Medical information processing device, medical information processing method, and medical information processing program
Patent Information
- Authority / Receiving Office
- JP · JP
- Patent Type
- Patents
- Current Assignee / Owner
- CANON KK
- Filing Date
- 2022-03-31
- Publication Date
- 2026-08-03
Smart Images
Figure 0007898899000001 
Figure 0007898899000002 
Figure 0007898899000003
Abstract
Description
Technical Field
[0004] , , ,
[0005] , , ,
[0001] The embodiments disclosed in this specification and the drawings relate to a medical information processing apparatus, a medical information processing method, and a medical information processing program.
Background Art
[0002] Due to the recent development of medical image diagnosis using deep learning technology, the diagnostic processing by a pre-trained machine learning model has been highly accurate. However, when processing high-definition images, the processing time of the pre-trained machine learning model often becomes extremely long. On the other hand, there is a problem that the accuracy decreases when the processing time is shortened. In particular, in the medical field, there are cases where urgency such as emergency is required and cases where highly accurate results are required.
Prior Art Documents
Patent Documents
[0003]
Patent Document 1
Summary of the Invention
Problems to be Solved by the Invention
[0004] One of the problems to be solved by the embodiments disclosed in this specification and the drawings is to provide processing according to needs. However, the problems to be solved by the embodiments disclosed in this specification and the drawings are not limited to the above problems. The problems corresponding to the respective effects of each configuration shown in the embodiments described later can also be regarded as other problems.
Means for Solving the Problems
[0005] The medical information processing device according to this embodiment includes an acquisition unit and a determination unit. The acquisition unit acquires patient information relating to a target patient, environmental information relating to the execution environment, and target data relating to the target patient. The determination unit determines a combination of trained models that satisfy conditions according to the patient information and the environmental information from a plurality of trained models with different speeds and accuracy. [Brief explanation of the drawing]
[0006] [Figure 1] Figure 1 is a block diagram showing the medical information document system according to this embodiment. [Figure 2] Figure 2 is a block diagram showing the details of the medical information processing device according to this embodiment. [Figure 3] Figure 3 is a flowchart showing the operation of the medical information processing device according to this embodiment. [Figure 4] Figure 4 illustrates the concept of variations in a trained model. [Figure 5] Figure 5 shows an example of a user interface that indicates whether speed or accuracy should be prioritized. [Figure 6] Figure 6 shows a first example of the user interface after the analysis process. [Figure 7] Figure 7 shows a second example of the user interface after the analysis process. [Figure 8] Figure 8 shows a third example of the user interface after the analysis process. [Figure 9] Figure 9 is a conceptual diagram illustrating the learning process of the configured model. [Figure 10] Figure 10 shows an example of a proposed model based on the settings. [Modes for carrying out the invention]
[0007] The medical information processing device, medical information processing method, and medical information processing program according to this embodiment will be described below with reference to the drawings. In the following embodiments, parts with the same reference numerals perform similar operations, and redundant explanations will be omitted as appropriate. One embodiment will be described below with reference to the drawings.
[0008] The medical information processing system according to this embodiment will be described with reference to Figure 1. The medical information processing system according to this embodiment includes a medical information processing device 1, an image server 2, an electronic medical record system 3, and a medical information management application 4. The medical information processing device 1, the image server 2, the electronic medical record system 3, and the medical information management application 4 are connected via a network 7. Although the medical information processing device 1 according to this embodiment is assumed to be separate from the image server 2, the electronic medical record system 3, and the medical information management application 4, it may be included in at least one of the image server 2, the electronic medical record system 3, and the medical information management application 4.
[0009] The medical information processing device 1 selects a combination of pre-trained models that meets the user's needs in terms of speed and accuracy, and then performs data processing on the target data. It also displays the selected combination of pre-trained models and information regarding the speed and accuracy of each pre-trained model. Further details of the medical information processing device 1 will be described later, referring to Figure 2.
[0010] Image server 2 is, for example, a PACS (Picture Archiving and Communication System), a system that stores and manages medical image data. Image server 2 stores and manages medical image data converted according to the DICOM (Digital Imaging and Communication Medicine) standard, for example.
[0011] Electronic medical record system 3 is a system that stores and manages electronic medical record data, including patient information. Patient information includes, for example, patient ID, patient name, gender, age, medical history, lifestyle, and other information related to the electronic medical record, such as findings, disease name, vital signs, examination stage information, clinical pathways, and treatment details. Clinical pathways represent standard treatment plans in chronological order.
[0012] Medical information management application 4 is an application that can integrate and manage medical information such as treatment information and test information related to patients over time, and can share medical information among multiple doctors, or among users represented by healthcare professionals such as doctors, technicians, and nurses.
[0013] Network 7 is, for example, the hospital's internal network. Network 7 can be either wired or wireless. Furthermore, as long as security is ensured, the connected lines are not limited to the hospital's internal network. For example, it is acceptable to connect to public communication lines such as the internet via a VPN (Virtual Private Network).
[0014] Next, the details of the medical information processing device 1 will be explained with reference to the block diagram in Figure 2. The medical information processing device 1 shown in Figure 2 includes a processing circuit 10, a memory 11, an input interface 12, a communication interface 13, and a display 14. The processing circuit 10, the memory 11, the input interface 12, the communication interface 13, and the display 14 are connected to each other in a way that allows them to communicate with one another, for example, via a bus.
[0015] The processing circuit 10 is a processor that functions as the central part of the medical information processing device 1. The processing circuit 10 includes an acquisition function 101, a decision function 102, a calculation function 103, an execution function 104, a learning function 105, and a display control function 106.
[0016] The acquisition function 101 acquires patient information regarding the target patient, environment information regarding the execution environment, and target data regarding the target patient. The environment information includes at least one of the system environment for executing analysis processing by the learned model and information regarding the scale of the hospital. The system environment includes information regarding the execution environment such as, for example, the number of workstations for executing analysis processing by the learned model and the specifications of the CPU (Central Processing Unit) and memory of the workstations. The scale of the hospital includes information such as, for example, the types of hospital facilities, the number of medical staff such as doctors, and the types of medical departments.
[0017] The determination function 102 determines a combination of learned models that satisfies the conditions according to the patient information and the environment information from a plurality of learned models with different speeds and accuracies. The calculation function 103 calculates the processing time and estimated accuracy assumed by the learned model.
[0018] The execution function 104 executes analysis processing according to the determined combination of learned models. Further, the execution function 104 inputs the patient information, environment information, and target data to be processed into the learned setting model described later, and outputs the combination of learned models. The learning function 105 uses the information regarding the learned models included in the combination, patient information, environment information, and target data as input data, and uses the feedback information regarding the feedback from the user regarding at least one of the execution of the analysis processing in the combination and the analysis result as correct data to train the network model and generate a learned setting model. The display control function 106 displays the information of the learned model on the display 14 or the like.
[0019] Memory 11 is a storage device such as ROM (Read Only Memory), RAM (Random Access Memory), HDD (Hard Disk Drive), SSD (Solid State Drive), and integrated circuit storage device that stores various types of information. Memory 11 may also be a drive device that reads and writes various types of information to and from portable storage media such as CD-ROM drives, DVD drives, and flash memory. Note that memory 11 does not necessarily have to be implemented by a single storage device. For example, memory 11 may be implemented by multiple storage devices. Furthermore, memory 11 may be located in another computer connected to the medical information processing device 1 via the network 7.
[0020] Memory 11 stores the processing program and the like according to this embodiment. This program may, for example, be pre-stored in memory 11. Alternatively, for example, it may be stored and distributed on a non-transient storage medium, read from the non-transient storage medium, and installed in memory 11.
[0021] The input interface 12 receives various input operations from the user, converts the received input operations into electrical signals, and outputs them to the processing circuit 10. In this embodiment, the input interface 12 is connected to input devices such as a mouse, keyboard, trackball, switch, button, joystick, touchpad, and touch panel, where instructions are input by touching the operating surface. Alternatively, the input device connected to the input interface 12 may be an input device provided on another computer connected via a network or the like.
[0022] The communication interface 13 performs data communication with the image server 2, the electronic medical record system 3, the medical information management application 4, and, although not shown, the hospital information system, the radiology department information system, etc. The communication interface 13 performs data communication in accordance with, for example, a pre-configured known standard. Communication between the medical information processing device 1 and the hospital information system, the electronic medical record system 3, the medical information management application 4, and the radiology department information system is, for example, conducted in accordance with HL7 (Health Level 7). Furthermore, communication between the medical information processing device 1 and the image server 2 and the medical information management application 4 is, for example, conducted in accordance with DICOM (Digital Imaging and Communications in Medicine).
[0023] The display 14 displays a GUI (Graphical User Interface) and the like for accepting various operations from the user. Any display can be used as appropriate, such as a CRT (Cathode Ray Tube) display, liquid crystal display, organic EL display, LED display, plasma display, or touch display capable of touch input. Alternatively, the medical information processing device 1 may not include a display, and the GUI may be displayed on an external display, or the GUI may be displayed via a projector or the like.
[0024] Next, an example of the operation of the medical information processing device 1 according to this embodiment will be explained with reference to the flowchart in Figure 3. Here, it is assumed that multiple analysis processes are executed sequentially to obtain a processing result for a single task. Furthermore, it is assumed that multiple pre-trained models are prepared in advance for each analysis process.
[0025] In step S301, the acquisition function 101 acquires patient information, environmental information, and target data to be processed. In this case, the target data is assumed to be medical images, but it may also be time-series data such as vital data such as heart rate and blood pressure. In step S302, the decision function 102 decides whether to prioritize speed or accuracy when applying machine learning to the target data and performing analysis. The decision of whether to prioritize speed or accuracy may be made based on user instructions, for example, or based on whether the various information and data meet the judgment conditions, using the patient information, environmental information, and target data acquired in step S301. For example, in the case of an emergency patient transported by ambulance, processing speed should often be prioritized, so if it can be determined from the patient information that the patient is in emergency condition, speed should be prioritized. On the other hand, in cases where the output results of a trained model are to be used as reference information for findings, estimation accuracy should often be prioritized, so for example, if supplementary information indicating that the target data is for reference purposes is attached, estimation accuracy should be prioritized if such supplementary information is present.
[0026] In step S303, the decision function 102 selects one pre-trained model from among multiple pre-trained models prepared for each analysis process, based on the priority items of step S302, and determines the combination of pre-trained models for the series of analysis processes. For example, if speed is the priority, the decision function 102 should select the pre-trained model with the shortest processing time for each analysis process. Similarly, if accuracy is the priority, the decision function 102 should select the pre-trained model with the highest estimation accuracy for each analysis process. In step S304, the calculation function 103 calculates the expected processing time and estimation accuracy of the trained model for each analysis process.
[0027] In step S305, the analysis process for the target data is started, and the display control function 106 displays a user interface regarding the status of the analysis process, for example, on the display 14. In step S306, the execution function 104 determines whether there is a change instruction from the user, in this case, a change instruction from speed priority to accuracy priority, or from accuracy priority to speed priority. If a change instruction from the user is received, the process proceeds to step S307. If there is no change instruction from the user, the analysis process continues until a change instruction is received from the user. In step S307, the combination of trained models is changed based on the change instruction. For example, the combination of trained models determined in the analysis processes from the currently processed analysis process onward is changed. Alternatively, the trained model in the currently processed analysis process is not changed, but the combination of trained models determined in the analysis processes from the next analysis process onward is changed. After the combination is changed, the process returns to step S304 and the same process is repeated.
[0028] Next, the concept of variations in the trained model assumed in this embodiment will be explained with reference to Figure 4. Figure 4 is a table showing model information for multiple pre-trained models that perform a single analysis process. Here, four different pre-trained models are shown, with information on input image size, processing time, and accuracy associated with each model. The input image size indicates the resolution (number of pixels, height x width) of the image that can be processed. Processing time indicates the ranking of the pre-trained models based on their processing time. Accuracy indicates the ranking of the pre-trained models' estimation accuracy. Here, ◎, ○, △, × indicate decreasing performance.
[0029] Specifically, for example, trained model ID "1", input image size "500 x 700" pixels, processing time "×", and accuracy "◎" are associated with each other. In other words, the model with trained model ID "1" takes a long time to process but has high estimation accuracy. On the other hand, trained model ID "4" has an input image size of "5 x 7" pixels, processing time "◎", and accuracy "×", indicating low estimation accuracy but fast processing time.
[0030] Next, an example of a user interface in step S302, which allows the user to choose between prioritizing speed or accuracy, will be explained with reference to Figure 5. Figure 5 shows a user interface window displayed by the display control function 106, which allows the user to select between processing that prioritizes speed or processing that prioritizes accuracy. To prioritize processing speed, the user presses the speed priority button 51; to prioritize estimation accuracy, the user presses the accuracy priority button 52. In this embodiment, "pressing a button" includes pressing a physical button to which a setting has been assigned, selecting and clicking a button such as the speed priority button 51 using a mouse cursor, and touching the area of a button when the display is a touch panel display.
[0031] The display method is not limited to a window as shown in Figure 5; any display method is acceptable, such as allowing the user to select text. Furthermore, the medical information processing device 1 according to this embodiment may accept the selection of priority items not only through input to an input device or touch panel display, but also through voice input via a microphone. In other words, any method of acquisition is acceptable as long as the user's specification of priority items can be received. In addition, in emergency cases, speed may be forcibly prioritized, in which case the accuracy priority button 52 may be hidden or made unselectable.
[0032] Next, a first example of the user interface provided by the display control function 106 after the analysis process in step S305 has started will be described with reference to Figure 6. Figure 6 illustrates a scenario where processing results for a single task are obtained through three analytical processes. Here, assuming a task to determine whether a tumor is benign or malignant, an example is shown in which the first analytical process 61 identifies the location of the tumor, the second analytical process 62 identifies the size of the tumor, and the third analytical process 63 determines whether the tumor is benign or malignant, in other words, whether the tumor is cancerous or not, and these processes are executed in order from the first analytical process 61.
[0033] The progress of the processing can also be represented by the cursor 64. Here, the cursor 64 indicates that the first analysis process 61, which identifies a location, has been completed up to the middle stage. This makes it easy to understand which analysis processes are currently being performed and to what extent for the task.
[0034] Furthermore, the button group 65 may be used to switch between setting the analysis process to "accuracy priority" or "speed priority." For example, pressing the "High Accuracy" button will apply the accuracy-prioritizing trained model to the currently ongoing analysis process or the next analysis process. On the other hand, pressing the "Accelerate" button will apply the speed-prioritizing trained model to the currently ongoing analysis process or the next analysis process. Of course, the text displayed on the buttons is not limited to the example in Figure 6; any text is acceptable as long as the user can select between accuracy priority and speed priority.
[0035] Note that, if the analysis process is in progress with the speed priority setting enabled, depending on the resolution of the image being processed, it may not be possible to switch to accuracy priority once the resolution is reduced. In such cases, the "High Precision" button should be disabled or hidden for processes that cannot be processed with accuracy priority enabled. Alternatively, the calculation function 103 may recalculate and display the remaining estimated processing time.
[0036] Additionally, pressing the "Output Intermediate Results" button may display intermediate data during the analysis process. For example, if the analysis process involves convolution for feature extraction on a medical image, the intermediate data after the convolution or a heatmap related to feature extraction may be displayed. Pressing the "Pause" button may interrupt the processing by the trained model and output the processing results at the time the button is pressed or after the currently ongoing analysis process is completed. For example, if a doctor wants to use the trained model to determine the size of a tumor and then make their own "cancer diagnosis," they can pause the process using the "Pause" button and perform the cancer diagnosis from the state where the tumor's location and size have been analyzed.
[0037] Next, a second example of a user interface using the display control function 106 will be explained with reference to Figure 7. Figure 7 shows an example where, in addition to the user interface in Figure 6, more detailed information such as the remaining time for each analysis process is displayed.
[0038] The processing time window 71 displays the total remaining processing time and the remaining processing time for each analysis process. It may also display accuracy information for the entire task, including each analysis process. This accuracy may be the average accuracy for each analysis process, or it may use other statistical methods, such as a weighted average of the accuracy values for each analysis process.
[0039] The slider window 72 displays sliders for processing time for each analysis process and for the overall accuracy of the task. Specifically, for example, if the user moves the "Location Identification" slider in the direction of decreasing time, a speed-prioritizing trained model is applied, and conversely, if the slider is moved in the direction of increasing time, an accuracy-prioritizing trained model is applied. In addition, when the processing time changes due to the adjustment of the sliders for each analysis process, the calculation function 103 recalculates the overall accuracy of the task, and the "Accuracy" slider is displayed in conjunction with this to the position corresponding to the recalculated value.
[0040] Furthermore, adjusting the "accuracy" slider may cause the sliders for each analysis process to move in conjunction. For example, if the user moves the slider towards higher accuracy, the sliders for each analysis process will move in conjunction towards longer processing times. In this way, by displaying the slider window 72 in addition to, or instead of, the button group 65, processing time and estimation accuracy can be set in more detail. The adjustment values for the slider bar mentioned above are set in stages according to the number of trained models. For example, in the case of the "location identification" analysis process, three trained models are prepared, so the processing time slider bar can be adjusted in three stages.
[0041] The target area window 73 is a checkbox that indicates the target area to be processed. For example, if there are multiple target areas in the captured medical image that you want to process, simply check the checkbox for the corresponding target area.
[0042] The progress window 74 includes icons 741 that indicate which process is currently running in each analysis process, and icons 742 that indicate multiple trained models capable of executing each analysis process. In the example in Figure 7, the trained models used in each analysis process are highlighted in the progress window 74 and connected by arrows as the analysis process progresses. Specifically, trained model "A" is connected for the "location identification" analysis process, trained model "E" for the "size determination" analysis process, and trained model "G" for the "cancer detection" analysis process. This display allows users to quickly understand which trained model is being used. In addition, each icon 741 may be displayed in a different way, such as by using different colors based on whether speed or accuracy is prioritized.
[0043] Next, a third example of a user interface using the display control function 106 will be explained with reference to Figure 8. Figure 8 shows an example of displaying the estimated accuracy (accuracy in Figure 8) and processing time (time in Figure 8) for each of the multiple pre-trained models available for each analysis process. In this way, the user can easily understand which pre-trained model is selected for each analysis process, as well as the estimated accuracy and processing time of each pre-trained model. The user can also select a pre-trained model different from the currently set pre-trained model, making the newly selected pre-trained model available for analysis. Furthermore, models that cannot be used in subsequent analysis processes based on their combination with the model currently in progress may be displayed with a dashed line or a light color to indicate that they cannot be selected.
[0044] Furthermore, feedback information regarding user actions, such as selecting accuracy priority or switching to a speed-prioritizing pre-trained model during accuracy-prioritizing analysis, may be used to train the system on combinations of pre-trained models and generate a pre-trained model that suggests combinations of pre-trained models for processing tasks. The training process for setting up a pre-trained model suitable for a task will be explained with reference to the conceptual diagram in Figure 9.
[0045] Figure 9 shows examples of input and output data for training the network model 90. The learning function 105 uses the trained model information, patient information, environmental information, and target image used in the analysis process as input data 91, and the feedback information in the context of the input data 91 as ground truth data 92 to train the network model 90. The trained model information includes the identifier of the trained model used in the analysis process, information on estimation accuracy and processing time, and information on estimation accuracy and processing time for the entire task. The feedback information indicates the user's actions in relation to the input data. Examples of feedback information include whether the estimation process for the task using the trained model was terminated midway, whether the priority items were switched (i.e., changed from accuracy priority to speed priority, or from speed priority to accuracy priority), the combination of trained models selected, and the evaluation of the task processing results (e.g., a level rating such as excellent, good, fair, or poor).
[0046] By training the network model 90 with the input data 91 and ground truth data 92 described above, a configuration model 95 is generated, which is a trained model that incorporates how users provided feedback on the execution of the trained model used in the analysis process. According to the configuration model 95, a combination of pre-trained models can be proposed that offers tasks with accuracy and processing time tailored to the user's preferences.
[0047] Next, Figure 10 shows an example of a proposal based on the setting model 95. Figure 10 shows a user interface similar to Figure 5, and is a user interface for selecting between speed priority and accuracy priority when using the configuration model 95. Note that Figure 5 assumes an example where the configuration model 95 is not used. Furthermore, the feedback information used as the ground truth data 92 assumes an example in which the user selects a trained model for the analysis process to switch from accuracy priority to speed priority midway through the process, and frequently stops the task execution midway through the process, in this case when the second analysis process for determining tumor size is completed.
[0048] The execution function 104 takes patient information, environmental information, and target data to be processed as input to the configuration model 95 and outputs a combination of trained models for processing the task. As shown in Figure 10, since the system has been trained to prioritize speed, the configuration model 95 will select a combination of trained models for each analysis process in a way that reduces processing time when accuracy is prioritized. Note that for speed-prioritized analysis processes, a modified combination that reduces processing time may also be presented. Furthermore, as shown in Figure 10, a new button labeled "Up to Tumor Size" may be generated and displayed. Thus, according to the configuration model, it is possible to present a combination of analysis processes tailored to the user's preferences.
[0049] According to the embodiments described above, based on patient information, environmental information, and target data, one pre-trained model is selected from multiple pre-trained models for each analysis process, depending on whether speed or accuracy is prioritized, and the combination of pre-trained models is determined. This allows for the provision of optimal processing tailored to the needs. Furthermore, based on feedback results derived from at least one of the user's priority item selection results and the pre-trained model selection results, the optimal combination of pre-trained models can be presented by using a pre-trained configuration model that has learned, for example, the trends in priority items for each doctor and hospital.
[0050] In the above description, the term "processor" refers to circuits such as a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), an Application Specific Integrated Circuit (ASIC), or a programmable logic device (e.g., a Simple Programmable Logic Device (SPLD), a Complex Programmable Logic Device (CPLD), and a Field Programmable Gate Array (FPGA)). When the processor is a CPU, for example, it performs its functions by reading and executing a program stored in a memory circuit. On the other hand, when the processor is an ASIC, for example, instead of the program being stored in a memory circuit, the function is directly incorporated as a logic circuit within the processor's circuit. In this embodiment, each processor is not limited to being configured as a single circuit; multiple independent circuits may be combined to form a single processor and perform its functions. Furthermore, multiple components shown in the figure may be integrated into a single processor to perform its functions.
[0051] In addition, each function according to the embodiment can also be realized by installing a program that performs the processing on a computer such as a workstation and loading it into memory. In this case, the program that can cause the computer to execute the method can also be stored and distributed on a storage medium such as a magnetic disk (hard disk, etc.), optical disk (CD-ROM, DVD, etc.), or semiconductor memory.
[0052] While several embodiments have been described, these embodiments are presented as examples only and are not intended to limit the scope of the invention. These embodiments can be implemented in a variety of other forms, and various omissions, substitutions, modifications, and combinations of embodiments can be made without departing from the spirit of the invention. These embodiments and their variations are included in the scope and spirit of the invention, as well as in the claims and their equivalents. [Explanation of Symbols]
[0053] 1. Medical Information Processing Device 2 Image Server 3. Electronic medical record system 4. Medical Information Management Application 7 Network 10 Processing Circuit 11 memory 12 Input Interfaces 13 Communication Interface 14 displays 51 Speed Priority Button 52. Accuracy Priority Button 61 First Analysis Process 62. Second Analysis Process 63 Third Analysis Process 64 Cursors 65 button group 71 Processing time window 72 Slide bar window 73 Target Area Window 74 Progress window 90 Network Models 91 Input data 92 Correct Data 95 Configuration Models 101 Acquisition function 102 Decision Function 103 Calculation Function 104 Execution Functions 105 Learning Functions 106 Display control function 741,742 icons
Claims
1. An acquisition unit that causes a computer to acquire patient information relating to the target patient, environmental information relating to the execution environment, and target data relating to the target patient, The computer includes a determination unit that causes it to determine a combination of trained models from a plurality of trained models with different speeds and accuracys stored in internal memory or an external device, which are suitable for application to the target data and satisfy speed or accuracy conditions according to the patient information and the environmental information. A medical information processing device equipped with the following features.
2. The aforementioned multiple trained models are prepared for each of the one or more analysis processes. The medical information processing apparatus according to claim 1, wherein, when the determination unit performs a task that includes multiple analysis processes, it selects one trained model from the multiple trained models provided to determine the combination for each analysis process.
3. The medical information processing apparatus according to claim 1 or 2, wherein the determination unit determines the combination based on whether to prioritize speed or accuracy.
4. The patient information includes the vital signs of the subject patient, The medical information processing device according to any one of claims 1 to 3, wherein the environmental information includes at least one of the system environment on which the trained model is executed and the size of the hospital.
5. A medical information processing apparatus according to any one of claims 1 to 4, further comprising a display control unit that displays information of a trained model relating to an analysis process in progress.
6. The medical information processing apparatus according to claim 5, further comprising a display control unit that displays information of a selected trained model for each of the aforementioned analysis processes.
7. The computer further comprises a calculation unit that causes the computer to calculate the processing time and estimation accuracy expected for the trained model, The medical information processing apparatus according to claim 5 or 6, wherein the display control unit displays the trained model in association with at least one of the processing time and the estimation accuracy.
8. The medical information processing apparatus according to any one of claims 1 to 7, further comprising a training unit that causes the computer to train a network model using as input data information about a trained model included in the combination, the patient information, the environmental information, and the target data, and as ground truth data feedback information regarding user feedback during the execution of analysis processing in the combination and on at least one of the analysis results, thereby generating a trained configuration model.
9. The medical information processing apparatus according to claim 8, further comprising an execution unit that inputs patient information, environmental information, and target data to be processed into the trained setting model, thereby outputting a combination of the trained model.
10. A medical information processing method performed by a computer, The acquisition means acquires patient information about the target patient, environmental information about the execution environment, and target data about the said target patient. The decision means determines, from a plurality of trained models with different speeds and accuracy, a combination of trained models that satisfies speed or accuracy conditions according to the patient information and the environmental information, for application to the target data. Medical information processing method.
11. On the computer, An acquisition function that acquires patient information about the target patient, environmental information about the execution environment, and target data about the said target patient, A decision function that determines a combination of trained models from multiple trained models with different speeds and accuracy stored in internal memory or an external device, which satisfies speed or accuracy conditions according to the patient information and environmental information, for application to the target data. A medical information processing program that makes this possible.