WO2021062885A1

WO2021062885A1 - Ct imaging method and apparatus, storage medium, and medical imaging system

Info

Publication number: WO2021062885A1
Application number: PCT/CN2019/111028
Authority: WO
Inventors: 葛永帅; 梁栋; 朱炯滔; 刘新; 郑海荣
Original assignee: 中国科学院深圳先进技术研究院
Priority date: 2019-09-30
Filing date: 2019-10-14
Publication date: 2021-04-08
Also published as: CN112581554A; CN112581554B

Abstract

Disclosed are a CT imaging method and apparatus, a storage medium, and a medical imaging system. The CT imaging method comprises: acquiring projection data of a target object, the projection data being collected when the target object is detected on the basis of a preset dose of X-ray photons, and the preset dose being smaller than a CT standard dose (S110); and sending the projection data to a pre-trained imaging model, and determining an output image of the imaging model as a CT image of the target object, wherein the imaging model is trained by a CT image at the standard dose and sample projection data which contains noise. The imaging model comprises a first sub-network module, a second sub-network module and a third sub-network module, the first sub-network module is used to perform noise reduction on projection data of a projection domain, the second sub-network module is used to transform the processed projection data so as to generate CT domain data, and the third sub-network module is used to perform noise reduction on the CT domain data so as to generate a CT image of the target object (S120).

Description

CT imaging method, device, storage medium and medical imaging system

This application claims the priority of the Chinese patent application filed with the Chinese Patent Office with application number 201910941729.1 on September 30, 2019. The entire content of this application is incorporated into this application by reference.

Technical field

The embodiments of the present application relate to medical imaging technology, for example, to a computer tomography (Computed Tomography, CT) imaging method, device, storage medium, and medical imaging system.

Background technique

X-ray-based CT imaging is a common way to assist diagnosis. X-ray imaging is radioactive and will increase the risk of cancer in the subject.

In order to reduce the radiation dose received by the subject, low-dose CT detection has become one of the research directions in the field of CT imaging. The low-dose CT reconstruction algorithm is developed based on iterative reconstruction technology, which simulates CT image reconstruction as a mathematical optimization problem. The iterative reconstruction algorithm simulates a comprehensive projection on the estimated image through forward projection, and the process simulates the reality as much as possible. The process of X-ray photons passing through the detected object to the detector in the CT system. The integrated projection is compared with the real measurement value collected by the detector, and the next update is determined based on the difference between the two. Correct the current estimated image.

However, the above iterative reconstruction algorithm has the problem of long reconstruction time, especially complex calculations cannot achieve real-time reconstruction.

Summary of the invention

The present application provides a CT imaging method, device, storage medium, and medical imaging system to improve the efficiency and quality of CT image reconstruction.

The embodiment of the present application provides a CT imaging method, including:

Acquiring projection data of the target object, where the projection data is collected when detecting the target object based on a preset dose of X-ray photons, and the preset dose is less than the CT standard dose;

The projection data is sent to a pre-trained imaging model, and the output image of the imaging model is determined as the CT image of the target object, wherein the imaging model is composed of a standard dose CT image and sample projection data containing noise After training, the imaging model includes a first sub-network module, a second sub-network module, and a third sub-network module. The first sub-network module is used to perform noise reduction processing on projection data in the projection domain, and the second The sub-network module is used to transform the processed projection data to generate CT domain data, and the third sub-network module is used to perform noise reduction processing on the CT domain data to generate a CT image of the target object.

An embodiment of the present application also provides a CT imaging device, which includes:

A projection data acquisition module configured to acquire projection data of a target object, wherein the projection data is acquired when the target object is detected based on a preset dose of X-ray photons, and the preset dose is less than the CT standard dose;

The CT image reconstruction module is configured to send the projection data to a pre-trained imaging model, and determine the output image of the imaging model as the CT image of the target object, wherein the imaging model consists of a standard dose of CT image And the sample projection data containing noise is obtained by training, the imaging model includes a first sub-network module, a second sub-network module and a third sub-network module, the first sub-network module is used to reduce the projection data of the projection domain Noise processing, the second sub-network module is used to transform the processed projection data to generate CT domain data, and the third sub-network module is used to perform noise reduction processing on the CT domain data to generate the target CT image of the subject.

The embodiment of the present application also provides a computer-readable storage medium that stores a computer program, and when the program is executed by a processor, the CT imaging method as provided in any embodiment of the present application is implemented.

The embodiment of the present application also provides a medical imaging system, which includes a medical imaging device and a computer device. The computer device includes a memory, one or more processors, and a computer program stored in the memory and running on the processor, When the program is executed by the processor, the CT imaging method as provided in any embodiment of the present application is realized.

Description of the drawings

FIG. 1 is a schematic flowchart of a CT imaging method provided in Embodiment 1 of the present application;

2 is a schematic structural diagram of an imaging model provided in Embodiment 1 of the present application;

3 is a schematic structural diagram of another imaging model provided in Embodiment 1 of the present application;

4 is a schematic flowchart of another CT imaging method provided in Embodiment 2 of the present application;

FIG. 5 is a schematic structural diagram of a CT imaging apparatus according to Embodiment 3 of the present application;

FIG. 6 is a schematic structural diagram of a medical imaging system provided by Embodiment 4 of the present application.

Detailed ways

The application will be described below with reference to the drawings and embodiments. The specific embodiments described herein are only used to explain the application, but not to limit the application. For ease of description, the drawings only show a part of the structure related to the present application instead of all of the structure.

Example one

Fig. 1 is a schematic flow chart of a CT imaging method provided in the first embodiment of this application. This embodiment can be applied to the situation of rapidly generating high-quality low-dose CT images based on an imaging model. The CT imaging method includes S110-S120.

S110. Obtain projection data of the target object, where the projection data is collected when detecting the target object based on a preset dose of X-ray photons, and the preset dose is less than the CT standard dose.

S120. Send the projection data to a pre-trained imaging model, and determine an output image of the imaging model as a CT image of the target object. Wherein, the imaging model is obtained by training a standard dose CT image and sample projection data containing noise. The imaging model includes a first sub-network module, a second sub-network module, and a third sub-network module. The network module is used to denoise the projection data in the projection domain, the second sub-network module is used to transform the processed projection data to generate CT domain data, and the third sub-network module is used to The CT domain data undergoes noise reduction processing to generate a CT image of the target object.

In this embodiment, the target object may be a human or an animal, and the detection of the target object by low-dose X-ray photons can reduce radiation damage to the target object by X-rays during the detection process, and reduce the risk of cancer of the target object. In this embodiment, the imaging model is pre-trained and has the function of reconstructing low-dose projection data and the function of reducing noise in the low-dose projection data. The target object will be detected by low-dose X-ray photons. , The collected projection data is input to the imaging model to generate a high-quality CT image of the target object after noise reduction. Since the imaging model is pre-trained, it can automatically process the input projection data and output the CT image of the target object. There is no need to iteratively calculate each set of projection data, which reduces the calculation and difficulty of CT image reconstruction, and improves CT The image reconstruction efficiency, meanwhile, the imaging model performs noise reduction processing on the projection data, which improves the clarity of the image generated by low-dose detection.

In this embodiment, the imaging model includes a first sub-network module, a second sub-network module, and a third sub-network module. For example, refer to FIG. 2, which is a schematic structural diagram of an imaging model provided by an embodiment of the present application. . The first sub-network module is used to perform convolution filtering processing on the input projection image, and the projection domain is a domain formed by all the projection images. The first sub-network module may include a first preset number of convolution modules connected in sequence. Each convolution module in the first preset number of convolution modules includes a convolution layer and an activation function layer. The network module can perform noise reduction and filtering processing on the projection data. In an embodiment, the first preset number may be 5, 6, 10, and so on. In an embodiment, the convolution kernel of the convolution layer in the first sub-network module is a×b, a and b are both positive integers greater than or equal to 1, and b>a. Exemplarily, the convolution kernel of the convolution layer in the first sub-network module may be 1×30, 3×30, 5×30, 3×33, 5×33, etc., by adopting the convolution of the form b>a The verification can simultaneously achieve the noise reduction and filtering functions of the projection data. Exemplarily, the activation function layer may be a Leaky Rectified Linear Unit (Leaky ReLU).

The second subconvolution model is used to convert projection data in the projection domain into CT domain data, which is a domain formed by all CT images. In an embodiment, the second sub-network module transforms the processed projection data according to the following formula:

In this formula, f(x, y) is the CT domain image output by the second sub-network module, x and y are the abscissa and ordinate in the CT domain image, respectively, and p(r, θ) is the second input The projection data of the sub-network module, r is the distance between the projection data and the origin, δ is the projection X-ray, and θ is the projection angle. The second sub-network module reconstructs the projection data input by the first sub-network module according to the above formula, and outputs a CT domain image of the same size as the expected size, and the CT domain image satisfies the CT image attributes.

The third sub-network module is used to perform convolution filtering operations on the CT domain image, and simultaneously perform image restoration processing and noise reduction processing on the obtained CT domain image. In an embodiment, restoring the image may include a second preset number of convolution modules connected in sequence, and each convolution module of the second preset number of convolution modules includes a convolution layer and an activation function layer. . In an embodiment, the second preset number may be the same as or different from the first preset number, the second preset number may be, for example, 5, 6, or 10, etc., and the second preset number and the first preset number may be Respectively determined according to the functional accuracy. Exemplarily, the activation function layer may be Leaky ReLU, the convolution kernel of the convolution layer in the third sub-network module is m×m, and the m is a positive integer greater than or equal to 1, for example, the third sub-network module The convolution kernel of the convolution layer in can be 3×3, 5×5, or 7×7.

Exemplarily, refer to FIG. 3, which is a schematic structural diagram of another imaging model provided in Embodiment 1 of the present application. The first sub-network module in Figure 3 includes 6 convolution modules. The convolution kernel of the convolution layer in the convolution module is 3×30 with a step size of 1. The third sub-network module includes 6 convolution modules. The convolution kernel of the convolution layer in the product module is 3×3, and the step size is 1. The size of the input projection data in Figure 3 is 900×848, and the size of the output CT image is 512×512. The above image size is only an example of Figure 3. The size of the input projection data can be determined according to the input layer of the imaging model. Previously, if the size of the collected projection data was inconsistent with the processing size of the input layer of the imaging module, the collected projection data was adjusted in advance to make the collected projection data meet the processing size of the input layer of the imaging module.

On the basis of the foregoing embodiment, the input end of the first sub-network module is short-circuited with the output end of the first sub-network module, and the input end of the third sub-network module is connected to the third sub-network module The output terminal is shorted. By short-circuiting the network layers of the first sub-network module and the third sub-network module, firstly, for any sub-network module, the input information is transmitted to the output terminal through the short connection, so as to avoid the loss of information in the processing process, Second, in the training process, it avoids the training failure caused by the gradient descent during the reverse training process, which improves the training efficiency and quality.

In the technical solution provided by this embodiment, by setting three sub-network modules, the projection domain data and the CT domain data are respectively denoised, so as to realize double denoising in the image reconstruction process, improve the denoising effect, and make When X-ray photons are detected, high-quality CT images are obtained. At the same time, the three sub-network modules are connected to each other and are independent of each other. Any module can be updated and replaced according to imaging requirements, which improves the configurability and applicability of the imaging model.

Example two

Fig. 4 is a schematic flow chart of a CT imaging method provided in the second embodiment of the present application. On the basis of the foregoing embodiment, a training method of an imaging model is provided. The CT imaging method includes S210-S250.

S210. Establish an initial imaging model.

S220: Input the sample projection data containing noise into the initial imaging model to obtain a reconstructed image.

S230: Determine a loss function according to the reconstructed image and the CT image of the standard dose corresponding to the noise-containing sample projection data, and adjust the network parameters in the initial imaging model according to the loss function to generate the imaging model.

S240. Obtain projection data of the target object, where the projection data is collected when detecting the target object based on a preset dose of X-ray photons, and the preset dose is less than the CT standard dose.

S250: Send the projection data to a pre-trained imaging model, and determine an output image of the imaging model as a CT image of the target object.

In this embodiment, the initial imaging model is trained based on sample data to obtain an imaging model with image reconstruction function and denoising function. In an embodiment, the sample data includes low-dose projection data containing noise and CT images of standard doses. The low-dose projection data containing noise can be collected when the sample object is detected by low-dose X-ray photons, or can be obtained by adding noise based on the standard-dose CT image.

In an embodiment, before inputting the noise-containing sample projection data into the initial imaging model, acquiring the noise-containing sample projection data may be: acquiring a standard dose CT image, and preprocessing the standard dose CT image, Obtain standard projection data; add noise data to the standard projection data to generate sample projection data containing noise. In one embodiment, the preprocessing of the standard dose CT image may be to perform Radon operation on the standard dose CT image, generate standard projection data, and add noise to the standard projection data, such as Poisson noise, to obtain noise-containing Sample projection data. In this embodiment, the standard-dose CT image may be a CT image obtained by clinical testing. The sample projection data is generated by adding noise to the standard projection data, so as to avoid the X-ray scanning of the sample object during the sample generation process, which will cause the sample object to be damaged. Risk of cancer. In one embodiment, different basic noises are added to the standard projection data, different sample projection data are generated, the variety of noise in the sample is increased, and the number of samples is increased. In an embodiment, the adding noise to the standard projection data to generate sample projection data containing noise includes: setting a Poisson function of at least two noise levels; and a pair of Poisson functions based on the at least two noise levels Noise is added to the standard projection data, and at least two noise-containing sample projection data corresponding to the CT image of the standard dose are generated. In one embodiment, the orthographic projection data of the CT image of the standard dose, namely the sine image (sino image), is obtained through the Radon operation, and the sample projection data containing noise is generated in the following manner, sino_ps=(poissrnd(exp(-sino).* I ₁ ,[del_num,angle_num])+1)./I ₁ , sino_ps is the sample projection data containing noise, del_num is the number of detectors, angle_num is the total number of projection angles, I ₁ ＝I ₀ ×k, k is noise Intensity coefficient, I ₀ is the basic photon number, I ₀ =1×10 ⁶ . In an embodiment, k≤1, for example, can be 1, 0.1, 0.2, 0.5, or 0.05, etc. The value of k is not limited, and the added noise level is adjusted by the value of k. The smaller the k, the stronger the noise. In this embodiment, by adding different levels of noise to the standard projection data, sample projection data with different levels of noise are formed. Based on the sample projection data of different levels of noise, the initial imaging model is trained to reduce noise, so as to improve the noise reduction effect of the imaging model on different levels of noise.

Example three

FIG. 5 is a schematic structural diagram of a CT imaging apparatus provided in Embodiment 3 of the present application. The apparatus includes: a projection data acquisition module 310 configured to acquire projection data of a target object, wherein the projection data is based on a preset dose X-ray photons are collected when the target object is detected, and the preset dose is less than the CT standard dose; the CT image reconstruction module 320 is configured to send the projection data to the pre-trained imaging model, and the imaging model The output image of is determined to be the CT image of the target object, wherein the imaging model is obtained by training a standard dose of CT image and sample projection data containing noise, and the imaging model includes a first sub-network module and a second sub-network Module and a third sub-network module. The first sub-network module is used for noise reduction processing on projection data in the projection domain, and the second sub-network module is used for transforming the processed projection data to generate CT domain data The third sub-network module is configured to perform noise reduction processing on the CT domain data to generate a CT image of the target object.

In an embodiment, the first sub-network module includes a first preset number of convolution modules connected in sequence, and each convolution module in the first preset number of convolution modules includes a convolution layer and an activation function Layer; the third sub-network module includes a second preset number of convolution modules connected in sequence, and each convolution module in the second preset number of convolution modules includes a convolution layer and an activation function layer.

In an embodiment, the convolution kernel of the convolution layer in the first sub-network module is a×b, where a and b are both positive integers greater than or equal to 1, and b>a; the third sub-network module The convolution kernel of the convolution layer in the network module is m×m, where m is a positive integer greater than or equal to 1.

In an embodiment, the input end of the first sub-network module is short-circuited with the output end of the first sub-network module, and the input end of the third sub-network module is connected to the output end of the third sub-network module Short.

In an embodiment, the second sub-network module transforms the processed projection data according to the following formula:

Among them, f(x, y) is the CT domain image output by the second sub-network module, x and y are the abscissa and ordinate in the CT domain image, respectively, and p(r, θ) is the input to the second sub-network The processed projection data of the module, r is the distance between the processed projection data and the origin, and θ is the projection angle.

In one embodiment, the device further includes: an initial imaging model establishment module configured to establish an initial imaging model; a reconstructed image generation module configured to input noise-containing sample projection data into the initial imaging model to obtain a reconstructed image; loss The function determining module is configured to determine a loss function according to the reconstructed image and the standard dose CT image corresponding to the sample projection data containing noise; the imaging model determining module is configured to adjust the initial imaging model according to the loss function Of the network parameters to generate the imaging model.

In an embodiment, the device further includes: a standard projection data acquisition module configured to acquire a standard dose CT image before inputting noise-containing sample projection data into the initial imaging model, and compare the standard dose CT image Preprocessing is performed to obtain standard projection data; the sample projection data acquisition module is configured to add noise data to the standard projection data to generate sample projection data containing noise.

In an embodiment, the sample projection data acquisition module is configured to: set a Poisson function of at least two noise levels; add noise to the standard projection data based on the Poisson function of the at least two noise levels to generate At least two noise-containing sample projection data corresponding to the CT image of the standard dose.

The CT imaging device provided in the embodiment of the present application can execute the CT imaging method provided in any embodiment of the present application, and has the corresponding functional modules and beneficial effects for executing the CT imaging method.

Example four

6 is a schematic structural diagram of a medical imaging system provided in the fourth embodiment of the present application. FIG. 6 shows a block diagram of an exemplary medical imaging system suitable for implementing the embodiments of the present application. The medical imaging system shown in FIG. 6 is only It is an example, and should not bring any limitation to the function and scope of use of the embodiments of the present application.

The medical imaging system includes a medical imaging device 500 and a computer 600.

The computer 600 may be configured to implement specific methods and devices disclosed in some embodiments of the present application. The specific device in this embodiment uses a functional block diagram to show a hardware platform including a display module. In some embodiments, the computer 600 may implement some embodiments of the present application through hardware devices, software programs, firmware, and combinations thereof. In some embodiments, the computer 600 may be a general purpose computer or a special purpose computer.

As shown in FIG. 6, the computer 600 may include an internal communication bus 601, a processor (processor) 602, a read-only memory (Read-Only Memory, ROM) 603, a random access memory (Random Access Memory, RAM) 604, and a communication port 605, input/output component 606, hard disk 607, and user interface 608. The internal communication bus 601 can implement data communication between the components of the computer 600. The processor 602 can make a judgment and issue a prompt. In some embodiments, the processor 602 may be composed of one or more processors. The communication port 605 can implement data communication between the computer 600 and other components (not shown in the figure), such as external devices, image acquisition devices, databases, external storage, and image processing workstations. In some embodiments, the computer 600 can send and receive information and data from the network through the communication port 605. The input/output component 606 supports the input/output data flow between the computer 600 and other components. The user interface 608 can implement interaction and information exchange between the computer 600 and the user. The computer 600 may also include different forms of program storage units and data storage units, such as hard disk 607, ROM603, RAM604, capable of storing various data files used for computer processing and/or communication, and possible program instructions executed by the processor 602 .

The processor 602 can be configured to execute a CT imaging method when executing a program, the method includes: acquiring projection data of a target object, wherein the projection data is based on a preset dose of X-ray photons on the target object It is acquired during detection, the preset dose is less than the CT standard dose; the projection data is sent to a pre-trained imaging model, and the output image of the imaging model is determined as the CT image of the target object, wherein The imaging model is obtained by training standard dose CT images and sample projection data containing noise. The imaging model includes a first sub-network module, a second sub-network module, and a third sub-network module. The first sub-network module uses To perform noise reduction processing on projection data in the projection domain, the second sub-network module is used to transform the processed projection data to generate CT domain data, and the third sub-network module is used to perform processing on the CT domain data. Perform noise reduction processing to generate a CT image of the target object.

This application uses specific words to describe the embodiments of this application. For example, "one embodiment", "an embodiment", and/or "some embodiments" mean a feature, structure, or characteristic related to at least one embodiment of the present application. Therefore, "an embodiment" or "an embodiment" or "an alternative embodiment" mentioned twice or more in different positions in this specification do not necessarily refer to the same embodiment. In addition, some features, structures, or characteristics in one or more embodiments of the present application can be appropriately combined.

One or more aspects of this application can be illustrated and described in multiple categories or situations, including any new and useful process, machine, product, or combination of substances, or any new and useful improvement to them. Correspondingly, one or more aspects of the present application may be completely executed by hardware, may be completely executed by software (including firmware, resident software, microcode, etc.), or may be executed by a combination of hardware and software. The above hardware or software can be called "data block", "module", "sub-network module", "engine", "unit", "sub-unit", "component" or "system". In addition, one or more aspects of the present application may be embodied as a computer product located in one or more computer-readable media, and the product includes computer-readable program codes.

Example five

The fifth embodiment of the present application provides a computer-readable storage medium that stores a computer program that, when executed by a processor, implements the CT imaging method as provided in all application embodiments of this application, and the method includes: acquiring a target object Projection data, wherein the projection data is collected when the target object is detected based on a preset dose of X-ray photons, the preset dose is less than the CT standard dose; the projection data is sent to the pre-trained imaging Model, the output image of the imaging model is determined as the CT image of the target object, wherein the imaging model is obtained by training a CT image of a standard dose and sample projection data containing noise, and the imaging model includes a first sub A network module, a second sub-network module, and a third sub-network module, the first sub-network module is used to perform noise reduction processing on projection data in the projection domain, and the second sub-network module is used to perform processing on the processed projection data Transformation is performed to generate CT domain data, and the third sub-network module is used to perform noise reduction processing on the CT domain data to generate a CT image of the target object.

The computer-readable signal medium may include a propagated data signal containing a computer program code, for example on a baseband or as part of a carrier wave. The propagated signal may have multiple manifestations, including electromagnetic forms, optical forms, etc., or suitable combinations. The computer-readable signal medium may be any computer-readable medium other than a computer-readable storage medium, and the medium may be connected to an instruction execution system, apparatus, or device to realize communication, propagation, or transmission of a program for use. The program code located on the computer-readable signal medium can be propagated through any suitable medium, including radio, cable, fiber optic cable, radio frequency signal, or similar medium, or any combination of the foregoing medium.

The computer program codes required for the operation of one or more parts of this application can be written in any one or more programming languages, including object-oriented programming languages such as Java, Scala, Smalltalk, Eiffel, JADE, Emerald, C++, C#, VB. NET, Python, etc., conventional programming languages such as C language, Visual Basic, Fortran 2003, Perl, COBOL 2002, PHP, ABAP, dynamic programming languages such as Python, Ruby and Groovy, or other programming languages. The program code can be run entirely on the user's computer, or run as an independent software package on the user's computer, or partly run on the user's computer and partly run on a remote computer, or run entirely on the remote computer or server. In the latter case, the remote computer can be connected to the user's computer through any network, such as a local area network (LAN) or a wide area network (WAN), or connected to an external computer (for example, via the Internet), or In the cloud computing environment, or used as a service such as Software-as-a-Service (SaaS).

Unless stated in the claims, the order of processing elements and sequences, the use of numbers and letters, or the use of other names in this application are not used to limit the order of the procedures and methods of this application. Although some application embodiments are discussed in the foregoing disclosure through various examples, such details are only for illustrative purposes. For example, although the system components described above can be implemented by hardware devices, they can also be implemented only by software solutions, such as installing the described system on a server or mobile device.

In order to simplify the expressions disclosed in this application, and to help the understanding of one or more application embodiments, in the description of the embodiments of this application, sometimes multiple features are combined into one embodiment, drawings, or the multiple features. In the description. In some embodiments, numbers describing the quantities of ingredients and attributes are used. Such numbers used in the description of the embodiments are modified by the modifiers "about", "approximately" or "substantially" in some examples. Unless otherwise stated, "approximately", "approximately" or "substantially" indicates that the number is allowed to vary by ±20%. Correspondingly, in some embodiments, the numerical parameters used are approximate values, and the approximate values can be changed according to the required characteristics of individual embodiments. In some embodiments, the numerical parameter should consider the prescribed effective digits and adopt the method of general digit retention. Although the numerical ranges and parameters used to confirm the breadth of the range in some embodiments of the present application are approximate values, in specific embodiments, the setting of such numerical values is as accurate as possible within the feasible range.

Claims

An electronic computed tomography CT imaging method, including:

Acquiring projection data of the target object, where the projection data is collected when detecting the target object based on a preset dose of X-ray photons, and the preset dose is less than the CT standard dose;

The projection data is sent to a pre-trained imaging model, and the output image of the imaging model is determined as the CT image of the target object, wherein the imaging model is composed of a standard dose CT image and sample projection data containing noise After training, the imaging model includes a first sub-network module, a second sub-network module, and a third sub-network module. The first sub-network module is used to perform noise reduction processing on projection data in the projection domain, and the second The sub-network module is used to transform the processed projection data to generate CT domain data, and the third sub-network module is used to perform noise reduction processing on the CT domain data to generate a CT image of the target object.
The method according to claim 1, wherein the first sub-network module comprises a first preset number of convolution modules connected in sequence, and each convolution module of the first preset number of convolution modules includes Convolutional layer and activation function layer;

The third sub-network module includes a second preset number of convolution modules connected in sequence, and each convolution module in the second preset number of convolution modules includes a convolution layer and an activation function layer.
The method according to claim 2, wherein the convolution kernel of the convolution layer in the first sub-network module is a×b, wherein a and b are both positive integers greater than or equal to 1, and b>a ；

The convolution kernel of the convolution layer in the third sub-network module is m×m, where m is a positive integer greater than or equal to 1.
The method according to any one of claims 1 to 3, wherein the input end of the first sub-network module is short-circuited with the output end of the first sub-network module, and the input end of the third sub-network module is Short-circuit with the output terminal of the third sub-network module.
The method according to any one of claims 1 to 4, wherein the second sub-network module transforms the processed projection data according to the following formula:

Wherein, f(x, y) is the CT domain image output by the second sub-network module, x and y are the abscissa and ordinate in the CT domain image, respectively, and p(r, θ) is the input For the processed projection data of the second sub-network module, r is the distance between the processed projection data and the origin, δ is the projection X-ray, and θ is the projection angle.
The method according to any one of claims 1 to 5, wherein the training method of the imaging model comprises:

Establish an initial imaging model;

Input the noise-containing sample projection data into the initial imaging model to obtain a reconstructed image;

A loss function is determined according to the reconstructed image and the CT image of the standard dose corresponding to the sample projection data containing noise, and the network parameters in the initial imaging model are adjusted according to the loss function to generate the imaging model.
The method according to claim 6, before said inputting said noise-containing sample projection data into said initial imaging model, further comprising:

Acquiring the CT image of the standard dose, and preprocessing the CT image of the standard dose to obtain standard projection data;

Noise data is added to the standard projection data to generate the noise-containing sample projection data.
8. The method according to claim 7, wherein said adding noise to said standard projection data to generate said noise-containing sample projection data comprises:

Set the Poisson function of at least two noise levels;

Noise is added to the standard projection data based on the Poisson function of the at least two noise levels, and at least two noise-containing sample projection data corresponding to the CT image of the standard dose are generated.
An electronic computed tomography CT imaging device, including:

A projection data acquisition module configured to acquire projection data of a target object, wherein the projection data is acquired when the target object is detected based on a preset dose of X-ray photons, and the preset dose is less than the CT standard dose;

The CT image reconstruction module is configured to send the projection data to a pre-trained imaging model, and determine the output image of the imaging model as the CT image of the target object, wherein the imaging model consists of a standard dose of CT image And the sample projection data containing noise is obtained by training, the imaging model includes a first sub-network module, a second sub-network module and a third sub-network module, the first sub-network module is used to reduce the projection data of the projection domain Noise processing, the second sub-network module is used to transform the processed projection data to generate CT domain data, and the third sub-network module is used to perform noise reduction processing on the CT domain data to generate the target CT image of the subject.
A computer-readable storage medium storing a computer program, which when executed by a processor, realizes the computerized tomography CT imaging method according to any one of claims 1-8.
A medical imaging system includes a medical imaging device and a computer device. The computer device includes a memory, at least one processor, and a computer program stored in the memory and running on the at least one processor. When a processor executes the program, it can be set to execute the computerized tomography CT imaging method of any one of claims 1-8.