Medical data processing device and data processing device

Abstract

To improve image quality.SOLUTION: A medical information processing device according to an embodiment comprises a processing unit. The processing unit applies linear operations with complex coefficients and non-linear activation independent of a complex argument to medical data with complex values.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The embodiments disclosed in the specification and drawings relate to a medical data processing device and a data processing device. [Background technology]

[0002] In machine learning using neural networks, real-valued neural networks are used as standard.

[0003] However, since signal processing using complex numbers is often used in medical data processing devices such as magnetic resonance imaging devices and ultrasound diagnostic devices, the use of complex neural networks is expected to lead to a variety of applications. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Special Publication No. 2019-525325 Summary of the Invention [Problem to be solved by the invention]

[0005] One of the problems to be solved by the embodiments disclosed in this specification and the drawings is to improve image quality. 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]

[0006] A medical data processing apparatus according to an embodiment includes a processing unit that applies linear computation of complex coefficients and nonlinear activation that does not depend on complex argument to medical data having complex values. [Brief description of the drawings]

[0007] [Figure 1] FIG. 1 is a diagram illustrating an example of a data processing device according to an embodiment. [Diagram 2] FIG. 2 is a diagram illustrating an example of a magnetic resonance imaging apparatus according to the embodiment. [Diagram 3] FIG. 3 is a diagram illustrating an example of an ultrasonic diagnostic apparatus according to an embodiment. [Figure 4] FIG. 4 is a diagram illustrating an example of a neural network according to the embodiment. [Diagram 5] FIG. 5 is a diagram illustrating a neural network according to the embodiment. [Figure 6] FIG. 6 is a diagram illustrating an example of a neural network according to the embodiment. [Figure 7] FIG. 7 is a diagram illustrating an example of a neural network according to the embodiment. [Figure 8] FIG. 8 is a diagram illustrating the background according to the embodiment. [Figure 9] FIG. 9 is a diagram illustrating an example of a neural network according to the embodiment. [Figure 10] FIG. 10 is a diagram illustrating an example of a process performed by the neural network according to the embodiment. [Figure 11] FIG. 11 is a diagram illustrating an example of a neural network according to the embodiment. [Figure 12] FIG. 12 is a diagram illustrating an example of a neural network according to the embodiment. [Figure 13] FIG. 13 is a diagram illustrating an example of a neural network according to the embodiment. [Figure 14] FIG. 14 is a diagram illustrating an example of a neural network according to the embodiment. [Figure 15] FIG. 15 is a diagram illustrating an example of a neural network according to the embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0008] (Embodiment) Hereinafter, embodiments of a medical data processing device and a data processing device will be described in detail with reference to the drawings.

[0009] The configuration of a data processing device 100 according to the embodiment will be described with reference to FIG.

[0010] The data processing device 100 is a device that generates data using machine learning. As an example, the data processing device 100 is connected to various medical image diagnostic devices, such as a magnetic resonance imaging device shown in FIG. 2 and an ultrasound diagnostic device shown in FIG. 3, and processes signals received from the medical image diagnostic devices, generates a trained model, executes the trained model, and so on. Note that examples of the medical image diagnostic devices to which the data processing device 100 is connected are not limited to a magnetic resonance imaging device and an ultrasound diagnostic device, and may be other devices, such as an X-ray CT device or a PET device.

[0011] The data processing device 100 is typically a medical data processing device that processes medical data. However, the embodiment is not limited to the case where the data processing device 100 is a medical data processing device. As an example, the data processing device 100 may be a device that processes magnetic resonance data that is not medical data.

[0012] The medical image processing device 100 includes a processing circuit 110, a memory 132, an input device 134, and a display 135. The processing circuit 110 includes a training data creation function 110a, a learning function 110b, an interface function 110c, a control function 110d, an application function 110e, and an acquisition function 110f.

[0013] In the embodiment, each processing function and the trained model (e.g., neural network) performed by the training data creation function 110a, the learning function 110b, the interface function 110c, the control function 110d, the application function 110e, and the acquisition function 110f are stored in the memory 132 in the form of a program executable by a computer. The processing circuit 110 is a processor that realizes the function corresponding to each program by reading the program from the memory 132 and executing it. In other words, the processing circuit 110 in a state in which each program has been read has each function shown in the processing circuit 110 of FIG. 1. Furthermore, the processing circuit 110 in a state in which a program corresponding to a trained model (neural network) has been read can perform processing according to the trained model. Note that, although the processing circuit 110 is described in FIG. 1 as being realized by a single processing circuit, the processing circuit 110 may be configured by combining multiple independent processors, and each processor may realize the function by executing a program. In other words, each of the above-mentioned functions may be configured as a program, and one processing circuit may execute each program. Also, a single processing circuit may realize two or more of the functions of the processing circuit 110. As another example, a specific function may be implemented in a dedicated, independent program execution circuit.

[0014] In addition, in FIG. 1, the processing circuit 110, the training data creation function 110a, the learning function 110b, the interface function 110c, the control function 110d, the application function 110e, and the acquisition function 110f are examples of a processing unit, a creation unit, an input unit (learning unit), a reception unit, a control unit, an application unit, and an acquisition unit, respectively.

[0015] The term "processor" used in the above description means a circuit such as a CPU (Central Processing Unit), a GPU (Graphical Processing Unit), an Application Specific Integrated Circuit (ASIC), a programmable logic device (e.g., a Simple Programmable Logic Device (SPLD), a Complex Programmable Logic Device (CPLD), and a Field Programmable Gate Array (FPGA)). The processor realizes functions by reading and executing a program stored in memory 132.

[0016] Also, instead of storing a program in the memory 132, the program may be directly embedded in the circuit of the processor. In this case, the processor realizes the function by reading and executing the program embedded in the circuit. Therefore, for example, instead of storing a trained model in the memory 132, a program related to the trained model may be directly embedded in the circuit of the processor.

[0017] In addition, when the processing circuitry 110 is incorporated into various medical image diagnostic devices, or when the processing circuitry 110 performs processing in cooperation with various medical image diagnostic devices, the processing circuitry 110 may have a function of simultaneously executing processing related to these devices.

[0018] The processing circuit 110 uses the training data generation function 110a to generate training data for learning based on the data and images acquired by the interface function 110c.

[0019] The processing circuit 110 performs learning using the training data generated by the training data creation function 110a through the learning function 110b, and generates a learned model.

[0020] The processing circuit 110 obtains data, images, etc. for image generation by the application function 110e from the memory 132 via the interface function 110c.

[0021] The processing circuit 110, by means of a control function 110d, controls the overall processing of the data processing device 100. Specifically, the processing circuit 110, by means of the control function 110d, controls the processing of the processing circuit 110 based on various setting requests input by an operator via the input device 134 and various control programs and various data read from the memory 132.

[0022] Furthermore, the processing circuit 110 generates an image based on the results of processing performed using the training data generation function 110a and the learning function 110b, using the application function 110e. Furthermore, the processing circuit 110 applies the trained model generated by the learning function 110b to the input image, using the application function 110e, and generates an image based on the application result of the trained model.

[0023] The memory 132 is composed of a semiconductor memory element such as a RAM (Random Access Memory), a flash memory, a hard disk, an optical disk, etc. The memory 132 is a memory that stores data such as image data for display and image data for training generated by the processing circuitry 110.

[0024] The memory 132 stores various data such as control programs for performing image processing and display processing as necessary.

[0025] The input device 134 receives various instructions and information input from an operator. The input device 134 is, for example, a pointing device such as a mouse or a trackball, a selection device such as a mode changeover switch, or an input device such as a keyboard.

[0026] The display 135 displays a GUI (Graphical User Interface) for receiving input of imaging conditions and images generated by the control function 110d under the control of the control function 110d, etc. The display 135 is, for example, a display device such as a liquid crystal display. The display 135 is an example of a display unit. The display 135 has a mouse, a keyboard, a button, a panel switch, a touch command screen, a foot switch, a trackball, a joystick, etc.

[0027] FIG. 2 shows an example of a magnetic resonance imaging apparatus 200 incorporating the data processing apparatus 100 according to the embodiment.

[0028] 2, the magnetic resonance imaging apparatus 200 includes a static magnetic field magnet 201, a static magnetic field power supply (not shown), a gradient magnetic field coil 203, a gradient magnetic field power supply 204, a bed 205, a bed control circuit 206, a transmission coil 207, a transmission circuit 208, a reception coil 209, a reception circuit 210, a sequence control circuit 220 (sequence control unit), and the data processing device 100 described in Fig. 1. Note that the magnetic resonance imaging apparatus 200 does not include a subject P (e.g., a human body). Also, the configuration shown in Fig. 2 is merely an example.

[0029] The static magnetic field magnet 201 is a magnet formed in a hollow, approximately cylindrical shape, and generates a static magnetic field in the internal space. The static magnetic field magnet 201 is, for example, a superconducting magnet, and is excited by receiving a current from a static magnetic field power supply. The static magnetic field power supply supplies a current to the static magnetic field magnet 201. As another example, the static magnetic field magnet 201 may be a permanent magnet, in which case the magnetic resonance imaging apparatus 200 may not include a static magnetic field power supply. Also, the static magnetic field power supply may be provided separately from the magnetic resonance imaging apparatus 100.

[0030] The gradient magnetic field coil 203 is a hollow coil formed in a substantially cylindrical shape, and is disposed inside the static magnetic field magnet 201. The gradient magnetic field coil 203 is formed by combining three coils corresponding to the mutually orthogonal X, Y, and Z axes, and these three coils are individually supplied with current from a gradient magnetic field power supply 204 to generate gradient magnetic fields whose magnetic field strength changes along the X, Y, and Z axes. The gradient magnetic fields of the X, Y, and Z axes generated by the gradient magnetic field coil 203 are, for example, a slice gradient magnetic field Gs, a phase encoding gradient magnetic field Ge, and a readout gradient magnetic field Gr. The gradient magnetic field power supply 204 supplies current to the gradient magnetic field coil 203.

[0031] The bed 205 includes a top plate 205a on which the subject P is placed, and under the control of a bed control circuit 206, the top plate 205a is inserted into the cavity (imaging port) of the gradient magnetic field coil 203 with the subject P placed thereon. The bed 205 is usually installed so that its longitudinal direction is parallel to the central axis of the static magnetic field magnet 201. Under the control of the data collecting device 100, the bed control circuit 206 drives the bed 205 to move the top plate 205a in the longitudinal direction and the up-down direction.

[0032] The transmission coil 207 is disposed inside the gradient magnetic field coil 203, and generates a high-frequency magnetic field upon receiving an RF pulse from a transmission circuit 208. The transmission circuit 208 supplies the transmission coil 207 with an RF pulse corresponding to a Larmor frequency determined by the type of target atom and the magnetic field strength.

[0033] The receiving coil 209 is disposed inside the gradient magnetic field coil 203, and receives a magnetic resonance signal (hereinafter, referred to as an “MR signal” as necessary) emitted from the subject P due to the influence of a high frequency magnetic field. When the receiving coil 209 receives a magnetic resonance signal, the receiving coil 209 outputs the received magnetic resonance signal to the receiving circuit 210.

[0034] The above-mentioned transmitting coil 207 and receiving coil 209 are merely examples. The coils may be configured by combining one or more of a coil having only a transmitting function, a coil having only a receiving function, or a coil having a transmitting and receiving function.

[0035] The receiving circuit 210 detects the magnetic resonance signal output from the receiving coil 209, and generates magnetic resonance data based on the detected magnetic resonance signal. Specifically, the receiving circuit 210 generates magnetic resonance data by digitally converting the magnetic resonance signal output from the receiving coil 209. The receiving circuit 210 also transmits the generated magnetic resonance data to the sequence control circuit 220. The receiving circuit 210 may be provided on the side of a gantry device including the static magnetic field magnet 201, the gradient magnetic field coil 203, and the like.

[0036] The sequence control circuit 220 drives the gradient magnetic field power supply 204, the transmission circuit 208, and the reception circuit 210 based on the sequence information to image the subject P. Here, the sequence information is information that defines a procedure for performing imaging. The sequence information defines the strength of the current that the gradient magnetic field power supply 204 supplies to the gradient magnetic field coil 203 and the timing of supplying the current, the strength of the RF pulse that the transmission circuit 208 supplies to the transmission coil 207 and the timing of applying the RF pulse, and the timing of the reception circuit 210 detecting a magnetic resonance signal. For example, the sequence control circuit 220 is an integrated circuit such as an ASIC (Application Specific Integrated Circuit) or an FPGA (Field Programmable Gate Array), or an electronic circuit such as a CPU (Central Processing Unit) or an MPU (Micro Processing Unit). The sequence control circuit 220 is an example of a scan unit.

[0037] Furthermore, when the sequence control circuit 220 receives magnetic resonance data from the receiving circuit 210 as a result of driving the gradient magnetic field power supply 204, the transmitting circuit 208, and the receiving circuit 210 to image the subject P, the sequence control circuit 220 transfers the received magnetic resonance data to the data processing device 100. The data processing device 100 performs the overall control of the magnetic resonance imaging apparatus 200 in addition to the processing described in FIG.

[0038] 1, to explain processing performed by the data processing device 100 and other than the processing described in Fig. 1, the processing circuit 110 transmits sequence information to the sequence control circuit 220 via the interface function 110c, and receives magnetic resonance data from the sequence control circuit 220. Furthermore, upon receiving the magnetic resonance data, the processing circuit 110 having the interface function 110c stores the received magnetic resonance data in the memory 132.

[0039] The magnetic resonance data stored in the memory 132 is arranged in k-space by the control function 110d, so that the memory 132 stores the k-space data.

[0040] The memory 132 stores magnetic resonance data received by the processing circuit 110 having the interface function 110c, k-space data arranged in k-space by the processing circuit 110 having the control function 110d, image data generated by the processing circuit 110 having the generation function (or application function 110e), etc.

[0041] The processing circuitry 110 performs overall control of the magnetic resonance imaging apparatus 200 using the control function 110d, and controls imaging, image generation, image display, etc. For example, the processing circuitry 110 having the control function 110d accepts input of imaging conditions (imaging parameters, etc.) on a GUI, and generates sequence information according to the accepted imaging conditions. In addition, the processing circuitry 200 having the control function 110d transmits the generated sequence information to the sequence control circuit 220.

[0042] The processing circuit 110 reads the k-space data from the memory 132 using a generation function (or application function 110e) not shown in FIG. 1, and generates a magnetic resonance image by performing reconstruction processing such as Fourier transform on the read k-space data.

[0043] 3 shows a configuration example of an ultrasonic diagnostic device 300 incorporating the data processing device 100 according to the embodiment. The ultrasonic diagnostic device according to the embodiment includes an ultrasonic probe 305 and an ultrasonic diagnostic device main body 300. The ultrasonic diagnostic device main body 300 includes a transmission circuit 309, a reception circuit 311, and the above-mentioned data processing device 100.

[0044] The ultrasonic probe 305 has a plurality of piezoelectric transducers, which generate ultrasonic waves based on a drive signal supplied from a transmission circuit 309 included in the ultrasonic diagnostic device main body 300 described later. The plurality of piezoelectric transducers included in the ultrasonic probe 305 receive reflected waves from the subject P and convert them into electrical signals (reflected wave signals). The ultrasonic probe 305 also has a matching layer provided on the piezoelectric transducers, a backing material that prevents ultrasonic waves from propagating backward from the piezoelectric transducers, and the like. The ultrasonic probe 305 is detachably connected to the ultrasonic diagnostic device main body 300. The ultrasonic probe 305 is an example of a scan unit.

[0045] When ultrasonic waves are transmitted from the ultrasonic probe 305 to the subject P, the transmitted ultrasonic waves are successively reflected by discontinuous surfaces of acoustic impedance in the internal tissues of the subject P, and are received as reflected waves by a plurality of piezoelectric transducers of the ultrasonic probe 305, and converted into reflected wave signals. The amplitude of the reflected wave signal 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 blood flow, heart wall, or the like, the reflected wave signal undergoes a frequency shift due to the Doppler effect depending on the velocity component of the moving body in the ultrasonic transmission direction.

[0046] The ultrasonic diagnostic device main body 300 is a device that generates ultrasonic image data based on a reflected wave signal received from the ultrasonic probe 305. The ultrasonic diagnostic device main body 300 is a device that can generate two-dimensional ultrasonic image data based on a two-dimensional reflected wave signal, and can generate three-dimensional ultrasonic image data based on a three-dimensional reflected wave signal. However, the embodiment is also applicable to the case where the ultrasonic diagnostic device 10 is a device dedicated to two-dimensional data.

[0047] As shown in FIG. 3, the ultrasound diagnostic apparatus 10 includes a transmission circuit 309, a reception circuit 311, and a medical image processing apparatus 100.

[0048] The transmission circuit 309 and the reception circuit 311 control the ultrasonic transmission and reception performed by the ultrasonic probe 305 based on instructions from the data processing device 110 having a control function. The transmission circuit 309 has a pulse generator, a transmission delay unit, a pulser, etc., and supplies a drive signal to the ultrasonic probe 305. The pulse generator repeatedly generates a rate pulse for forming a transmission ultrasonic wave at a predetermined pulse repetition frequency (PRF: Pulse Repetition Frequency). In addition, the transmission delay unit focuses the ultrasonic waves generated from the ultrasonic probe 305 into a beam shape and provides a delay time for each piezoelectric transducer required to determine the transmission directivity to each rate pulse generated by the pulse generator. In addition, the pulser applies a drive signal (drive pulse) to the ultrasonic probe 305 at a timing based on the rate pulse.

[0049] That is, the transmission delay unit arbitrarily adjusts the transmission direction of the ultrasonic waves transmitted from the piezoelectric transducer surface by changing the delay time given to each rate pulse. Also, the transmission delay unit controls the position of the focal point (transmission focus) in the depth direction of the ultrasonic transmission by changing the delay time given to each rate pulse.

[0050] The receiving circuit 311 has an amplifier circuit, an A / D (Analog / Digital) converter, a receiving delay circuit, an adder, a quadrature detection circuit, etc., and performs various processes on the reflected wave signal received from the ultrasonic probe 305 to generate a receiving signal (reflected wave data). The amplifier circuit amplifies the reflected wave signal for each channel and performs gain correction processing. The A / D converter A / D converts the gain-corrected reflected wave signal. The receiving delay circuit provides the digital data with a receiving delay time required to determine the receiving directivity. The adder performs an adding process of the reflected wave signal to which the receiving delay time has been provided by the receiving delay circuit. The adding process of the adder emphasizes the reflected component from the direction according to the receiving directivity of the reflected wave signal. The quadrature detection circuit converts the output signal of the adder into an in-phase signal (I signal, I: In-phase) and a quadrature signal (Q signal, Q: Quadrature-phase) of the baseband. Then, the quadrature detection circuit transmits the I signal and Q signal (hereinafter, referred to as IQ signal) as a received signal (reflected wave data) to the processing circuit 110. Note that the quadrature detection circuit may convert the output signal of the adder into an RF (Radio Frequency) signal and transmit it to the processing circuit 110. The IQ signal and RF signal become received signals having phase information.

[0051] When scanning a two-dimensional region in the subject P, the transmission circuit 309 causes the ultrasonic probe 305 to transmit an ultrasonic beam for scanning the two-dimensional region. Then, the reception circuit 311 generates a two-dimensional reception signal from the two-dimensional reflected wave signal received from the ultrasonic probe 305. When scanning a three-dimensional region in the subject P, the transmission circuit 309 causes the ultrasonic probe 305 to transmit an ultrasonic beam for scanning the three-dimensional region. Then, the reception circuit 311 generates a three-dimensional reception signal from the three-dimensional reflected wave signal received from the ultrasonic probe 305. The reception circuit 311 generates a reception signal based on the reflected wave signal, and transmits the generated reception signal to the processing circuit 110.

[0052] The transmission circuit 309 causes the ultrasonic probe 305 to transmit an ultrasonic beam from a predetermined transmission position (transmission scanning line). The reception circuit 311 receives a signal due to a reflected wave of the ultrasonic beam transmitted by the transmission circuit 309 from the ultrasonic probe 305 at a predetermined reception position (reception scanning line). When parallel simultaneous reception is not performed, the transmission scanning line and the reception scanning line are the same scanning line. On the other hand, when parallel simultaneous reception is performed, when the transmission circuit 309 causes the ultrasonic probe 305 to transmit one ultrasonic beam at one transmission scanning line, the reception circuit 311 simultaneously receives the signal due to the reflected wave originating from the ultrasonic beam transmitted by the transmission circuit 309 to the ultrasonic probe 305 as multiple reception beams at multiple predetermined reception positions (reception scanning lines) through the ultrasonic probe 305.

[0053] The data processing device 100 is connected to a transmission circuit 309 and a reception circuit 311, and in addition to the functions already shown in Fig. 1, processes signals received from the reception circuit 311, controls the transmission circuit 309, generates trained models, executes the trained models, and executes various image processing. In addition to the functions already shown in Fig. 1, the processing circuit 110 has a B-mode processing function, a Doppler processing function, a generation function, etc. Returning to Fig. 1, the following describes configurations that the data processing device 100 incorporated in the ultrasound diagnostic device 10 may have in addition to the configurations already shown in Fig. 1.

[0054] Each processing function and learned model performed in the B-mode processing function, Doppler processing function, and generation function is stored in the memory 132 in the form of a program executable by a computer. The processing circuitry 110 is a processor that realizes the function corresponding to each program by reading and executing the program from the memory 132. In other words, the processing circuitry 110 in a state in which each program has been read has each of these functions.

[0055] The B-mode processing function and the Doppler processing function are examples of a B-mode processing section and a Doppler processing section.

[0056] The processing circuit 110 performs various signal processing on the received signal received from the receiving circuit 311 .

[0057] Using its B-mode processing function, the processing circuit 110 receives data from the receiving circuit 311 and performs logarithmic amplification processing, envelope detection processing, logarithmic compression processing, etc. to generate data (B-mode data) in which the signal intensity is expressed as brightness.

[0058] In addition, the processing circuit 110 uses a Doppler processing function to perform frequency analysis of velocity information from the received signal (reflected wave data) received from the receiving circuit 311, and generates data (Doppler data) extracted from multiple points of moving object information such as velocity, dispersion, and power due to the Doppler effect.

[0059] In addition, the B-mode processing function and the Doppler processing function are capable of processing both two-dimensional reflected wave data and three-dimensional reflected wave data.

[0060] Furthermore, the processing circuitry 110 controls the overall processing of the ultrasonic diagnostic apparatus by the control function 110d. Specifically, the processing circuitry 110 controls the processing of the transmission circuitry 309, the reception circuitry 311, and the processing circuitry 110 by the control function 110d based on various setting requests input by the operator via the input device 134 and various control programs and various data read from the memory 132. Furthermore, the processing circuitry 110 controls the display 135 to display ultrasonic image data for display stored in the memory 132 by the control function 110d.

[0061] Furthermore, the processing circuitry 110 generates ultrasound image data from data generated by the B-mode processing function and the Doppler processing function using a generating function (not shown). The processing circuitry 110 generates, using the generating function, 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 function. The processing circuitry 110 also generates, using the generating function, two-dimensional Doppler image data representing moving object information from the two-dimensional Doppler data generated by the Doppler processing function 110b. The two-dimensional Doppler image data is velocity image data, variance image data, power image data, or image data that is a combination of these.

[0062] Furthermore, the processing circuit 110 uses a generation function to convert (scan convert) a scan line signal sequence of an ultrasonic scan into a scan line signal sequence of a video format represented by a television or the like, and generates ultrasonic image data for display. Furthermore, the processing circuit 110 uses a generation function to perform various image processing other than scan conversion, such as image processing (smoothing processing) for regenerating an average brightness image using multiple image frames after scan conversion, and image processing (edge ​​enhancement processing) using a differential filter within an image. Furthermore, the processing circuit 110 uses a generation function to perform various rendering processing on the volume data in order to generate two-dimensional image data for displaying the volume data on the display 135.

[0063] The memory 132 can also store data generated in the B-mode processing function and the Doppler processing function. The B-mode data and Doppler data stored in the memory 132 can be called up by an operator after a diagnosis, for example, and become ultrasonic image data for display via the processing circuitry 110. The memory 132 can also store the reception signal (reflected wave data) output by the reception circuitry 311.

[0064] In addition, the memory 132 stores, as necessary, control programs for performing ultrasound transmission and reception, image processing, and display processing, as well as various data such as diagnostic information (e.g., patient ID, doctor's findings, etc.), diagnostic protocols, and various body marks.

[0065] Next, the configuration of the neural network according to the embodiment will be described with reference to FIGS.

[0066] FIG. 4 shows an example of inter-layer interconnections in a neural network 7 used for machine learning by a processing circuit 110 having a learning function 110b. The neural network 7 is composed of an input layer 1, an output layer 2, and intermediate layers 3, 4, 5, etc. between the input layer 1 and the output layer 2. Each intermediate layer is composed of a layer related to each input (hereinafter referred to as the input layer in each layer), a linear layer, and a layer related to processing using an activation function (hereinafter referred to as the activation layer). For example, the intermediate layer 3 is composed of an input layer 3a, a linear layer 3b, and an activation layer 3c, the intermediate layer 4 is composed of an input layer 4a, a linear layer 4b, and an activation layer 4c, and the intermediate layer 5 is composed of an input layer 5a, a linear layer 5b, and an activation layer 5c. Each layer is composed of a plurality of nodes (neurons).

[0067] Here, the data processing device 100 according to the embodiment applies a linear layer of complex coefficients and a nonlinear activation (activation function) to medical data having complex values. That is, the processing circuit 110 generates a trained model by training the neural network 7 that applies a linear layer of complex coefficients and a nonlinear activation (activation function) to medical data having complex values ​​by the learning function 110b. The processing circuit 100 stores the generated trained model in, for example, the memory 132 as necessary.

[0068] The data input to the input layer 1 is typically a medical image or medical image data acquired from a medical image diagnostic device. When the medical image diagnostic device is a magnetic resonance imaging device 200, the data input to the input layer 1 is, for example, a magnetic resonance image. When the medical image diagnostic device is, for example, an ultrasound diagnostic device 300, the data input to the input layer 1 is, for example, an ultrasound image.

[0069] In addition, the input data input to the input layer 1 may be a medical image, or may be various image data, projection data, intermediate data, or raw data before a medical image is generated. For example, when the medical image diagnostic device is a PET device, the input data input to the input layer 10 may be a PET image, or may be various data before the reconstruction of a PET image, such as time-series data of simultaneous coefficient information.

[0070] The data output from the output layer 2 is, for example, a medical image or medical image data, and may be various projection data, intermediate data, or raw data prior to the generation of a medical image, similar to the data input to the input layer 1. When the purpose of the neural network 7 is denoising, the data output from the output layer 2 is, for example, an image with noise removed and with higher image quality than the input image.

[0071] In addition, when the neural network 7 according to the embodiment is, for example, a convolutional neural network (CNN), the data input to the input layer 1 is data represented by a two-dimensional array having a size of, for example, 32×32, and the data output from the output layer 2 is data represented by a two-dimensional array having a size of, for example, 32×32. The size of the data input to the input layer 1 and the size of the data output from the output layer 2 may be the same or different. Similarly, the number of nodes in an intermediate layer may be the same as or different from the number of nodes in the layers before and after it.

[0072] Next, the generation of a trained model according to the embodiment, that is, the training step, will be described. The processing circuit 110 generates a trained model by performing, for example, machine learning on the neural network 7 using the training function 110b. Here, performing machine learning means, for example, determining weights in the neural network 7 consisting of the input layer 1, intermediate layers 3, 4, 5, and output layer 2, and specifically, determining a set of coefficients that characterize the connection between the input layer 1 and the intermediate layer 3, a set of coefficients that characterize the connection between the intermediate layer 3 and the intermediate layer 4, ..., a set of coefficients that characterize the connection between the intermediate layer 5 and the output layer 2. The processing circuit 150 determines these sets of coefficients using the training function 110b, for example, by backpropagation.

[0073] The processing circuit 110 performs machine learning using the learning function 110b based on training data, which is teacher data consisting of data input to the input layer 1 and data output to the output layer 12, determines weighting between each layer, and generates a learned model with the weighting determined.

[0074] In addition, in deep learning, autoencoders can be used, and in this case, the data required for machine learning does not need to be supervised data.

[0075] Next, the process of applying the trained model according to the embodiment will be described. First, the processing circuit 110 inputs an input medical image, for example a clinical image, to the trained model by the application function 110e. For example, the processing circuit 110 inputs the input medical image, which is a clinical image, to the input layer 1 of the neural network 7, which is a trained model, by the application function 110e. Next, the processing circuit 110 acquires data output from the output layer 2 of the neural network 7, which is a trained model, as an output medical image by the application function 110e. The output medical image is a medical image that has been subjected to a predetermined process, such as noise removal. In this way, the processing circuit 150 generates an output medical image that has been subjected to a predetermined process, such as noise removal, by the application function 150e. If necessary, the processing circuit 150 may display the obtained output medical image on the display 135 by the control function 110d.

[0076] Returning to the explanation of the activation function and the activation layer, the activation function in the neural network 7 will be explained with reference to FIG. 5. In FIG. 5, nodes 10a, 10b, 10c, and 10d are cut-outs of some nodes in an input layer in a certain layer. Meanwhile, node 11 is one of the nodes in the linear layer, and node 12 is one of the nodes in the activation layer, which is a layer related to processing (activation) using the activation function. Note that in the embodiment, a complex neural network is used, but here, the case of a real neural network will be explained first.

[0077] Here, the output values ​​of nodes 10a, 10b, 10c, and 10d are real numbers x1 and x 2、 x 3、 Consider the case of x4, the output to node 11 in the linear layer is Σ i=1 m (ω i x i +b), where ω iis the weighting coefficient between the i-th input layer and node 11, m is the number of nodes to which node 11 is connected, and b is a predetermined constant. Next, if the output result output to node 12, which is the activation layer, is y, y is expressed as the following equation (1) using the activation function f.

[0078]

number

[0079] Here, the activation function f is usually a nonlinear function, and for example, a sigmoid function, a tanh function, ReLu (Rectified Linear Unit), or the like is selected as the activation function f.

[0080] FIG. 6 shows a process using such an activation function. In FIG. 6, the hidden layer 5 is the nth layer of the neural network 7, and is composed of an input layer 5a, a linear layer 5b, and an activation layer 5c. The input layer 6a is the (n+1)th layer of the neural network. The input layer 5a has nodes 20a, 20b, 20c, 20d, etc., the linear layer 5b has nodes 21a, 21b, 21c, 21d, etc., and the linear layer 5c has nodes 22a, 22b, 22c, etc. FIG. 6 also shows a real-valued neural network in which each node has a real value, and the input result x n,i and the output result x of the input layer 6a n+1、i is a real value.

[0081] Here, a predetermined weighted sum is performed for each node of the input layer 5a, and an output result to the linear layer 5b is calculated. For example, the output result to the j-th node 21b in the linear layer 5b is Σ i=1 m ω i,j x n,i +b n,j where ω i,j is the weight coefficient between the i-th input layer and the j-th linear layer, b n,j is a predetermined constant known as the bias term. i,j is the first-order coefficient in the linear operation of the linear layer, bn,j can be said to be the zeroth coefficient in the linear operation in the linear layer. Next, the activation function f is applied to each node of the linear layer 5b, and the output result to the activation layer 5c is calculated. For example, the output to the j-th node 22b in the activation layer 5c is calculated by applying the activation function f n,j Using f n,j (Σ i=1 m ω i,j x n,i +b n,j ) The processing in the linear layer is not limited to the above processing, and various known methods may be used. For example, a convolution layer may be used as the linear layer.

[0082] Next, the value of each node of the n-th layer, the input layer 6a, is determined based on the value output by the node of the activation layer 5c. As an example, the value of each node of the activation layer 5c is directly input to each node of the input layer 6a. As another example, each node of the input layer 6a may be determined by applying a further nonlinear function to the activation layer 5c.

[0083] Next, a complex neural network will be described. Fig. 7 shows a configuration in which the neural network 7 used for machine learning by the processing circuit 110 having the learning function 110b is a complex neural network. The complex neural network shown in Fig. 7 has a similar configuration to the real neural network shown in Fig. 6, but the value z of each node n,j is a complex number. Also, the activation function f n,j is also defined in the domain of complex numbers and is a function that takes complex values. The input value input to the j-th node of the j-th input layer 6a of the n+1-th input layer is z n+1、j Then, for example, the following formula (2) holds: In this way, for example, by making each node of the neural network 7 a complex number node, it is possible to generate a neural network 7 having complex number values.

[0084]

number

[0085] Next, the background of the embodiment will be described.

[0086] In machine learning using neural networks, real-valued neural networks are the standard. However, in medical data processing devices such as magnetic resonance imaging devices and ultrasound diagnostic devices, signal processing using complex numbers is often used, so various applications are expected by using complex-valued neural networks.

[0087] Here, as a method for handling complex numbers in a neural network, for example, there is a method for handling complex numbers in a neural network by dividing the complex numbers into real and imaginary parts and considering each part as each node of a standard real neural network. As an example, a method for handling complex numbers in a neural network can be considered by applying ReLU as an activation function to each of the real and imaginary parts of the complex numbers.

[0088] As another example, there is a method for handling complex numbers in a neural network by expressing the complex numbers using absolute value (or signed absolute value) and phase, and considering each of them as each node of a standard real-valued neural network.

[0089] Here, in medical images such as magnetic resonance images and ultrasound images, the phase information of the image, for example, the phase gradient, is often important, while the absolute value of the phase is relatively rarely important. For example, in a magnetic resonance imaging device, a slight difference in the center frequency appears as a phase modulation of the entire image, but in many cases, the importance of the absolute value of the phase itself is relatively low. Therefore, in the application of a complex neural network to medical images, such as noise removal and area extraction, it is desirable to configure a neural network that does not ignore the phase information of the input image, while at the same time, does not greatly fluctuate the output result with respect to the phase modulation of the entire image.

[0090] As an example, when considering denoising processing using complex-convolution and ReLU for the real and imaginary parts, if the distribution of teacher images in a complex neural network is biased toward the real part, for example, coefficients that emphasize the real part of the input image and ignore the imaginary part may be learned.

[0091] If an image biased toward the imaginary part is applied as an input image to a trained model generated in this way, it is expected that the model will not perform as expected.

[0092] Such a situation is shown in FIG. 8, for example. FIG. 8 is a graph showing the denoising performance of a trained model in the case where the trained model is generated by using a complex neural network 7 in which six layers of Complex-convolution and CReLU are stacked. FIG. 8 plots the mean square error (MSE) of the output image after the trained model is applied as a function of phase when training is performed using 38 teacher images when generating the trained model, and a test image is applied to the generated trained model while modulating only the phase. As can be seen from FIG. 8, the quality of the output image after the trained model is applied is not constant with respect to the phase of the input image.

[0093] One method to solve the problem that the image quality of the output image after applying the trained model is not constant with respect to the phase of the input image is to randomly apply phase modulation to the teacher image to achieve statistical stabilization (phase argumentation). However, phase argumentation may not be efficient for learning in some cases.

[0094] The medical image data device according to the embodiment is based on this background, and constructs a neural network by combining an activation function that is angle-independent on the complex plane with a linear layer whose coefficients are complex numbers.

[0095] In other words, the data processing device 100 according to the embodiment has a processing circuit 110 that applies a linear layer of complex coefficients and a nonlinear activation that is independent of the complex argument (i.e., the gain does not change depending on the complex argument) to medical data having complex values.

[0096] Such a situation is shown in Fig. 9. Fig. 9 shows an extracted configuration of one hidden layer 5 shown in Fig. 4 in the neural network 7 trained by the data processing device 100 according to the embodiment. As already explained in Fig. 4, the hidden layer 5 is composed of an input layer 5a, a linear layer 5b, and an activation layer 5c (activation function), but in the data processing device 100 according to the embodiment, the activation layer 5c has an angle-independent activation function 5c1, which is a nonlinear activation independent of the complex argument.

[0097] That is, the processing circuit 110 generates a learned model by training a neural network 7 using the learning function 110b to apply a linear layer 5b with complex coefficients and an angle-independent activation function 5c1, which is a nonlinear activation that does not depend on a complex argument, to medical data having complex values.

[0098] Here, examples of the angle-independent activation function 5c1 include a function system such as those shown in formulas (3) to (8). Note that in formulas (3) to (8), z represents a complex number.

[0099]

number

[0100]

number

[0101]

number

[0102]

number

[0103] In equation (3), λ is a real number, and ReLU(x) (Rectified Linear Unit) is max(0, x) when x is a real number. The right-hand side of equation (3) is 0 when the absolute value of the complex number z is smaller than λ, and is a complex number whose absolute value is equal to |z|- λ and whose argument is equal to z when the absolute value of the complex number z is larger than λ. Therefore, equation (3) is a nonlinear activation that does not depend on the complex argument. Equation (3) can also be considered as an extension of the soft-shrink function defined for real numbers to complex numbers.

[0104] In addition, in formula (4), z * represents the complex conjugate of z. Equation (4) is a function similar to equation (3), but while equation (3) is a function constructed based on the first power of the absolute value of the complex number z, equation (4) is a function constructed based on the square of the absolute value of the complex number z. Equation (4) has the advantage that the value on the right hand side can be evaluated quickly when performing numerical calculations.

[0105] In addition, in equation (5), β is a real number. The absolute value of the right-hand side of equation (5) is also expressed as a nonlinear function of the input signal, and its argument is equal to the argument of the input signal. Therefore, equation (5) is a nonlinear activation that does not depend on the complex argument. Equation (5) can also be considered as an extension of the tanh-shrink function defined for real numbers to complex numbers.

[0106] In addition, in equation (6), p is a real number. The right-hand side of equation (6) also does not selectively increase the contribution of a specific complex argument direction, so equation (6) is a nonlinear activation that does not depend on the complex argument.

[0107] As a variation of equation (3), for example, the function system of equation (7) or equation (8) below can be considered.

[0108]

number

[0109]

number

[0110] The nonlinear activation independent of the complex argument in the above formulas (3) to (8) is merely an example, and various other function systems are possible. As an example, a function that can be expressed as f(z)=g(|z|)z using a real function g(x), that is, a function including a real function related to the absolute value of an input complex number, for example, a function obtained by multiplying a real function related to the absolute value of an input complex number by the input complex number, is an example of a nonlinear activation independent of a complex argument. As a real function constituting the nonlinear activation, for example, a function including a soft-shrink function, a tanh-shrink function, a power function, or a ReLU function, as described above, can be given.

[0111] FIG. 10 shows a situation in which activation according to the embodiment is applied. FIG. 10 is a graph showing the denoising performance of a trained model in the case where a trained model is generated by performing denoising processing using a complex-convolution with a bias term fixed to 0 and a complex neural network 7 in which a complex tanh-shrink function represented by formula (5) is stacked in six layers. That is, the processing circuit 110 further applies a complex convolution process with a bias term fixed to 0. In other words, the linear operation of complex coefficients that the processing circuit 110 applies to medical data having complex values ​​includes a process with a bias term fixed to 0, and also includes complex convolution. Note that the complex convolution process is not limited to the case where the bias term is fixed to 0. FIG. 10 shows a plot of the mean square error (MSE) of the output image after applying the trained model as a function of phase, in which a certain test image is applied to the trained model generated while modulating only the phase. As can be seen from FIG. 10, an output image with stable image quality was obtained even if the phase of the input image changed.

[0112] The embodiment is not limited to the above example, and the processing circuit 110 may generate a trained model using two or more nonlinear activations that do not depend on the complex argument by the learning function 110b. Such a situation is shown in FIG. 11. In FIG. 11, the intermediate layer 5 has an input layer 5a, a linear layer 5b, a first angle-independent activation function 5c2, and a second angle-independent activation function 5c3. That is, the processing circuit 110 applies the first angle-independent activation function 5c2 to some nodes in the linear layer 5b to generate an output result for the next layer, and applies the second angle-independent activation function 5c3 to some nodes other than the some nodes in the linear layer 5b to generate an output result for the next layer.

[0113] Here, as a first example of using nonlinear activation independent of two or more complex argument angles, there is a case where nonlinear activation independent of complex argument angles of different function systems is used. For example, the processing circuit 110 performs learning by applying nonlinear activation of multiple different function systems to medical data, such as a complex soft-shrink function as shown in formula (3) and a complex tanh-shrink function as shown in formula (5), using the learning function 110b. In the example of FIG. 10, the processing circuit 110 performs learning by applying a complex soft-shrink function as a first angle-independent activation function 5c2 to some nodes of the linear layer 5b, and applying a complex tanh-shrink function as a second angle-dependent activation function 5c3 to some nodes of the linear layer 5b using the learning function 110b.

[0114] As a second example of using nonlinear activation that is independent of two or more complex argument angles, there is a case where nonlinear activation is used using a plurality of functions that are the same function system but have different function parameter values, such as λ in Equation (3) and Equation (4), β in Equation (5), and p in Equation (6). In other words, the processing circuit 110 performs learning by applying nonlinear activation to medical data using a plurality of functions that are the same function system but have different function parameters, using the learning function 110b. As an example, the processing circuit 110 performs learning by applying a complex soft-shrink function given by Equation (3) where λ=λ1 as a first angle-independent activation function 5c2 to some nodes of the linear layer 5b, and applying a complex soft-shrink function given by Equation (3) where λ=λ2 as a second angle-dependent activation function 5c3 to some nodes of the linear layer 5b.

[0115] As another example, an identity function may be applied as an activation function to some nodes.

[0116] The function parameters may be fixed or may vary during the machine learning process, and may be a mixture of fixed function parameters and parameterized and variable function parameters in one neural network 7.

[0117] Such function parameters may also be trainable by machine learning. An example of such a configuration is shown in Figure 12.

[0118] As shown in FIG. 12, the processing circuit 110 includes a first neural network 7, which is a neural network that outputs an output image / output data for an input image / input data, and a second neural network 8 for adjusting an activation function in the first neural network 7. The second neural network 8 is connected to the activation layers 3c, 4c, and 5c of the first neural network 7, and controls the parameters of the activation function in the activation layers. That is, the processing circuit 110 includes the neural network 7 that applies a nonlinear activation that does not depend on a complex argument to medical data, and a calculation unit (not shown in FIG. 1) that optimizes function parameters related to the nonlinear activation that does not depend on the complex argument. The second neural network 8 is an example of the calculation unit described above.

[0119] As an example, in the case where the complex soft-shrink function given by the formula (3) is used as the activation function f, the value of the parameter λ of the complex soft-shrink function, which is the activation function in the activation layers 3c, 4c, and 5c of the first neural network 7, is expressed as follows, where i is the i-th layer: i It is defined as follows. i has a constant value for each layer. i The value of is optimized by the calculation unit through learning.

[0120] As an example of such a parameter optimization method, the processing circuit 110 may alternately execute a first learning, which is learning of weighting coefficients in the first neural network 7 executed by the learning function 110b, and a second learning, which is learning of parameter values ​​of the activation function of the first neural network 7 executed by the calculation unit.

[0121] As another example, the processing circuit 110 may perform a second learning, which is learning of the parameter value of the activation function of the first neural network 7, executed by the calculation unit, and then use the parameter value to perform a first learning, which is learning of the weighting coefficient in the first neural network 7, executed by the learning function 110b.

[0122] Furthermore, the processing circuit 110 may perform the first learning and the second learning simultaneously.

[0123] The configuration of the calculation unit is not limited to a neural network, and the parameter values ​​of the activation function of the first neural network 7 may be optimized using, for example, linear regression or the like.

[0124] In addition, the above-mentioned embodiment has been described as using parameters that are different for each layer and common for each node, but the embodiment is not limited to this, and the parameters may be common to each layer, or different parameters may be used for each layer and node.

[0125] In the above embodiments, the case where only an angle-independent activation function is used as the activation function of the neural network 7 has been described, but the embodiments are not limited to this, and a complex argument-sensitive activation function (CPSAF: Complex Sensitive Activation Function) may also be used as the activation function of the neural network 7. That is, the processing circuit 110 may further apply, to the neural network 7, an activation whose gain changes according to the complex argument by the learning function 110b.

[0126] For example, in cases where it is desired to remove a specific complex argument component, such as in phase denoising, an activation function whose gain varies with the complex argument can be used in combination with a nonlinear activation that is independent of the complex argument, i.e., an activation function whose gain does not vary with the complex argument, thereby enabling efficient denoising.

[0127] Specifically, as shown in, for example, Figures 13 and 14, the processing circuit 110 sequentially applies a nonlinear activation that is independent of the complex argument and an activation whose gain changes depending on the complex argument by using the learning function 110b.

[0128] 13, the intermediate layer 5 includes an input layer 5a, a linear layer 5b, an activation layer 5c using an angle-independent activation function, and an activation layer 5d using an activation function sensitive to a complex argument. The processing circuit 110 outputs the output result of the input layer 5a to the linear layer 5b, outputs the output result of the linear layer 5b to the activation layer 5c using an angle-independent activation function, outputs the output result of the activation layer 5c using the angle-independent activation function to the activation layer 5d using an activation function sensitive to a complex argument, and outputs the output result of the activation layer 5d using the activation function sensitive to a complex argument to the next layer, the input layer, using the learning function 110b.

[0129] 14, the intermediate layer 5 includes an input layer 5a, a linear layer 5b, an activation layer 5c using an angle-independent activation function, a linear layer 5e, and an activation layer 5d using an activation function sensitive to a complex argument. The processing circuit 110 outputs the output result of the input layer 5a to the linear layer 5b, outputs the output result of the linear layer 5b to the activation layer 5c using an angle-independent activation function, outputs the output result of the activation layer 5c using the angle-independent activation function to the linear layer 5e, outputs the output result of the linear layer 5e to the activation layer 5d using an activation function sensitive to a complex argument, and outputs the output result of the activation layer 5d using the activation function sensitive to a complex argument to the next layer, the input layer, by the learning function 110b.

[0130] In addition, in Figures 13 and 14, the order in which the activation layer 5c using the angle-independent activation function and the activation layer 5d using the activation function sensitive to a complex argument are executed is not limited to the order shown in these figures. For example, the activation layer 5d using the activation function sensitive to a complex argument may be executed first, and then the activation layer 5c using the angle-independent activation function may be executed.

[0131] As another example, as shown in FIG. 15, the processing circuit 110 may selectively apply a nonlinear activation independent of a complex argument and an activation whose gain changes according to the complex argument. In the example of FIG. 15, the intermediate layer 5 has an input layer 5a, a linear layer 5b, an activation layer 5c1 using an angle-independent activation function, an activation layer 5c2 using an activation function sensitive to a complex argument, and an activation layer 5c3 using another activation function. Here, the processing circuit 110 outputs the output result of the input layer 5a to the linear layer 5b by the learning function 110b, outputs the output results of some nodes among the output results of the linear layer 5b to the activation layer 5c1 using the angle-independent activation function, and outputs the output result of the activation layer 5c1 using the angle-independent activation function to the corresponding node 6a1 in the input layer of the next layer. Meanwhile, the processing circuit 110 outputs the output results of another part of the nodes among the output results of the linear layer 5b to the activation layer 5c2 using an activation function sensitive to a complex argument, and outputs the output results of the activation layer 5c2 using the activation function sensitive to a complex argument to the corresponding node 6a2 in the next input layer. Also, the processing circuit 110 outputs the output results of yet another part of the nodes among the output results of the linear layer 5b to the activation layer 5c3 using another activation function, and outputs the output results of the activation layer 5c3 using the other activation function to the corresponding node 6a3 in the next input layer.

[0132] 15, the processing circuit 110 applies different non-selective activations to each node in the linear layer 5b. This allows efficient machine learning that reflects the properties of the nodes in the linear layer 5b, improving image quality.

[0133] Returning to the explanation of the activation function sensitive to the complex argument, a specific example of the above-mentioned activation function sensitive to the complex argument (CPSFA) is the function f1 given by the following equation (9).

[0134]

number

[0135] Here, z represents a complex number, phase(z) represents the complex argument of the complex number z, and α and β represent real parameters. β (x) is a function defined on a real number x, for example, a function that extracts an angle near x=0 by a method characterized by a parameter β. For example, the gain control function W β Here, we will explain the example of a function that has a maximum value at x=0 and decreases as the value moves away from x=0. Note that angles that differ by a constant multiple of 2π can be regarded as the same, so for example, the gain control function W β , which has a periodicity of 2π, and W β (x+2nπ)=W β We can also choose a periodic function for which (x) holds.

[0136] The activation function f1(z) is the complex number z with the gain control function W β (phase(z)-α). Therefore, the activation function f1(z) has a large gain (signal value) when the complex argument of z is relatively close to α, and the magnitude of the gain is controlled by the parameter β. Therefore, the activation function f1(z) expressed by the formula (9) αβ can be considered as a function expressed by the product of a gain control function that extracts a signal component in a predetermined angular direction and an input complex number, and is an example of an activation function that is sensitive to a complex argument.

[0137] Another example of an activation function that is sensitive to the complex argument is the activation function f2(z) given by the following equation (10).

[0138]

number

[0139] Here, the activation function f2(z) is expressed by the gain control function W β This is a special case where (x) is given by the following equation (11).

[0140]

number

[0141] Here, the wrap function on the right side of equation (11) is given by the following equation (12), where n is a natural number.

[0142]

number

[0143] That is, the gain control function W on the left side of equation (11) β (x) is a function that returns 1 if the angle x is within the range of β with 0 as the base, and 0 if not. In other words, the activation function f2 αβ (z) is a function that extracts a complex region in the range of angle β from the direction of angle α. In other words, the activation function f2 expressed by equation (10) αβ (z) can be considered as a function that extracts signal components within a range from a predetermined angle α to a predetermined angle β, and is an example of an activation function that is sensitive to a complex argument.

[0144] Further, as another example of activation functions sensitive to a complex argument, there are activation functions f3(z) to f5(z) given by the following equations (13) to (15).

[0145]

number

[0146]

number

[0147]

number

[0148] Here, the activation function f3 given by equation (13) αβ (z) is the activation function f1 αβ (z) is rotated right by an angle α, the real part is taken, and then rotated in the opposite direction to the previous rotation by an angle α. That is, the activation function f3 αβ (z) is a function corresponding to an operation including a rotation operation around the origin, an operation taking the real part of a complex number, and an operation of rotation in the opposite direction to the rotation operation.

[0149] In addition, the activation function f4 given by Eq. (14) αβ (z) is the activation function f3 αβ (z) is obtained by rotating the complex number z to the right by an angle α, taking the imaginary part, and then rotating it by an angle α in the opposite direction to the previous rotation operation.

[0150] In addition, in formula (15), A legacy is a standard activation function that returns a real value for a given real value. legacy Examples of the activation function f5 given by Eq. (15) are the sigmoid function, the soft sine function, the soft plus function, the tanh function, the ReLU, the truncated power function, the polynomial, the radial basis function, the wavelet, etc. αβ (z) is basically the activation function f3 αβ This is the same function as (z), but after taking the real part, we use the activation function A defined in real numbers. legacy The method further includes an operation of applying

[0151] To summarize the above, f1 expressed by equation (9) αβ (z), f2 expressed by Eq. (10) αβ (z), f3 expressed by equation (13) αβ (z), f4 represented by equation (14) αβ (z), f5 represented by equation (15) αβ(z) is a specific example of the activation function used in the activation layer 5d in FIGS. 13 and 14 and the activation layer 5c2 in FIG.

[0152] According to at least one of the embodiments described above, image quality can be improved.

[0153] 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, modifications, and combinations of embodiments 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]

[0154] 110 Processing circuit 132 Memory 134 Input Devices 135 Display

Claims

1. The medical data includes a processing unit that applies linear calculation of complex coefficients and nonlinear activation that does not depend on complex argument to medical data having complex values. Medical data processing equipment.

2. The medical data processing apparatus according to claim 1 , wherein the nonlinear activation is a function including a real function related to an absolute value of an input complex number.

3. The medical data processing device according to claim 2 , wherein the nonlinear activation is a function obtained by multiplying the real function by the input complex number.

4. 4. The medical data processing device according to claim 2, wherein the real function constituting the nonlinear activation is any one of a soft-shrink function, a tanh-shrink function, a power function, and a ReLU function.

5. The medical data processing apparatus according to claim 1 , wherein the processing unit applies the nonlinear activation of a plurality of different function systems to the medical data.

6. The medical data processing device according to claim 1 , wherein the processing unit applies the nonlinear activation to the medical data using a plurality of functions in the same function system but with different function parameters.

7. The processing unit includes a neural network that applies the nonlinear activation to the medical data; a calculation unit for optimizing a function parameter related to the nonlinear activation; The medical data processing device according to claim 1 .

8. The medical data processing device according to claim 1 , wherein the processing unit further applies activation in which a gain changes depending on the complex argument.

9. The medical data processing device according to claim 8 , wherein the processing unit sequentially applies a nonlinear activation that does not depend on the complex argument and an activation whose gain changes depending on the complex argument.

10. The medical data processing device according to claim 8 , wherein the processing unit selectively applies, to an output result of the linear calculation, a nonlinear activation that does not depend on the complex argument and an activation whose gain changes depending on the complex argument.

11. The medical data processing device according to claim 1 , wherein the linear calculation in the processing unit includes a process in which a bias term is fixed to zero.

12. The medical data processing device according to claim 1 , wherein the linear operation in the processing unit includes a complex convolution.

13. A processing unit that applies linear arithmetic of complex coefficients and nonlinear activation that is independent of complex argument to the complex-valued magnetic resonance data. Data processing device.

Images