Medical image processing device and medical image processing program

Abstract

To provide a medical image processing device capable of causing a result of inference using a learning model to become closer to a user's desire when performing inference for a medical image using a learning model.SOLUTION: A learning model 30 performs learning for outputting output data indicating an inference result on an ultrasonic image on the basis of input data indicating the ultrasonic image, which is a medical image. An inference parameter changing part 36 changes inference parameters, which are parameters to be adjusted in learning processing of the learning model 30, and parameters that influence output data, according to a user's instructions. An inference processing part 34 outputs an inference result on the object ultrasonic image on the basis of output data when inputting the object ultrasonic image to be subjected to inference processing to the learning model 30 having inference parameters after the change.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] This specification discloses an improvement to a medical image processing device and a medical image processing program. [Background technology]

[0002] Medical images, which are images showing the internal structure and function of a subject, are used mainly for the diagnosis and treatment of diseases. Examples of medical images include X-ray images, computed tomography (CT) images, magnetic resonance imaging (MRI), and ultrasound images. On the other hand, (machine) learning models are known in which internal parameters are adjusted by performing learning processing using learning data. A learning model that has been sufficiently trained becomes able to predict (infer) data corresponding to each of various input data with high accuracy and output the inference results as output data.

[0003] Conventionally, it has been proposed to incorporate a learning model into a medical image processing device for processing medical images.

[0004] For example, Patent Document 1 discloses an ultrasound image forming device that constructs a trained model that learns the correlation between ultrasound image data and an evaluation value of the ultrasound image data, and changes parameters related to the acquisition of ultrasound image data when the evaluation value output by the trained model using the ultrasound image data to be processed as input is less than a threshold value. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] Patent Publication No. 2020-92739 Summary of the Invention [Problem to be solved by the invention]

[0006] Now, consider training a learning model to output output data showing an inference result regarding a medical image. One example of a learning model that outputs an inference result regarding a medical image is a learning model that infers the contour (lumen) of the heart when the heart is shown in the medical image.

[0007] In this case, the learning model is trained using a large number of medical images as training data, and its internal parameters (referred to as "inference parameters" in this specification) are set. However, the inference results of the trained learning model may not completely match the results desired by a particular user. For example, in the above example, a user may have wanted to know the outline of the heart's contour roughly, but the learning model may have output the outline of the heart in great detail, resulting in a loss of smoothness.

[0008] The purpose of the medical image processing device disclosed in this specification is to provide a medical image processing device that, when performing inference on medical images using a learning model, is capable of bringing the inference results using the learning model closer to the user's needs. [Means for solving the problem]

[0009] The medical image processing device disclosed in this specification is characterized in that it comprises a learning model trained to output output data indicating an inference result regarding a medical image based on input data representing the medical image, an inference parameter change unit that changes inference parameters, which are parameters to be adjusted in the learning process of the learning model and which affect the output data, in response to a user's instructions, and an inference processing unit that outputs an inference result regarding a target medical image based on output data when a target medical image, which is a medical image to be subjected to inference processing, is input into the learning model having the changed inference parameters.

[0010] The system may further include a display control unit that displays the target medical image and the inference results related to the target medical image on a display, and changes the display content of the inference results on the display in response to changes in the inference parameters.

[0011] When a user modifies an inference result regarding the target medical image displayed on the display, the inference parameter modification unit may modify the inference parameters based on the modified inference result.

[0012] The system further includes a parameter change history database that stores the inference parameters after they have been changed in accordance with user instructions, and when performing inference processing on a target medical image in accordance with user instructions, the inference parameter change unit changes the inference parameters based on the changed inference parameters stored in the parameter change history database, regardless of instructions from the user.

[0013] The parameter change history database stores in association with a user identifier that identifies the user who instructed the change to the inference parameter, the inference parameter after the change in accordance with the user's instruction, and when the inference parameter related to the user is stored in the parameter change history database, the inference parameter change unit changes the inference parameter based on the changed inference parameter associated with the user identifier of the user in the parameter change history database, regardless of the instruction from the user, when performing inference processing on the target medical image in accordance with the user's instruction.

[0014] The medical image is an ultrasound image formed by irradiating ultrasound toward the subject's heart, and the inference processing unit outputs information to be used for measuring the heart using the ultrasound image, which is the target medical image, as an inference result for the target medical image.

[0015] In addition, the medical image processing program disclosed in this specification is characterized in that it causes a computer to function as: a learning model trained to output output data indicating an inference result regarding a medical image based on input data representing the medical image; an inference parameter change unit that changes inference parameters, which are parameters to be adjusted in the learning process of the learning model and which affect the output data, in response to a user's instructions; and an inference processing unit that outputs an inference result regarding a target medical image based on output data when a target medical image, which is a medical image to be subjected to inference processing, is input into the learning model having the changed inference parameters. Effect of the Invention

[0016] According to the medical image processing device disclosed in this specification, when performing inference on medical images using a learning model, it is possible to bring the inference results using the learning model closer to the user's needs. [Brief description of the drawings]

[0017] [Figure 1] 1 is a schematic diagram illustrating the configuration of an ultrasonic diagnostic apparatus according to an embodiment of the present invention. [Diagram 2] FIG. 13 is a diagram showing an example display of an inference result regarding a target ultrasound image. [Diagram 3] FIG. 13 is a diagram showing an example of an inference parameter change screen. [Figure 4] A figure showing an example of the display of inference results using changed inference parameters. [Diagram 5] FIG. 13 is a conceptual diagram showing an example of the contents of a parameter change history DB. [Figure 6] 4 is a flowchart showing a process flow of the ultrasound diagnostic apparatus according to the present embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0018] 1 is a schematic diagram of the configuration of an ultrasound diagnostic apparatus 10 as a medical image processing apparatus according to this embodiment. The ultrasound diagnostic apparatus 10 is a medical apparatus installed in a medical institution such as a hospital.

[0019] The ultrasonic diagnostic device 10 is a device that scans an ultrasonic beam on a subject and generates an ultrasonic image as a medical image based on the received signal obtained by the scanning. In particular, in this embodiment, the ultrasonic diagnostic device 10 forms an ultrasonic tomographic image (B-mode image) in which the amplitude intensity of the reflected wave from the scanning surface is converted into brightness based on the received signal. Note that the ultrasonic diagnostic device 10 can also form other ultrasonic images, such as a Doppler image that represents the motion speed of tissue in the subject and is formed based on the difference (Doppler shift) between the frequencies of the transmitted wave and the received wave.

[0020] The transmission / reception unit 14, the signal processing unit 16, the image forming unit 18, the display control unit 20, the inference processing unit 34, and the inference parameter changing unit 36 ​​of the ultrasound diagnostic device 10 are configured by a processor. The processor includes at least one of a general-purpose processing device (e.g., a central processing unit (CPU)) and a dedicated processing device (e.g., a graphics processing unit (GPU), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA), or a programmable logic device). The processor may not be a single processing device, but may be configured by the cooperation of multiple processing devices located at physically separate locations. Furthermore, each of the above units may be realized by the cooperation of hardware such as a processor and software.

[0021] The ultrasonic probe 12 is a device that transmits and receives ultrasonic waves to and from a subject. The ultrasonic probe 12 has a transducer array that includes a plurality of transducer elements that transmit and receive ultrasonic waves to and from the subject.

[0022] The transmitting / receiving unit 14 transmits a transmission signal to the ultrasonic probe 12 (specifically, each transducer element of the transducer element array) under the control of a control unit 26 (described later). As a result, ultrasonic waves are transmitted from each transducer element to the subject.

[0023] The transmitting / receiving unit 14 also receives a reception signal from each transducer element that has received a reflected wave from the subject. The transmitting / receiving unit 14 has an adder and a plurality of delays corresponding to each transducer element, and performs a phasing addition process in which the adder and the plurality of delays align and add the phases of the reception signals from each transducer element. This forms a reception beam signal in which information indicating the signal strength of the reflected wave from the subject is aligned in the depth direction of the subject.

[0024] The signal processing unit 16 performs various signal processing on the received beam signal from the transmitting / receiving unit 14, including filtering using a bandpass filter and detection processing.

[0025] The image forming unit 18 forms an ultrasonic tomographic image (B-mode image) based on the received beam signal that has been signal-processed by the signal processing unit 16. First, the image forming unit 18 converts the received beam signal into data on the coordinate space of an ultrasonic image. Then, the image forming unit 18 forms an ultrasonic tomographic image based on the coordinate conversion signal.

[0026] The display control unit 20 controls the display of the ultrasonic tomographic image formed by the image forming unit 18 and various other information on the display 22. The display 22 is a display device constituted by, for example, a liquid crystal display or an organic EL (Electro Luminescence) display.

[0027] The input interface 24 is composed of, for example, a button, a trackball, a touch panel, etc. The input interface 24 is used to input a user's command to the ultrasound diagnostic apparatus 10. In this embodiment, the display 22 is a touch panel, and the display 22 also functions as the input interface 24.

[0028] The control unit 26 includes at least one of a general-purpose processor (e.g., a CPU (Central Processing Unit)) and a dedicated processor (e.g., a GPU (Graphics Processing Unit), an ASIC (Application Specific Integrated Circuit), an FPGA (Field-Programmable Gate Array), or a programmable logic device). The control unit 26 may not be a single processing device, but may be configured by the cooperation of multiple processing devices located at physically separate locations. The control unit 26 controls each part of the ultrasound diagnostic apparatus 10 according to a medical image processing program stored in a memory 28 described later.

[0029] The memory 28 includes a hard disk drive (HDD), a solid state drive (SSD), an embedded multi media card (eMMC), a read only memory (ROM), or the like. A medical image processing program for operating each unit of the ultrasound diagnostic apparatus 10 is stored in the memory 28. The medical image processing program can also be stored in a computer-readable non-transitory storage medium such as a universal serial bus (USB) memory or a CD-ROM. The ultrasound diagnostic apparatus 10 can read and execute the medical image processing program from such a storage medium.

[0030] As shown in FIG. 1, the memory 28 stores a learning model 30. The learning model 30 receives data representing a medical image as input data, performs inference on the medical image related to the input data, and outputs the inference result. In this specification, an example is described in which the learning model 30 receives an ultrasonic tomographic image representing a cross section of the heart as input, and outputs information used for measuring the heart using the ultrasonic tomographic image as an inference result. In particular, in this specification, an example is described in which the learning model 30 infers the contour line of the heart and characteristic positions of the heart (e.g., the apex position and the annulus position) as information used for measuring the heart, and outputs the inference result. The input data representing the medical image input to the learning model 30 is not limited to the medical image itself. For example, when the medical image is an ultrasonic image, it may be a received beam signal before the ultrasonic image is formed. In addition, the medical image is not limited to an ultrasonic image, and may be, for example, an X-ray image, a CT image, or an MRI. Furthermore, the inference content of the learning model 30 may be any content as long as it is related to the input medical image.

[0031] The learning model 30 is, for example, a model that uses an ASM (Active Shape Model). The ASM is a model that expresses a target shape (in this embodiment, the contour of the heart) as a set of points. The ASM incorporated in the learning model 30 in this embodiment calculates the average shape of multiple heart contours as learning data, and performs principal component analysis on the multiple heart contours to obtain multiple combinations of eigenvalues ​​and eigenvectors of the contours. Then, a weight is added to each eigenvector. The shape of the heart contour changes by changing the weight of each eigenvector. For example, by increasing the weight of a high-ranked principal component (e.g., the first principal component) and decreasing the weight of a low-ranked principal component, the learning model 30 infers the rough shape of the heart contour (in other words, the heart contour becomes smooth). On the other hand, by decreasing the weight of a high-ranked principal component and increasing the weight of a low-ranked principal component, the learning model 30 infers the details of the heart contour while emphasizing them. Thus, the weight of each eigenvector is a parameter that influences the output data of the learning model 30 .

[0032] The weight of each eigenvector in the ASM is a parameter to be adjusted in the learning process of the learning model 30. The learning model 30 uses a combination of an ultrasound image showing the heart and the contour of the heart in the ultrasound image as learning data, and adjusts inference parameters including the weight of each eigenvector in the ASM described above so that the contour and feature position of the heart shown in the ultrasound image can be inferred. Note that the inference parameter in this specification is a parameter to be adjusted in the learning process of the learning model 30, and means a parameter that affects the inference result of the learning model 30 (i.e., the output data of the learning model 30). The learning model 30 that has been sufficiently trained becomes able to output the contour and feature position of the heart in the ultrasound image with high accuracy based on the ultrasound image showing the heart.

[0033] 1, the memory 28 stores a parameter change history DB (DataBase) 32. Details of the parameter change history DB 32 and how it is used will be described later together with details of the inference parameter change unit 36.

[0034] The inference processing unit 34 performs inference processing on a target medical image (a target ultrasound image in this embodiment) which is a medical image that is the target of the inference processing, and outputs the inference result. Specifically, the inference processing unit 34 inputs the target ultrasound image to the trained learning model 30, and outputs the inference result on the target ultrasound image based on the output data of the learning model 30 for the input data. In this embodiment, as described above, the inference processing unit 34 infers the contour line and feature positions of the heart for the target ultrasound image showing the heart, and outputs the inference result.

[0035] The display control unit 20 causes the inference results of the inference processing unit 34 regarding the target ultrasound image to be displayed on the display 22. Fig. 2 is a diagram showing an example of a display of the inference results regarding the target ultrasound image USI. The target ultrasound image USI shown in Fig. 2 shows a four-chamber cross section of the heart, and the contour line OL and feature position P of the heart (particularly the left ventricle) are shown as the inference results regarding the target ultrasound image USI.

[0036] The control unit 26 may have a function as a measurement unit, and may perform measurements related to the heart shown in the target ultrasound image USI based on at least one of the contour line OL or the characteristic position P inferred by the inference processing unit 34. For example, the control unit 26 may measure the end-diastolic long axis length, end-systolic long axis length, end-diastolic volume, end-systolic volume, ejection fraction, etc. of the heart. The display control unit 20 may display the measurement results R of the control unit 26 on the display 22 together with the inference results of the inference processing unit 34.

[0037] The inference parameter change unit 36 ​​changes the inference parameters of the learning model 30 in response to a user's instruction. As described above, the inference parameters of the learning model 30 that performs inference regarding the target ultrasound image are adjusted by the learning process, but in this embodiment, the inference parameter change unit 36 ​​can change at least a part of the inference parameters of the learning model 30 in response to a user's instruction. In other words, the user can change at least a part of the inference parameters of the learning model 30. As described above, the inference parameters are parameters that affect the inference results of the learning model 30, so when the inference parameter change unit 36 ​​changes the inference parameters, the inference results of the learning model 30 will change.

[0038] 3 is a diagram showing an example of an inference parameter change screen. In this embodiment, the display control unit 20 causes a button B for the user to change the inference parameters to be displayed on the display 22, which is a touch panel, together with the inference result of the inference processing unit 34. The user operates the button B to input an instruction to change the inference parameters of the learning model 30 to the inference parameter change unit 36. The inference parameter change unit 36 ​​changes the inference parameters of the learning model 30 in response to the input change instruction.

[0039] The learning model 30 generally has a plurality of inference parameters, and it may be difficult for the user to grasp how each inference parameter affects the inference result of the learning model 30. Therefore, it is preferable to store in advance in the memory 28 inference parameter information that associates the inference parameters of the learning model 30 with the items of the inference result of the learning model 30 that are affected by the inference parameters, and to have the display control unit 20 associate the items included in the inference parameter information with the button B for changing the inference parameter corresponding to the item, and display them on the display 22.

[0040] For example, in the example of Fig. 3, the display control unit 20 displays a button B for the item "Smoothness (of the heart contour OL)", a button B for the item "Apex position", and a button B for the item "Outside / Inside". The button B for each item is configured as a set of a button B (in the example of Fig. 3, the button B indicated with a left-pointing arrow) for changing the inference parameter for the item in a first direction (e.g., a direction to decrease) and a button B (in the example of Fig. 3, the button B indicated with a right-pointing arrow) for changing the inference parameter for the item in a second direction opposite to the first direction (e.g., a direction to increase).

[0041] Furthermore, the display control unit 20 may display an index value representing the current inference parameter value for each item (in the example of FIG. 3, the index value for each item is "5"). The index value may take a value from 0 to 10, for example, and the value for each item determined by the learning process is set to the intermediate value "5".

[0042] For example, the inference parameter for the item "smoothness" is a set of weights for each eigenvector in the ASM. For example, the closer the user specifies the index value for the item "smoothness" to "0", the more the inference parameter change unit 36 ​​increases the weights of the principal components with higher ranks and decreases the weights of the principal components with lower ranks for the weights of each eigenvector in the ASM, and the closer the user specifies the index value for the item "smoothness" to "10", the more the inference parameter change unit 36 ​​decreases the weights of the principal components with higher ranks and increases the weights of the principal components with lower ranks for the weights of each eigenvector in the ASM.

[0043] The item "outside / inside" is an adjustment item for the position of the heart contour line OL. For example, the closer to "0" the index value of the item "outside / inside" is set, the more the contour line OL moves inward in the direction of decreasing the volume of the heart (the left ventricle in the example of FIG. 3), and the closer to "10" the index value of the item "outside / inside" is set, the more the contour line OL moves outward in the direction of increasing the volume of the heart.

[0044] The inference parameters are parameters adjusted by a learning process so that inference results regarding medical images (ultrasound images in this embodiment) can be output with high accuracy. Therefore, if the inference parameters are changed significantly, there is a risk that the learning model 30 will not be able to output inference results regarding medical images with high accuracy. Therefore, it is preferable that the inference parameter change unit 36 ​​limits the amount by which the inference parameters can be changed by a user instruction to a predetermined amount so that the inference accuracy of the learning model 30 is maintained at a certain level.

[0045] The inference processing unit 34 again inputs the target ultrasound image to the learning model 30 having the inference parameters changed by the inference parameter changing unit 36. Then, the inference processing unit 34 outputs an inference result regarding the target ultrasound image based on the output data of the learning model 30 for the input data.

[0046] The display control unit 20 changes the display content of the inference result of the inference processing unit 34 on the display 22 in response to the change in the inference parameters. Fig. 4 is a diagram showing an example of the display of the inference result using the changed inference parameters. This allows the user to get the inference result about the medical image using the learning model 30 closer to his / her own desire (preference).

[0047] Each time the user operates button B and the inference processing unit 34 performs inference processing based on the changed inference parameters, the display control unit 20 may cause the inference result based on the changed inference parameters to be displayed on the display 22. This allows the user to easily obtain a preferred inference result while fine-tuning the inference parameters.

[0048] The user may be able to directly modify the inference results for the medical image (in this example, the contour line OL and the feature position P) instead of operating button B. In this case, the inference parameter modification unit 36 ​​may modify the inference parameters of the learning model 30 based on the modified inference results. For example, when the user manually modifies the shape of the contour line OL, the inference parameter modification unit 36 ​​quantifies the smoothness of the contour line OL, and modifies the inference parameter corresponding to the item "smoothness" (i.e., the weight of each eigenvector in the ASM) based on this.

[0049] When the inference parameter change unit 36 ​​changes an inference parameter in response to a user instruction, the inference parameter change unit 36 ​​may store the changed inference parameter in the parameter change history DB 32. In particular, when the inference parameter change unit 36 ​​changes an inference parameter in response to a user instruction, the inference parameter change unit 36 ​​may store in the parameter change history DB 32 a user identifier that uniquely identifies the user who instructed the change of the inference parameter and the changed inference parameter in association with each other.

[0050] 5 is a conceptual diagram showing an example of the contents of the parameter change history DB 32. In this embodiment, the parameter change history DB 32 stores a user ID as a user identifier, an inference parameter (item in FIG. 3 or 4), and an index value designated by the user for the item, in association with each other. Note that the user ID of the user can be acquired by the control unit 26 authenticating the user by a method such as a password before the user starts using the ultrasound diagnostic device 10.

[0051] When performing an inference process on the target ultrasound image USI (see FIG. 3 or FIG. 4) in response to an instruction from a user, the inference parameter change unit 36 ​​may change the inference parameters of the learning model 30 based on the changed inference parameters stored in the parameter change history DB 32 without depending on the instruction from the user. For example, when the contents of the parameter change history DB 32 are as shown in FIG. 5, the inference parameter change unit 36 ​​determines the inference parameters corresponding to the item "smoothness" of the learning model 30 in the current inference without depending on an instruction from the user, in other words automatically, based on, for example, index values ​​"3" and "5" for the item "smoothness" set by the user in the past. For example, the inference parameter change unit 36 ​​first calculates a representative value (for example, an average value) of multiple values ​​for the item "smoothness" stored in the parameter change history DB 32. Here, "4", which is the average value of the multiple values ​​"3" and "5" for the item "smoothness", is calculated. Then, the inference parameter change unit 36 ​​changes the inference parameters of the learning model 30 to a value corresponding to the calculated index value "4". The inference parameter change unit 36 ​​performs similar processing for the other items "apex position" and "outside / inside".

[0052] In particular, when the parameter change history DB 32 stores the user ID and the changed inference parameters in association with each other, the inference parameter change unit 36 ​​may change the inference parameters of the learning model 30 based on the changed inference parameters associated with the user ID of the user in the parameter change history DB 32, without depending on the instruction from the user, when performing inference processing on the target medical image in response to an instruction from the user. For example, when the content of the parameter change history DB 32 is as shown in FIG. 5, the inference parameter change unit 36 ​​changes the inference parameters of the learning model 30 to values ​​corresponding to the index value "3" for the item "apex position", the index value "8" for the item "outside / inside", without depending on the instruction from the user, when performing inference processing on the target ultrasound image USI in response to an instruction from the user indicated by the user ID "U0001". On the other hand, when performing inference processing on the target ultrasound image USI in response to instructions from the user indicated by user ID "U0002", the inference parameter change unit 36 ​​changes the inference parameters of the learning model 30 to values ​​corresponding to the index value "5" for the item "smoothness", the index value "6" for the item "apex position", and the index value "6" for the item "outside / inside", regardless of instructions from the user.

[0053] The outline of the configuration of the ultrasound diagnostic device 10 according to this embodiment has been described above. The flow of processing by the ultrasound diagnostic device 10 will be described below with reference to the flowchart shown in Fig. 6. It is assumed that the learning model 30 has already been trained at the start of the flowchart shown in Fig. 6.

[0054] In step S10, the control unit 26 authenticates the user and identifies the user ID of the user.

[0055] In step S12, the image forming unit 18 forms an ultrasonic tomographic image based on the received beam signals that have been signal-processed by the signal processing unit 16. It is assumed that the ultrasonic tomographic image shows a four-chamber cross section of the heart.

[0056] In step S14, the type of inference is set by the user. In this example, it is assumed that inference of the heart contour and feature positions is set.

[0057] In step S16, the inference processing unit 34 identifies the frame to be processed (target ultrasound image).

[0058] In step S18, the inference parameter change unit 36 ​​determines whether or not data related to the user ID identified in step S10 is stored in the parameter change history DB 32. If data related to the user ID is stored in the parameter change history DB 32, the process proceeds to step S20, and if data related to the user ID is not stored, the process bypasses step S20 and proceeds to step S22.

[0059] In step S20, the inference parameter change unit 36 ​​changes the inference parameters of the learning model 30 based on the changed inference parameters associated with the user ID of the user in the parameter change history DB 32.

[0060] In step S22, the inference processing unit 34 inputs the target ultrasound image to the learning model 30, and outputs an inference result regarding the target ultrasound image based on the output data of the learning model 30 for the input data.

[0061] In step S24, the display control unit 20 causes the display 22 to display the inference result of the inference processing unit 34 regarding the target ultrasound image.

[0062] In step S26, the inference parameter change unit 36 ​​determines whether or not an instruction to change the inference parameters of the learning model 30 has been received from the user. If an instruction to change the inference parameters has not been received, the process ends, and if an instruction to change the inference parameters has been received, the process proceeds to step S28.

[0063] In step S28, the inference parameter change unit 36 ​​changes the inference parameters of the learning model 30 in response to the user instruction received in step S26. At the same time, the inference parameter change unit 36 ​​associates the user ID of the user with the changed inference parameters and stores them in the parameter change history DB 32.

[0064] In step S22 after step S28, the inference processing unit 34 inputs the target ultrasound image to the learning model 30 having the inference parameters changed in step S28, and outputs an inference result regarding the target ultrasound image based on the output data of the learning model 30 for the input data.

[0065] In step S24 again, the display control unit 20 changes the display content of the inference result of the inference processing unit 34 on the display 22 in response to the change in the inference parameters.

[0066] Every time the user issues an instruction to change the inference parameters, the processes of steps S22 to S28 are repeated.

[0067] Although the embodiment of the present invention has been described above, the present invention is not limited to the above embodiment, and various modifications are possible without departing from the spirit of the present invention.

[0068] For example, in the above embodiment, the medical image processing device is the ultrasound diagnostic device 10, but the medical image processing device is not limited to the ultrasound diagnostic device 10. For example, the medical image processing device may be a personal computer (PC (Personal Computer)) or a server. In this case, the PC or server as the medical image processing device has a display control unit 20, a display 22, an input interface 24, a control unit 26, a memory 28, an inference processing unit 34, and an inference parameter change unit 36, and the memory 28 stores a learning model 30 and a parameter change history DB 32. The PC or server as the medical image processing device acquires a target medical image from an X-ray device, a CT device, an MRI device, an ultrasound diagnostic device, or the like, and performs inference processing on the acquired target medical image. [Explanation of symbols]

[0069] 10 ultrasound diagnostic device, 12 ultrasound probe, 14 transmitter / receiver unit, 16 signal processing unit, 18 image formation unit, 20 display control unit, 22 display, 24 input interface, 26 control unit, 28 memory, 30 learning model, 32 parameter change history DB, 34 inference processing unit, 36 inference parameter change unit.

Claims

1. A learning model that is trained to output output data indicating an inference result regarding a medical image based on input data representing the medical image; an inference parameter change unit that changes an inference parameter, which is a parameter to be adjusted in a learning process of the learning model and which has an effect on the output data, in response to a user's instruction; an inference processing unit that outputs an inference result regarding a target medical image, which is a medical image to be inferred, based on output data when the target medical image, which is a medical image to be inferred, is input to the learning model having the changed inference parameters; A medical image processing apparatus comprising:

2. a display control unit that displays a target medical image and an inference result related to the target medical image on a display and changes the display content of the inference result on the display in response to a change in the inference parameter; Further comprising:

2. The medical image processing apparatus according to claim 1.

3. the inference parameter change unit changes the inference parameters based on the inference result after a user makes corrections to the inference result related to the target medical image displayed on the display.

3. The medical image processing apparatus according to claim 2.

4. a parameter change history database for storing the inference parameters after the inference parameters have been changed in response to a user's instruction; Further equipped with the inference parameter change unit, when performing inference processing on a target medical image in response to an instruction from a user, changes the inference parameters based on the changed inference parameters stored in the parameter change history database, without depending on the instruction from the user; 2. The medical image processing apparatus according to claim 1.

5. the parameter change history database stores, in association with each other, a user identifier for identifying a user who has instructed a change to the inference parameter and the inference parameter after the change in response to the instruction of the user; When the inference parameters related to the user are stored in the parameter change history database, the inference parameter change unit changes the inference parameters based on the changed inference parameters associated with the user's user identifier in the parameter change history database, regardless of instructions from the user, when performing inference processing on a target medical image in response to instructions from the user.

5. The medical image processing apparatus according to claim 4.

6. the medical image is an ultrasound image formed by irradiating an ultrasound wave toward a heart of a subject; the inference processing unit outputs, as an inference result for the target medical image, information to be used for measuring the heart using an ultrasound image which is the target medical image.

2. The medical image processing apparatus according to claim 1.

7. Computer, A learning model that is trained to output output data indicating an inference result regarding a medical image based on input data representing the medical image; an inference parameter change unit that changes an inference parameter, which is a parameter to be adjusted in a learning process of the learning model and which has an effect on the output data, in response to a user's instruction; an inference processing unit that outputs an inference result regarding a target medical image, which is a medical image to be inferred, based on output data when the target medical image, which is a medical image to be inferred, is input to the learning model having the changed inference parameters; A medical image processing program characterized by causing the program to function as follows.

Images