Ultrasonic diagnostic device and ultrasonic signal generation method

Abstract

To yield ultrasonic signals in which influence of saturation has been robustly reduced and deterioration of image quality is suppressed.SOLUTION: An ultrasonic diagnostic device includes a storage unit and a generation unit. The storage unit stores a learned model which has been learned using a first ultrasonic signal containing a saturated signal as input data and a second ultrasonic signal with influence of saturation reduced relative to the first ultrasonic signal as teacher data. The generation unit inputs a third ultrasonic signal containing a saturated signal to the learned model, and acquires a fourth ultrasonic signal which is output from the learned model and has the influence of saturation reduced relative to the third ultrasonic signal, thus generating the fourth ultrasonic signal.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The embodiments disclosed in this specification and the drawings relate to an ultrasound diagnostic device and a method for generating an ultrasound signal. [Background technology]

[0002] In order to obtain ultrasound image data with fewer artifacts, tissue harmonic imaging (THI) using harmonic components (nonlinear signals) generated during ultrasound propagation is widely used. Usually, the second harmonic component is used, but the use of the third harmonic component has also been proposed. Note that the Nth (N is an integer equal to or greater than 2) harmonic component is also simply called the Nth component or the Nth harmonic.

[0003] The analog circuit of an ultrasound diagnostic device has a finite input dynamic range. In addition, to protect the analog circuit, if the signal value of the input signal is greater than a certain value, a clamp circuit operates on the input signal, and the waveform of the input signal is clipped so that the signal value of the input signal becomes a constant value. This causes the input signal to saturate. The clipped signal is then superimposed with odd-multiple harmonic components. For this reason, when imaging is performed using third-order harmonic components, the imaging band and the band of harmonics generated by saturation overlap, degrading image quality.

[0004] Although some configurations and algorithms such as pre-filters have been proposed to suppress the effects of such saturation, these may result in an increase in circuit size, require signal processing at an earlier stage, or may lack robustness. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] JP 2016-112400 A [Patent Document 2] JP 2017-55845 A Summary of the Invention [Problem to be solved by the invention]

[0006] One of the problems to be solved by the embodiments disclosed in this specification and the drawings is to obtain an ultrasound signal in which the effect of saturation is robustly reduced and degradation of image quality is suppressed. However, the problems to be solved by the embodiments disclosed in this specification and the drawings are not limited to the above problem. Problems corresponding to the effects of each configuration shown in the embodiments described later can also be positioned as other problems. [Means for solving the problem]

[0007] An ultrasound diagnostic device according to an embodiment includes a storage unit and a generation unit. The storage unit stores a trained model trained by using a first ultrasound signal including a saturated signal as input data and a second ultrasound signal in which the influence of saturation on the first ultrasound signal is reduced as teacher data. The generation unit inputs a third ultrasound signal including a saturated signal to the trained model, and acquires a fourth ultrasound signal in which the influence of saturation on the third ultrasound signal is reduced, which is output from the trained model, thereby generating the fourth ultrasound signal. [Brief description of the drawings]

[0008] [Figure 1] FIG. 1 is a block diagram showing an example of the configuration of an ultrasonic diagnostic apparatus according to the first embodiment. [Diagram 2] FIG. 2 is a diagram illustrating an example of a configuration of a receiving circuit according to the first embodiment. [Diagram 3] FIG. 3 is a diagram illustrating an example of the operation of the clamp circuit according to the first embodiment. [Figure 4] FIG. 4 is a diagram for explaining an example of a method for generating a trained model according to the first embodiment. [Diagram 5] FIG. 5 is a diagram for explaining an example of a method for generating input data and teacher data according to the first embodiment. [Figure 6]FIG. 6 is a diagram illustrating an example of input data and teacher data according to the first embodiment. [Figure 7] FIG. 7 is a diagram for explaining an example of processing executed by the ultrasound diagnostic apparatus according to the first embodiment during inference using a trained model. [Figure 8] FIG. 8 is a flowchart showing an example of the flow of processing executed by the ultrasound diagnostic apparatus according to the first embodiment. [Figure 9A] FIG. 9A is a diagram showing an example of a B-mode image based on B-mode image data obtained from input data used during learning according to the first embodiment. [Figure 9B] FIG. 9B is a diagram showing an example of a B-mode image based on B-mode image data obtained from teacher data used during learning according to the first embodiment. [Figure 9C] Figure 9C is a figure showing an example of a B-mode image based on B-mode image data obtained from output data output from a trained model during inference relating to the first embodiment. [Figure 10] FIG. 10 is a diagram for explaining an example of a method for generating input data and teacher data according to the second embodiment. [Figure 11] FIG. 11 is a diagram illustrating an example of input data and teacher data according to the second embodiment. [Figure 12] FIG. 12 is a diagram for explaining an example of processing executed by the ultrasound diagnostic apparatus according to the second embodiment during inference using a trained model. [Figure 13] FIG. 13 is a diagram for explaining an example of a method for generating input data and teacher data according to the third embodiment. [Figure 14] FIG. 14 is a diagram illustrating an example of input data and teacher data according to the third embodiment. [Figure 15] FIG. 15 is a diagram for explaining an example of a process executed by the ultrasound diagnostic apparatus 1 according to the third embodiment during inference using a trained model. [Figure 16]FIG. 16 is a diagram for explaining an example of a method for generating input data and teacher data according to the fourth embodiment. [Figure 17] FIG. 17 is a diagram illustrating an example of input data and teacher data according to the fourth embodiment. [Figure 18] FIG. 18 is a diagram for explaining an example of processing executed by the ultrasound diagnostic apparatus according to the fourth embodiment during inference using a trained model. [Figure 19] FIG. 19 is a diagram for explaining an example of a method for generating input data and teacher data according to the fifth embodiment. [Figure 20] FIG. 20 is a diagram illustrating an example of input data and teacher data according to the fifth embodiment. [Figure 21] FIG. 21 is a diagram for explaining an example of processing executed by the ultrasound diagnostic apparatus according to the fifth embodiment during inference using a trained model. [Figure 22] FIG. 22 is a flowchart showing the flow of an example of processing executed by the ultrasound diagnostic device 1 according to the modified example. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0009] Hereinafter, an ultrasound diagnostic apparatus and an ultrasound signal generating method according to each embodiment and each modified example will be described with reference to the drawings.

[0010] (First embodiment) Fig. 1 is a block diagram showing an example of the configuration of an ultrasonic diagnostic device 1 according to the first embodiment. As shown in Fig. 1, the ultrasonic diagnostic device 1 according to the first embodiment includes a device main body 100, an ultrasonic probe 101, an input device 102, and a display 103.

[0011] The ultrasonic probe 101 has, for example, a plurality of elements (piezoelectric vibrators, piezoelectric elements). These elements generate ultrasonic waves based on a drive signal supplied from a transmission circuit 111 of a transmission / reception circuit 110 included in the device body 100. Specifically, the elements generate ultrasonic waves having a waveform corresponding to the transmission drive voltage when a voltage (transmission drive voltage) is applied by the transmission circuit 111. The waveform of the transmission drive voltage indicated by the drive signal is the waveform of the voltage applied to the plurality of elements. That is, the ultrasonic probe 101 transmits ultrasonic waves corresponding to the magnitude of the applied transmission drive voltage. The ultrasonic probe 101 also receives a reflected wave from the subject P, converts the received reflected wave into a reflected wave signal, which is an electrical signal, and outputs the reflected wave signal to the device body 100. The reflected wave signal is an example of an ultrasonic signal. The ultrasonic probe 101 also has, for example, a matching layer provided on the element, and a backing material that prevents the ultrasonic wave from propagating backward from the element. The ultrasonic probe 101 is detachably connected to the device body 100.

[0012] When ultrasonic waves are transmitted from the ultrasonic probe 101 to the subject P, the transmitted ultrasonic waves are successively reflected by discontinuous surfaces of acoustic impedance in the tissues of the subject P, and are received as reflected waves by multiple elements of the ultrasonic probe 101. The amplitude of the received reflected waves depends on the difference in acoustic impedance at the discontinuous surfaces where the ultrasonic waves are reflected. When the transmitted ultrasonic pulse is reflected by the surface of a moving body such as a moving blood flow or a heart wall, the reflected waves undergo a frequency shift due to the Doppler effect depending on the velocity component of the moving body in the ultrasonic transmission direction. Then, the ultrasonic probe 101 outputs the reflected wave signal to a receiving circuit 112 of a transmitting / receiving circuit 110 described later.

[0013] The ultrasonic probe 101 is provided detachably to the device body 100. When scanning a two-dimensional area inside the subject P (two-dimensional scanning), the operator connects, for example, a 1D array probe in which a plurality of elements are arranged in a row to the device body 100 as the ultrasonic probe 101. Types of 1D array probes include a linear ultrasonic probe, a convex ultrasonic probe, and a sector ultrasonic probe. When scanning a three-dimensional area inside the subject P (three-dimensional scanning), the operator connects, for example, a mechanical 4D probe or a 2D array probe to the device body 100 as the ultrasonic probe 101. The mechanical 4D probe is capable of two-dimensional scanning using a plurality of elements arranged in a row like the 1D array probe, and is also capable of three-dimensional scanning by swinging the plurality of elements at a predetermined angle (swing angle). The 2D array probe is capable of three-dimensional scanning using a plurality of elements arranged in a matrix, and is also capable of two-dimensional scanning by focusing and transmitting ultrasonic waves.

[0014] The input device 102 is realized by input means such as a mouse, a keyboard, a button, a panel switch, a touch command screen, a foot switch, a trackball, a joystick, etc. The input device 102 receives various setting requests from the operator of the ultrasound diagnostic apparatus 1, and transfers the received various setting requests to the apparatus main body 100.

[0015] The display 103 displays, for example, a GUI (Graphical User Interface) for the operator of the ultrasound diagnostic apparatus 1 to input various setting requests using the input device 102, and displays ultrasound images and the like based on ultrasound image data generated in the apparatus body 100. The display 103 is realized by a liquid crystal monitor, an OLED (Organic Light Emitting Diode) monitor, or the like. The display 103 is an example of a display unit.

[0016] The device body 100 generates ultrasonic image data based on a reflected wave signal transmitted from the ultrasonic probe 101. The ultrasonic image data is an example of an ultrasonic signal and an example of image data. The device body 100 can generate two-dimensional ultrasonic image data based on a reflected wave signal corresponding to a two-dimensional region of the subject P transmitted from the ultrasonic probe 101. The device body 100 can also generate three-dimensional ultrasonic image data based on a reflected wave signal corresponding to a three-dimensional region of the subject P transmitted from the ultrasonic probe 101. As shown in FIG. 1, the device body 100 has a transmission / reception circuit 110, a buffer memory 120, a B-mode processing circuit 130, a Doppler processing circuit 140, an image generation circuit 150, an image memory 160, a storage circuit 170, and a control circuit 180.

[0017] The transmission / reception circuit 110, under the control of the control circuit 180, causes the ultrasonic probe 101 to transmit ultrasonic waves and causes the ultrasonic probe 101 to receive reflected waves of the ultrasonic waves. That is, the transmission / reception circuit 110 executes scanning via the ultrasonic probe 101. Note that scanning is also called scanning, ultrasonic scanning, or ultrasonic scanning. The transmission / reception circuit 110 is an example of a transmission / reception unit. The transmission / reception circuit 110 has a transmission circuit 111 and a reception circuit 112. The transmission circuit 111 is an example of a transmission unit, and the reception circuit 112 is an example of a reception unit.

[0018] The transmission circuit 111, under the control of the control circuit 180, supplies a drive signal to the ultrasonic probe 101, causing the ultrasonic probe 101 to transmit ultrasonic waves. The transmission circuit 111 has a rate pulser generating circuit, a transmission delay circuit, and a transmission pulser. When scanning a two-dimensional region within the subject P, the transmission circuit 111 causes the ultrasonic probe 101 to transmit an ultrasonic beam for scanning the two-dimensional region. When scanning a three-dimensional region within the subject P, the transmission circuit 111 causes the ultrasonic probe 101 to transmit an ultrasonic beam for scanning the three-dimensional region.

[0019] The rate pulser generating circuit is controlled by the control circuit 180 and repeatedly generates rate pulses for forming a transmission ultrasonic wave (transmission beam) at a predetermined pulse repetition frequency (PRF). The rate pulse passes through the transmission delay circuit, so that a voltage having a different transmission delay time is applied to the transmission pulser. For example, the transmission delay circuit provides each rate pulse generated by the rate pulser generating circuit with a transmission delay time for each element required for focusing the ultrasonic waves generated from the ultrasonic probe 101 into a beam shape and determining the transmission directivity. The transmission pulser supplies a drive signal (drive pulse) to the ultrasonic probe 101 at a timing based on the rate pulse. That is, the transmission pulser applies a voltage (transmission drive voltage) having a waveform indicated by the drive signal to the ultrasonic probe 101 at a timing based on the rate pulse. The transmission delay circuit arbitrarily adjusts the transmission direction of the ultrasonic wave from the element surface by changing the transmission delay time provided to each rate pulse.

[0020] The driving pulse is transmitted from the transmitting pulser via a cable to an element in the ultrasonic probe 101, and then converted from an electrical signal to mechanical vibration in the element. That is, when a voltage is applied to an element, the element vibrates mechanically. Ultrasound generated by this mechanical vibration is transmitted inside the living body (inside the subject P). Here, the ultrasonic waves having different transmission delay times for each element are focused and propagated in a predetermined direction.

[0021] The transmission circuit 111 has a function of instantaneously changing the transmission frequency, transmission drive voltage, etc., in order to execute a predetermined scanning sequence under the control of the control circuit 180. In particular, the change in the transmission drive voltage is realized by a linear amplifier type oscillation circuit capable of instantaneously switching the value of the transmission drive voltage, or a mechanism for electrically switching multiple power supply units. The transmission frequency is, for example, the center frequency of the ultrasound to be transmitted.

[0022] The reflected wave of the ultrasound transmitted by the ultrasonic probe 101 reaches an element inside the ultrasonic probe 101, where it is converted from mechanical vibration to an electrical signal (reflected wave signal), and the reflected wave signal is input to the receiving circuit 112.

[0023] The receiving circuit 112 generates reflected wave data by performing various processes on the reflected wave signal transmitted from the ultrasonic probe 101. Then, the receiving circuit 112 stores the generated reflected wave data in the buffer memory 120. The reflected wave data is an example of an ultrasonic signal.

[0024] Fig. 2 is a diagram showing an example of the configuration of the receiving circuit 112 according to the first embodiment. As shown in Fig. 2, the receiving circuit 112 includes an analog front-end 113, an A / D (Analog to Digital) conversion circuit (A / D converter) 114, a beamformer 115, and a saturation reduction processing circuit 116.

[0025] A reflected wave signal is input to the analog front end 113 from the ultrasonic probe 101. The analog front end 113 is an analog circuit that performs known analog processing (analog signal processing) on ​​the input reflected wave signal and outputs the reflected wave signal, which is an analog signal that has been subjected to the analog processing, to the A / D conversion circuit 114. For example, to explain a part of the processing executed by the analog front end 113, the analog front end 113 includes, for example, a preamplifier, and the preamplifier amplifies the reflected wave signal for each channel and performs gain adjustment (gain correction).

[0026] Moreover, the analog front end 113 includes a clamp circuit 113a. FIG. 3 is a diagram showing an example of the operation of the clamp circuit 113a according to the first embodiment. For example, as shown in FIG. 3, the clamp circuit 113a converts the reflected wave signal 113b to a reflected wave signal 113d so that the amplitude of the reflected wave signal 113b input to the preamplifier falls within a predetermined allowable range 113c, and inputs the reflected wave signal 113d to the preamplifier. For example, the clamp circuit 113a generates the reflected wave signal 113d by clipping (fixing) the amplitude of the reflected wave signal 113b equal to or greater than a certain value to a certain value. Then, the clamp circuit 113a outputs the reflected wave signal 113d to the preamplifier. This protects the preamplifier, which is a circuit in the analog front end 113, from an excessively high voltage pulse of the reflected wave signal.

[0027] In this way, the clamp circuit 113a performs clamp processing to suppress, to a constant value, a signal value equal to or greater than a certain value of the reflected wave signal 113b obtained by transmitting and receiving ultrasonic waves to and from the subject P. A reflected wave signal 113d is obtained by the clamp processing. Such clamp processing is an example of processing included in analog processing.

[0028] Here, clipping the amplitude of the reflected wave signal 113b above a certain value to a certain value is equivalent to convolving a square wave. Therefore, odd-order harmonic components (e.g., third-order harmonic components, fifth-order harmonic components, etc.) are superimposed on the reflected wave signal 113d. In other words, if the center frequency (frequency of the fundamental wave) included in the reflected wave signal 113b is f0, the reflected wave signal 113d includes odd-order harmonic components (e.g., third-order harmonic components corresponding to a frequency of 3f0, fifth-order harmonic components corresponding to a frequency of 5f0, etc.).

[0029] Therefore, when imaging is performed using the third harmonic based on such a reflected wave signal 113d, the imaging band overlaps with the band of the harmonic generated by saturation, resulting in degradation of image quality. That is, since the reflected wave signal 113d is affected by saturation, the image quality of the ultrasound image data based on the reflected wave signal 113d is degraded. Therefore, the ultrasound diagnostic device 1 according to the first embodiment is configured to generate an ultrasound signal that does not degrade image quality even when imaging is performed using the third harmonic, as described below.

[0030] 2, the reflected wave signal output from the analog front end 113 is input to the A / D conversion circuit 114. The A / D conversion circuit 114 converts the reflected wave signal into a digital signal by A / D converting the reflected wave signal, and outputs the reflected wave signal converted into a digital signal to the beamformer 115.

[0031] The beamformer 115 receives the reflected wave signal output from the A / D conversion circuit 114. The beamformer 115 performs phasing and summing on the input reflected wave signal. For example, the beamformer 115 provides a reception delay time required to determine the reception directivity to the reflected wave signal, which is a digital signal. The beamformer 115 then performs summing on the reflected wave signal to which the reception delay time has been provided. The summing process of the beamformer 115 emphasizes the reflected component from a direction corresponding to the reception directivity of the reflected wave signal.

[0032] The beamformer 115 then converts the reflected wave signal that has been subjected to the phasing and summing process into an in-phase signal (I signal, I: In-phase) and a quadrature signal (Q signal, Q: Quadrature-phase) in the baseband. The beamformer 115 then stores the I signal and Q signal (IQ signal) as reflected wave data in the buffer memory 120. The beamformer 115 is realized by, for example, a processing circuit including a processor.

[0033] The saturation reduction processing circuit 116 is a processing circuit that operates when input data and teacher data used when a trained model 170a is generated by a learning device 200 described later are generated, that is, during learning. During learning, the saturation reduction processing circuit 116 receives reflected wave data output from the B-mode processing circuit 130 via the beamformer 115. The reflected wave data input to the saturation reduction processing circuit 116 is signal data including a third harmonic component described later. The saturation reduction processing circuit 116 generates reflected wave data in which the influence of saturation is reduced more than that of the input reflected wave data, and transmits the generated reflected wave data to the learning device 200 described later as teacher data. Details of the processing performed by the saturation reduction processing circuit 116 will be described later.

[0034] 1, the receiving circuit 112 generates two-dimensional reflected wave data from the two-dimensional reflected wave signal transmitted from the ultrasonic probe 101. The receiving circuit 112 also generates three-dimensional reflected wave data from the three-dimensional reflected wave signal transmitted from the ultrasonic probe 101.

[0035] In this embodiment, the ultrasound diagnostic device 1 can perform various processes in real time. For example, the ultrasound probe 101 transmits one frame's worth of reflected wave signals one after another to the receiving circuit 112. Every time the receiving circuit 112 receives one frame's worth of reflected wave signals transmitted from the ultrasound probe 101, the receiving circuit 112 generates one frame's worth of reflected wave data from the one frame's worth of reflected wave signals. Every time the receiving circuit 112 generates one frame's worth of reflected wave data, the receiving circuit 112 stores one frame's worth of reflected wave data in the buffer memory 120.

[0036] The buffer memory 120 is a memory that temporarily stores the reflected wave data generated by the transmission / reception circuit 110. For example, the buffer memory 120 is configured to be able to store a predetermined number of frames of reflected wave data. Then, when a new frame of reflected wave data is generated by the reception circuit 112 while the buffer memory 120 has stored the predetermined number of frames of reflected wave data, the buffer memory 120, under the control of the reception circuit 112, discards the oldest one frame of reflected wave data and stores the newly generated one frame of reflected wave data. For example, the buffer memory 120 is realized by a semiconductor memory element such as a RAM (Random Access Memory) or a flash memory.

[0037] The B-mode processing circuit 130 reads the reflected wave data from the buffer memory 120, performs various signal processing on the read reflected wave data, and outputs the reflected wave data subjected to various signal processing as B-mode data to the image generating circuit 150. The B-mode processing circuit 130 is realized by, for example, a processor. The B-mode processing circuit 130 is an example of a B-mode processing unit. The B-mode data is also an example of an ultrasonic signal.

[0038] For example, every time one frame of reflected wave data is newly stored in the buffer memory 120, the B-mode processing circuit 130 reads out one frame of reflected wave data newly stored in the buffer memory 120. Then, the B-mode processing circuit 130 generates one frame of B-mode data by performing various signal processing on the read one frame of reflected wave data. Then, every time the B-mode processing circuit 130 generates one frame of B-mode data, it outputs the newly generated one frame of B-mode data to the image generating circuit 150. An example of various signal processing executed by the B-mode processing circuit 130 will be described below.

[0039] For example, the B-mode processing circuit 130 performs quadrature detection, envelope detection processing, logarithmic amplification processing, and the like on the reflected wave data read from the buffer memory 120 to generate B-mode data in which the signal strength (amplitude strength) of each sample point is expressed as luminance. Then, the B-mode processing circuit 130 outputs the generated B-mode data to the image generating circuit 150.

[0040] Here, using the function of the B-mode processing circuit 130, the ultrasound diagnostic device 1 can extract the third harmonic component from the reflected wave data (received signal) and generate B-mode image data based on the extracted third harmonic component. For example, the ultrasound diagnostic device 1 extracts the third harmonic component from the reflected wave data using the technology described in JP 2016-112400 A. For example, the transmission circuit 111 causes the ultrasound probe 101 to transmit ultrasound three times in which the phases of the transmitted ultrasound (the phases of the center frequency components included in the transmitted ultrasound) differ from each other by 120 degrees. The reception circuit 112 generates three pieces of reflected wave data related to a common reception scanning line based on a plurality of reflected wave signals obtained by the three ultrasound transmissions. The B-mode processing circuit 130 extracts the second harmonic component by performing processing including a phase rotation processing on two or more pieces of reflected wave data out of the three pieces of reflected wave data, and adds the three pieces of reflected wave data to extract the third harmonic component. For example, the transmission circuit 111 causes the ultrasonic probe 101 to transmit an ultrasonic wave of a first phase, an ultrasonic wave of a second phase whose phase is advanced by 120 degrees from the first phase, and an ultrasonic wave of a third phase whose phase is advanced by 240 degrees from the first phase. In other words, the transmission circuit 111 causes the ultrasonic probe to transmit a first ultrasonic wave having a single center frequency component of the first phase, a second ultrasonic wave having a single center frequency component of the second phase whose phase is substantially advanced by 120 degrees from the first phase, and a third ultrasonic wave having a single center frequency component of the third phase whose phase is substantially advanced by 240 degrees from the first phase. Here, "substantially" includes any of the following: (1) allowing for error, (2) allowing for a phase advance in the negative direction (including, for example, a case in which the ultrasonic probe 101 transmits a first ultrasonic wave with a first phase, a second ultrasonic wave with a second phase delayed by 120 degrees from the first phase, and a third ultrasonic wave with a phase delayed by 240 degrees from the first phase), and (3) allowing for a phase rotation of N degrees to be equated with a phase rotation of N+360 degrees (for example, allowing a phase rotation of 120 degrees, a phase rotation of 480 degrees, and a phase rotation of -240 degrees).

[0041] Then, for example, the receiving circuit 112 generates first reflected wave data corresponding to the ultrasonic transmission of the first phase, second reflected wave data corresponding to the ultrasonic transmission of the second phase, and third reflected wave data corresponding to the ultrasonic transmission of the third phase. The B-mode processing circuit 130 adds the first reflected wave data, reflected wave data in which the phase of the second reflected wave data is advanced by 120 degrees, and reflected wave data in which the phase of the third reflected wave data is advanced by 240 degrees to extract the second harmonic component. The B-mode processing circuit 130 also adds the first reflected wave data, the second reflected wave data, and the third reflected wave data to extract the third harmonic component. In other words, the B-mode processing circuit 130 adds the first reflected wave data, the second reflected wave data, and the third reflected wave data in which the phases of the second harmonic components are substantially aligned, to extract the second harmonic component. In addition, the B-mode processing circuit 130 adds the first reflected wave data, the second reflected wave data, and the third reflected wave data, in which the phases of the third harmonic components are substantially aligned, to extract the third harmonic components. Here, for example, "substantially aligned" means that a small error is allowed.

[0042] Here, the third harmonic components extracted by the B-mode processing circuit 130 are reflected wave data used as input data to be described later, in detail. The B-mode processing circuit 130 obtains reflected wave data in which the influence of saturation is reduced by applying a trained model 170a to such input data to be described later. That is, the B-mode processing circuit 130 extracts third harmonic components in which the influence of saturation is reduced from the input data. The B-mode processing circuit 130 may extract the third harmonic components by subtracting fourth reflected wave data obtained by transmitting ultrasonic waves of a fourth phase, which is the inverted phase of the first phase, from first reflected wave data obtained by transmitting ultrasonic waves of a first phase. The third harmonic components extracted in this manner may be used as input data to be described later. The third harmonic components may also be extracted by applying a trained model 170a to the reflected wave data generated by the beamformer 115 to be described later.

[0043] The B-mode processing circuit 130 extracts the second harmonic component and the third harmonic component using the method described above. The B-mode processing circuit 130 having the function of extracting such second harmonic component and third harmonic component can be said to be an example of an extraction unit.

[0044] Then, the B-mode processing circuit 130 generates B-mode data based on the extracted second harmonic component, and outputs the generated B-mode data to the image generating circuit 150. The B-mode processing circuit 130 also generates B-mode data based on the extracted third harmonic component, and outputs the generated B-mode data to the image generating circuit 150.

[0045] The Doppler processing circuit 140 reads the reflected wave data from the buffer memory 120, performs various signal processing on the read reflected wave data, and outputs the reflected wave data that has been subjected to various signal processing as Doppler data to the image generating circuit 150. The Doppler processing circuit 140 is realized by, for example, a processor. The Doppler processing circuit 140 is an example of a Doppler processing unit.

[0046] For example, every time one frame of reflected wave data is newly stored in the buffer memory 120, the Doppler processing circuit 140 reads out one frame of reflected wave data newly stored in the buffer memory 120. Then, the Doppler processing circuit 140 generates one frame of Doppler data by performing various signal processing on the read one frame of reflected wave data. Then, every time the Doppler processing circuit 140 generates one frame of Doppler data, it outputs the newly generated one frame of Doppler data to the image generating circuit 150. An example of various signal processing executed by the Doppler processing circuit 140 will be described below.

[0047] For example, the Doppler processing circuit 140 performs frequency analysis on the reflected wave data read from the buffer memory 120 to extract motion information of a moving body (blood flow, tissue, contrast agent echo components, etc.) based on the Doppler effect from the reflected wave data, and generates Doppler data indicating the extracted motion information. For example, the Doppler processing circuit 140 extracts average speed, average variance, average power value, etc. over multiple points as motion information of the moving body, and generates Doppler data indicating the extracted motion information of the moving body. The Doppler processing circuit 140 outputs the generated Doppler data to the image generation circuit 150.

[0048] Using the above-mentioned function of the Doppler processing circuit 140, the ultrasound diagnostic device 1 can execute a color Doppler method, also called a color flow mapping (CFM) method. In the color flow mapping method, ultrasonic waves are transmitted and received multiple times on multiple scanning lines. In the color flow mapping method, a moving target indicator (MTI) filter is applied to a data sequence at the same position to suppress signals (clutter signals) originating from stationary tissue or slow-moving tissue from the data sequence at the same position, and a signal (blood flow signal) originating from blood flow is extracted. In the color flow mapping method, blood flow information such as the blood flow speed (average speed), blood flow variance (average variance value), and blood flow power (average power value) are estimated from the blood flow signal. The Doppler processing circuit 140 outputs color Doppler data indicating the blood flow information estimated by the color flow mapping method to the image generating circuit 150. Note that the color Doppler data is an example of Doppler data.

[0049] The B-mode processing circuitry 130 and the Doppler processing circuitry 140 are capable of processing both two-dimensional reflected wave data and three-dimensional reflected wave data.

[0050] The image generating circuit 150 generates various types of ultrasound image data from the B-mode data, the second harmonic component, the third harmonic component output from the B-mode processing circuit 130, and the Doppler data output from the Doppler processing circuit 140. The image generating circuit 150 is realized by a processor.

[0051] For example, the image generating circuit 150 generates two-dimensional B-mode image data in which the intensity of the reflected wave is expressed as brightness from the two-dimensional B-mode data generated by the B-mode processing circuit 130. The image generating circuit 150 also generates two-dimensional Doppler image data or two-dimensional color image data in which motion information or blood flow information is visualized from the two-dimensional Doppler data or Doppler color data generated by the Doppler processing circuit 140. The two-dimensional Doppler image data in which motion information is visualized and the two-dimensional color image data in which blood flow information is visualized are velocity image data, variance image data, power image data, or image data that is a combination of these.

[0052] Here, the image generating circuit 150 generally converts (scan converts) a scan line signal sequence of ultrasonic scanning into a scan line signal sequence of a video format represented by a television or the like, and generates ultrasonic image data for display. For example, the image generating circuit 150 generates ultrasonic image data for display by performing coordinate conversion on the data output from the B-mode processing circuit 130 and the Doppler processing circuit 140 according to the ultrasonic scanning form of the ultrasonic probe 101. In addition to the scan conversion, the image generating circuit 150 may perform various image processing such as image processing (smoothing processing) for regenerating an average brightness image using a plurality of image frames after the scan conversion, and image processing (edge ​​enhancement processing) using a differential filter in the image. In addition, the image generating circuit 150 may combine character information of various parameters, scales, body marks, etc. with the ultrasonic image data.

[0053] Furthermore, the image generating circuit 150 generates three-dimensional B-mode image data by performing coordinate conversion on the three-dimensional B-mode data generated by the B-mode processing circuit 130. The image generating circuit 150 also generates three-dimensional Doppler image data by performing coordinate conversion on the three-dimensional Doppler data generated by the Doppler processing circuit 140. That is, the image generating circuit 150 generates "three-dimensional B-mode image data and three-dimensional Doppler image data" as "three-dimensional ultrasound image data (volume data)". Then, the image generating circuit 150 performs various rendering processes on the volume data to generate various types of two-dimensional image data for displaying the volume data on the display 103.

[0054] The rendering process performed by the image generating circuit 150 includes, for example, a process of generating MPR image data from volume data using a multi-planar reconstruction method (MPR). In addition, the rendering process performed by the image generating circuit 150 includes, for example, a volume rendering (VR) process of generating two-dimensional image data reflecting three-dimensional information. The image generating circuit 150 is an example of an image generating unit.

[0055] The B-mode data and the Doppler data are ultrasonic image data before the scan conversion process, and the data generated by the image generating circuit 150 is ultrasonic image data for display after the scan conversion process. The B-mode data and the Doppler data are also called raw data.

[0056] Furthermore, when the image generating circuit 150 receives B-mode data based on the second harmonic component from the B-mode processing circuit 130, the image generating circuit 150 generates B-mode image data based on the B-mode data based on the second harmonic component. Similarly, when the image generating circuit 150 receives B-mode data based on the third harmonic component from the B-mode processing circuit 130, the image generating circuit 150 generates B-mode image data based on the B-mode data based on the third harmonic component.

[0057] The image memory 160 is a memory that stores various image data generated by the image generating circuit 150. The image memory 160 also stores data generated by the B-mode processing circuit 130 and the Doppler processing circuit 140. The B-mode data and Doppler data stored in the image memory 160 can be called up by an operator after diagnosis, for example, and become ultrasound image data for display via the image generating circuit 150. For example, the image memory 160 is realized by a semiconductor memory element such as a RAM (Random Access Memory) or a flash memory, a hard disk, or an optical disk.

[0058] The memory circuitry 170 stores various data such as control programs for performing scanning (transmission and reception of ultrasonic waves), image processing, and display processing, diagnostic information (e.g., patient ID, doctor's findings, etc.), diagnostic protocols, and various body marks. The memory circuitry 170 is also used for storing data stored in the image memory 160 as necessary. For example, the memory circuitry 170 is realized by a semiconductor memory element such as a flash memory, a hard disk, or an optical disk.

[0059] Moreover, the memory circuitry 170 according to this embodiment stores the trained model 170a. The memory circuitry 170 may store the trained model 170a when the ultrasound diagnostic device 1 is delivered, or may store the trained model 170a acquired from an external device or the like after the delivery of the ultrasound diagnostic device 1. The memory circuitry 170 is, for example, an example of a memory unit. The trained model 170a will be described later.

[0060] The control circuit 180 controls the overall processing of the ultrasound diagnostic device 1. Specifically, the control circuit 180 controls the processing of the transmission circuit 111, the reception circuit 112, the B-mode processing circuit 130, the Doppler processing circuit 140, and the image generation circuit 150 based on various setting requests input by the operator via the input device 102 and various control programs and various data read from the storage circuit 170. The control circuit 180 also controls the display 103 to display an ultrasound image based on ultrasound image data for display stored in the image memory 160. For example, the control circuit 180 controls the display 103 to display a B-mode image based on the B-mode image data or a color image based on the color image data. The control circuit 180 also controls the display 103 to display a color image superimposed on the B-mode image.

[0061] The control circuit 180 also generates a composite image by combining a B-mode image based on B-mode image data generated based on B-mode data based on the second harmonic component and a B-mode image based on B-mode image data generated based on B-mode data based on the third harmonic component. The control circuit 180 then controls the display 103 to display the generated composite image.

[0062] The control circuit 180 is an example of a display control unit or a control unit. The control circuit 180 is realized by, for example, a processor.

[0063] Moreover, the control circuit 180 controls the ultrasonic probe 101 via the transmission and reception circuit 110, thereby controlling ultrasonic scanning.

[0064] The term "processor" used in the description means a circuit such as a central processing unit (CPU), a graphics processing unit (GPU), 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), or a field programmable gate array (FPGA)). The processor reads out a program stored in the memory circuit 170 and executes the read out program to realize its function. Instead of storing the program in the memory circuit 170, the processor may be configured to directly incorporate the program in the circuit. In this case, the processor reads out and executes the program incorporated in the circuit to realize its function. Each processor in the present embodiment is not limited to being configured as a single circuit for each processor, and may be configured as a single processor by combining multiple independent circuits to realize its function. 1 (for example, the B-mode processing circuit 130, the Doppler processing circuit 140, the image generating circuit 150, and the control circuit 180) may be integrated into one processor to realize the functions. That is, the B-mode processing circuit 130, the Doppler processing circuit 140, the image generating circuit 150, and the control circuit 180 may be integrated into one processing circuit realized by a processor. Note that the transmission / reception circuit 110, the B-mode processing circuit 130, the Doppler processing circuit 140, the image generating circuit 150, and the control circuit 180 may be integrated into one processing circuit including a processor.

[0065] The above has described the overall configuration of the ultrasound diagnostic device 1 according to the first embodiment. With the above-mentioned configuration, the ultrasound diagnostic device 1 executes the processing described below so as to obtain an ultrasound signal in which the influence of saturation is robustly reduced and degradation of image quality is suppressed.

[0066] 4 is a diagram for explaining an example of a method for generating the trained model 170a according to the first embodiment. The trained model 170a is a trained machine learning model obtained by making a machine learning model perform machine learning according to a model learning program based on input data and teacher data. The trained model 170a is generated by the learning device 200.

[0067] The learning device 200 includes a machine learning model such as a convolution neural network (CNN). The learning device 200 generates a trained model 170a by performing learning (supervised learning) based on input data and teacher data related to an ultrasound examination of the same position (same cross section, same part) of a subject. The trained model 170a is provided with a function to output data (output data) corresponding to the teacher data when data corresponding to the input data is input during inference. Note that the ultrasound diagnostic device 1 may have a function similar to that of the learning device 200, and the ultrasound diagnostic device 1 may generate the trained model 170a instead of the learning device 200.

[0068] A case where the machine learning model is CNN will be described. In this case, in the learning device 200, input data is input to the CNN, which is a machine learning model. The learning device 200 applies the CNN to the input data to generate output data. Then, the output data is output from the CNN. In the learning device 200, the output data is input to an evaluation function. In the learning device 200, the teacher data is also input to the evaluation function. The learning device 200 evaluates the output data generated by the CNN based on the input data and the teacher data by the evaluation function. For example, the evaluation function compares the generated output data with the teacher data and corrects the CNN coefficients (network parameters such as weights and biases) by the error backpropagation method. The evaluation by the evaluation function is fed back to the CNN. The learning device 200 repeats such a series of supervised learning based on the input data and the teacher data acquired for the same position of the subject, until, for example, the error between the output data and the teacher data becomes equal to or less than a predetermined threshold. The learning device 200 can output the learned machine learning model as a trained model.

[0069] For example, in a CNN, when input data and training data are given, coefficients are generated that convert the input data into training data from the characteristics of the input data. The more input data and training data used in machine learning, the better, and for example, several thousand or more pieces of data are desirable.

[0070] A method for generating input data and teacher data used in machine learning in the learning device 200 will be described with reference to Fig. 5. Fig. 5 is a diagram for explaining an example of a method for generating input data and teacher data according to the first embodiment. Note that in Fig. 5, the B-mode processing circuit 130 connected to the rear stage of the beamformer 115 is omitted from the illustration.

[0071] As shown in Fig. 5, a reflected wave signal obtained by scanning a predetermined position of a subject P with an ultrasonic probe 101 passes through an analog front end 113 and an A / D conversion circuit 114, and is converted into reflected wave data by a beamformer 115. Then, as described above, the first reflected wave data, the second reflected wave data, and the third reflected wave data are added by the B-mode processing circuit 130 to obtain a third harmonic component. This third harmonic component as the reflected wave data (data after phase adjustment and addition processing) becomes input data. In addition, a saturation reduction processing circuit 116 performs saturation reduction processing to reduce the influence of saturation on the reflected wave data, which is the input data.

[0072] An example of saturation reduction processing performed by the saturation reduction processing circuit 116 according to the first embodiment will be described below. For example, the saturation reduction processing circuit 116 performs saturation reduction processing to reduce the amplitude value of a signal affected by saturation by applying a negative gain to a signal (part) affected by saturation, which is included in the reflected wave data (input data) after the delay and sum processing.

[0073] In the reflected wave data that has been subjected to the saturation reduction process, the influence of saturation is reduced compared to the reflected wave data as input data. The reflected wave data that has been subjected to such saturation reduction process becomes teacher data. That is, the saturation reduction processing circuit 116 generates teacher data by executing the saturation reduction process.

[0074] Then, the ultrasound diagnostic device 1 transmits the input data and teacher data thus generated to the learning device 200.

[0075] The input data and the teacher data are not limited to the reflected wave data described above. For example, the reflected wave data output from the beamformer 115 becomes B-mode data by the B-mode processing circuit 130. This B-mode data may become the input data. For example, the input data is B-mode data based on the third harmonic component. In this case, the saturation reduction processing circuit 116 applies saturation reduction processing to the reflected wave data that is the source of the input data. Then, the reflected wave data that has been subjected to the saturation reduction processing becomes B-mode data by the B-mode processing circuit 130. In such B-mode data based on the reflected wave data that has been subjected to the saturation reduction processing, the influence of saturation is reduced more than that of the B-mode data as the input data. The B-mode data based on the reflected wave data obtained by applying such saturation reduction processing becomes the teacher data.

[0076] Also, for example, the reflected wave data output from the beamformer 115 passes through the B-mode processing circuit 130 and becomes B-mode image data by the image generating circuit 150. This B-mode image data may be input data. For example, the input data is B-mode image data based on the third harmonic component. In this case, the saturation reduction processing circuit 116 applies saturation reduction processing to the reflected wave data that is the source of the input data. Then, the reflected wave data that has been subjected to saturation reduction processing passes through the B-mode processing circuit 130 and becomes B-mode image data by the image generating circuit 150. In such B-mode image data based on the reflected wave data that has been subjected to saturation reduction processing, the influence of saturation is reduced more than in the B-mode image data as input data. The B-mode image data based on the reflected wave data obtained by applying such saturation reduction processing becomes teacher data.

[0077] The reflected wave data, B-mode data, and B-mode image data, which are input data, are examples of a first ultrasonic signal including a saturated signal. The reflected wave data, B-mode data, and B-mode image data, which are teacher data, are examples of a second ultrasonic signal in which the influence of saturation on the input data is reduced. Each circuit (each unit) that generates the first ultrasonic signal and the second ultrasonic signal is, for example, an example of a generating unit. The same applies to embodiments other than this embodiment.

[0078] The learning device 200 uses the input data and teacher data generated by the above-mentioned method to train the learning model and generate a trained model 170a. At this time, the learning device 200 generates the trained model 170a for each part to be scanned. The ultrasound diagnostic device 1 then acquires the trained model 170a generated for each part from the learning device 200, and stores the acquired trained model 170a for each part in the memory circuitry 170.

[0079] 6 is a diagram showing an example of input data and teacher data according to the first embodiment. The input data 20 includes a signal (saturation signal) 20a that is affected by saturation. Meanwhile, teacher data 21 is obtained by reducing the signal value of the saturation signal 20a of the input data 20. The signal value of the signal 21a of the teacher data 21 is reduced more than the signal value of the saturation signal 20a.

[0080] 7 is a diagram for explaining an example of a process executed by the ultrasound diagnostic device 1 according to the first embodiment during inference using a trained model. During inference, the ultrasound diagnostic device 1 acquires a trained model 170a corresponding to a part to be scanned from the storage circuitry 170, infers output data corresponding to input data using the acquired trained model 170a, and outputs the inferred output data.

[0081] 7, for example, the ultrasound diagnostic device 1 generates input data used during inference in a manner similar to that for the input data used during learning. The input data used during inference is data that has been subjected to delay-and-sum processing, and is an example of a third ultrasound signal that includes a saturated signal.

[0082] When input data used during inference is input, the trained model 170a generates output data corresponding to the input data and outputs the generated output data. The output data is, for example, data in which the influence of saturation is reduced compared to the input data used during inference. That is, the output data is data in which the influence of saturation is reduced compared to the input data used during inference. The output data is an example of a fourth ultrasonic signal.

[0083] Each circuit (each unit) that generates the third ultrasonic signal and the fourth ultrasonic signal is, for example, an example of a generating unit. The same applies to embodiments other than this embodiment.

[0084] Here, the input data used during inference is reflected wave data, B-mode data or B-mode image data as described above, but below, a case will be described in which the input data used during inference is reflected wave data.

[0085] In this case, the B-mode processing circuit 130 acquires the trained model 170a corresponding to the part to be scanned from the trained models 170a for each part stored in the storage circuit 170. Then, the B-mode processing circuit 130 inputs the input data used during inference to the acquired trained model 170a. For example, as described above, the B-mode processing circuit 130 adds three pieces of reflected wave data obtained by three ultrasonic transmissions to extract a third harmonic component, and inputs the extracted third harmonic component to the trained model 170a as input data.

[0086] Then, the B-mode processing circuit 130 acquires the output data output from the trained model 170a. Here, the output data output from the trained model 170a is signal data including a third harmonic component in which the influence of saturation is reduced. Then, the B-mode processing circuit 130 generates B-mode data based on the output data, and the image generating circuit 150 generates B-mode image data from the B-mode data as ultrasound image data.

[0087] Therefore, according to the ultrasound diagnostic apparatus 1 according to the first embodiment, it is possible to obtain ultrasound image data in which the effects of saturation are robustly reduced and degradation of image quality is suppressed.

[0088] When the input data used during inference is B-mode data, the B-mode processing circuit 130 performs the same processing. Specifically, the B-mode processing circuit 130 inputs the B-mode data as input data to the trained model 170a. Then, the B-mode processing circuit 130 acquires output data output from the trained model 170a. Here, the output data output from the trained model 170a is B-mode data in which the influence of saturation has been reduced. Then, the B-mode processing circuit 130 transmits the acquired output data (B-mode data in which the influence of saturation has been reduced) to the image generation circuit 150. The image generation circuit 150 generates B-mode image data in which the influence of saturation has been reduced as ultrasound image data from the B-mode data in which the influence of saturation has been reduced.

[0089] Moreover, when the input data used during inference is B-mode image data, the image generation circuit 150 performs the same process. Specifically, the image generation circuit 150 inputs the B-mode image data as input data to the trained model 170a. Then, the image generation circuit 150 acquires the output data output from the trained model 170a. Here, the output data output from the trained model 170a is B-mode image data in which the influence of saturation has been reduced. In this way, the image generation circuit 150 uses the trained model 170a to generate B-mode image data in which the influence of saturation has been reduced from the input data.

[0090] Fig. 8 is a flowchart showing a flow of an example of processing executed by the ultrasound diagnostic device 1 according to the first embodiment. The processing shown in Fig. 8 is processing in which the ultrasound diagnostic device 1 extracts third harmonic components, generates B-mode image data based on the third harmonic components, and displays a B-mode image based on the generated B-mode image data.

[0091] As shown in FIG. 8, the ultrasound diagnostic device 1 uses the trained model 170a to generate B-mode image data in which the influence of saturation is robustly reduced and degradation of image quality is suppressed (step S101).

[0092] Then, the ultrasonic diagnostic apparatus 1 causes the display 103 to display a B-mode image based on the B-mode image data (step S102).

[0093] Then, the control circuit 180 of the ultrasound diagnostic apparatus 1 determines whether or not to continue the scan (step S103). If the scan is to be continued (step S103: Yes), the ultrasound diagnostic apparatus 1 returns to step S101 and executes the processes of steps S101 to S103 again. Note that the processes of steps S101 to S103 are executed for each frame of the B-mode image displayed on the display 103. As a result, the B-mode image is displayed on the display 103 as a moving image.

[0094] If the scan is not to be continued (step S103: No), the ultrasound diagnostic apparatus 1 ends the process shown in FIG.

[0095] Fig. 9A is a diagram showing an example of a B-mode image based on B-mode image data obtained from input data used during learning according to the first embodiment. Fig. 9B is a diagram showing an example of a B-mode image based on B-mode image data obtained from teacher data used during learning according to the first embodiment. Fig. 9C is a diagram showing an example of a B-mode image based on B-mode image data obtained from output data output from the trained model 170a during inference according to the first embodiment.

[0096] The area (saturated area) indicated by the arrow in the B-mode image 22 based on the input data used during learning shown in Fig. 9A is affected by saturation. In the saturated area, the contrast is excessive compared to the surrounding area of ​​the saturated area.

[0097] In the B-mode image 23 based on the teacher data used during learning shown in FIG. 9B, the luminance of the area corresponding to the saturated area in FIG. 9A is lowered, reducing the visual discomfort.

[0098] In the B-mode image 24 based on the output data output from the trained model 170a during inference shown in FIG. 9C, the effects of saturation are reduced, similar to the B-mode image 23.

[0099] The above describes the ultrasound diagnostic device 1 according to the first embodiment. As described above, the ultrasound diagnostic device 1 can robustly reduce the effects of saturation and obtain ultrasound image data with reduced degradation in image quality.

[0100] Second embodiment Next, an ultrasonic diagnostic device 1 according to a second embodiment will be described. In the description of the ultrasonic diagnostic device 1 according to the second embodiment, differences from the ultrasonic diagnostic device 1 according to the first embodiment will be mainly described, and a description of the same configuration as the ultrasonic diagnostic device 1 according to the first embodiment may be omitted.

[0101] A method for generating input data and teacher data used for machine learning in the learning device 200 in the second embodiment will be described with reference to Fig. 10. Fig. 10 is a diagram for explaining an example of a method for generating input data and teacher data according to the second embodiment. Note that in Fig. 10, the B-mode processing circuit 130 connected to the rear stage of the beamformer 115 is omitted from the illustration.

[0102] As shown in FIG. 10, the method of generating input data used during learning according to the second embodiment is similar to the method of generating input data used during learning according to the first embodiment.

[0103] On the other hand, the method of generating teacher data used during learning according to the second embodiment is different from the method of generating teacher data used during learning according to the first embodiment. This point will be specifically described.

[0104] For example, as shown in FIG. 10, a reflected wave signal obtained by scanning a predetermined position of a subject P with an ultrasonic probe 101 passes through an analog front end 113 and is converted into a digital signal by an A / D conversion circuit 114. Then, a saturation reduction processing circuit 116 performs a saturation reduction process for reducing the influence of saturation on the reflected wave signal, which is a digital signal. Then, the reflected wave signal with the reduced influence of saturation passes through a beamformer 115 and a B-mode processing circuit 130, and is subjected to a process similar to the process for obtaining a third harmonic component in the above-described first embodiment, thereby obtaining a third harmonic component. In the second embodiment, the reflected wave data based on the third harmonic component, the B-mode data based on the third harmonic component, or the B-mode image data based on the third harmonic component is used as teacher data. That is, the teacher data is obtained by reducing the influence of saturation on the reflected wave signal after it has been converted into a digital signal and before it has been subjected to a delay-and-sum process.

[0105] An example of the saturation reduction processing executed by the saturation reduction processing circuit 116 according to the second embodiment will be described. For example, the saturation reduction processing circuit 116 reduces the influence of saturation of the reflected wave signal by using a technique similar to that described in Japanese Patent Application Laid-Open No. 2017-55845. For example, the saturation reduction processing circuit 116 determines whether the reflected wave signal is saturated for each channel. To explain by taking a specific example, the saturation reduction processing circuit 116 detects saturation by utilizing the fact that the value of the reflected wave signal output from the A / D conversion circuit 114 is a digital positive upper limit or negative lower limit. Note that the saturation reduction processing circuit 116 may determine that the reflected wave signal is saturated when the value of the reflected wave signal is equal to or greater than a predetermined threshold, and may determine that the reflected wave signal is not saturated when the value of the reflected wave signal is less than the predetermined threshold. Then, the saturation reduction processing circuit 116 multiplies the value of the reflected wave signal of the channel determined to be saturated by a coefficient less than 1, and outputs the reflected wave signal obtained as a result of the multiplication (the reflected wave signal of the channel determined to be saturated) to the beamformer 115. On the other hand, the saturation reduction processing circuit 116 outputs the reflected wave signals of the channels that are determined to be not saturated as they are to the beamformer 115. Therefore, the teacher data is obtained by determining for each channel whether or not the reflected wave signal is saturated, and using the coefficient for each channel according to the determination result as a weight to weight the reflected wave signal for each channel.

[0106] The saturation reduction processing circuit 116 may also perform other saturation reduction processing on the reflected wave signal. For example, the saturation reduction processing circuit 116 determines whether or not the reflected wave signal is saturated for each channel. Then, the saturation reduction processing circuit 116 estimates the value of the reflected wave signal of the channel determined to be saturated by an interpolation process using the value of the reflected wave signal of the channel determined to be not saturated. Then, the saturation reduction processing circuit 116 replaces the value of the reflected wave signal of the channel determined to be saturated with the estimated value of the reflected wave signal. Then, the saturation reduction processing circuit 116 outputs the replaced value of the reflected wave signal of the channel determined to be saturated to the beamformer 115. On the other hand, the saturation reduction processing circuit 116 outputs the reflected wave signal of the channel determined to be not saturated to the beamformer 115 as it is. Therefore, the teacher data is obtained by determining whether or not the reflected wave signal is saturated for each channel, and replacing the value of the reflected wave signal of the channel determined to be saturated with the value of the reflected wave signal estimated using the value of the reflected wave signal of the channel determined to be not saturated.

[0107] In the reflected wave signal that has been subjected to the saturation reduction process, the influence of saturation is reduced more than in the reflected wave signal that has not been subjected to the saturation reduction process. The teacher data based on the reflected wave signal that has been subjected to such saturation reduction process is an example of a second ultrasonic signal.

[0108] Then, the ultrasound diagnostic device 1 transmits the input data and teacher data thus generated to the learning device 200.

[0109] The learning device 200 uses the input data and teacher data generated by the above-mentioned method to train a learning model and generates a trained model 170b (see FIG. 12). At this time, the learning device 200 generates a trained model 170b for each part to be scanned. Then, the ultrasound diagnostic device 1 acquires the trained model 170b generated for each part from the learning device 200, and stores the acquired trained model 170b for each part in the memory circuitry 170.

[0110] 11 is a diagram showing an example of input data and teacher data according to the second embodiment. The input data 25a includes a signal (saturation signal) 25c that is affected by saturation. The reflected wave signal 25, which is the source of the input data 25a, includes a saturation signal that corresponds to the saturation signal 25c.

[0111] On the other hand, teacher data 26b is obtained by reducing the signal value of the saturation signal of the reflected wave signal 25. Specifically, the reflected wave signal 26 is generated by reducing the signal value of the saturation signal of the reflected wave signal 25. The effect of saturation is reduced in signal 26a of the reflected wave signal 26 corresponding to the saturation signal of the reflected wave signal 25. Therefore, the signal value of signal 26c of teacher data 26b is reduced more than the signal value of saturation signal 25c.

[0112] 12 is a diagram for explaining an example of a process executed by the ultrasound diagnostic device 1 according to the second embodiment during inference using the trained model 170b. During inference, the ultrasound diagnostic device 1 acquires the trained model 170b corresponding to the part to be scanned from the storage circuitry 170, infers output data corresponding to input data using the acquired trained model 170b, and outputs the inferred output data.

[0113] As shown in FIG. 12, for example, the ultrasound diagnostic device 1 generates input data used during inference in a similar manner to the input data used during learning.

[0114] When input data used during inference is input, the trained model 170b generates output data corresponding to the input data and outputs the generated output data. The output data is, for example, data in which the influence of saturation is reduced compared to the input data used during inference. That is, the output data is data in which the influence of saturation is reduced compared to the input data used during inference. The output data is an example of a fourth ultrasonic signal.

[0115] During inference, the ultrasound diagnostic apparatus 1 according to the second embodiment executes the same processing as that executed by the ultrasound diagnostic apparatus 1 according to the first embodiment.

[0116] Therefore, according to the ultrasound diagnostic apparatus 1 of the second embodiment, similarly to the first embodiment, it is possible to obtain ultrasound image data in which the effects of saturation are robustly reduced and degradation of image quality is suppressed.

[0117] (Third embodiment) Next, an ultrasonic diagnostic device 1 according to a third embodiment will be described. In the description of the ultrasonic diagnostic device 1 according to the third embodiment, differences from the ultrasonic diagnostic device 1 according to each of the above-mentioned embodiments will be mainly described, and a description of the same configuration as the ultrasonic diagnostic device 1 according to each of the above-mentioned embodiments may be omitted.

[0118] A method for generating input data and teacher data used in machine learning in the learning device 200 in the third embodiment will be described with reference to Fig. 13. Fig. 13 is a diagram for explaining an example of a method for generating input data and teacher data according to the third embodiment.

[0119] As shown in FIG. 13, the method of generating input data used during learning according to the third embodiment is similar to the method of generating input data used during learning according to the first embodiment.

[0120] On the other hand, the method of generating teacher data used during learning according to the third embodiment is different from the method of generating teacher data used during learning according to the first embodiment. This point will be specifically described.

[0121] 13, for example, various analog processes are performed by the analog front end 113 on the reflected wave signal obtained by scanning a predetermined position of the subject P with the ultrasonic probe 101. Then, the saturation reduction processing circuit 116 performs saturation reduction processing for reducing the influence of saturation on the reflected wave signal, which is an analog signal that has been subjected to various analog processes.

[0122] An example of the saturation reduction processing performed by the saturation reduction processing circuit 116 according to the third embodiment will be described. For example, the saturation reduction processing circuit 116 uses a known analog filter such as a prefilter to reduce the influence of saturation in the reflected wave signal. For example, the saturation reduction processing circuit 116 uses a prefilter to perform processing to suppress signal components (saturation-derived components) caused by saturation included in the reflected wave signal as the saturation reduction processing. Then, the saturation reduction processing circuit 116 outputs the reflected wave signal that has been subjected to the saturation reduction processing to the A / D conversion circuit 114.

[0123] Furthermore, the saturation reduction processing circuit 116 may perform other saturation reduction processing on the reflected wave signal. For example, the saturation reduction processing circuit 116 determines whether or not the reflected wave signal is saturated. Then, the saturation reduction processing circuit 116 applies an analog gain to the reflected wave signal determined to be saturated to reduce the amplitude value of the reflected wave signal. Then, the saturation reduction processing circuit 116 outputs the reflected wave signal that has been applied with the analog gain to the A / D conversion circuit 114. On the other hand, the saturation reduction processing circuit 116 outputs the reflected wave signal determined to be not saturated to the A / D conversion circuit 114 as is.

[0124] The reflected wave signal output from the saturation reduction processing circuit 116 is input to the A / D conversion circuit 114. The A / D conversion circuit 114 converts the input reflected wave signal into a digital signal, and outputs the reflected wave signal, which is a digital signal in which the effects of saturation have been reduced, to the beamformer 115.

[0125] Then, the reflected wave signal, which is a digital signal in which the influence of saturation has been reduced, passes through the beamformer 115 and the B-mode processing circuit 130, and is subjected to a process similar to that for obtaining the third harmonic component in the above-described first embodiment, thereby obtaining the third harmonic component. In the third embodiment, the reflected wave data based on the third harmonic component, the B-mode data based on the third harmonic component, or the B-mode image data based on the third harmonic component is used as teacher data.

[0126] Such teacher data is an example of a second ultrasonic signal. In this way, the teacher data is data obtained after the reflected wave signal, which is an analog signal that has been subjected to saturation reduction processing to reduce the effect of saturation, is converted into a digital signal. Also, the teacher data is data obtained after the reflected wave signal, which has been subjected to saturation reduction processing to apply an analog gain according to the determination result of the saturation of the reflected wave signal, is converted into a digital signal.

[0127] Then, the ultrasound diagnostic device 1 transmits the input data and teacher data thus generated to the learning device 200.

[0128] The learning device 200 uses the input data and teacher data generated by the above-mentioned method to train the learning model and generates a trained model 170c (see FIG. 15). At this time, the learning device 200 generates a trained model 170c for each part to be scanned. Then, the ultrasound diagnostic device 1 acquires the trained model 170c generated for each part from the learning device 200, and stores the acquired trained model 170c for each part in the memory circuitry 170.

[0129] Fig. 14 is a diagram showing an example of input data and teacher data according to the third embodiment. As shown in Fig. 14, the analog front end 113 generates a reflected wave signal 27a by clipping the amplitude of the reflected wave signal 27 that is equal to or greater than a certain value to a certain value. Then, the A / D conversion circuit 114 converts the reflected wave signal 27a, which is an analog signal, into a reflected wave signal 27b, which is a digital signal. Then, the ultrasound diagnostic device 1 generates input data 27c from the reflected wave signal 27b.

[0130] The input data 27c includes a signal (saturated signal) 27d that is affected by saturation.

[0131] On the other hand, the saturation reduction processing circuit 116 uses a prefilter to perform saturation reduction processing to suppress signal components due to saturation contained in the reflected wave signal 27a. This generates a reflected wave signal 28 that has been subjected to saturation reduction processing. Then, the A / D conversion circuit 114 converts the reflected wave signal 28, which is an analog signal, into a reflected wave signal 28a, which is a digital signal. Then, the ultrasound diagnostic device 1 generates teacher data 28b from the reflected wave signal 28.

[0132] The signal value of the signal 28c of the teacher data 28b is lower than the signal value of the signal 27d corresponding to the signal 28c of the input data 27c.

[0133] 15 is a diagram for explaining an example of a process executed by the ultrasound diagnostic device 1 according to the third embodiment during inference using the trained model 170c. During inference, the ultrasound diagnostic device 1 acquires the trained model 170c corresponding to the part to be scanned from the storage circuitry 170, infers output data corresponding to input data using the acquired trained model 170c, and outputs the inferred output data.

[0134] As shown in FIG. 15, for example, the ultrasound diagnostic device 1 generates input data used during inference in a similar manner to the input data used during learning.

[0135] When input data used during inference is input, the trained model 170c generates output data corresponding to the input data and outputs the generated output data. The output data is, for example, data in which the influence of saturation is reduced compared to the input data used during inference. That is, the output data is data in which the influence of saturation is reduced compared to the input data used during inference. The output data is an example of a fourth ultrasonic signal.

[0136] During inference, the ultrasound diagnostic apparatus 1 according to the third embodiment executes the same processing as that executed by the ultrasound diagnostic apparatus 1 according to the first embodiment.

[0137] Therefore, according to the ultrasound diagnostic apparatus 1 of the third embodiment, similarly to the first embodiment and the like, it is possible to obtain ultrasound image data in which the effects of saturation are robustly reduced and degradation of image quality is suppressed.

[0138] (Fourth embodiment) Next, an ultrasonic diagnostic device 1 according to a fourth embodiment will be described. In the description of the ultrasonic diagnostic device 1 according to the fourth embodiment, differences from the ultrasonic diagnostic device 1 according to each of the above-mentioned embodiments will be mainly described, and a description of the same configuration as the ultrasonic diagnostic device 1 according to each of the above-mentioned embodiments may be omitted.

[0139] A method for generating input data and teacher data used in machine learning in the learning device 200 in the fourth embodiment will be described with reference to Fig. 16. Fig. 16 is a diagram for explaining an example of a method for generating input data and teacher data according to the fourth embodiment.

[0140] 16, a reflected wave signal obtained by scanning a predetermined position of a subject P with an ultrasonic probe 101 passes through an analog front end 113 and is converted into a digital signal by an A / D conversion circuit 114. This digital reflected wave signal becomes input data. Such a reflected wave signal as input data is an example of a first ultrasonic signal including a saturated signal. Moreover, such reflected wave data as input data is data after being converted into a digital signal and is data before being subjected to phasing addition processing.

[0141] In addition, saturation reduction processing for reducing the influence of saturation on the reflected wave signal of the digital signal, which is input data, is performed by the saturation reduction processing circuit 116. The saturation reduction processing circuit 116 according to the fourth embodiment performs saturation reduction processing on the reflected wave signal similar to the saturation reduction processing performed by the saturation reduction processing circuit 116 according to the second embodiment.

[0142] In the reflected wave signal that has been subjected to the saturation reduction process, the influence of saturation is reduced compared to the reflected wave signal as input data. Such a reflected wave signal that has been subjected to the saturation reduction process is used as teacher data. The reflected wave signal that has been subjected to the saturation reduction process is an example of a second ultrasonic signal in which the influence of saturation is reduced with respect to the input data.

[0143] Then, the ultrasound diagnostic device 1 transmits the input data and teacher data thus generated to the learning device 200.

[0144] The learning device 200 uses the input data and teacher data generated by the above-mentioned method to train the learning model and generates a trained model 170d (see FIG. 18). At this time, the learning device 200 generates a trained model 170d for each part to be scanned. Then, the ultrasound diagnostic device 1 acquires the trained model 170d generated for each part from the learning device 200, and stores the acquired trained model 170d for each part in the memory circuitry 170.

[0145] 17 is a diagram showing an example of input data and teacher data according to the fourth embodiment. The input data 29 includes a saturated signal that is affected by saturation.

[0146] On the other hand, the signal value of the saturation signal of the input data 29 is reduced to obtain teacher data 30. The effect of saturation is reduced in signal 30a of teacher data 30 corresponding to the saturation signal of input data 29. Therefore, the signal value of signal 30a of teacher data 30 is reduced more than the signal value of the saturation signal of input data 29.

[0147] 18 is a diagram for explaining an example of a process executed by the ultrasound diagnostic device 1 according to the fourth embodiment at the time of inference using the trained model 170d. At the time of inference, the ultrasound diagnostic device 1 acquires the trained model 170d corresponding to the part to be scanned from the storage circuitry 170, infers output data corresponding to input data using the acquired trained model 170d, and outputs the inferred output data.

[0148] As shown in Fig. 18, for example, the ultrasound diagnostic device 1 generates input data used during inference in a similar manner to the input data used during learning. Here, the input data is generated for each channel. The input data used during inference is an example of a third ultrasound signal including a saturated signal. In addition, such input data is data after being converted into a digital signal and before being subjected to delay-and-sum processing.

[0149] When input data used during inference is input, the trained model 170d generates output data corresponding to the input data and outputs the generated output data. The output data is, for example, data in which the influence of saturation is reduced compared to the input data used during inference. That is, the output data is data in which the influence of saturation is reduced compared to the input data used during inference. The output data is an example of a fourth ultrasonic signal.

[0150] The A / D conversion circuit 114 acquires a trained model 170d corresponding to the part to be scanned from trained models 170d for each part stored in the storage circuit 170. Then, the A / D conversion circuit 114 inputs input data used during inference for each channel to the acquired trained model 170d.

[0151] The A / D conversion circuit 114 acquires the output data output from the trained model 170d, and generates output data for each channel as a reflected wave signal. Then, the A / D conversion circuit 114 outputs the generated reflected wave signal to the beamformer 115. The beamformer 115 generates reflected wave data by performing phasing and summing processing on the input reflected wave signal.

[0152] When the ultrasound diagnostic device 1 extracts the third harmonic component, the receiving circuit 112 generates three pieces of reflected wave data related to a common receiving scan line based on a plurality of reflected wave signals obtained by three ultrasound transmissions, using a method using the trained model 170d as described above. Then, the B-mode processing circuit 130 adds the three pieces of reflected wave data to extract the third harmonic component. That is, the B-mode processing circuit 130 extracts the third harmonic component as a high-frequency component of a multiple of three from the reflected wave data based on the output data output from the trained model 170d.

[0153] Here, the three reflected wave data used in extracting the third harmonic component are data in which the influence of saturation is reduced. Therefore, according to the ultrasonic diagnostic device 1 according to the fourth embodiment, similarly to the first embodiment, etc., it is possible to obtain ultrasonic image data in which the influence of saturation is robustly reduced and deterioration of image quality is suppressed.

[0154] Fifth embodiment Next, an ultrasonic diagnostic device 1 according to a fifth embodiment will be described. In the description of the ultrasonic diagnostic device 1 according to the fifth embodiment, differences from the ultrasonic diagnostic device 1 according to each of the above-mentioned embodiments will be mainly described, and a description of the same configuration as that of the ultrasonic diagnostic device 1 according to each of the above-mentioned embodiments may be omitted.

[0155] A method for generating input data and teacher data used in machine learning in the learning device 200 in the fifth embodiment will be described with reference to Fig. 19. Fig. 19 is a diagram for explaining an example of a method for generating input data and teacher data according to the fifth embodiment.

[0156] As shown in FIG. 19, the method of generating input data used during learning according to the fifth embodiment is similar to the method of generating input data used during learning according to the fourth embodiment.

[0157] A method for generating teacher data used during learning according to the fifth embodiment will be described.

[0158] 19, for example, various analog processes are performed by the analog front end 113 on the reflected wave signal obtained by scanning a predetermined position of the subject P with the ultrasonic probe 101. Then, the saturation reduction processing circuit 116 performs saturation reduction processing for reducing the influence of saturation on the reflected wave signal, which is an analog signal that has been subjected to various analog processes.

[0159] An example of the saturation reduction processing performed by the saturation reduction processing circuit 116 according to the fifth embodiment will be described. For example, the saturation reduction processing circuit 116 performs saturation reduction processing similar to the saturation reduction processing performed by the saturation reduction processing circuit 116 according to the third embodiment on the reflected wave signal. Then, the saturation reduction processing circuit 116 outputs the reflected wave signal that has been subjected to the saturation reduction processing to the A / D conversion circuit 114.

[0160] The reflected wave signal output from the saturation reduction processing circuit 116 is input to the A / D conversion circuit 114. The A / D conversion circuit 114 converts the input reflected wave signal into a digital signal.

[0161] In the reflected wave signal, which is a digital signal that has been subjected to saturation reduction processing, the influence of saturation is reduced more than in the reflected wave signal as input data. Such a reflected wave signal, which is a digital signal that has been subjected to saturation reduction processing, becomes teacher data. Such a reflected wave signal that becomes teacher data is an example of a second ultrasonic signal. In this way, the teacher data is a reflected wave signal that is an analog signal that has been subjected to saturation reduction processing that reduces the influence of saturation. In addition, the teacher data is a reflected wave signal that has been subjected to saturation reduction processing that applies an analog gain according to the determination result of the saturation of the reflected wave signal.

[0162] Then, the ultrasound diagnostic device 1 transmits the input data and teacher data thus generated to the learning device 200.

[0163] The learning device 200 uses the input data and teacher data generated by the above-mentioned method to train the trained model and generate trained model 170e (see FIG. 21). At this time, the learning device 200 generates trained model 170e for each part to be scanned. Then, the ultrasound diagnostic device 1 acquires the trained model 170e generated for each part from the learning device 200, and stores the acquired trained model 170e for each part in the memory circuitry 170.

[0164] Fig. 20 is a diagram showing an example of input data and teacher data according to the fifth embodiment. As shown in Fig. 20, the analog front end 113 generates a reflected wave signal 31a by clipping the amplitude of the reflected wave signal 31 that is equal to or greater than a certain value to a certain value. Then, the A / D conversion circuit 114 converts the reflected wave signal 31a, which is an analog signal, into a reflected wave signal 31b, which is a digital signal. This reflected wave signal 31b becomes input data. The input data includes a saturated signal that is affected by saturation.

[0165] On the other hand, the saturation reduction processing circuit 116 uses a prefilter to perform saturation reduction processing to suppress signal components due to saturation contained in the reflected wave signal 31a. This generates a reflected wave signal 32 that has been subjected to saturation reduction processing. Then, the A / D conversion circuit 114 converts the reflected wave signal 32, which is an analog signal, into a reflected wave signal 32a, which is a digital signal. This reflected wave signal 32a becomes teacher data.

[0166] The signal value of the reflected wave signal 32a (a signal corresponding to the saturation signal of the reflected wave signal 31b) which is the teacher data is reduced below the signal value of the saturation signal of the reflected wave signal 31b which is the input data.

[0167] 21 is a diagram for explaining an example of a process executed by the ultrasound diagnostic device 1 according to the fifth embodiment during inference using the trained model 170e. During inference, the ultrasound diagnostic device 1 acquires the trained model 170e corresponding to the part to be scanned from the storage circuitry 170, infers output data corresponding to input data using the acquired trained model 170e, and outputs the inferred output data.

[0168] As shown in FIG. 21, for example, the ultrasound diagnostic device 1 generates input data used during inference in a similar manner to the input data used during learning.

[0169] When input data used during inference is input, the trained model 170e generates output data corresponding to the input data and outputs the generated output data. The output data is, for example, data in which the influence of saturation is reduced compared to the input data used during inference. That is, the output data is data in which the influence of saturation is reduced compared to the input data used during inference. The output data is an example of a fourth ultrasonic signal.

[0170] The A / D conversion circuit 114 acquires a trained model 170e corresponding to the part to be scanned from trained models 170e for each part stored in the storage circuit 170. Then, the A / D conversion circuit 114 inputs input data used during inference to the acquired trained model 170d for each channel.

[0171] The A / D conversion circuit 114 acquires the output data output from the trained model 170d, and generates output data for each channel as a reflected wave signal. Then, the A / D conversion circuit 114 outputs the generated reflected wave signal to the beamformer 115. The beamformer 115 generates reflected wave data by performing phasing and summing processing on the input reflected wave signal.

[0172] When the ultrasound diagnostic device 1 extracts a third harmonic component, the receiving circuit 112 generates three pieces of reflected wave data related to a common receiving scan line based on a plurality of reflected wave signals obtained by three ultrasound transmissions, using a method using the trained model 170e as described above. Then, the B-mode processing circuit 130 adds the three pieces of reflected wave data to extract the third harmonic component. That is, the B-mode processing circuit 130 extracts the third harmonic component as a high-frequency component of a multiple of three from the reflected wave data based on the output data output from the trained model 170e.

[0173] Here, the three reflected wave data used in extracting the third harmonic component are data in which the influence of saturation is reduced. Therefore, according to the ultrasound diagnostic device 1 according to the fifth embodiment, similarly to the first embodiment, etc., it is possible to obtain ultrasound image data in which the influence of saturation is robustly reduced and deterioration of image quality is suppressed.

[0174] (Modification) Next, an ultrasonic diagnostic device 1 according to a modification will be described. FIG. 22 is a flowchart showing a flow of an example of a process executed by the ultrasonic diagnostic device 1 according to the modification. In each of the above-mentioned embodiments, as shown in FIG. 22, when extracting a third harmonic component, the control circuit 180 of the ultrasonic diagnostic device 1 judges whether or not the input data used at the time of inference is affected by saturation (step S201). Then, when it is judged that the input data used at the time of inference is affected by saturation (step S201: Yes), the ultrasonic diagnostic device 1 may execute the above-mentioned processes of steps S101 to S103. For example, in step S101, the ultrasonic diagnostic device 1 extracts a third harmonic component by performing the above-mentioned processes using a trained model (trained models 170a, 170b, 170c, 170d, and 170e), and generates B-mode image data based on the extracted third harmonic component.

[0175] On the other hand, when it is determined that the input data used during inference is not affected by saturation (step S201: No), the ultrasound diagnostic device 1 extracts the third harmonic components without using a trained model, and generates B-mode image data based on the extracted third harmonic components (step S202). For example, in step S202, the three reflected wave data used when the B-mode processing circuit 130 extracts the third harmonic components are the reflected wave data generated by the beamformer 115 without using a trained model. Then, the ultrasound diagnostic device 1 proceeds to step S102.

[0176] The program executed by the processor is provided in advance in a ROM (Read Only Memory) or a storage circuit. The program may be provided in a format that can be installed in these devices or in a format that can be executed, recorded on a non-transient computer-readable recording medium such as a CD (Compact Disk)-ROM, a FD (Flexible Disk), a CD-R (Recordable), or a DVD (Digital Versatile Disk). The program may be provided or distributed by being stored on a computer connected to a network such as the Internet and downloaded via the network. For example, the program is composed of modules including each of the above-mentioned processing functions. As actual hardware, a CPU reads out and executes the program from a recording medium such as a ROM, and each module is loaded onto a main storage device and generated on the main storage device.

[0177] According to at least one of the embodiments and at least one of the modified examples described above, it is possible to obtain an ultrasound signal in which the influence of saturation is robustly reduced and degradation of image quality is suppressed.

[0178] Although some embodiments have been described, these embodiments are presented as examples and are not intended to limit the scope of the invention. These embodiments can be implemented in various other forms, and various omissions, substitutions, and modifications can be made without departing from the spirit of the invention. These embodiments and their modifications are included in the scope of the invention and its equivalents as described in the claims, as well as in the scope and spirit of the invention. [Explanation of symbols]

[0179] 1. Ultrasound diagnostic equipment 112 Receiving circuit 130 B-mode processing circuit 150 Image generation circuit 170 Memory circuit 170a Pre-trained model

Claims

1. A storage unit that stores a trained model trained by using a first ultrasonic signal including a saturated signal as input data and a second ultrasonic signal in which the influence of saturation on the first ultrasonic signal is reduced as teacher data; A generation unit that generates the fourth ultrasonic signal by inputting a third ultrasonic signal including a saturated signal to the trained model and acquiring a fourth ultrasonic signal output from the trained model in which an effect of saturation on the third ultrasonic signal is reduced; An ultrasound diagnostic device comprising:

2. the generating unit generates the first ultrasonic signal and the third ultrasonic signal by executing analog signal processing including a clamping process for suppressing a signal value of an ultrasonic signal obtained by transmitting and receiving ultrasonic waves to a subject, the signal value being equal to or greater than a certain value, to a certain value. The ultrasonic diagnostic apparatus according to claim 1 .

3. The first ultrasonic signal and the third ultrasonic signal are data after a delay and sum process is performed. The ultrasonic diagnostic apparatus according to claim 1 .

4. The first ultrasonic signal and the third ultrasonic signal are data after being converted into digital signals and before being subjected to a delay-and-sum process. The ultrasonic diagnostic apparatus according to claim 1 .

5. The teacher data is the second ultrasonic signal in which the influence of saturation is reduced with respect to the first ultrasonic signal after the delay and sum processing is performed. The ultrasonic diagnostic apparatus according to claim 1 .

6. The second ultrasonic signal is generated by applying a negative gain to a signal affected by saturation included in the first ultrasonic signal to reduce an amplitude value of the signal affected by saturation. The ultrasonic diagnostic apparatus according to claim 5 .

7. The second ultrasonic signal as the teacher data is obtained by reducing the effect of saturation on the ultrasonic signal after being converted into a digital signal and before being subjected to a delay-and-sum process. The ultrasonic diagnostic apparatus according to claim 1 .

8. The second ultrasonic signal as the teacher data is obtained by determining for each channel whether the ultrasonic signal is saturated or not, and weighting the ultrasonic signal for each channel using a weight for each channel according to the determination result. The ultrasonic diagnostic apparatus according to claim 7.

9. The second ultrasonic signal as the teacher data is obtained by determining whether or not the ultrasonic signal is saturated for each channel, and replacing the value of the ultrasonic signal of the channel determined to be saturated with the value of the ultrasonic signal estimated using the value of the ultrasonic signal of the channel determined not to be saturated. The ultrasonic diagnostic apparatus according to claim 7.

10. The second ultrasonic signal as the teacher data is data obtained by converting an ultrasonic signal, which is an analog signal that has been subjected to processing to reduce the effect of saturation, into a digital signal. The ultrasonic diagnostic apparatus according to claim 1 .

11. The process of reducing the effect of saturation is a process of suppressing signal components caused by saturation using an analog filter. The ultrasonic diagnostic apparatus according to claim 10.

12. The second ultrasonic signal as the teacher data is data obtained by converting the ultrasonic signal, which has been subjected to the process of applying an analog gain according to a result of the determination of saturation of the ultrasonic signal, into a digital signal. The ultrasonic diagnostic apparatus according to claim 10.

13. The generating unit further extracts odd-order harmonic components from the fourth ultrasonic signal, and generates ultrasonic image data based on the extracted odd-order harmonic components. The ultrasonic diagnostic apparatus according to claim 1 .

14. The generation unit determines whether the third ultrasonic signal is affected by saturation, and when it is determined that the third ultrasonic signal is affected by saturation, inputs the third ultrasonic signal to the trained model, and acquires the fourth ultrasonic signal output from the trained model, thereby generating the fourth ultrasonic signal. The ultrasonic diagnostic apparatus according to claim 1 .

15. A trained model is trained by using a first ultrasonic signal including a saturated signal as input data and a second ultrasonic signal obtained by reducing the effect of saturation on the first ultrasonic signal as teacher data, and inputting a third ultrasonic signal including a saturated signal into the trained model; generating the fourth ultrasonic signal by acquiring a fourth ultrasonic signal in which an effect of saturation is reduced with respect to the third ultrasonic signal output from the trained model; Ultrasonic signal generation method.

Images