit 101 determines whether the recommended detector corresponding to the fourth observation target ID 4 matches the previous detector corresponding to the observation target ID 4 in the past inspection information.

[0133] In Figure 3, the identification information 305 for the recommended detector in the observation target column 309 of the database under observation is BS_0005, and in Figure 5(a), the detector identification information 505 in the observation target column 509 is BS_0003. Therefore, it is determined that the recommended detector and the previous detector do not match, and the process proceeds to step S407.

[0134] In step S407, the control unit 101 uses the previous detector BS_0003 to detect the state of the first object to be inspected. Assume the area of ​​damage extending beyond the subcutaneous tissue is 4 square centimeters.

[0135] In step S408, the control unit 101 uses the recommended detector BS_0005 to detect the state of the first object to be inspected. The area of ​​damage extending beyond the subcutaneous tissue is assumed to be 3.5 square centimeters.

[0136] In step S409, the control unit 101 determines whether the difference between the detection result in step S407 and the detection result in step S408 is less than a threshold. The threshold is set to 1 square centimeter. Since the detection result in step S407 is 4 square centimeters and the detection result in step S408 is 3.5 square centimeters, the difference is 0.5 square centimeters, which is determined to be less than the threshold of 1 square centimeter, and the process proceeds to step S410.

[0137] In step S410, the control unit 101 determines whether to use the recommended detector and the detection result from the recommended detector for the next inspection. In this embodiment, the difference between the detection result of the recommended detector and the detection result of the previous detector is 0.5 square centimeters, and the user determines that this is unacceptable and presses the "No" button, so the process proceeds to step S411.

[0138] In step S411, the control unit 101 stores the previously detected result from the detector in the work memory 103, associating it with the observation target ID. The observation target ID is 4, the inspection purpose is pressure ulcers, the inspection target is damage extending beyond the subcutaneous tissue, and the detector identification information is BS_0003. The detected region and area are then stored in the work memory 103. The stored inspection information corresponds to the record in the observation target column 513 in Figure 5(b).

[0139] In step S413, the control unit 101 increments counter i and returns the process to step S404. Counter i becomes 5.

[0140] <When performing state detection using the recommended detector for unprocessed inspection targets (NO in step S4054)> In step S404, the control unit 101 determines whether the counter i is less than the number of items to be inspected. Since the counter i is 5 and is less than or equal to 6, the process proceeds to step S405.

[0141] In step S405, the control unit 101 loads the fifth observation target ID of the target examination purpose in the follow-up observation target DB into the work memory 103 and determines whether or not it exists in past examination information. In the follow-up observation target DB, the fifth observation target ID for the examination purpose of pressure ulcers is 5, and there is no matching observation target ID in the observation target columns 506 to 509 in Figure 5, so the process proceeds to step S414.

[0142] In step S414, the control unit 101 uses the recommended detector to detect the state of the fifth object to be examined. The control unit 101 uses the recommended detector BS_0005 to detect the area and size of the granulation tissue of the pressure ulcer.

[0143] In step S412, the control unit 101 stores the detection result from the recommended detector in the work memory 103, associating it with the observation target ID. The observation target ID is 5, the inspection purpose is pressure ulcers, the inspection target is granulation tissue, and the detector identification information is BS_0005. The detected region and area are then stored in the work memory 103. The stored inspection information corresponds to the record in the observation target column 514 in Figure 5(b).

[0144] In step S413, the control unit 101 increments counter i and returns the process to step S404. Counter i becomes 6.

[0145] <Perform a state detection using the recommended detector for unprocessed test targets. If the recommended detector differs from that used for other test targets (NO in S405)> In step S404, the control unit 101 determines whether the counter i is less than the number of items to be inspected. Since the counter i is 6, and therefore less than or equal to 6, the process proceeds to step S405.

[0146] In step S405, the control unit 101 loads the sixth observation target ID of the target examination in the follow-up target DB into the work memory 103 and determines whether or not it exists in past examination information. In the follow-up target DB, the fifth observation target ID for the examination target of pressure ulcers is 6, and since there is no observation target ID that matches in the observation target column 506~509 in Figure 5, the process proceeds to step S414.

[0147] In step S414, the control unit 101 performs condition detection of the sixth object to be examined using the recommended detector. The control unit 101 detects the area and size of the necrotic tissue of the pressure ulcer using the recommended detector BS_0006.

[0148] In step S412, the control unit 101 stores the detection result from the recommended detector in the work memory 103, associating it with the observation target ID. The observation target ID is 6, the examination purpose is pressure ulcers, the examination target is necrotic tissue, and the detector identification information is BS_0006. The detected region and area are then stored in the work memory 103. The stored examination information corresponds to the record in the observation target column 515 in Figure 5(b).

[0149] In step S413, the control unit 101 increments counter i and returns the process to step S404. Counter i becomes 7.

[0150] In step S404, the control unit 101 determines whether the counter i is less than the number of items to be inspected. Since the counter i is 7, which is greater than 6, the process proceeds to step S416.

[0151] In step S416, the control unit 101 determines whether to use the detector and detection results stored in the work memory 103 for the next inspection. In this embodiment, assuming the user chooses to use them for the next inspection, the process proceeds to step S417.

[0152] In step S417, the control unit 101 stores the observation target ID, detection result, and detector identification information in the storage device 104, associating them with the inspection image. As a result, the inspection information shown in Figure 5(b) is recorded in association with the inspection image and detection result.

[0153] In step S410, we explained an example of displaying the message, "Do you want to use the detection results of the recommended detector?", but if the frequency of use of the previous detector is below a threshold, the message may be changed to encourage switching to the recommended detector. The threshold is set at less than 25% of the frequency of use of the previous detector for the same inspection purpose. This allows for a gradual reduction in the frequency of use of the previous detector, leading to a shift to the recommended detector and reducing the maintenance costs for the service provider.

[0154] Furthermore, although step S409 explained that the threshold is uniquely determined for each subject being examined, it may also be determined by the ratio to the surface area of ​​the patient's body, or by the location of the affected area. It may also be determined by the difference in points when the previous detector detection result is set to 100%.

[0155] Furthermore, although it was explained that a recommended detector is determined for each inspection target, this embodiment can also be applied when a learning model is determined for each inspection target and the inference processing engine is common, as the detector identification information is simply replaced with information that identifies the learning model. In addition, even when detection is performed by combining multiple learning models for each inspection target, this embodiment can be applied as long as the detector identification information is information that can identify the combination.

[0156] Furthermore, while we have described an example in which state detection is performed on the inspection image using each detector in steps S407 and S408, and in step S409 it is determined whether the difference in detection results is greater than or equal to a threshold, it is also possible to compare the detection result from the previous detector on a past inspection image with the detection result from the recommended detector on the same inspection image, and if the difference in detection results is less than a threshold, perform state detection on the new inspection image using the selected detector after going through the process in step S410. This makes it possible to use the detection results from past inspection images in selecting a detector.

[0157] <Explanation of effects> According to this embodiment, for an injury or illness under observation, the previous detector is used to detect the state of the injury or illness. If the recommended detector differs from the previous detector, the recommended detector is also used to detect the state of the injury or illness. If the difference in the detection results is below a threshold, the recommended detector and its detection results can be selected to be used in the next examination. This reduces the likelihood of being unable to correctly observe the changes in the injury or illness over time while transitioning to the recommended detector. Furthermore, if the difference between the detection results of the previous detector and the recommended detector is greater than a threshold, the previous detector and its detection results are used in the next examination, thus reducing the likelihood of being unable to correctly observe the changes in the subject under examination over time.

[0158] Furthermore, according to this embodiment, if the recommended detector and the previous detector are different, and the difference in detection results is less than a threshold, the user can choose whether or not to use the recommended detector and its detection results for the next inspection. This allows the user to be aware of the detector update and avoids situations where the detector is automatically updated against the user's intention.

[0159] Furthermore, according to this embodiment, the state of injuries and illnesses detected in the past can be used in the next examination after confirming that there is no significant difference between the results detected with the previous detector and the results detected with the recommended detector for the same examination purpose. Moreover, if the number of subjects to be examined by the recommended detector increases for the same examination purpose, the recommended detector and its detection results can also be used in the next examination. Therefore, even for injuries and illnesses that are subject to observation over time, it is possible to enjoy the additional functions and performance of the new detector, and the possibility of being unable to correctly observe the changes in the subject over time can be reduced. In addition, since the detector information is saved as examination information for each subject to be examined, rather than treating injuries and illnesses for the purpose of examination as a single entity, the possibility of being unable to correctly observe the changes in the subject over time can be reduced.

[0160] Furthermore, according to this embodiment, if the user chooses to use the recommended detector and its detection results in the next examination, the identification information of the recommended detector is registered in association with the examination image. As a result, when examining an injury or illness that is subject to follow-up observation in a subsequent examination, the recommended detector from the previous examination is registered, which reduces the possibility of being unable to correctly observe the changes in the subject over time, while also enabling state detection using the recommended detector used in the previous examination, and furthermore, enabling detection using the previous detector. In this embodiment, in subsequent examinations, state detection will be performed using three types of detectors, as shown in Figure 5(b), with detector identification information BS_003, BS_005, and BS_008.

[0161] [Embodiment 3] In this embodiment, in step S410 or S416 of Figure 4 in Embodiment 2, when it is decided to use the recommended detector and the detection results of the recommended detector for the next examination, the recommended detector is used to perform detection on past examination images used for detecting the state of injury or illness for the same examination purpose, and a notification is given if a difference of note is found in the detection results.

[0162] Figure 6 is a flowchart illustrating the process of detecting the state of an object to be inspected from past inspection images according to this embodiment.

[0163] The process shown in Figure 6 is executed after the process shown in Figure 4 of Embodiment 2.

[0164] In step S601, the control unit 101 loads past examination information and images of the same illness / injury and the same examination site into the work memory 103, using the patient management number of the patient information as the key. If the past examination information spans several years, it is acquired in time periods. In this embodiment, it is assumed that examination images taken during the most recent year are acquired.

[0165] In step S602, the control unit 101 assigns 1 to counter j, which corresponds to the number of examination images. Counter j is a number assigned to each examination image in order to process all examination images.

[0166] In step S603, the control unit 101 determines whether the counter j is less than or equal to the number of inspection images. If the counter j is less than or equal to the number of inspection images, the control unit 101 proceeds to step S604. If the counter j is greater than the number of inspection images, the control unit 101 determines that the detection process for all inspection images has been completed and terminates the process.

[0167] In step S604, the control unit 101 uses the recommended detector to detect the state of the injury or illness for the same purpose as the past examination information, targeting the j-th examination image.

[0168] In step S605, the control unit 101 determines whether there is a difference of note between the detection result of the j-th inspection image and the detection result of the previous detector. A difference of note is defined as a difference of 20 points or more in the area of ​​the injury or disease when the previous result is set to 100%. If the control unit 101 determines that there is a difference of note, it proceeds to step S606; if it determines that there is no difference of note, it proceeds to step S607.

[0169] In step S606, the control unit 101 notifies the user that there is a difference that warrants attention.

[0170] In step S607, the control unit 101 increments counter j and returns the process to step S603.

[0171] In this embodiment, a difference requiring attention was described as a case where the difference in the area of ​​the injury differs by 20 points or more when the previous result is set to 100%, but it may also be described as a case where a lesion requiring attention is found. For example, if necrotic tissue that was not previously detectable is detected in the examination image of a pressure ulcer, notification shall be given.

[0172] According to this embodiment, the recommended detector can be used to detect the state of past examination images, and if a condition requiring attention is detected by the recommended detector, the user can recognize it. This reduces the likelihood of being unable to properly observe the temporal changes of the examination target, even for conditions that are subject to observation, while allowing medical professionals to notice lesions requiring attention and provide appropriate treatment to the patient.

[0173] [Other embodiments] The present invention can also be realized by supplying a program that implements one or more of the functions of the above-described embodiments to a system or device via a network or storage medium, and by having one or more processors in the computer of that system or device read and execute the program. It can also be realized by a circuit (e.g., an ASIC) that implements one or more functions.

[0174] The invention is not limited to the embodiments described above, and various modifications and variations are possible without departing from the spirit and scope of the invention. Accordingly, claims are attached to disclose the scope of the invention.

[0175] The disclosures herein include the following information processing devices, control methods, and programs. [Item 1] An information processing device that detects the state of an object to be inspected from an image of the object to be inspected, An acquisition means for acquiring information of a first detector used for detecting the state of the object to be inspected, and information of a second detector used when state detection was performed on the object to be inspected in the past. An information processing apparatus characterized by having control means for controlling the apparatus to use the second detector to detect the state of the object being inspected when the object being inspected is subject to observation and the first detector and the second detector are different. [Item 2] If the acquisition means has inspection information from which a state detection has been performed on the object to be inspected in the past, it acquires information from the second detector based on the inspection information. The information processing device according to item 1, characterized in that the control means compares the information of the first detector with the information of the second detector. [Item 3] The information processing apparatus according to item 2, characterized in that the control means controls the first detector to detect the state of the object to be inspected when the information from the first detector and the information from the second detector are the same. [Item 4] The information processing apparatus according to item 2, characterized in that the control means controls the first detector to detect the state of the object to be inspected if the inspection information is not available. [Item 5] The information processing apparatus according to any one of items 1 to 4, characterized in that the control means controls the inspection target to use the first detector to detect the state of the inspection target when the inspection target is not subject to ongoing observation. [Item 6] The information processing apparatus according to any one of items 1 to 4, characterized in that the control means controls the first detector to perform state detection of the object to be inspected if a predetermined period of time has elapsed since the previous state detection of the object to be inspected. [Item 7] A selection means that allows selection of a detector and detection result to be used for the next state detection of the object being inspected, using the first detector and the detection result from the first detector or the second detector and the detection result from the second detector, The information processing apparatus according to any one of items 1 to 6, characterized by having a storage means for storing information of a detector selected by the selection means, the detection result by the selected detector, and the image of the object to be inspected in association with each other. [Item 8] The information processing apparatus according to any one of items 1 to 5, characterized in that, when the first detector and the second detector are different, the control means controls the apparatus to use the second detector to detect the state of the object to be inspected and to use the first detector to detect the state of the object to be inspected. [Item 9] The control means compares the difference between the detection result from the first detector and the detection result from the second detector with a threshold value. The information processing apparatus according to item 8, characterized in that, if the difference is less than a threshold, it has a first detector and a detector that uses the detection result of the first detector for the next state detection of the object to be inspected, and a selection means that can be selected as the detection result. [Item 10] The information processing apparatus according to item 9, characterized in that, if the difference is greater than or equal to a threshold, the first detector and the detection result by the first detector are not selected as the detector and detection result to be used for the next state detection of the object to be inspected. [Item 11] A selection means for selecting the first detector and the detection result from the first detector or the second detector and the detection result from the second detector to be used as the detector and detection result for the next state detection of the object to be inspected, The information processing apparatus according to any one of items 8 to 10, characterized by having a storage means for storing information of a detector selected by the selection means, the detection result by the selected detector, and the image of the object to be inspected in association with each other. [Item 12] The information processing apparatus according to any one of items 8 to 11, characterized in that, when there are multiple objects to be inspected, the control means performs state detection for all objects to be inspected using the first detector and the second detector. [Item 13] The information processing apparatus according to item 8, characterized in that, when the usage frequency of the second detector is less than a threshold, the control means has a detector and a selection means that can be selected as a detection result to use the first detector and the detection result of the first detector for the next state detection of the object to be inspected. [Item 14] The information processing apparatus according to item 10 or 11, characterized in that when the control means selects the first detector and the detection result of the first detector as detectors to be used for the next state detection of the object to be inspected, it performs state detection on past images of the object to be inspected that were used for state detection of the same object to be inspected using the first detector, and provides notification based on the difference between the detection result of the image of the object to be inspected and the detection result of the past images of the object to be inspected. [Item 15] The subject of the aforementioned examination is the patient's condition of injury or illness. The aforementioned test information is associated with patient information and includes, for each test subject, the subject to follow-up, the disease or injury targeted by the test, the detection result of the test subject, and identification information of the detector used to detect the state of the test subject. The information processing device according to item 2, characterized in that the acquisition means acquires the examination information based on the patient information. [Item 16] The first detector and the second detector include a learning model generated by machine learning using training data. The information processing device according to any one of items 1 to 15, characterized in that the first detector includes a learning model generated by training the learning model of the second detector with training data. [Item 17] A control method for an information processing device that detects the state of an object to be inspected from an image of the object to be inspected, The steps include obtaining information of a first detector used for detecting the state of the object to be inspected, and information of a second detector used when state detection was performed on the object to be inspected in the past, A control method characterized by comprising the step of controlling the system to use the second detector to detect the state of the object being inspected, if the object being inspected is subject to observation and the first detector and the second detector are different. [Item 18] A program for causing a computer to function as one of the means of an information processing device described in any of items 1 through 16. [Explanation of Symbols]

[0176] 100... Information processing device, 101... Control unit

Claims

1. An information processing device that detects the state of an object to be inspected from an image of the object to be inspected, An acquisition means for acquiring information of a first detector used for detecting the state of the object to be inspected, and information of a second detector used when state detection was performed on the object to be inspected in the past. An information processing apparatus characterized by having control means for controlling the apparatus to use the second detector to detect the state of the object being inspected when the object being inspected is an object being observed over time and the first detector and the second detector are different.

2. If the acquisition means has inspection information from which a state has been detected in the past for the object to be inspected, it acquires information from the second detector based on the inspection information. The information processing apparatus according to claim 1, characterized in that the control means compares the information of the first detector with the information of the second detector.

3. The information processing apparatus according to claim 2, characterized in that the control means controls the first detector to detect the state of the object to be inspected when the information from the first detector and the information from the second detector are the same.

4. The information processing apparatus according to claim 2, characterized in that the control means controls the first detector to detect the state of the object to be inspected if there is no inspection information.

5. The information processing apparatus according to claim 1, characterized in that the control means controls the first detector to detect the state of the object being inspected when the object being inspected is not subject to ongoing observation.

6. The information processing apparatus according to claim 1, characterized in that the control means controls the first detector to perform state detection of the object to be inspected if a predetermined period of time has elapsed since the previous state detection of the object to be inspected.

7. A selection means that allows selection of a detector and detection result to be used for the next state detection of the object being inspected, using the first detector and the detection result from the first detector or the second detector and the detection result from the second detector, The information processing apparatus according to claim 1, further comprising a storage means for storing information of a detector selected by the selection means, the detection result by the selected detector, and the image of the object to be inspected in association with each other.

8. The information processing apparatus according to claim 1, characterized in that the control means controls the system to use the second detector to detect the state of the object to be inspected and to use the first detector to detect the state of the object to be inspected when the first detector and the second detector are different.

9. The control means compares the difference between the detection result from the first detector and the detection result from the second detector with a threshold value. The information processing apparatus according to claim 8, characterized in that, if the difference is less than a threshold, it has a first detector and a detector that uses the detection result of the first detector for the next state detection of the object to be inspected, and a selection means that can be selected as the detection result.

10. The information processing apparatus according to claim 9, characterized in that, if the difference is greater than or equal to a threshold, the first detector and the detection result by the first detector are not selected as the detector and detection result to be used for the next state detection of the object to be inspected.

11. A selection means for selecting the first detector and the detection result from the first detector or the second detector and the detection result from the second detector to be used as the detector and detection result for the next state detection of the object to be inspected, The information processing apparatus according to claim 8, further comprising a storage means for storing information of a detector selected by the selection means, the detection result by the selected detector, and the image of the object to be inspected in association with each other.

12. The information processing apparatus according to claim 8, characterized in that, when there are multiple objects to be inspected, the control means performs state detection for all objects to be inspected using the first detector and the second detector.

13. The information processing apparatus according to claim 8, characterized in that, when the usage frequency of the second detector is less than a threshold, the control means has a detector and a selection means that can be selected as a detection result to be used for the next state detection of the object to be inspected, which are the first detector and the detection result of the first detector.

14. The information processing apparatus according to claim 10, characterized in that when the control means selects the first detector and the detection result of the first detector as detectors to be used for the next state detection of the object to be inspected, it performs state detection on past images of the object to be inspected that were used for state detection of the same object to be inspected using the first detector, and provides notification based on the difference between the detection result of the image of the object to be inspected and the detection result of the past images of the object to be inspected.

15. The subject of the aforementioned examination is the patient's condition of injury or illness. The aforementioned test information is associated with patient information and includes, for each test subject, the subject to follow-up, the disease or injury targeted by the test, the detection result of the test subject, and identification information of the detector used to detect the state of the test subject. The information processing apparatus according to claim 2, characterized in that the acquisition means acquires the examination information based on the patient information.

16. The first detector and the second detector include a learning model generated by machine learning using training data. The information processing apparatus according to claim 1, characterized in that the first detector includes a learning model generated by training a learning model of the second detector with training data.

17. A control method for an information processing device that detects the state of an object to be inspected from an image of the object to be inspected, The steps include obtaining information of a first detector used for detecting the state of the object to be inspected, and information of a second detector used when state detection was performed on the object to be inspected in the past, A control method characterized by comprising the step of controlling the system to use the second detector to detect the state of the object being inspected, if the object being inspected is an object being observed over time and the first detector and the second detector are different.

18. A program for causing a computer to function as one of the means of an information processing apparatus according to any one of claims 1 to 16.

Images