A conductivity profile image prediction method, system, device, medium and product

Abstract

The application discloses a conductivity distribution image prediction method, system, device, medium and product, relates to the field of magnetic acoustic electric imaging, and comprises the following steps: randomly generating multiple conductivity distribution images with different geometric shapes, and generating an excitation sound field with a rotation angle of 0 to plus or minus 88 degrees based on a limited-angle rotation scanning mode; training a conductivity model by taking the magnetic acoustic electric signal corresponding to the excitation sound field as input and taking the conductivity distribution image as output, and determining a trained conductivity network; the conductivity model comprises multiple sub-networks, and each sub-network comprises a denoising diffusion probability model; acquiring a current limited-angle magnetic acoustic electric signal based on the limited-angle rotation scanning mode; inputting the current limited-angle magnetic acoustic electric signal into the trained conductivity model, and outputting a predicted conductivity distribution image, so that the effectiveness of imaging an object with irregular conductivity distribution and the credibility of reconstructing the conductivity distribution image are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of magnetic acoustic electrical imaging, and in particular to a conductivity distribution image prediction method, system, device, medium and product. BACKGROUND

[0002] Magnetic Acoustic Electrical Tomography (MAET) is an emerging functional imaging technique used to detect the electrical conductivity properties of soft tissues, which are closely related to the physiological and pathological states of the tissues. The imaging process of MAET generally includes three main steps. First, ultrasonic waves are generated in a static magnetic field using ultrasonic sensors, which are used for imaging within the sample. Then, while the ultrasonic waves propagate within the sample, the Magnetic Acoustic Electric Effect (MAE) signals related to tissue conductivity are measured using surrounding electrodes. Next, the ultrasonic transducer is moved to different positions mechanically or electronically to collect a series of MAE signals. Finally, by processing these MAE signals, images containing spatial conductivity information, i.e., so-called MAET images, can be generated. These images provide valuable information to researchers about tissue conductivity and other physiological properties.

[0003] A model-based MAET imaging method using linear scanning is disclosed in the prior art. This method successfully reconstructed the conductivity distribution image of a regular phantom for the first time, and its accuracy and stability have been improved compared with traditional techniques. However, the fixed limitations of the linear scanning strategy limit its effectiveness in imaging targets with irregular conductivity distribution. SUMMARY

[0004] The purpose of the present application is to provide a conductivity distribution image prediction method, system, device, medium and product to solve the problem that the fixed limitations of the linear scanning strategy limit the effectiveness of imaging targets with irregular conductivity distribution.

[0005] To achieve the above-mentioned purpose, the present application provides the following solutions.

[0006] In a first aspect, the present application provides a conductivity distribution image prediction method, comprising the following steps.

[0007] Randomly generate multiple conductivity distribution images with different geometric shapes, and generate excitation sound fields with rotation angles of 0~±88 degrees based on a limited-angle rotational scanning method.

[0008] The conductivity model includes a plurality of sub-networks, and each sub-network includes a denoising diffusion probability model.

[0009] The current limited-angle magnetic acoustic electric signal is obtained based on a limited-angle rotation scanning mode.

[0010] The current limited-angle magnetic acoustic electric signal is obtained based on a limited-angle rotation scanning mode.

[0011] In a second aspect, the present application provides a conductivity distribution image prediction system, comprising the following modules.

[0012] The conductivity distribution image and excitation sound field generation module is configured to randomly generate a plurality of conductivity distribution images with different geometrical shapes, and simultaneously generate excitation sound fields with rotation angles of 0 to ±88 degrees based on a limited-angle rotation scanning mode.

[0013] The training module is configured to train a conductivity model by taking the magnetic acoustic electric signal corresponding to the excitation sound field as input and taking the conductivity distribution image as output, and determine a trained conductivity network; the conductivity model includes a plurality of sub-networks, and each sub-network includes a denoising diffusion probability model.

[0014] The scanning module is configured to obtain a current limited-angle magnetic acoustic electric signal based on a limited-angle rotation scanning mode.

[0015] The prediction module is configured to input the current limited-angle magnetic acoustic electric signal into the trained conductivity model, and output a predicted conductivity distribution image.

[0016] In a third aspect, the present application provides a computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the conductivity distribution image prediction method described above.

[0017] In a fourth aspect, the present application provides a computer readable storage medium having a computer program stored thereon, wherein the computer program is executed by a processor to implement the conductivity distribution image prediction method described above.

[0018] In a fifth aspect, the present application provides a computer program product comprising a computer program, wherein the computer program is executed by a processor to implement the conductivity distribution image prediction method described above.

[0019] According to the specific embodiments provided by the present application, the present application has the following technical effects:

[0020] The application uses a limited-angle rotation scanning method to reconstruct conductivity, overcomes the problem that linear scanning cannot image irregular shapes by dividing multiple angles; at the same time, the limited-angle rotation scanning method scans fewer angles than the full-angle rotation scanning method, and the signal acquisition time is shorter, which improves the effectiveness of imaging targets with irregular conductivity distribution, finally, by training a conductivity model including a de-noise diffusion probabilistic model (DDPM), realistic details and textures can be generated, and DDPM can rely on learned knowledge to "imagine" the most reasonable image information, improving the credibility of the reconstructed conductivity distribution image. BRIEF DESCRIPTION OF DRAWINGS

[0021] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed in the embodiments will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0022] Figure 1 A conductivity distribution image prediction method flowchart is provided for an embodiment of the present application.

[0023] Figure 2 A 0~±32 degree limited-angle conductivity boundary image diagram is provided for an embodiment of the present application.

[0024] Figure 3 A 0~±88 degree full-angle conductivity boundary image diagram is provided for an embodiment of the present application.

[0025] Figure 4 A non-artifact noise conductivity boundary image diagram is provided for an embodiment of the present application.

[0026] Figure 5 A conductivity distribution image diagram is provided for an embodiment of the present application.

[0027] Figure 6 A conductivity distribution image reconstruction step diagram is provided for an embodiment of the present application. DETAILED DESCRIPTION

[0028] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some embodiments of the present application, not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0029] In order to make the purposes, characteristics and advantages of the present application more obvious and easy to understand, the present application will be further described in detail below in combination with the drawings and specific embodiments.

[0030] As shown in the Figure 1 The present application provides a conductivity distribution image prediction method, which comprises the following steps.

[0031] S1: randomly generating a plurality of conductivity distribution images with different geometric shapes, and generating an excitation sound field with a rotation angle of 0~±88 degrees based on a limited angle rotation scanning mode.

[0032] S2: training a conductivity model by taking the magnetoacoustic electric signal corresponding to the excitation sound field as input and taking the conductivity distribution image as output, and determining a trained conductivity network; the conductivity model comprises a plurality of sub-networks, and each sub-network comprises a denoising diffusion probability model.

[0033] S3: acquiring a current limited angle magnetoacoustic electric signal based on the limited angle rotation scanning mode.

[0034] S4: inputting the current limited angle magnetoacoustic electric signal into the trained conductivity model to output a predicted conductivity distribution image.

[0035] In an exemplary embodiment, the excitation sound field generates an excitation sound field with a rotation angle every 4 degrees.

[0036] In actual application, 500 conductivity distribution images with different geometric shapes are randomly generated, and an excitation sound field with a rotation angle of 0~±88 degrees is generated, and the excitation sound field generates a sound field with an angle every 4 degrees.

[0037] Wherein, the 0~±88 degree magnetoacoustic electric signal is obtained through the measurement formula positive process of magnetoacoustic electric imaging; and the training data is obtained by using the filtered back projection algorithm and the untrained conductivity model, and the training data comprises a limited angle conductivity boundary image and a full angle conductivity boundary image.

[0038] Specifically, the limited angle conductivity boundary image is generated by using the filtered back projection algorithm on the 0~±32 degree limited angle magnetoacoustic electric signal; the full angle conductivity boundary image is generated by using the filtered back projection algorithm on the 0~±88 degree full angle magnetoacoustic electric signal; and the high-pass filtered gold standard conductivity distribution image is generated to obtain the artifact-free noise conductivity boundary image, wherein the full angle boundary image is the boundary image of the gold standard conductivity distribution image.

[0039] In an exemplary embodiment, the plurality of sub-networks comprises an angle completion network, an artifact removal network and a conductivity reconstruction network, and S2 specifically comprises the following steps.

[0040] S21: Based on the excitation sound field, the magnetoacoustic signal is determined using the measurement formula of magnetoacoustic imaging.

[0041] S22: Using a filtered back projection algorithm, the magnetoacoustic signal with a finite angle in the magnetoacoustic signal is processed to generate a finite-angle conductivity boundary image.

[0042] S23: Input the finite-angle conductivity boundary image into the angle completion network and output the full-angle conductivity boundary image.

[0043] S24: Input the full-angle conductivity boundary image into the artifact removal network and output an artifact-free and noise-free conductivity boundary image.

[0044] S25: Input the artifact-free noise conductivity boundary image into the conductivity reconstruction network and output the conductivity distribution image.

[0045] In practical applications, this application requires training a total of 3 sub-networks, each of which is trained using DDPM.

[0046] The first sub-network is the angle completion network, used to complete the angle of a finite-angle conductivity boundary image. The input to the angle completion network is a finite-angle conductivity boundary image ranging from 0 to ±32 degrees, such as... Figure 2 As shown, the output of the angle completion network is a full-angle conductivity boundary image from 0 to ±88 degrees, as... Figure 3 As shown.

[0047] The second sub-network is the artifact removal network, used to denoise and remove artifacts from the full-angle conductivity boundary image. The input to the artifact removal network is the full-angle conductivity boundary image from 0 to ±88 degrees, and the output is an artifact-free and noise-free conductivity boundary image, i.e., a high-pass filtered conductivity boundary image, such as... Figure 4 As shown.

[0048] The third sub-network is the conductivity reconstruction network, used to reconstruct the conductivity distribution image from the conductivity boundary image. The input to the conductivity reconstruction network is the high-pass filtered conductivity boundary image, and the output is the conductivity distribution image, such as... Figure 5 As shown.

[0049] In one exemplary embodiment, the magnetoacoustic signal is a magnetoacoustic signal with a range of 0 to ±88 degrees.

[0050] The finite-angle magnetoacoustic signal is a magnetoacoustic signal with a finite angle of 0 to ±32 degrees.

[0051] The full-angle conductivity boundary image is a full-angle conductivity boundary image of 0 to ±88 degrees.

[0052] In an example embodiment, the measurement formula of magnetoacoustic electrical imaging is:

[0053]

[0054] wherein, is a magnetoacoustic electrical signal; is a sound vibration velocity field, r is an arbitrary point in the field space, t is time; B is a static magnetic field; is a reciprocal current density; is an integral region.

[0055] In an example embodiment, the prediction process of the conductivity distribution image is the same as the training process of the conductivity model.

[0056] In actual application, as shown in Figure 6 , the prediction process is the process of reconstructing the conductivity distribution image.

[0057] In the experiment, the limited-angle magnetoacoustic electrical signals are obtained by limited-angle rotational scanning, and the limited-angle conductivity boundary image is obtained by filtering back-projection of the magnetoacoustic electrical signals. The limited-angle conductivity boundary image is input into the angle completion network, and the completed full-angle conductivity boundary image is output. The completed full-angle conductivity boundary image is input into the artifact removal network, and the artifact-free noise conductivity boundary image, i.e., the clean and noise-free artifact-removed full-angle conductivity boundary image, is output. The artifact-free noise conductivity boundary image is input into the conductivity reconstruction network, and the conductivity distribution image is output.

[0058] The application also provides a conductivity distribution image prediction system, comprising the following modules.

[0059] The conductivity distribution image and excitation sound field generation module is configured to randomly generate multiple conductivity distribution images with different geometric shapes, and generate excitation sound fields with rotation angles of 0~±88 degrees based on a limited-angle rotational scanning mode.

[0060] The training module is configured to train a conductivity model by taking the magnetoacoustic electrical signals corresponding to the excitation sound fields as input and taking the conductivity distribution images as output, and determine a trained conductivity network; the conductivity model comprises multiple sub-networks, and each sub-network comprises a denoising diffusion probability model.

[0061] The scanning module is configured to obtain current limited-angle magnetoacoustic electrical signals based on a limited-angle rotational scanning mode.

[0062] The prediction module is configured to input the current limited-angle magnetoacoustic electrical signals into the trained conductivity model, and output a predicted conductivity distribution image.

[0063] In an example embodiment, a computer device is provided, including a memory and a processor, the memory storing a computer program, and the processor implementing the steps in the above method embodiments when executing the computer program. The computer device can be a server or a terminal. The computer device includes a processor, a memory, an input / output interface (I / O) and a communication interface. The processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. The processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operating system and the computer program in the non-volatile storage medium to run. The database of the computer device is configured to store to-be-processed data. The input / output interface of the computer device is configured to exchange information between the processor and external devices. The communication interface of the computer device is configured to communicate with external terminals through a network connection. The computer program is executed by the processor to implement the above method.

[0064] In an example embodiment, a computer device is provided, including a memory and a processor, the memory storing a computer program, and the processor implementing the steps in the above method embodiments when executing the computer program.

[0065] In an example embodiment, a computer readable storage medium is provided, storing a computer program, and the computer program is executed by a processor to implement the steps in the above method embodiments.

[0066] In an example embodiment, a computer program product is provided, including a computer program, and the computer program is executed by a processor to implement the steps in the above method embodiments.

[0067] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data authorized by the user or authorized by all parties, and the collection, use and processing of related data need to comply with relevant regulations.

[0068] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by computer program instructions related to hardware, and the computer program can be stored in a non-volatile computer readable storage medium. When the computer program is executed, it can include the processes of the above-mentioned embodiments of each method. Among them, any reference to memory, database or other medium used in each embodiment provided by the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (Read-Only Memory, ROM), magnetic tape, floppy disk, flash memory, optical storage, high-density embedded non-volatile memory, resistive memory (ReRAM), magnetoresistive random access memory (Magnetoresistive Random Access Memory, MRAM), ferroelectric memory (Ferroelectric Random Access Memory, FRAM), phase change memory (Phase Change Memory, PCM), graphene memory, etc. Volatile memory can include random access memory (Random Access Memory, RAM) or external cache memory, etc. As an illustration but not limitation, RAM can be in various forms, such as static random access memory (Static Random Access Memory, SRAM) or dynamic random access memory (Dynamic Random Access Memory, DRAM), etc.

[0069] The database involved in each embodiment provided by the present application can include at least one of a relational database and a non-relational database. The non-relational database can include a distributed database based on blockchain, etc., without being limited thereto. The processor involved in each embodiment provided by the present application can be a general-purpose processor, a central processing unit, a graphics processing unit, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., without being limited thereto.

[0070] Each technical feature of the above embodiments can be combined arbitrarily. In order to make the description simple, not all possible combinations of each technical feature in the above embodiments are described, but as long as the combination of these technical features does not exist, it should be considered as the scope of the present application.

[0071] The principles and implementation modes of the present application are described by applying specific examples herein, and the above embodiment descriptions are only used to help understand the method and its core idea of the present application; at the same time, for those skilled in the art, according to the idea of the present application, the specific implementation mode and application range will be changed. In conclusion, the content of the present application should not be understood as a limitation.

Claims

1. An electrical conductivity profile image prediction method characterized by, The method comprises the steps of: randomly generating multiple conductivity distribution images with different geometric shapes, and generating excitation sound fields with rotation angles of 0~±88 degrees based on a limited-angle rotation scanning mode; training a conductivity model by taking magnetic acoustic electric signals corresponding to the excitation sound fields as input and taking the conductivity distribution images as output, and determining a trained conductivity model; the conductivity model comprises multiple sub-networks, each of which comprises a denoising diffusion probability model; the multiple sub-networks comprise an angle completion network, an artifact removal network, and a conductivity reconstruction network; training a conductivity model by taking magnetic acoustic electric signals corresponding to the excitation sound fields as input and taking the conductivity distribution images as output, and determining a trained conductivity model, specifically comprising: determining the magnetic acoustic electric signals based on the excitation sound fields through a magnetic acoustic electric imaging measurement formula; processing the magnetic acoustic electric signals with limited angles in the magnetic acoustic electric signals by using a filtered back-projection algorithm to generate limited-angle conductivity boundary images; inputting the limited-angle conductivity boundary images into the angle completion network by using the filtered back-projection algorithm to output full-angle conductivity boundary images; inputting the full-angle conductivity boundary images into the artifact removal network to output artifact-free noise conductivity boundary images; inputting the artifact-free noise conductivity boundary images into the conductivity reconstruction network to output conductivity distribution images; acquiring current limited-angle magnetic acoustic electric signals based on a limited-angle rotation scanning mode; inputting the current limited-angle magnetic acoustic electric signals into the trained conductivity model to output predicted conductivity distribution images.

2. The electrical conductivity profile image prediction method according to claim 1, characterized by, The excitation sound field generates an excitation sound field with a rotation angle every 4 degrees.

3. The electrical conductivity profile image prediction method according to claim 1, characterized by, The magnetic acoustic electric signals are 0~±88-degree magnetic acoustic electric signals. The limited-angle magnetic acoustic electric signals are 0~±32-degree limited-angle magnetic acoustic electric signals. The full-angle conductivity boundary images are 0~±88-degree full-angle conductivity boundary images.

4. The electrical conductivity profile image prediction method according to claim 1, characterized by, The measurement formula of the magnetic acoustic electric imaging is: wherein is the magnetoacoustic electric signal; is the acoustic vibration velocity field, r is an arbitrary point in the field space, t is time; B is the static magnetic field; is the reciprocal current density; is the integration region.

5. The conductivity profile image prediction method according to claim 1, characterized by, The prediction process of the conductivity distribution image is the same as the training process of the conductivity model.

6. An electrical conductivity profile image prediction system characterized by, The conductivity distribution image prediction system executes the conductivity distribution image prediction method of any one of claims 1-5, and the conductivity distribution image prediction system comprises: a conductivity distribution image and excitation sound field generation module, configured to randomly generate multiple conductivity distribution images with different geometric shapes, and generate excitation sound fields with rotation angles of 0~±88 degrees based on a limited-angle rotation scanning mode; a training module, configured to train a conductivity model by taking magnetic acoustic electric signals corresponding to the excitation sound fields as input and taking the conductivity distribution images as output, and determine a trained conductivity model; the conductivity model comprises multiple sub-networks, each of which comprises a denoising diffusion probability model; a scanning module, configured to acquire current limited-angle magnetic acoustic electric signals based on a limited-angle rotation scanning mode; a prediction module, configured to input the current limited-angle magnetic acoustic electric signals into the trained conductivity model to output predicted conductivity distribution images.

7. A computer device comprising: A memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that the processor executes the computer program to implement the conductivity distribution image prediction method of any one of claims 1-5.

8. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the conductivity distribution image prediction method of any one of claims 1-5.

9. A computer program product comprising a computer program, characterized in that, The computer program is executed by the processor to implement the conductivity distribution image prediction method of any one of claims 1-5.

Images