KR20220107912A - Method and apparatus for multi-variable quantitative imaging using ultrasound data - Google Patents

Abstract

The present invention relates to a method for operating an imaging apparatus operated by at least one processor, comprising the steps of: receiving ultrasound data obtained from a tissue; and extracting a feature corresponding to a target variable selected from among a plurality of variables, from the ultrasound data, and restoring the feature, to output a quantitative image of the target variable.

Description

Multivariate quantitative imaging method and apparatus using ultrasound data

The present invention relates to ultrasound imaging.

Since cancer is difficult to detect early, periodic diagnosis is required, and the size and characteristics of the lesion must be continuously monitored. Representative imaging equipment for this purpose includes X-rays, magnetic resonance imaging (MRI), computed tomography (CT), ultrasound, and the like. While X-rays, MRIs, and CTs have the risk of radiation exposure, long measurement times, and high costs, ultrasound imaging equipment is safe, relatively inexpensive, and provides real-time images so that users can monitor the lesion in real time and obtain the desired image. can be obtained

Currently, as the most commercially available ultrasound imaging equipment, there is a B-mode (brightness mode) imaging system. The B-mode imaging method is a method of determining the position and size of an object through the time and intensity of ultrasonic waves reflected from the object surface. Since it finds the location of the lesion in real time, the user can efficiently obtain the desired image while monitoring the lesion in real time, and it is safe and relatively inexpensive, so accessibility is good. However, there are disadvantages in that image quality cannot be kept constant depending on the skill level of the user and quantitative characteristics cannot be imaged. That is, since the B-mode technique provides only the morphological information of the tissue, the sensitivity and specificity may be low in the differential diagnosis for distinguishing benign and malignant tumors, which are distinguished by histological characteristics.

Pathological changes in tissues cause structural changes in cells, and accordingly, imaging changes in ultrasound properties in the tissue. Representatively, Elastography and Ultrasound Computed Tomography (USCT) are used. have. Elastic imaging technique can quantitatively image the elasticity and stiffness of tissue, but it requires an additional dedicated device and consumes a lot of energy. Therefore, the elastic imaging technique can be applied only to expensive ultrasound equipment, and the frame rate is low, so it is not suitable for imaging dynamically moving tissues. The ultrasound tomography technique can obtain high-resolution quantitative images, but it is limited to mammography because the ultrasound sensor must surround the object, and there is a limitation in measuring various organs. In addition, in the ultrasound tomography technique, it takes minutes for imaging, so real-time movement cannot be seen, and the size of the system is very large, making it impossible to move.

The present disclosure provides a method and apparatus for imaging quantitative characteristics of various variables of tissue by inputting ultrasound signals beamformed into different beam patterns into tissue, and using multi-pattern ultrasound data that is reflected from the tissue and returned. will provide

The present disclosure provides an Attenuation Coefficient (AC), a Speed of Sound (SoS), and an Effective Scatterer Concentration (ESC) using multi-pattern ultrasound data acquired from a tissue through a single ultrasound probe. , to provide a neural network that complexly reconstructs quantitative characteristics of variables including scatterer size (Effective Scatterer Diameter, ESD).

The present disclosure provides a conditional encoder that variably encodes ultrasound data by changing a network parameter so as to optimally extract a target variable from ultrasound data for quantitative characteristics of each variable to be restored, thereby providing various quantitative characteristics of tissue. An object of the present invention is to provide a method and apparatus for improving image restoration performance for variables.

According to one embodiment, there is provided a method of operating an imaging device operated by at least one processor, comprising: receiving ultrasound data obtained from a tissue; and outputting a quantitative image of the target variable by extracting and reconstructing the feature.

The ultrasound data may be multi-pattern RF (Radio Frequency) data in which ultrasound signals incident to the tissue in different beam patterns are reflected back from the tissue.

The plurality of variables is an Attenuation Coefficient (AC), a Speed of Sound (SoS), an Effective Scatterer Concentration (ESC), and two or more of an Effective Scatterer Diameter (ESD). may include

The outputting of the quantitative image of the target variable may include extracting a feature corresponding to the target variable from the ultrasound data by changing a network parameter of the neural network according to the target variable selected from among the plurality of variables.

The neural network may perform conditional encoding for normalizing to a value selected according to the target variable.

The outputting of the quantitative image of the target variable may include outputting the quantitative image of the target variable through a decoder of the neural network that restores features output from the encoder of the neural network to the same path.

The decoder may be configured as a network that restores features of each variable while preserving structural similarity in a path for restoring features of different variables.

As an imaging apparatus according to another embodiment, a memory for loading a computer program, and executing the computer program, extracts a feature corresponding to a target variable selected from among a plurality of variables from ultrasound data obtained from a tissue, the feature and a processor for outputting a quantitative image of the target variable by restoring .

The ultrasound data may be multi-pattern RF (radio frequency) data in which ultrasound signals incident to the tissue in different beam patterns are reflected from the tissue and returned.

The plurality of variables is an Attenuation Coefficient (AC), a Speed of Sound (SoS), an Effective Scatterer Concentration (ESC), and two or more of an Effective Scatterer Diameter (ESD). may include

The processor may output a quantitative image corresponding to a target variable selected from among a plurality of variables from the ultrasound data through a neural network including an encoder and a decoder. The encoder may be configured as a network for extracting a feature corresponding to the target variable from the ultrasound data by changing a network parameter according to the target variable selected from among the plurality of variables. The decoder may be configured as a network that restores the features output from the encoder to the same path.

The neural network may perform conditional encoding for normalizing to a value selected according to the target variable.

The decoder may be configured as a network that restores features of each variable while preserving structural similarity in a path for restoring features of different variables.

According to another embodiment, a computer program including instructions stored in a computer-readable storage medium and executed by a processor, which receives ultrasound data obtained from a tissue, and a purpose among a plurality of variables recoverable from the ultrasound data When a variable is selected, an encoder for extracting a feature corresponding to the target variable from the ultrasound data by changing a network parameter according to the selected target variable, and restoring a feature output from the encoder to obtain a quantitative image of the target variable Contains instructions for executing the output decoder.

The ultrasound data may be multi-pattern RF (radio frequency) data in which ultrasound signals incident to the tissue in different beam patterns are reflected from the tissue and returned.

The plurality of variables is an Attenuation Coefficient (AC), a Speed of Sound (SoS), an Effective Scatterer Concentration (ESC), and two or more of an Effective Scatterer Diameter (ESD). may include

The encoder may include a conditional encoding block that normalizes to a value selected according to the target variable.

The decoder may be configured as a network that restores features of each variable while preserving structural similarity in a path for restoring features of different variables.

According to an embodiment, it is possible to image various quantitative variables in real time from ultrasound data of a tissue through a single neural network.

According to an embodiment, a high-quality reconstructed image may be generated through information shared in a process of reconstructing a quantitative image of variables having structural similarity.

According to the embodiment, since the ultrasound probe and imaging device for B-mode (Brightness mode) imaging can be used as it is, imaging is simple, and various organs that can be measured with the existing ultrasound imaging device can be measured.

According to an embodiment, a high-quality quantitative image may be reconstructed regardless of a user's skill level using a single ultrasound probe.

1 is a diagram conceptually illustrating a multivariate quantitative imaging apparatus using ultrasound data.
2 is a diagram exemplarily illustrating a neural network for extracting quantitative characteristics of tissue from multi-pattern ultrasound data.
3 is a diagram for explaining conditional normalization.
4 is an example of an encoder of a neural network that reconstructs multivariate quantitative images.
5 is an illustration of a decoder of a neural network that reconstructs multivariate quantitative images.
6 is a flowchart of a multivariate quantitative imaging method using ultrasound data according to an exemplary embodiment.
7 is an example of a phantom for multivariate quantitative simulation.
8 is a diagram illustrating a multivariate quantitative imaging result of a breast-mimicking phantom.
9 is a view showing the results of ex-vivo multivariate quantitative imaging.
10 is a configuration diagram of a computing device according to an embodiment.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings so that those of ordinary skill in the art to which the present invention pertains can easily implement them. However, the present invention may be embodied in several different forms and is not limited to the embodiments described herein. And in order to clearly explain the present invention in the drawings, parts irrelevant to the description are omitted, and similar reference numerals are attached to similar parts throughout the specification.

Throughout the specification, when a part "includes" a certain component, it means that other components may be further included, rather than excluding other components, unless otherwise stated. In addition, terms such as “…unit”, “…group”, and “module” described in the specification mean a unit that processes at least one function or operation, which may be implemented as hardware or software or a combination of hardware and software. have.

The neural network of the present invention is an artificial intelligence model for learning at least one task, and may be implemented as software/program executed in a computing device. The program is stored in a storage medium (non-transitory storage media), and includes instructions described for executing the operations of the present invention by a processor. The program may be downloaded over the network or sold as a product.

1 is a diagram conceptually explaining a multivariate quantitative imaging apparatus using ultrasound data, FIG. 2 is a diagram exemplarily illustrating a neural network for extracting quantitative characteristics of tissue from multi-pattern ultrasound data, and FIG. 3 is conditional normalization. It is a diagram explaining (conditional normalization).

Referring to FIG. 1 , a multivariate quantitative imaging device (referred to simply as an 'imaging device') 100 is a computing device operated by at least one processor, and is obtained from a tissue through a single ultrasound probe 10 . Multi-pattern ultrasound data is input. The imaging apparatus 100 is loaded with a computer program for the operations described in the present disclosure, and the computer program is executed by a processor.

The imaging apparatus 100 may complexly generate multi-variable quantitative images of a tissue by using the neural network 200 that extracts quantitative features of the tissue from the multi-pattern ultrasound data. Here, the imaging device 100 includes an attenuation coefficient (AC), a speed of sound (SoS), an effective scatterer concentration (ESC) indicating a density distribution in a tissue, which are quantitative variables of a tissue, and a tissue. It is possible to restore the size of the scatterer (Effective Scatterer Diameter, ESD) indicating the size of the inner cell. The neural network 200 may be called SQI-Net.

Meanwhile, attenuation coefficient, sound velocity, scatterer density, and scatterer size are parameters known as biomarkers for lesion extraction, and are closely related to the biomechanical properties of tissues. Therefore, the more variables are used, the more comprehensive analysis of the lesion is possible, thereby increasing diagnostic sensitivity and specificity. However, since each variable has a distinct effect on ultrasound propagation, it is not easy to simultaneously reconstruct several variables from ultrasound data.

To solve this problem, the neural network 200 performs conditional encoding for variably extracting quantitative features from ultrasound data according to a target variable to be restored. Through this, the neural network 200 may complexly generate multi-variable quantitative images of the tissue. Here, the conditional encoding conditionally changes the network parameter of the encoding path according to the selected variable so that the quantitative feature of the variable to be restored from the ultrasound data is optimally extracted, thereby improving image restoration performance for the corresponding variable.

While the neural network 200 reconstructs the quantitative image, the characteristics of the independent variables provide complementary information. Accordingly, even if the neural network 200 individually reconstructs the quantitative image of each variable, the independent variables provide complementary information, so that the quantitative value can be more accurately reconstructed.

The features of each variable extracted from the conditional encoder are reconstructed through the same decoding path. Thus, although each quantitative image is individually reconstructed according to a variable, the quantitative images may have structural similarity. Accordingly, since the neural network 200 reconstructs multivariate quantitative images while maintaining structural similarity, the imaging apparatus 100 may generate a high-quality reconstructed image.

Multi-pattern ultrasound data used to reconstruct multivariate quantitative images of tissue is obtained from the ultrasound probe 10 . The ultrasound probe 10 is a probe capable of emitting an ultrasound signal and obtaining ultrasound data reflected from a tissue. The ultrasound signal radiated to the tissue may be a plane wave. In the ultrasonic probe 10 , N (eg, 128) ultrasonic sensors are arranged, and there may be various types according to the arrangement shape. The sensors may be implemented with piezoelectrical elements. Also, the ultrasonic probe 10 may be a phased array probe that generates an ultrasonic signal by applying an electric signal to each piezoelectric element at regular time intervals. For reference, the ultrasound probe 10 may be a typical B-mode imaging probe. In this case, the imaging apparatus 100 receives RF data used for B-mode imaging, which is referred to as ultrasound data in the description.

The ultrasound probe 10 may input ultrasound signals of different beam patterns (Tx patterns #1 to #k) to the tissue, and may acquire radio frequency (RF) data reflected back from the tissue. In the description, RF data obtained by the ultrasound probe 10 for an ultrasound signal emitted in a specific beam pattern is referred to as ultrasound data of a specific beam pattern. By bundling ultrasound data obtained from a plurality of beam patterns, it may be simply referred to as multi-pattern ultrasound data or beamformed ultrasound data. The multi-pattern ultrasound data is, for example, ultrasound data obtained from 7 different beam patterns (θ _1: :θ ₇ ), and the different beam patterns have, for example, incident angles of -15°, -10°, Can be set to -5°, 0°, 5°, 10°, 15°.

Meanwhile, the ultrasound data obtained by the ultrasound probe 10 includes delay time information for receiving the reflected ultrasound signal for each sensor of the ultrasound probe 10 . Accordingly, the ultrasound data may be expressed as an image representing delay time information for each sensor. The ultrasound data expressed as an image may be referred to as an ultrasound data image for convenience, and the ultrasound data image may be used as an input image of the neural network 200 .

Referring to FIG. 2 , the neural network 200 is an artificial intelligence model capable of learning at least one task, and may be implemented as software/program executed in a computing device. The neural network 200 generates a high-resolution quantitative image I _q of the target variable by increasing the resolution of the encoder 210 that encodes the multi-pattern ultrasound data U according to the target variable q to be restored, and the encoded low-resolution feature map. It includes a decoder (decoder) (230).

The encoder 210 may receive multi-pattern ultrasound data U represented by the multi-pattern ultrasound data images 300 . The decoder 230 outputs a high-resolution quantitative image I _q corresponding to the objective variable, an attenuation coefficient image 400-1, a sound velocity image 400-2, a scatterer density image 400-3, and a scatterer. In addition to the size image 400 - 4 , images in which a plurality of variables are combined according to a target variable may be output.

The encoder 210 receives ultrasound data images indicating a delay time for each sensor of the ultrasound data obtained from the ultrasound probe 10 , and extracts quantitative features included in the ultrasound data images. In this case, the encoder 210 may be a conditional encoder that variably extracts quantitative features of input images according to a target variable to be restored.

The encoder 210 performs conditional encoding for optimally extracting quantitative features of a variable to be restored by changing network parameters of an encoding path according to the target variable q.

행렬에서 선택된 값인

및

으로, 정규화된 값

을 변환(스케일링 및 이동)할 수 있다.

는 목적 변수 q에 대한 스케일링(scaling) 값이고,

는 목적 변수 q에 대한 이동(shift) 값이다. 이를 통해, 인코더(210)는 목적 변수에 따라 입력 이미지들의 정량적 특징을 가변적으로 추출할 수 있다. 인코더(210)는, 예를 들면, 조건부 인스턴스 정규화(conditional instance normalization)을 수행해서, 목적 변수에 대응하는 정량적 특징을 추출할 수 있다. 목적 변수 q는 조직에서 추출하고자 하는 적어도 하나의 정량적 변수를 나타낸다.Referring to Figure 3, the encoder 210, according to the objective variable q,

the value selected from the matrix

and

, the normalized value

can be transformed (scaled and moved).

is the scaling value for the objective variable q,

is the shift value for the objective variable q. Through this, the encoder 210 may variably extract quantitative features of the input images according to the target variable. The encoder 210 may, for example, perform conditional instance normalization to extract a quantitative feature corresponding to the target variable. The objective variable q represents at least one quantitative variable to be extracted from the tissue.

Referring back to FIG. 2 , the decoder 230 restores features extracted by the encoder 210 to the same path. The decoder 230 learns to restore the features of each variable while preserving structural similarity in the process of reconstructing the features of each variable extracted from the encoder 210 through the same decoding path. Accordingly, the decoder 230 outputs an image having structural similarity with other variables even when the characteristic of a single objective variable is restored, so that it is possible to generate a quantitative image with much higher resolution than the conventional quantitative imaging method.

The objective function G ^* used for learning of the neural network 200 may be defined as a loss function as in Equation 1. The neural network 200 may learn to minimize restoration loss by using the objective function G ^* .

In Equation 1, the objective function G ^* is a function that minimizes the difference between the output G(U,q) inferred from the input U and the objective variable q, and the ground truth Y _q for the objective variable q. The objective function G ^* can undergo L ₂ regularization (λ = 10 ⁶ ) to avoid overfitting.

In order for the neural network 200 to learn about the variables at a similar level, the target variable q may be changed for each iteration.

4 is an example of an encoder of a neural network that reconstructs multivariate quantitative images, and FIG. 5 is an example of a decoder of a neural network that reconstructs multivariate quantitative images.

Referring to FIG. 4 , the encoder 210 of the neural network 200 receives multi-pattern ultrasound data images 300 . Each ultrasound data image U _i ∈R ^128X3018 means i-th RF data obtained from a beamformed ultrasound signal, where the image includes the number of probe sensors (N=128) and time axis indexes (eg For example, it can be expressed as 3018).

The encoder 210 may configure an encoding path with various network models/network blocks.

The encoder 210 convolutionally filters and integrates each of the multi-pattern ultrasound data images U, constructs an encoding network that compresses the feature value according to the objective variable q, and a feature map f extracted from the input U through the encoding network. Prints _q (U). The encoding network may be configured in various ways.

The encoder 210 performs convolutional filtering (2d-Convolution, 3×3 kernel size) and activation (eg, activation function is ReLU) block 211, down-sampling (eg, For example, 1x2 stride) blocks 212 can be configured independently.

행렬에서 선택된 값인

및

을 스케일링(scaling) 팩터 및 이동(shift) 팩터로 입력받고, 목적 변수에 따라 특징값을 변환할 수 있다. 조건부 인코딩 블록(213)은 예를 들면, 컨볼루션 필터링(2d-Convolution, 3×3 Kernel size) 블록(214), 조건부 인스턴스 정규화(conditional instance normalization) 블록(215), 그리고 활성화 블록(ReLU)(216)을 포함할 수 있다. After the feature values filtered in each channel are integrated for each channel, they are input to the conditional encoding block 213 . Conditional encoding block 213 according to the objective variable q,

the value selected from the matrix

and

may be input as a scaling factor and a shift factor, and a feature value may be transformed according to a target variable. The conditional encoding block 213 includes, for example, a 2d-Convolution (3×3 Kernel size) block 214, a conditional instance normalization block 215, and an activation block (ReLU) ( 216) may be included.

Subsequently, the encoder 210 extracts feature values in the temporal and spatial domains of the quantitative variables included in the multi-pattern ultrasound data by using the subsequent conditional encoding blocks 213 .

For example, the encoder 210 iteratively (eg, three times) concatenates a down-sampling block (eg, 1x2 stride) 217 with a subsequent conditional encoding block 213 , such that features in the time domain It is possible to construct a time domain encoding block that extracts values. Through an encoding path that repeatedly passes through the down-sampling block 217 and the conditional encoding block 213 , a correlation in the time domain of the quantitative variable included in the multi-pattern ultrasound data may be extracted. After passing through the time domain encoding block, 7 ultrasound data R ^128X3018 may be encoded into 128 feature maps R ^128X186X128 of (128X186) size.

The encoder 210 repeatedly (eg, twice) concatenates the down-sampling block (eg, 2x2 stride) 218 and the subsequent conditional encoding block 213 to extract features in the spatial domain. , it is possible to configure a spatial domain encoding block for extracting feature values in the spatial domain. While passing through the spatial domain encoding block, the extracted feature map of R ^32X47X128 may be encoded into the feature map f _q (U) of R ^16X16X512 while passing through the down-sampling block (eg, 2x3 stride) 219 .

The encoder 210 outputs a feature map f _q (U) of the multi-pattern ultrasound data U extracted according to the objective variable q. The output f _q (U) is input to the decoder 230 .

Referring to FIG. 5 , the decoder 230 of the neural network 200 receives the output f _q (U) of the encoder 210 and generates a high-resolution quantitative image I _q corresponding to the objective variable q. For example, the decoder 230 may generate a high-resolution quantitative image I _q ∈R ^128X128 from f _q (U)∈R ^16X16X512 .

The decoder 230 may configure a decoding network with various network models/network blocks. For example, the decoder 230 may generate a high-resolution quantitative image I _q by an upsampling method. Alternatively, the decoder 230 may generate a high-resolution quantitative image I _q using parallel multi-resolution subnetworks based on a recently known High-Resolution Network (HRNet). .

The decoder 230 composed of parallel multi-resolution subnetworks sequentially performs multi-resolution convolution from the low-resolution subnetwork to increase the image resolution, and finally generates a high-resolution quantitative image I _q . can do. The output layer of the decoder 230 may be merged into the highest resolution representation, and a high-resolution quantitative image I _q synthesized through 1×1 convolution may be generated. Meanwhile, the unit subnetwork may be implemented as, for example, four residual blocks, thereby ensuring learning stability.

Multi-resolution convolution can preserve the low-resolution feature map f _q (U) while minimizing information loss while integrating various quantitative profiles to generate high-resolution quantitative images. Multi-resolution fusion is implemented at all nodes of subnetworks to exchange information through multi-resolution representations, which play an important role in extracting high-resolution quantitative images I _q .

6 is a flowchart of a multivariate quantitative imaging method using ultrasound data according to an exemplary embodiment.

Referring to FIG. 6 , the imaging apparatus 100 receives multi-pattern ultrasound data obtained from tissue ( S110 ). The multi-pattern ultrasound data is RF data obtained with respect to ultrasound signals emitted in different beam patterns, and may be referred to as beamformed ultrasound data.

The imaging apparatus 100 changes a network parameter according to a target variable to be restored among a plurality of variables, extracts a feature corresponding to the selected target variable from the multi-pattern ultrasound data, and restores the feature corresponding to the target variable to achieve high resolution outputs a quantitative image of (S120). To this end, the imaging apparatus 100 may use the learned neural network 200 to receive multi-pattern ultrasound data and output a quantitative image corresponding to the selected variable. The neural network 200 changes a network parameter according to a target variable to be restored among a plurality of variables, and extracts a feature corresponding to the selected target variable from the multi-pattern ultrasound data. A conditional encoder 210, and a conditional encoder 210 It may include a decoder 230 that restores the features of the target variable extracted from .

A target variable to be restored may be selected from among a number of variables including attenuation coefficient, sound velocity, scatterer density, scatterer size, and the like. The target variable may be a single variable, or may be a plurality of variables. For example, when the target variable is selected as the attenuation coefficient, the imaging apparatus 100 may extract a quantitative feature of the attenuation coefficient from the input multi-pattern ultrasound data and output a quantitative image in which the attenuation coefficient is expressed. When the target variable is selected as the attenuation coefficient and the speed of sound, the imaging apparatus 100 may extract quantitative features of the attenuation coefficient and the speed of sound from the input multi-pattern ultrasound data, and output a quantitative image in which the attenuation coefficient and the speed of sound are expressed. have.

7 is an example of a phantom for multivariate quantitative simulation.

Referring to FIG. 7 , the learning data of the neural network 200 may be composed of ultrasound data obtained from various human environments, and may be collected using an ultrasound simulation tool. It can be confirmed that it is possible to extract quantitative information about tissue through ultrasound data measured through sophisticated simulation.

(a) is the simulated sound velocity distribution, (b) is the simulated damping coefficient distribution, and (c) the simulated density distribution.

Simulation phantoms representing organs and lesions are modeled by sound velocity distribution (a), damping coefficient distribution (b), and density distribution (c) so as to imitate the human body and not lose its simplicity and generality. can be

For example, in the simulation, a region of interest (RoI) is set to 45 mm x 45 mm, and ellipses with a radius of 2 to 30 mm may be randomly arranged. The background and the ellipse may have a speed of sound in the range of 1400 to 1700 m/s, an attenuation coefficient in the range of 0 to 1.5 dB/cm/MHz, and a density value in the range of 0.9 to 1.1 kg/m ³ . A square speckle having a size of 25 to 150 μm may be arranged at a concentration of 0 to 10/wavelength ² to express the scatterer density and the size of the scatterer.

8 is a diagram showing the results of multivariate quantitative imaging of a breast-mimicking phantom, and FIG. 9 is a diagram illustrating the results of ex-vivo multivariate quantitative imaging.

Referring to FIG. 8 , in a breast-mimicking phantom including a cyst mass and cancer tissue, the B-mode imaging technique, which is a conventional image restoration technique, does not distinguish between cyst/cancer tissue (lower left).

On the other hand, the imaging apparatus 100 may restore the attenuation coefficient (AC), the speed of sound (SoS), the scattering body density (ESC), and the scattering body size (ESD). It can distinguish and restore quantitative variables with a high accuracy of 95% or more. For example, the imaging device 100 may have an attenuation coefficient (AC) greater than normal, a speed of sound (SoS) faster than normal, a scatterer density (ESC) higher than normal, or a scatterer size (ESD). is smaller than normal, it can be identified as cancerous tissue. The imaging apparatus 100 may determine the cancer tissue by synthesizing the results of multiple variables in order to increase accuracy.

Referring to FIG. 9 , benign (a), malignant (b), and cancer-like substances (c) are inserted into bovine skeletal muscle, and quantitative variables are restored through the imaging device 100 to differentiate cancer. was verified for clinical significance.

Looking at the B-mode image (second column) obtained by the B-mode imaging technique, which is a conventional image restoration technique, it can be seen that it is very difficult to make a differential diagnosis of the implanted tissue. On the other hand, the imaging apparatus 100 may restore the attenuation coefficient (AC), the speed of sound (SoS), the scattering body density (ESC), and the scattering body size (ESD), which are closely related to the biomechanical properties of the tissue. Accordingly, by analyzing the speed of sound (SoS) and the tissue density (ESC) restored through the imaging apparatus 100 , it is possible to easily distinguish a cyst from a cancerous tissue. In addition, by analyzing the attenuation coefficient (AC) and the tissue cell size (ESD) restored through the imaging apparatus 100 , it is possible to determine whether the tissue is positive or negative.

10 is a configuration diagram of a computing device according to an embodiment.

Referring to FIG. 10 , the imaging device 100 may be a computing device 500 operated by at least one processor, and is connected to the ultrasound probe 10 or a device providing data obtained from the ultrasound probe 10 . do.

The computing device 500 includes one or more processors 510 , a memory 530 for loading a program executed by the processor 510 , a storage 550 for storing programs and various data, a communication interface 570 , and these It may include a bus 590 for connecting. In addition, the computing device 500 may further include various components. The program may include instructions that, when loaded into the memory 530 , cause the processor 510 to perform a method/operation according to various embodiments of the present disclosure. That is, the processor 510 may perform methods/operations according to various embodiments of the present disclosure by executing instructions. An instruction refers to a set of computer readable instructions grouped on the basis of a function, which is a component of a computer program and is executed by a processor.

The processor 510 controls the overall operation of each component of the computing device 500 . The processor 510 includes at least one of a central processing unit (CPU), a micro processor unit (MPU), a micro controller unit (MCU), a graphic processing unit (GPU), or any type of processor well known in the art of the present disclosure. may be included. In addition, the processor 510 may perform an operation on at least one application or program for executing the method/operation according to various embodiments of the present disclosure.

The memory 530 stores various data, commands and/or information. The memory 530 may load one or more programs from the storage 550 to execute methods/operations according to various embodiments of the present disclosure. The memory 530 may be implemented as a volatile memory such as RAM, but the technical scope of the present disclosure is not limited thereto.

The storage 550 may non-temporarily store a program. The storage 550 is a non-volatile memory, such as a read only memory (ROM), an erasable programmable ROM (EPROM), an electrically erasable programmable ROM (EEPROM), a flash memory, a hard disk, a removable disk, or well in the art to which the present disclosure pertains. It may be configured to include any known computer-readable recording medium.

The communication interface 570 supports wired/wireless communication of the computing device 500 . To this end, the communication interface 570 may be configured to include a communication module well-known in the technical field of the present disclosure.

The bus 590 provides a communication function between components of the computing device 500 . The bus 590 may be implemented as various types of buses, such as an address bus, a data bus, and a control bus.

As such, according to the embodiment, it is possible to image various quantitative variables in real time from ultrasound data of a tissue through a single neural network.

According to an embodiment, a high-quality reconstructed image may be generated through information shared in a process of reconstructing a quantitative image of variables having structural similarity.

According to the embodiment, since the ultrasound probe and imaging device for B-mode (Brightness mode) imaging can be used as it is, imaging is simple, and various organs that can be measured with the existing ultrasound imaging device can be measured.

According to an embodiment, a high-quality quantitative image may be reconstructed regardless of a user's skill level using a single ultrasound probe.

The embodiment of the present invention described above is not implemented only through the apparatus and method, but may be implemented through a program for realizing a function corresponding to the configuration of the embodiment of the present invention or a recording medium in which the program is recorded.

Although the embodiments of the present invention have been described in detail above, the scope of the present invention is not limited thereto. is within the scope of the right.

Claims (18)

A method of operating an imaging device operated by at least one processor, the method comprising:
receiving ultrasound data obtained from the tissue; and
extracting a feature corresponding to a target variable selected from among a plurality of variables from the ultrasound data, restoring the feature, and outputting a quantitative image of the target variable;
A method of operation comprising a. In claim 1,
The ultrasound data is multi-pattern RF (Radio Frequency) data in which ultrasound signals incident on the tissue in different beam patterns are reflected back from the tissue. In claim 1,
Many of the variables are
an Attenuation Coefficient (AC), a Speed of Sound (SoS), an Effective Scatterer Concentration (ESC), and an Effective Scatterer Diameter (ESD). . In claim 1,
The step of outputting a quantitative image of the target variable is
An operating method of extracting a feature corresponding to the target variable from the ultrasound data by changing a network parameter of a neural network according to the target variable selected from among the plurality of variables. In claim 4,
and the neural network performs conditional encoding for normalizing to a value selected according to the target variable. In claim 4,
The step of outputting a quantitative image of the target variable is
An operation method of outputting a quantitative image of the target variable through a decoder of the neural network that restores features output from the encoder of the neural network to the same path. In claim 6,
wherein the decoder is configured with a network that restores the characteristics of each variable while preserving structural similarity in the path of restoring the characteristics of the different variables. memory for loading a computer program; and
A processor that executes the computer program, extracts a feature corresponding to a target variable selected from among a plurality of variables from ultrasound data obtained from a tissue, restores the feature, and outputs a quantitative image of the target variable
comprising, an imaging device.
In claim 8,
The ultrasound data is multi-pattern RF (Radio Frequency) data in which ultrasound signals incident to the tissue in different beam patterns are reflected from the tissue and returned. In claim 8,
Many of the variables are
An imaging device comprising at least two of an Attenuation Coefficient (AC), a Speed of Sound (SoS), an Effective Scatterer Concentration (ESC), and an Effective Scatterer Diameter (ESD). . In claim 8,
the processor
Outputting a quantitative image corresponding to a target variable selected from among a plurality of variables from the ultrasound data through a neural network composed of an encoder and a decoder,
The encoder is configured as a network for extracting a feature corresponding to the target variable from the ultrasound data by changing a network parameter according to the target variable selected from among the plurality of variables,
The decoder is configured as a network that restores the features output from the encoder to the same path. In claim 11,
and the neural network performs conditional encoding of normalizing to a value selected according to the objective variable. In claim 11,
wherein the decoder is configured with a network that restores features of each variable while preserving structural similarity in a path for restoring features of different variables. A computer program stored in a computer-readable storage medium and comprising instructions executed by a processor, the computer program comprising:
When ultrasound data obtained from tissue is input and a target variable is selected from among a plurality of variables that can be restored from the ultrasound data, a network parameter is changed according to the selected target variable, and the ultrasound data corresponds to the target variable. an encoder to extract, and
A computer program comprising instructions for executing a decoder that outputs a quantitative image of the target variable by restoring the feature output from the encoder. 15. In claim 14,
The ultrasound data is multi-pattern RF (Radio Frequency) data in which ultrasound signals incident on the tissue in different beam patterns are reflected from the tissue and returned. 15. In claim 14,
Many of the variables are
A computer program comprising two or more of Attenuation Coefficient (AC), Speed of Sound (SoS), Effective Scatterer Concentration (ESC), and Effective Scatterer Diameter (ESD). . 15. In claim 14,
wherein the encoder comprises a conditional encoding block for normalizing to a selected value according to the objective variable. 15. In claim 14,
wherein the decoder consists of a network that restores the characteristics of each variable while preserving structural similarity in the path of restoring the characteristics of the different variables.

Images