AI-assisted focused ultrasound focal region temperature measurement method and system

Abstract

The invention discloses an AI-assisted focused ultrasound focal region temperature measurement method and system, and belongs to the technical field of ultrasonic lossless temperature measurement, and the method comprises the steps: collecting medical ultrasonic RF data in a high intensity focused ultrasound (HIFU) treatment process, generating a B-mode ultrasonic image, and combining thermocouple temperature data to construct a training set and a test set. The method comprises the steps that training is carried out on the basis of a preset space-frequency domain intelligent network (SFInsight Net, so that an AI regression model is obtained; the SFInsight Net comprises a spatial feature analysis network and a frequency domain feature analysis network, the spatial feature analysis network extracts transverse and longitudinal spatial information of a temperature measurement area, the frequency domain feature analysis network learns frequency domain features from the global perspective, the two modules supplement each other, and the temperature measurement precision is improved. And finally, optimizing the test set, and improving an AI regression model so as to realize accurate in-vivo fat single-point ultrasonic temperature measurement. The method not only has innovativeness, but also improves the real-time performance and the accuracy of a clinical ultrasonic temperature measurement technology, and is beneficial exploration in the technical field of medical ultrasonic.

Description

Technical Field

[0001] The present invention belongs to the technical field of ultrasonic non-destructive temperature measurement, and specifically relates to an AI-assisted focused ultrasound focal-area temperature measurement method and system. Background Art

[0002] High-intensity focused ultrasound (HIFU) is a new non-invasive cancer treatment technology. Compared with traditional treatment technologies, it is highly favored as a non-invasive treatment method. During HIFU treatment, it is very important to monitor the focal temperature field, on the one hand to ensure the ablation effect of necrotic cells, and on the other hand to avoid collateral damage to healthy tissues during treatment as much as possible. However, during the current HIFU treatment, achieving non-invasive, accurate and real-time temperature detection is still a major challenge in the field of medical engineering.

[0003] Since the last century, a variety of temperature monitoring methods for HIFU treatment have emerged, but they all have certain limitations: the most common temperature measurement technologies are thermocouples and magnetic resonance imaging (MRI), but thermocouples are contrary to the "non-invasive" concept of HIFU and can easily cause cancer cell metastasis; MRI equipment is bulky and expensive, difficult to popularize, and the generated temperature field image frame rate is low (0.1-1Hz), which cannot provide real-time guidance for HIFU treatment. In summary, whether it is thermocouple temperature measurement or MRI temperature measurement, it is difficult to achieve real-time and accurate monitoring of the focal temperature field of tissues during HIFU treatment.

[0004] Ultrasonic non-destructive temperature measurement technology is also a classic non-destructive temperature measurement technology. This technology has low instrument cost, wide audience, and is easy to popularize. The probe that collects ultrasound data can be directly integrated with the HIFU treatment probe to achieve integration. It is currently the most promising temperature monitoring technology for clinical HIFU treatment. However, ultrasonic non-destructive temperature measurement technology still has problems such as insufficient accuracy and low frame rate of temperature field images.

[0005] With the development of artificial intelligence (AI) technology and its combination with ultrasonic non-destructive temperature measurement technology, many existing technologies based on artificial intelligence to achieve ultrasonic non-destructive temperature measurement have made significant progress in improving accuracy, eliminating motion artifacts, and increasing temperature measurement speed. However, there are still many areas in which ultrasonic non-destructive temperature measurement technology needs to be improved. For example, in terms of temperature measurement objects, ultrasonic temperature measurement research is mainly focused on muscle tissue, and there are relatively few studies on adipose tissue temperature measurement technology. Fat cancer, namely liposarcoma, is also a common cancer. Liposarcoma accounts for 20% of tissue tumors, which has brought great physical and mental pain to countless patients. Considering the differences in the structure and acoustic properties of muscle and fat tissues, the two temperature measurement technologies are not completely interoperable, so further research is needed on adipose tissue temperature measurement technology. Summary of the invention

[0006] In order to solve the technical problems existing in the background technology, the present invention aims to provide an AI-assisted focused ultrasound focal area temperature measurement method and system, so as to provide more technical support to medical staff during HIFU treatment and benefit the majority of patients.

[0007] In order to solve the technical problem, the technical solution of the present invention is:

[0008] An AI-assisted focused ultrasound focal-area temperature measurement method, the method comprising:

[0009] Collect and process the medical ultrasound RF data of the focal area during HIFU treatment to obtain a B-mode ultrasound image;

[0010] Using the B-mode ultrasound images and the corresponding thermocouple recorded temperature data obtained in the experiment to construct a training set and a test set, and training an AI regression model based on a space-frequency domain intelligent network model;

[0011] The AI ​​regression model is optimized based on the test set, and the spatial feature analysis network and the frequency domain feature analysis network are functionally coupled to fully utilize the information of the tissues surrounding the measured point and the grayscale change trend of the measured point to achieve single-point ultrasonic temperature measurement of body fat; wherein, the spatial feature analysis module extracts the horizontal and vertical spatial information of the temperature measurement area, and the frequency domain feature analysis module learns the frequency domain features from a global perspective.

[0012] Furthermore, the spatial feature analysis network in the space-frequency domain intelligent network model includes three stages, each of which includes multiple trans blocks, a convolution block and an average pooling layer;

[0013] The trans block in the spatial feature analysis network model includes four channels: an H auxiliary channel for extracting inter-layer spatial correlation; a V auxiliary channel for extracting intra-layer spatial correlation; an H2V main channel for fusing the features of the two auxiliary channels; and a V2H main channel for fusing the features of the two auxiliary channels.

[0014] Furthermore, the input of the spatial feature analysis network model is an ultrasound B-mode image; the ultrasound B-mode image is convolved to achieve downsampling; the downsampled ultrasound B-mode image is embedded in three ways: (1) the image is divided into multiple 16×16 blocks to form single-piece visual words; (2) the image is divided into multiple layers along the depth H direction to form inter-layer visual words; (3) the image is divided into multiple layers along the horizontal V direction to form intra-layer visual words; the single-piece visual words, inter-layer visual words and intra-layer visual words are flattened to obtain z0, z h and z v .

[0015] Furthermore, the four channel inputs of the lth trans block in the sth stage of the spatial feature analysis network model are and The output of the previous layer is the input of the next layer. The formula is as follows:

[0016]

[0017] The final output of the s-1th stage of the spatial feature analysis network model is the input of the first trans block of the sth stage, and the formula is as follows:

[0018]

[0019] The initial inputs of the two auxiliary channels of the trans block are z h 、z v , the initial inputs of both main channels are z0;

[0020] The input of the auxiliary channel of the trans block is output through the transformer neural network, and the formula is as follows:

[0021]

[0022] The main channel input of the trans block passes through the transformer neural network, is added to the output of the auxiliary channel, and is input into the transformer neural network again to obtain the output. The formula is as follows:

[0023]

[0024] Among them, the final output of the trans block in each stage is and Input convolution block: First and The one-dimensional vector is restored to a two-dimensional tensor and connected end to end, and then three convolutions are performed. After each convolution, batch normalization is performed to accelerate the convergence of the model. The three convolutions all use RELU as the activation function. The convolution block outputs z s+1 Finally, the average pooling layer is input to get the output of each stage.

[0025] Further, the frequency domain feature analysis neural network includes: a frequency domain channel learning block and a frequency domain time learning block;

[0026] The frequency domain channel learning block includes three parts: the first part is a domain converter based on Fourier transform, which is used to transform the time domain signal into the frequency domain signal; the second part is a frequency domain multilayer perceptron, which is used to share learnable weights between different channels and capture the interaction and correlation between different variables; the third part is a domain inverter based on inverse Fourier transform, which is used to transform the frequency domain signal into the time domain signal;

[0027] The frequency-domain time learning block includes three parts: the first part is a domain converter based on Fourier transform, which is used to convert the time domain signal into a frequency domain signal; the second part is a frequency domain multilayer perceptron, which is used to study the frequency domain features in the same channel; the third part is a domain inverter based on inverse Fourier transform, which is used to convert the frequency domain signal into a time domain signal.

[0028] Furthermore, the training of the AI ​​regression model based on the spatial feature analysis network model and the frequency domain feature analysis network model includes:

[0029] The frequency domain feature analysis network converts the time domain information into the frequency domain to learn global features from a global perspective, and combines it with the spatial features learned by the spatial feature analysis network to perform temperature regression calculation;

[0030] The input of the frequency domain feature analysis network is the B-mode ultrasound image. After selecting 10 consecutive frames of images from time t, a 16×16 small image block {p t-9 ,p t-8, ...,p t}, after straightening the image block, we get the tensor {x t-9 ,x t-8, ...,x t |x i ∈R N}, X t =[x t-9 ,x t-8, ...,x t ]∈R N×L , where L = 10, input frequency domain characteristic analysis network to estimate the temperature Y corresponding to time t t ;

[0031] For the tensor X t Expand the dimension and multiply it by the learnable vector Used to obtain more hidden information, the formula is as follows:

[0032]

[0033] The tensor H t ∈R N×L×dAs the input of the frequency domain channel learning block; it contains Fourier transform to convert time domain information into frequency domain information, and inverse Fourier transform to convert frequency domain information into time domain information. The formulas are as follows:

[0034]

[0035] Where f is the frequency, v is the integral variable, j is an imaginary number, H(v)cos(2πfv) is the real part, and H(v)sin(2πfv) is the imaginary part;

[0036]

[0037] Where, f is the integral variable Re(h(f)) is the real part, and Im(h(f)) is the imaginary part;

[0038] When considering time series prediction, channel dependency can provide interactions and correlations between different variables, thereby improving prediction accuracy. Therefore, the frequency domain channel learning block shares weights between L timestamps to understand the association of channels: For input H t ∈R N×L×d , consider the lth timestamp corresponding to This is fed into the frequency domain channel learning block:

[0039]

[0040] Among them, fourier() and Fourier() represent Fourier transform and inverse Fourier transform respectively. express The frequency domain components of , and the weights and biases satisfy:

[0041]

[0042] The output Z∈R of the frequency domain channel learning block N×L×d As the input of the frequency-domain time learning block, consider the nth channel corresponding to Feed this into the frequency-domain time learning block:

[0043]

[0044] S t ∈R N×L×d After being straightened into a one-dimensional vector, it is connected with the output features of the spatial feature analysis network and then input into the fully connected layer for regression calculation, which is used for AI regression calculation of the temperature at time t.

[0045] Furthermore, the features output by the spatial feature analysis network model and the frequency domain feature analysis network model are concat-connected and input into the fully connected layer for regression calculation to infer the temperature of the measured point; the spatial feature analysis network model and the frequency domain feature analysis network model calculate the loss function based on the mean square error, and the formula of the loss function is as follows:

[0046]

[0047] Where n represents the batch size of the model; Indicates the temperature recorded by the corresponding thermocouple; i =f(x i ), represents the predicted temperature result, x i For input data.

[0048] An AI-assisted focused ultrasound focal area temperature measurement system, characterized in that the system is used to perform any of the above methods, and the system includes:

[0049] A data acquisition module, used to collect and process the medical ultrasound RF data of the focal area during HIFU treatment to obtain a B-mode ultrasound image;

[0050] A model training module, used to construct a training set and a test set using the B-mode ultrasound image and the corresponding thermocouple recorded temperature data obtained in the experiment, and to train an AI regression model based on a space-frequency domain intelligent network model;

[0051] The model optimization module is used to optimize the AI ​​regression model based on the test set, couple the spatial feature analysis network with the frequency domain feature analysis network function, and make full use of the information of the tissue around the measured point and the grayscale change trend of the measured point to realize single-point ultrasonic temperature measurement of body fat; wherein, the spatial feature analysis module extracts the horizontal and vertical spatial information of the temperature measurement area, and the frequency domain feature analysis module learns the frequency domain features from a global perspective.

[0052] A computer device comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, an AI-assisted focused ultrasound focal area temperature measurement method described above is implemented.

[0053] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the above-mentioned AI-assisted focused ultrasound focal-area temperature measurement methods.

[0054] Compared with the prior art, the advantages of the present invention are:

[0055] The present invention collects RF data in the body, processes the data to obtain B-mode ultrasound images, constructs training sets and test sets based on the ultrasound data and the temperature recorded by the hotspot couples in the experiment, and obtains the optimal AI regression model based on deep learning model training to achieve single-point ultrasound temperature measurement of body fat. It can effectively save human resources, shorten temperature measurement time, and improve temperature measurement accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 It is a structural schematic diagram of the AI-assisted focused ultrasound focal-area temperature measurement system disclosed in the present invention;

[0057] Figure 2 It is a structural schematic diagram of a space-frequency domain intelligent network model 200 of the AI-assisted focused ultrasound focal-area temperature measurement method disclosed in the present invention;

[0058] Figure 3 It is a structural schematic diagram of the spatial feature analysis network and frequency domain feature analysis network model 300 disclosed in the present invention. DETAILED DESCRIPTION

[0059] The specific implementation mode of the present invention is described below in conjunction with embodiments:

[0060] It should be noted that the structures, proportions, sizes, etc. shown in this specification are only used to match the contents disclosed in the specification so that people familiar with this technology can understand and read them, and are not used to limit the conditions under which the present invention can be implemented. Any structural modification, change in proportional relationship or adjustment of size should still fall within the scope of the technical content disclosed in the present invention without affecting the effects and purposes that can be achieved by the present invention.

[0061] At the same time, the terms such as "upper", "lower", "left", "right", "middle" and "one" cited in this specification are only for the convenience of description and are not used to limit the scope of implementation of the present invention. Changes or adjustments to their relative relationships should be regarded as the scope of implementation of the present invention without substantially changing the technical content.

[0062] Embodiment 1:

[0063] In the existing ultrasonic non-destructive temperature measurement, there is a lack of temperature measurement technology for body fat during HIFU treatment. To this end, this embodiment proposes an AI-assisted focused ultrasound focal area temperature measurement system and method.

[0064] Figure 1 1 is a schematic diagram of the structure of an AI-assisted focused ultrasound focal temperature measurement system 100 according to the first embodiment of the present disclosure. The AI-assisted focused ultrasound focal temperature measurement system for body fat in HIFU treatment includes the following functional modules: a data acquisition module 101, a model training module 102, and a model optimization module 103.

[0065] The data acquisition module 101 is used to acquire ultrasonic RF data of the HIFU treatment area and process the data to obtain a B-mode ultrasonic image. In addition, the data acquisition module 101 also acquires temperature data of the focal area using a thermocouple.

[0066] The original RF data must first be filtered to remove noise and artifacts to improve signal quality. Then, envelope detection (such as Hilbert transform) is used to extract the envelope signal in the RF signal, that is, the echo intensity of the tissue, and then the extracted envelope data is converted into grayscale values ​​for imaging processing to obtain a B-mode ultrasound image.

[0067] The model training module 102 is used to construct a training set based on a deep learning model using temperature data recorded from B-mode ultrasound images and corresponding thermocouples to train an AI regression model.

[0068] In this embodiment, the B-mode ultrasound image obtained by the data acquisition module 101 and the temperature data recorded by the corresponding thermocouple are divided into a training set and a test set in a ratio of 4:1. The test set is used to train the AI ​​regression model. Based on the calculation results of the mean square error loss function, the back propagation algorithm is used to update the parameters of the AI ​​regression model, thereby realizing the training process of the AI ​​regression model.

[0069] The model optimization module 103 is used to optimize the AI ​​regression model from a test set constructed from the B-mode ultrasound image and the corresponding thermocouple recording data, so that the model training module can train the best AI regression model.

[0070] In this example, the performance of the AI ​​regression model is tested on a test set constructed from B-mode ultrasound images and temperature data recorded by corresponding thermocouples. The difference between the results inferred by the AI ​​regression model and the results recorded by the thermocouples is mainly evaluated based on the mean square error loss function. The closer the loss function difference is to 0, the stronger the performance of the AI ​​regression training model. The test results can be used to guide model training and debug the hyperparameters of the neural network model to achieve the best performance, so that the AI ​​regression model can infer a temperature as close as possible to the temperature recorded by the thermocouple at time t after inputting the B-mode ultrasound image at time t and its previous 9 frames.

[0071] Figure 2 It is a structural schematic diagram of the space-frequency domain intelligent network model 200 of the AI-assisted focused ultrasound focal-area temperature measurement method according to an embodiment of the present disclosure.

[0072] In this example, the spatial-frequency domain intelligent neural network model 200 includes two main parts: one is the spatial feature analysis network 301, which is used to extract spatial information from B-mode ultrasound images; the other is the frequency domain feature analysis network 302, which converts time domain signals into frequency domain signals, so that the neural network can learn features from a global perspective. Finally, the outputs of the two parallel neural networks will be combined and input into the fully connected layer to realize the AI ​​regression model's inference function for temperature.

[0073] Figure 3 It is a structural schematic diagram of the spatial feature analysis network model 301 and the frequency domain feature analysis network 302 of the AI-assisted focused ultrasound focal-area temperature measurement method according to an embodiment of the present disclosure.

[0074] In this embodiment, the spatial feature analysis network and the frequency domain feature analysis network are combined, wherein the spatial feature analysis network starts from the spatial information, taking into account that the B-mode ultrasound image centered on fat tissue may contain information about the superficial epidermis and deeper muscles, and the acoustic parameters such as the acoustic attenuation coefficient and the speed of sound change with temperature between different tissues are different, and it is possible to try to calculate the temperature based on the information of the association between layers. Therefore, the spatial feature analysis network not only includes traditional single-piece visual words, but also adds intra-layer visual words and inter-layer visual words divided horizontally and vertically. Intra-layer visual words can be used to extract spatial information at the same depth, and inter-layer visual words are used to extract the association information between different tissues corresponding to different depths. The single-piece visual words, inter-layer visual words, and intra-layer visual words are straightened (flattened) to obtain z0, z h and z v , where the initial inputs of the two auxiliary channels of the trans block are z h 、z v , the initial inputs of the two main channels are both z0. Considering the four channel inputs of the lth trans block in the sth stage of the spatial feature analysis network model, they are and The output of the previous layer is the input of the next layer. The formula is as follows:

[0075]

[0076] Where f() represents the role of the trans block. In this spatial feature analysis network model, the final output of the s-1th stage is the input of the first trans block of the sth stage, and the formula is as follows:

[0077]

[0078] In order to make up for the deficiency of the spatial feature analysis network that does not consider the change trend of the measured point, the frequency domain feature analysis network is introduced to convert the time domain information into the frequency domain to learn the global features from a global perspective, and combine it with the spatial features learned by the spatial feature analysis network to perform temperature regression calculation. The input of the frequency domain feature analysis network is 10 frames of processed continuous B-mode ultrasound images: a 16×16 B-mode ultrasound image centered on the measured temperature point is cropped to obtain {p t-9 ,p t-8, ...,p t}, and then straighten it to get 10 one-dimensional vectors {x t-9 ,x t-8, ...,x t |x i ∈R N}, X t =[x t-9 ,x t-8, ...,x t ]∈R N×L (here L = 10) as the input of the frequency domain feature analysis network. t Expand the dimension and multiply it by the learnable vector Used to obtain more hidden information, the formula is as follows:

[0079]

[0080] Get the tensor H t ∈R N×L×d As the input of the frequency domain channel learning block, it includes Fourier transform to convert time domain information into frequency domain information, and inverse Fourier transform to convert frequency domain information into time domain information. The formulas are as follows:

[0081]

[0082] Where f is the frequency, v is the integration variable, j is an imaginary number, H(v)cos(2πfv) is the real part, and H(v)sin(2πfv) is the imaginary part.

[0083]

[0084] Where f is the integral variable, Re(h(f)) is the real part, and Im(h(f)) is the imaginary part.

[0085] When considering time series prediction, channel dependency can provide interactions and correlations between different variables, thereby improving prediction accuracy. Therefore, the frequency domain channel learning block shares weights between L timestamps to understand the association of channels: For input H t ∈R N×L×d , consider the lth timestamp corresponding to This is fed into the frequency domain channel learning block:

[0086]

[0087] Where fourier() and Fourier() represent Fourier transform and inverse Fourier transform respectively. express The frequency domain components of , and the weights and biases satisfy:

[0088]

[0089] The output Z∈R of the frequency domain channel learning block N×L×d As the input of the frequency-domain time learning block, consider the nth channel corresponding to Feed this into the frequency-domain time learning block:

[0090]

[0091] S t ∈R N×L×d After being straightened into a one-dimensional vector, it is connected with the output features of the spatial feature analysis network and then input into the fully connected layer for regression calculation, which is used for AI regression calculation of the temperature at time t.

[0092] According to a preferred embodiment, the neural network model combining spatial feature analysis and frequency domain feature analysis calculates the loss function based on the mean square error. The formula of the loss function is as follows:

[0093]

[0094] Where n represents the batch size of the model; Indicates the temperature recorded by the corresponding thermocouple; i =f(x i ), represents the predicted temperature result, x i For input data.

[0095] The AI-assisted focused ultrasound focal area temperature measurement system and method disclosed in the present invention have good temperature monitoring capabilities, can provide technical reference for the field of ultrasonic non-destructive temperature measurement technology, and contribute to the wider application of HIFU treatment technology.

[0096] Embodiment 2:

[0097] This embodiment provides an AI-assisted focused ultrasound focal area temperature measurement system, which is used to execute the above-mentioned AI-assisted focused ultrasound focal area temperature measurement method. Specifically, the system includes:

[0098] A data acquisition module, used to collect and process the medical ultrasound RF data of the focal area during HIFU treatment to obtain a B-mode ultrasound image;

[0099] A model training module is used to construct a training set and a test set using the B-mode ultrasound image and the corresponding thermocouple recorded temperature data obtained in the experiment, and to obtain an AI regression model based on a preset spatial feature analysis network model and a frequency domain feature analysis network model;

[0100] The model optimization module is used to optimize the AI ​​regression model based on the test set to obtain the optimal AI regression model, that is, to achieve single-point ultrasonic temperature measurement of body fat.

[0101] Embodiment 3:

[0102] This embodiment provides a terminal device, which includes a processor and a memory, wherein the memory is used to store a computer program, the computer program includes program instructions, and the processor is used to execute the program instructions stored in the computer storage medium. The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. It is the computing core and control core of the terminal, which is suitable for implementing one or more instructions, and is specifically suitable for loading and executing one or more instructions to implement the corresponding method flow or corresponding function; the processor described in the embodiment of the present invention can be used for the operation of an AI-assisted focused ultrasound focal area temperature measurement method, including the following steps:

[0103] Collect and process the medical ultrasound RF data of the focal area during HIFU treatment to obtain a B-mode ultrasound image;

[0104] Using the B-mode ultrasound images and the corresponding thermocouple recorded temperature data obtained in the experiment to construct a training set and a test set, and training an AI regression model based on a spatial feature analysis network model and a frequency domain feature analysis network model;

[0105] The AI ​​regression model is optimized based on the test set, thereby achieving single-point ultrasonic temperature measurement of body fat.

[0106] Embodiment 4:

[0107] This embodiment provides a storage medium, specifically a computer-readable storage medium (Memory), which is a memory device in a terminal device for storing programs and data. It is understandable that the computer-readable storage medium here can include both the built-in storage medium in the terminal device and the extended storage medium supported by the terminal device. The computer-readable storage medium provides a storage space, which stores the operating system of the terminal. In addition, one or more instructions suitable for being loaded and executed by the processor are also stored in the storage space, and these instructions can be one or more computer programs (including program codes). It should be noted that the computer-readable storage medium here can be a high-speed RAM memory or a non-volatile memory, such as at least one disk memory.

[0108] One or more instructions stored in a computer-readable storage medium may be loaded and executed by a processor to implement the corresponding steps of an AI-assisted focused ultrasound focal area temperature measurement method in the above embodiment; one or more instructions in the computer-readable storage medium are loaded and executed by the processor as follows:

[0109] Collect and process the medical ultrasound RF data of the focal area during HIFU treatment to obtain a B-mode ultrasound image;

[0110] Using the B-mode ultrasound images and the corresponding thermocouple recorded temperature data obtained in the experiment to construct a training set and a test set, and training an AI regression model based on a spatial feature analysis network model and a frequency domain feature analysis network model;

[0111] The AI ​​regression model is optimized based on the test set, thereby achieving single-point ultrasonic temperature measurement of body fat.

[0112] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0113] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0114] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0115] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0116] The preferred embodiments of the present invention are described in detail above, but the present invention is not limited to the above embodiments, and various changes can be made within the knowledge scope of ordinary technicians in this field without departing from the purpose of the present invention.

[0117] Many other changes and modifications may be made without departing from the concept and scope of the present invention.It should be understood that the present invention is not limited to the specific embodiments, and the scope of the present invention is defined by the appended claims.

Claims

1. An AI-assisted focused ultrasound focal area temperature measurement method, characterized in that: The method comprises: Collect and process the medical ultrasound RF data of the focal area during HIFU treatment to obtain a B-mode ultrasound image; Using the B-mode ultrasound images and the corresponding thermocouple recorded temperature data obtained in the experiment to construct a training set and a test set, and training an AI regression model based on a space-frequency domain intelligent network model; The AI ​​regression model is optimized based on the test set, and the spatial feature analysis network and the frequency domain feature analysis network are functionally coupled to fully utilize the information of the tissues surrounding the measured point and the grayscale change trend of the measured point to achieve single-point ultrasonic temperature measurement of body fat; wherein, the spatial feature analysis module extracts the horizontal and vertical spatial information of the temperature measurement area, and the frequency domain feature analysis module learns the frequency domain features from a global perspective.

2. The AI-assisted focused ultrasound focal area temperature measurement method according to claim 1, characterized in that: The spatial feature analysis network in the space-frequency domain intelligent network model includes three stages, each of which includes multiple trans blocks, a convolution block and an average pooling layer; The trans block in the spatial feature analysis network model includes four channels: an H auxiliary channel for extracting inter-layer spatial correlation; a V auxiliary channel for extracting intra-layer spatial correlation; and an H2V main channel for fusing the features of the two auxiliary channels. The V2H main channel is used to fuse the features of the two auxiliary channels.

3. The AI-assisted focused ultrasound focal area temperature measurement method according to claim 1, characterized in that: The input of the spatial feature analysis network model is an ultrasound B-mode image; the ultrasound B-mode image is convolved to achieve downsampling; the downsampled ultrasound B-mode image is embedded in three ways: (1) the image is divided into multiple 16×16 blocks to form single-piece visual words; (2) the image is divided into multiple layers along the depth H direction to form inter-layer visual words; (3) Divide the image into multiple layers along the horizontal V direction to form intra-layer visual words; perform a flatten operation on the single-piece visual words, inter-layer visual words, and intra-layer visual words to obtain z0, z h and z v .

4. The AI-assisted focused ultrasound focal area temperature measurement method according to claim 1, characterized in that: The four channel inputs of the lth trans block in the sth stage of the spatial feature analysis network model are and The output of the previous layer is the input of the next layer. The formula is as follows: The final output of the s-1th stage of the spatial feature analysis network model is the input of the first trans block of the sth stage, and the formula is as follows: The initial inputs of the two auxiliary channels of the trans block are z h 、z v , the initial inputs of both main channels are z0; The input of the auxiliary channel of the trans block is output through the transformer neural network, and the formula is as follows: The main channel input of the trans block passes through the transformer neural network, is added to the output of the auxiliary channel, and is input into the transformer neural network again to obtain the output. The formula is as follows: Among them, the final output of the trans block in each stage is and Input convolution block: First and The one-dimensional vector is restored to a two-dimensional tensor and connected end to end, and then three convolutions are performed. After each convolution, batch normalization is performed to accelerate the convergence of the model. The three convolutions all use RELU as the activation function. The convolution block outputs z s+1 Finally, the average pooling layer is input to get the output of each stage.

5. The AI-assisted focused ultrasound focal area temperature measurement method according to claim 1, characterized in that: The frequency domain feature analysis neural network includes: a frequency domain channel learning block and a frequency domain time learning block; The frequency domain channel learning block includes three parts: the first part is a domain converter based on Fourier transform, which is used to transform the time domain signal into the frequency domain signal; the second part is a frequency domain multilayer perceptron, which is used to share learnable weights between different channels and capture the interaction and correlation between different variables; the third part is a domain inverter based on inverse Fourier transform, which is used to transform the frequency domain signal into the time domain signal; The frequency-domain time learning block includes three parts: the first part is a domain converter based on Fourier transform, which is used to convert the time domain signal into a frequency domain signal; the second part is a frequency domain multilayer perceptron, which is used to study the frequency domain features in the same channel; the third part is a domain inverter based on inverse Fourier transform, which is used to convert the frequency domain signal into a time domain signal.

6. The AI-assisted focused ultrasound focal area temperature measurement method according to claim 1, characterized in that: The training is based on the AI ​​regression model of the spatial feature analysis network model and the frequency domain feature analysis network model, including: The frequency domain feature analysis network converts the time domain information into the frequency domain to learn global features from a global perspective, and combines it with the spatial features learned by the spatial feature analysis network to perform temperature regression calculation; The input of the frequency domain feature analysis network is the B-mode ultrasound image. After selecting 10 consecutive frames of images from time t, a 16×16 small image block {p t-9 ,p t-8 ,...,p t }, after straightening the image block, we get the tensor {x t-9 ,x t-8 ,...,x t |x i ∈R N }, X t =[x t-9 ,x t-8 ,...,x t ]∈R N×L , where L = 10, input frequency domain characteristic analysis network to estimate the temperature Y corresponding to time t t ; For the tensor X t Expand the dimension and multiply it by the learnable vector Used to obtain more hidden information, the formula is as follows: The tensor H t ∈R N×L×d As the input of the frequency domain channel learning block; it contains Fourier transform to convert time domain information into frequency domain information, and inverse Fourier transform to convert frequency domain information into time domain information. The formulas are as follows: Where f is the frequency, v is the integral variable, j is an imaginary number, H(v)cos(2πfv) is the real part, and H(v)sin(2πfv) is the imaginary part; Where, f is the integral variable Re(h(f)) is the real part, and Im(h(f)) is the imaginary part; When considering time series prediction, channel dependency can provide interactions and correlations between different variables, thereby improving prediction accuracy. Therefore, the frequency domain channel learning block shares weights between L timestamps to understand the association of channels: For input H t ∈R N×L×d , consider the lth timestamp corresponding to This is fed into the frequency domain channel learning block: Among them, fourier() and Fourier() represent Fourier transform and inverse Fourier transform respectively. express The frequency domain components of , and the weights and biases satisfy: The output Z∈R of the frequency domain channel learning block N×L×d As the input of the frequency-domain time learning block, consider the nth channel corresponding to Feed this into the frequency-domain time learning block: S t ∈R N×L×d After being straightened into a one-dimensional vector, it is connected with the output features of the spatial feature analysis network and then input into the fully connected layer for regression calculation, which is used for AI regression calculation of the temperature at time t.

7. The AI-assisted focused ultrasound focal-area temperature measurement method according to claim 1, characterized in that: The features output by the spatial feature analysis network model and the frequency domain feature analysis network model are concat-connected and input into the fully connected layer for regression calculation to infer the temperature of the measured point; the spatial feature analysis network model and the frequency domain feature analysis network model calculate the loss function based on the mean square error, and the formula of the loss function is as follows: Where n represents the batch size of the model; Indicates the temperature recorded by the corresponding thermocouple; i =f(x i ), represents the predicted temperature result, x i For input data.

8. An AI-assisted focused ultrasound focal-area temperature measurement system, characterized in that: The system is used to execute the method described in any one of claims 1 to 7, and the system comprises: A data acquisition module, used to collect and process the medical ultrasound RF data of the focal area during HIFU treatment to obtain a B-mode ultrasound image; A model training module, used to construct a training set and a test set using the B-mode ultrasound image and the corresponding thermocouple recorded temperature data obtained in the experiment, and to train an AI regression model based on a space-frequency domain intelligent network model; The model optimization module is used to optimize the AI ​​regression model based on the test set, couple the spatial feature analysis network with the frequency domain feature analysis network function, and make full use of the information of the tissue around the measured point and the grayscale change trend of the measured point to realize single-point ultrasonic temperature measurement of body fat; wherein, the spatial feature analysis module extracts the horizontal and vertical spatial information of the temperature measurement area, and the frequency domain feature analysis module learns the frequency domain features from a global perspective.

9. A computer device, characterized in that: The invention comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, an AI-assisted focused ultrasound focal area temperature measurement method according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, which, when executed by a processor, implements an AI-assisted focused ultrasound focal-area temperature measurement method according to any one of claims 1 to 7.

Images