JP6722322B1 - Ultrasonic device and its control program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an ultrasonic device capable of acquiring a large number of pair images for constructing a learned model by a method with the least effort. A function of a processor of an ultrasonic diagnostic apparatus to acquire data of a first ultrasonic image and data of a second ultrasonic image whose image quality is deteriorated as one pair used as learning data for machine learning. This function, after reading the first condition from the memory, transmits and receives the first ultrasonic wave and creates a first ultrasonic image based on the echo signal obtained thereby. The first function of displaying and storing the first ultrasonic image by using the first condition and the second condition after reading the second condition from the memory And a second function of creating a second ultrasonic image based on the echo signal using the second condition and further storing the second ultrasonic image in the memory. [Selection diagram] Figure 2

Description

The present invention relates to an ultrasonic device that acquires an ultrasonic image of a subject and a control program therefor.

The ultrasonic device transmits an ultrasonic wave to the subject from the ultrasonic probe and acquires an ultrasonic image based on the echo signal. For example, in an ultrasonic diagnostic apparatus which is an example of an ultrasonic apparatus, various diagnoses are performed using ultrasonic images.

There are various conditions such as the transmission frequency, the transmission sound pressure, the depth, the gain, and the dynamic range, which are used for acquiring the ultrasonic image (see, for example, Patent Document 1). These conditions are set so that an ultrasonic image with a quality as good as possible can be obtained.

JP, 2017-143973, A

Here, when an ultrasonic image of another subject is acquired under the condition that an ultrasonic image of good image quality is obtained for a certain subject, the image quality may deteriorate. For example, the greater the subcutaneous fat thickness, the greater the attenuation of ultrasonic waves, and the lower the SN ratio of the acquired ultrasonic image, particularly in the deep region. Therefore, if an ultrasonic image is acquired for another subject having a thicker subcutaneous fat thickness under the condition that an ultrasonic image of good image quality is acquired for a certain subject, the image quality may deteriorate. In this case, for example, the operator lowers the transmission frequency of ultrasonic waves and acquires an ultrasonic image in which the SN ratio of the deep region is improved.

However, if the transmission frequency is lowered, the spatial resolution of the ultrasonic image deteriorates. Further, the operator needs to change the conditions, which is complicated. In particular, it may be difficult for an operator who is unfamiliar with the operation. Moreover, when the image quality of the ultrasonic image displayed after the inspection is deteriorated, the ultrasonic image cannot be acquired again. Therefore, the inventor of the present application is studying a method of improving image quality using machine learning in order to more easily acquire an ultrasonic image with better image quality.

The general principle will be described. First, machine learning requires input data and teacher data. Here, the input data is a tomographic image (hereinafter referred to as “low image quality image”) including noise and artifacts, and the teacher data is an image obtained by capturing the same region of interest as the low image quality image. It is an ideal image that is not included (hereinafter referred to as "high-quality image"). After learning by using the data of the low quality image and the data of the high quality image to build a learned model, if another newly obtained low quality image is input to this learned model, it is possible to obtain a high quality image. As a result, a virtual high quality image as if the image was taken is obtained as an output. According to such a method, it is possible to acquire an ultrasonic image with better image quality without changing the conditions.

By the way, in this machine learning, it is desirable to acquire as many pairs of "low-quality images" and "high-quality images" as possible during learning. Therefore, it is an issue to acquire a large number of pair images for constructing a trained model by a method with the least effort.

One aspect of the invention made to solve the above problems is an ultrasonic device including a memory, a user interface, a processor, an ultrasonic probe, and a display, wherein the memory is a first ultrasonic image of a subject. The first condition used to acquire the second ultrasonic image used to acquire the second ultrasonic image for the same position as the position where the first ultrasonic image is acquired in the subject. In the condition, the second ultrasonic image is configured to store a second condition that is a condition in which the image quality is deteriorated compared to the first ultrasonic image, and the user interface is stored. When the user interface receives an input by the operator, the processor is configured to receive the input of the operator, and machine-learns the data of the first ultrasonic image and the data of the second ultrasonic image. Is configured to execute a function of acquiring as one pair to be used as learning data of, and the function reads and transmits the first ultrasonic wave by the ultrasonic probe after reading the first condition from the memory. , Creating the first ultrasonic image based on the echo signal obtained by transmitting and receiving the first ultrasonic wave using the first condition, and further displaying the first ultrasonic image on the display. A first function of displaying and storing the data of the first ultrasonic image in the memory, and transmitting and receiving a second ultrasonic wave by the ultrasonic probe after reading the second condition from the memory And, the creation of the second ultrasonic image based on the echo signal obtained by the transmission and reception of the second ultrasonic is performed using the second condition, further data of the second ultrasonic image An ultrasonic device including a second function of storing in the memory.

According to the invention of the above aspect, when the user interface receives an input from the operator, the first ultrasonic image displayed on the display as one pair used as learning data for machine learning and the second ultrasonic image are displayed. An ultrasonic image is acquired and the data of these first and second ultrasonic images are stored. The second ultrasonic image is an image acquired using the second condition, which is a condition in which the image quality is lower than that of the first ultrasonic image. Therefore, for example, after acquiring the first ultrasonic image acquired for the purpose of diagnosing the subject, a second ultrasonic image forming a pair with this as the learning data for machine learning can be acquired. This makes it possible to acquire the paired images for constructing the trained model with the least effort.

It is a block diagram showing composition of an ultrasonic diagnostic equipment in a first embodiment of the present invention. 6 is a flowchart showing a process for acquiring learning data for machine learning in the embodiment. It is a flowchart which shows the process of the image quality improvement of the ultrasonic image using the learned model. 9 is a flowchart showing a process for acquiring learning data for machine learning in a modified example of the embodiment. It is a block diagram which shows the outline of the system of 2nd embodiment. It is a block diagram which shows the structure of the ultrasonic diagnosing device in 2nd embodiment of this invention. It is a block diagram which shows the structure of the server in 2nd embodiment. It is a flow chart which shows operation of a second embodiment.

Embodiments of the present invention will be described below with reference to the drawings. In the following embodiments, as an example of the ultrasonic apparatus according to the present invention, an ultrasonic diagnostic apparatus that displays an ultrasonic image of a subject for the purpose of diagnosis will be described.
(First embodiment)
First, the first embodiment will be described. The ultrasonic diagnostic apparatus 1 shown in FIG. 1 includes a transmission beamformer 3 and a transmitter 4 that drive elements 2a arranged in an ultrasonic probe 2 and emit a pulse ultrasonic signal to a subject (not shown). Including. The pulsed ultrasonic signal generates an echo that is reflected in the subject and returned to the element 2a. The echo is converted into an electric signal by the element 2a, and the electric signal is received by the receiver 5. An electric signal representing the received echo, that is, an echo signal is input to the reception beamformer 6, and reception beamforming is performed in the reception beamformer 6. The reception beam former 6 outputs ultrasonic wave data after reception beam forming.

The receive beamformer 6 may be a hardware beamformer or a software beamformer. If the receive beamformer 6 is a software beamformer, the receive beamformer 6 may perform a graphics processing unit (GPU), a microprocessor, a central processing unit (CPU), a digital signal processor (DSP), or a logical operation. May include one or more processors, including any one or more of other types of processors capable of The processor configuring the reception beamformer 6 may be configured by a processor different from the processor 7 described later, or may be configured by the processor 7.

The ultrasonic probe 2 can include an electric circuit for performing all or a part of the transmit beamforming and/or the receive beamforming. For example, all or a part of the transmission beamformer 3, the transmitter 4, the receiver 5, and the reception beamformer 6 may be provided in the ultrasonic probe 2.

The ultrasonic diagnostic apparatus 1 also includes a processor 7 for controlling the transmit beamformer 3, the transmitter 4, the receiver 5, and the receive beamformer 6. The processor 7 is in electronic communication with the ultrasound probe 2. The processor 7 can control the ultrasonic probe 2 to acquire ultrasonic data. The processor 7 controls which of the elements 2 a are active and the shape of the beam emitted from the ultrasound probe 2. Processor 7 is also in electronic communication with display 8, which can process the ultrasound data into an ultrasound image for display on display 8. The term "electronic communication" can be defined to include both wireline and wireless communication. The processor 7 may include a central processing unit (CPU) according to one embodiment. According to other embodiments, the processor 7 may perform other processing functions such as a digital signal processor, a field programmable gate array (FPGA), a graphics processing unit (GPU), or other type of processor. Electronic components can be included. According to other embodiments, the processor 7 may include multiple electronic components capable of performing processing functions. For example, the processor 7 may include two or more electronic components selected from a list of electronic components including a central processing unit, a digital signal processor, a field programmable gate array, and a graphics processing unit.

The processor 7 may also include a composite demodulator (not shown) that demodulates the RF data and produces raw data. In another embodiment, demodulation can be performed early in the processing chain.

The processor 7 is configured to perform on the data one or more processing operations according to a plurality of selectable ultrasound modalities. When the echo signal is received, the data can be processed in real time during the scanning session. For the purposes of this disclosure, the term "real time" is defined to include procedures that occur without any intentional delay.

Also, the data can be temporarily stored in a buffer (not shown) during ultrasound scanning and processed in real-time or off-line operation rather than real-time. In this disclosure, the term "data" may be used in this disclosure to refer to one or more data sets acquired with an ultrasound device.

The ultrasound data may be processed by the processor 7 in other or different mode-related modules (eg, B mode, color Doppler, M mode, color M mode, spectral Doppler, elastography, TVI, distortion, strain rate, etc.). Ultrasonic image data can be created. For example, one or more modules generate ultrasound images of B-mode, color Doppler, M-mode, color M-mode, spectral Doppler, elastography, TVI, distortion, strain rate, and combinations thereof. You can The image beam and/or image frame may be saved and may record timing information indicating when the data was acquired in memory. The module may include, for example, a scan conversion module that performs scan conversion operations to convert an image frame from coordinate beam space to display space coordinates. A video processor module may be provided that reads the image frames from the memory while the procedure is being performed on the subject and displays the image frames in real time. The video processor module can store the image frames in the image memory and the ultrasound image is read from the image memory and displayed on the display 8.

When the processor 7 includes a plurality of processors, a plurality of processors may be responsible for the above-described processing tasks that the processor 7 is responsible for. For example, a first processor can be used to demodulate and decimate an RF signal, and a second processor can be used to display an image after further processing the data.

Further, for example, when the reception beamformer 6 is a software beamformer, its processing function may be executed by a single processor or a plurality of processors.

The display 8 is an LED (Light Emitting Diode) display, an LCD (Liquid Crystal Display), an organic EL (Electro-Luminescence) display, or the like.

The memory 9 is any known data storage medium, including non-transitory storage media and transient storage media. The non-transitory storage medium is a non-volatile storage medium such as an HDD (Hard Disk Drive) or a ROM (Read Only Memory). The non-transitory storage medium may include a portable storage medium such as a CD (Compact Disk) or a DVD (Digital Versatile Disk). The program executed by the processor 7 is stored in a non-transitory storage medium.

The non-transitory storage medium forming the memory 9 has a first condition used to acquire a first ultrasonic image of the subject and a first ultrasonic image of the subject. The second condition used to acquire the second ultrasonic image for the same position as the recorded position is stored. The second condition is a condition in which the second ultrasonic image has an image quality deteriorated as compared with the first ultrasonic image. Details of the first condition and the second condition will be described later.

Furthermore, a machine learning algorithm is stored in the non-transitory storage medium that constitutes the memory 9.

The transient storage medium is a volatile storage medium such as a RAM (Random Access Memory).

The user interface 10 can receive an operator's input for controlling the operation of the ultrasonic apparatus 1. For example, the user interface 10 receives an instruction or information input from a user. The user interface 10 includes a keyboard, a hard key, a trackball, a rotary control, a softkey, and the like. The user interface 10 may include a touch screen that displays soft keys and the like.

Next, the operation of the ultrasonic diagnostic apparatus of this example will be described. First, acquisition of learning data for machine learning will be described based on the flowchart of FIG.

First, in step S1, the user interface 10 receives an operator's input. Here, the operator inputs an instruction to display an ultrasonic image.

In step S1, when the user interface 10 receives an input from the operator, in step S2 and subsequent steps, the processor 7 converts the data of the first ultrasonic image and the data of the second ultrasonic image into learning data for machine learning. The function to acquire as one pair is used. Specifically, first, in step S2, the processor 7 reads the first condition stored in the memory 9. The first condition includes a condition for transmitting/receiving the first ultrasonic wave, a condition for processing an echo signal obtained by transmitting/receiving the first ultrasonic wave, and a condition for image processing.

Next, in step S3, the processor 7 causes the ultrasonic probe 2 to transmit and receive the first ultrasonic wave using the conditions for transmitting and receiving the first ultrasonic wave. Next, in step S4, the processor 7 creates a first ultrasonic image based on the echo signal obtained by transmitting and receiving the first ultrasonic wave using the first condition. Further, the processor 7 displays the first ultrasonic image on the display 8 and stores the data of the first ultrasonic image in the memory 9. The functions of steps S1 to S4 are an example of the embodiment of the first function of the present invention.

For example, the first ultrasonic image is displayed for the operator to make a diagnosis on the subject.

Next, in step S5, the processor 7 reads the second condition stored in the memory 9. The second condition includes a second ultrasonic wave transmission/reception condition, an echo signal processing condition obtained by the second ultrasonic wave transmission/reception, and an image processing condition.

Here, the first condition and the second condition will be described. The first condition is a condition that is set so that the first ultrasonic image having the best possible image quality can be acquired in order to diagnose the subject. On the other hand, the second condition is a condition in which the image quality of the second ultrasonic image is lower than that of the first ultrasonic image. For example, the second condition may include a condition that the SN ratio of the second ultrasonic image is lower than the SN ratio of the first ultrasonic image. The second condition may include a condition that the depth sensitivity of the second ultrasonic image is lower than the depth sensitivity of the first ultrasonic image.

More specifically, the first condition and the second condition may include the transmission sound pressure of ultrasonic waves. In this case, the second transmission sound pressure under the second condition is lower than the first transmission sound pressure under the first condition.

Further, the first condition and the second condition may include the transmission frequency of ultrasonic waves. In this case, the second transmission frequency under the second condition is higher than the first transmission frequency under the first condition.

Furthermore, the first condition and the second condition may include a reception frequency band of ultrasonic waves. In this case, the second reception frequency band under the second condition is a narrower band than the first reception frequency band under the first condition, and the second center frequency in the second reception frequency band is the first Is higher than the first center frequency in the reception frequency band.

Next, in step S6, the processor 7 causes the ultrasonic probe 2 to transmit and receive the second ultrasonic wave using the condition of transmitting and receiving the second ultrasonic wave. For example, the processor 7 causes the second ultrasonic wave to be transmitted and received at the second transmission sound pressure lower than the first transmission sound pressure under the first condition. Further, the processor 7 may cause the second ultrasonic wave to be transmitted and received at a second transmission frequency higher than the first transmission frequency under the first condition. Further, the processor 7 may cause the second ultrasonic wave to be transmitted and received in the second reception frequency band having the second center frequency. The second reception frequency band is narrower than the first reception frequency band under the first condition, and the second center frequency is higher than the first center frequency in the first reception frequency band.

Next, in step S7, the processor 7 creates a second ultrasonic image based on the echo signals obtained by transmitting and receiving the second ultrasonic wave, using the second condition, and the second ultrasonic wave is generated. The image data is stored in the memory 9. The functions of steps S5 to S7 are an example of the embodiment of the second function of the present invention.

Here, in the flowchart of FIG. 2, after the first ultrasonic image for one frame is acquired, the second ultrasonic image for one frame is acquired. However, the present invention is not limited to this, for example, the transmission and reception of the first ultrasonic wave and the transmission and reception of the second ultrasonic wave are alternately performed for each sound ray, so that the first and A second ultrasound image may be acquired.

As described above, in the present example, when the user interface 10 receives an operator's input, in addition to the first ultrasonic image displayed on the display 8 as one pair used as learning data for machine learning, A second ultrasound image is acquired and the data for these first and second ultrasound images is stored. The second ultrasonic image is an image acquired using the second condition, which is a condition in which the image quality is lower than that of the first ultrasonic image. Therefore, for example, while acquiring the first ultrasonic image acquired for the purpose of diagnosing the subject, the second ultrasonic image forming a pair with this as the learning data for machine learning can also be acquired. As a result, it is possible to acquire a large number of pair images for constructing the learned model with the least effort.

When the first ultrasonic image and the second ultrasonic image are acquired as described above, the processor 7 reads the machine learning algorithm stored in the memory 9, and the data of the first ultrasonic image and Machine learning is performed using the data of the second ultrasonic image as learning data. The machine learning algorithm is supervised learning. In this supervised learning, the first ultrasonic image is used as correct answer data.

As a specific algorithm for machine learning, for example, the algorithm described in Non-Patent Document 1 below may be used.
Non-Patent Document 1: Kenji Suzuki, Junchi Liu et al. Neural Network Convolution (NNC) for Converting Ultra-Low-Dose to “Virtual” High-Dose CT Images. MLMI 2017: Machine Learning in Medical Imaging pp 334-343.

The machine learning may be performed each time one pair of the first ultrasonic image data and the second ultrasonic image data used as the learning data is acquired, or a required number of plural data may be acquired. May be performed each time a pair of is acquired.

According to this example, it is possible to acquire a second ultrasonic image that is paired with the data of the first ultrasonic image as the learning data for machine learning while performing diagnosis and the like, and continuously perform machine learning. Can be done. Thereby, the accuracy of the learned model can be further improved.

The learned model obtained by machine learning using the data of the first ultrasonic image and the data of the second ultrasonic image is stored in the memory 9.

Next, the image quality improvement of the ultrasonic image using the learned model obtained by machine learning will be described with reference to the flowchart of FIG. First, in step S11, the processor 7 causes the ultrasonic probe 2 to transmit and receive ultrasonic waves, and creates an ultrasonic image based on the obtained echo signal. Further, the processor 7 causes the display 8 to display the obtained ultrasonic image.

As the condition used for acquiring the ultrasonic image in step S11, for example, the first condition may be used.

Next, in step S12, the processor 7 determines whether or not an instruction to convert the ultrasonic image acquired in step S11 using the learned model is input in the user interface 10. More specifically, for example, the operator determines whether or not conversion is necessary based on the image quality of the ultrasonic image displayed on the display 8 in step S11. When the operator determines that the image quality is deteriorated and needs to be converted, the operator inputs an instruction for conversion on the user interface 10. The processor 7 determines whether or not the user interface 10 has received an input instructing this conversion.

Here, “conversion” refers to converting an ultrasonic image having a further deteriorated image quality into an ultrasonic image having a better image quality. The quality of the image quality is determined, for example, from the viewpoint of the SN ratio and the depth sensitivity. Therefore, by the conversion of this example, an ultrasonic image having a lower SN ratio and deeper sensitivity becomes an ultrasonic image having a higher SN ratio and deeper sensitivity.

For example, in a subject whose subcutaneous fat thickness is not relatively thick, the ultrasound image acquired using the first condition has a relatively good image quality that does not require conversion. On the other hand, even if the same first condition is used, for a subject having a thicker subcutaneous fat thickness, an ultrasonic image with deteriorated image quality as if it were acquired using the second condition is acquired. When such an ultrasonic image with deteriorated image quality is acquired, the operator inputs an instruction for conversion on the user interface 10.

In addition, even if an inexperienced operator cannot adjust the conditions so that an ultrasonic image with good image quality can be acquired and an ultrasonic image with deteriorated image quality is acquired, the operator is Input for instructing conversion is performed on the user interface 10.

When it is determined in step S12 that the user interface 10 has received the input (“YES” in step S12), the process proceeds to step S13. In step S13, the processor 7 uses the learned model stored in the memory 9 to convert the ultrasonic image obtained in step S11. More specifically, the ultrasonic image data obtained in step S11 is input to the learned model, and an ultrasonic image with improved image quality is obtained as an output.

Next, in step S14, the processor 7 causes the display 8 to display the ultrasonic image obtained as the output of the learned model in step S13. If the converted ultrasonic image is displayed in step S14, and if the user interface 10 does not accept the input of conversion in step S12 (“NO” in step S12), the process ends.

Next, a modified example of the first embodiment will be described. FIG. 4 is a flowchart showing a process for acquiring learning data for machine learning in this modified example. Steps S11 to S13 in FIG. 4 are the same as the steps S1 to S3 in the flowchart of FIG. Further, steps S16 to S18 in FIG. 4 are the same as the processes of steps S5 to S7 in the flowchart of FIG. Therefore, steps S14 and S15 will be described.

In step S14, the first ultrasonic image is created and displayed on the display 8 as in step S4. In step S15, the user interface 10 receives an input of an instruction to store the first ultrasonic image displayed in step S14. This input is performed by the operator. When the user interface 10 receives the input of the instruction to store the first ultrasonic image, the processor 7 stores the data of the first ultrasonic image in the memory 9.

Incidentally, the first ultrasonic images of a plurality of frames may be acquired and displayed until the input is performed in step S15.

When the user interface 10 receives the input in step S15, the processor 7 reads the second condition in step S16.

(Second embodiment)
Next, a second embodiment will be described. In this example, the system 100 shown in FIG. 5 will be described. In the first embodiment, both the acquisition of the learning data for machine learning and the construction of the learned model are performed in the ultrasonic diagnostic apparatus 1, but in the second embodiment, the learning model is constructed by the server. Performed at 101. The details will be described below.

The system 100 has a plurality of ultrasonic diagnostic apparatuses 20A, 20B, and 20C as the ultrasonic diagnostic apparatus 20, and also has a server 101 and a network 102. The ultrasonic diagnostic apparatuses 20A, 20B, and 20C are connected to the server 101 via the network 102.

As shown in FIG. 6, the ultrasonic diagnostic apparatus 20 has a network interface 21 that is connected to the network 102 and inputs and outputs data to and from the network 102. The other configurations of the ultrasonic diagnostic apparatus 20 are the same as those of the ultrasonic diagnostic apparatus 1 of the first embodiment.

The server 101 has a known configuration. FIG. 7 illustrates the configuration of the main part of the server 101 in this example, which includes a processor 103, a memory 104, and a network interface 105. In the present example, the machine learning algorithm is stored in the memory 104 of the server 101 instead of the memory 9 of the ultrasonic diagnostic apparatus 20.

The operation of this example will be described based on the flowchart of FIG. First, in step S31, the ultrasonic diagnostic apparatus 20 acquires the first ultrasonic image data and the second ultrasonic image data. The data of the first ultrasonic image and the data of the second ultrasonic image are acquired by the ultrasonic diagnostic apparatus 20 according to the flowchart of FIG. 2 as in the first embodiment.

Next, in step S32, the processor 7 of the ultrasonic diagnostic apparatus 20 transmits the data of the first ultrasonic image and the data of the second ultrasonic image acquired in step S31 via the network interface 21 and the network 102. Transmitted to the server 101. The processor 7 may perform transmission to the server 101 each time one pair of data of the first ultrasonic image and data of the second ultrasonic image is acquired, or a required number of plural data may be transmitted. The transmission to the server 101 may be performed each time a pair is acquired.

Next, in step S33, the processor 103 of the server 101 reads the machine learning algorithm stored in the memory 104, performs machine learning, and constructs a learned model. The processor 103 performs machine learning using the first ultrasonic image data and the second ultrasonic image data transmitted from the ultrasonic diagnostic apparatus 20 as learning data. Also in this example, the machine learning may be performed every time one pair of the first ultrasonic image data and the second ultrasonic image data is transmitted, or a required number of plural data may be transmitted. It may be done each time a pair is sent.

In this example, the server 101 can acquire learning data from a plurality of ultrasonic diagnostic apparatuses 20A, 20B, and 20C and perform machine learning, and thus acquire more learning data without trouble. be able to.

Next, in step S34, the processor 103 of the server 101 transmits the learned model constructed in step S33 to the ultrasonic diagnostic apparatus 20 via the network interface 105 and the network 102.

In this example, using the learned model transmitted from the server 101, conversion for improving the image quality of the ultrasonic image acquired by the ultrasonic diagnostic apparatus 20 is performed as necessary. Also in this example, the image quality improvement in the ultrasonic diagnostic apparatus 20 is performed according to the flowchart of FIG.

Although the present invention has been described above with reference to the above-described embodiment, it is needless to say that the present invention can be variously modified and implemented without departing from the spirit of the invention. For example, in the second embodiment, the ultrasonic image conversion using the learned model may be performed in the server 101. In this case, the ultrasonic image data acquired by the ultrasonic diagnostic apparatus 20 is transmitted to the server 101, and this data is converted using the learned model. The converted ultrasonic image is transmitted from the server 101 to the ultrasonic diagnostic apparatus 20 and displayed on the display 8.

The server 101 may be configured by a workstation, and the converted ultrasonic image using the learned model may be displayed on the display (not shown) of the server 101. In this case, even if the image quality of the ultrasonic image acquired by the ultrasonic diagnostic apparatus 20 is deteriorated, the ultrasonic image with the improved image quality is displayed on the server 101 without acquiring the ultrasonic image again. You can

In addition, the above-mentioned embodiment,
A method for controlling an ultrasonic device including a memory, a user interface, a processor, an ultrasonic probe, and a display,
The memory is a first condition used for acquiring a first ultrasonic image of the subject, and a second condition for the same position as the position where the first ultrasonic image is acquired in the subject. A second condition which is a second condition used for acquiring an ultrasonic image, wherein the second ultrasonic image is a condition in which the image quality is deteriorated as compared with the first ultrasonic image. Is configured to be stored,
The user interface is configured to accept operator input,
In the control method, the processor, when the user interface receives an input by the operator, sets the data of the first ultrasonic image and the data of the second ultrasonic image as learning data for machine learning. A method of executing a function to be acquired as one pair to be used,
The processor is
Reading the first condition from the memory,
Transmission and reception of the first ultrasonic wave by the ultrasonic probe, and creation of the first ultrasonic image based on the echo signal obtained by the transmission and reception of the first ultrasonic wave is performed using the first condition. ,
Displaying the first ultrasonic image on the display,
Storing the data of the first ultrasonic image in the memory,
Reading the second condition from the memory,
Transmission and reception of the second ultrasonic wave by the ultrasonic probe, and creation of the second ultrasonic image based on the echo signal obtained by the transmission and reception of the second ultrasonic wave is performed using the second condition. ,
A method of controlling the ultrasonic device may be one in which the data of the second ultrasonic image is stored in the memory.

1, 20 Ultrasonic Diagnostic Device 2 Ultrasonic Probe 7 Processor 8 Display 9 Memory 10 User Interface 101 Server 102 Network 103 Processor

Claims (11)

An ultrasonic device comprising a memory, a user interface, a processor, an ultrasonic probe and a display,
The memory is a first condition used for acquiring a first ultrasonic image of the subject, and a second condition for the same position as the position where the first ultrasonic image is acquired in the subject. A second condition which is a second condition used for acquiring an ultrasonic image, wherein the second ultrasonic image is a condition in which the image quality is deteriorated as compared with the first ultrasonic image. Is configured to be stored,
The user interface is configured to accept operator input,
When the user interface receives an input from the operator, the processor forms one pair of the first ultrasonic image data and the second ultrasonic image data as learning data for machine learning. Is configured to perform the function to get,
The function is
After reading the first condition from the memory, transmission and reception of the first ultrasonic wave by the ultrasonic probe, and the first ultrasonic image based on the echo signal obtained by the transmission and reception of the first ultrasonic wave And the first function of displaying the first ultrasonic image on the display and storing the data of the first ultrasonic image in the memory. ,
After reading the second condition from the memory, transmission and reception of the second ultrasonic wave by the ultrasonic probe, and the second ultrasonic image based on the echo signal obtained by the transmission and reception of the second ultrasonic wave Is performed using the second condition, and a second function of storing the data of the second ultrasonic image in the memory is included. The first condition includes at least one of the first ultrasonic transmission/reception condition, an echo signal processing condition and an image processing condition obtained by the first ultrasonic transmission/reception,
The second condition includes at least one of a condition for transmitting and receiving the second ultrasonic wave, a condition for processing an echo signal obtained by transmitting and receiving the second ultrasonic wave, and an image processing condition. Item 1. The ultrasonic device according to item 1. The ultrasonic device according to claim 1, wherein the second condition is a condition that an SN ratio of the second ultrasonic image is lower than an SN ratio of the first ultrasonic image. The first condition and the second condition include the transmission sound pressure of ultrasonic waves, and the second transmission sound pressure in the second condition is the first transmission sound pressure in the first condition. The ultrasonic device according to any one of claims 1 to 3, which is lower than the above. The ultrasonic device according to claim 1, wherein the second condition is a condition in which the depth sensitivity of the second ultrasonic image is lower than the depth sensitivity of the first ultrasonic image. The first condition and the second condition include an ultrasonic transmission frequency, and the second transmission frequency in the second condition is higher than the first transmission frequency in the first condition. The ultrasonic device according to any one of claims 1, 2 and 5. The first condition and the second condition include a reception frequency band of ultrasonic waves, the second reception frequency band in the second condition, the first reception frequency band in the first condition. 6. The second center frequency in the second reception frequency band is higher than the first center frequency in the first reception frequency band, and the second center frequency is higher than the first center frequency in the first reception frequency band. The ultrasonic device according to. The ultrasonic device according to any one of claims 1 to 7, wherein the memory further stores the machine learning algorithm. The processor further reads the machine learning algorithm, performs machine learning using the data of the first ultrasonic image and the data of the second ultrasonic image as learning data, and acquires a learned model. The ultrasonic device according to claim 8. A system comprising the ultrasonic device according to any one of claims 1 to 7 and a server connected via a network, wherein the server includes a memory in which the machine learning algorithm is stored and a processor. Prepare,
The processor of the ultrasonic device performs control of transmitting the data of the first ultrasonic image and the data of the second ultrasonic image to the server via the network,
The processor of the server reads out the machine learning algorithm, performs machine learning using the first ultrasonic image data and the second ultrasonic image data as learning data, and acquires a learned model. Yes, the system. A control program for an ultrasonic device including a memory, a user interface, a processor, an ultrasonic probe, and a display,
The memory is a first condition used for acquiring a first ultrasonic image of the subject, and a second condition for the same position as the position where the first ultrasonic image is acquired in the subject. A second condition which is a second condition used for acquiring an ultrasonic image, wherein the second ultrasonic image is a condition in which the image quality is deteriorated as compared with the first ultrasonic image. Is configured to be stored,
The user interface is configured to accept operator input,
When the user interface receives an input from the operator, the control program uses the data of the first ultrasonic image and the data of the second ultrasonic image as one of learning data for machine learning. Configured to cause the processor to perform the function of obtaining as a pair,
The function is
After reading the first condition from the memory, transmission and reception of the first ultrasonic wave by the ultrasonic probe, and the first ultrasonic image based on the echo signal obtained by the transmission and reception of the first ultrasonic wave And the first function of displaying the first ultrasonic image on the display and storing the data of the first ultrasonic image in the memory. ,
After reading the second condition from the memory, transmission and reception of the second ultrasonic wave by the ultrasonic probe, and the second ultrasonic image based on the echo signal obtained by the transmission and reception of the second ultrasonic wave Is performed using the second condition, and a second function of storing the data of the second ultrasonic image in the memory, and a control program for the ultrasonic device.