Pulmonary artery CT angiography image generation method and device, equipment and medium

Abstract

The invention discloses a pulmonary artery CT angiography image generation method and device, equipment and a medium, and relates to the technical field of image processing, and the method comprises the steps: obtaining a chest CT plain scanning image of a target object; inputting the chest CT plain scanning image of the target object into the image generation model to obtain a pulmonary artery CT angiography image of the target object; wherein the image generation model is constructed based on a deep learning method, the pulmonary artery CT angiography image is generated by adopting the chest CT plain scanning image, an iodine-containing contrast agent does not need to be used, the cost is low, safety is achieved, and the application range is wide.

Description

Technical Field

[0001] The present application relates to the field of image processing technology, and in particular to a method, device, equipment and medium for generating pulmonary artery CT angiography images. Background Art

[0002] Computed Tomography Pulmonary Angiography (CTPA) is a non-invasive examination technique for imaging the pulmonary artery system. It is of great value in diagnosing pulmonary thromboembolism (PTE). In a CTPA scan, iodinated contrast agent (ICA) is first injected intravenously by a high-pressure syringe, and then the pulmonary artery is scanned rapidly and continuously using spiral CT. The morphology and location of the thrombus in the pulmonary artery and the degree of vascular blockage can be clearly observed. CTPA scanning has high sensitivity and specificity. It is a non-invasive and convenient examination method and has become the preferred examination method for diagnosing PTE.

[0003] The method of using CTPA scanning to generate CTPA images has the following limitations and disadvantages: first, CTPA scanning requires the use of iodine-containing contrast agents to enhance image clarity, which will inevitably increase the economic burden of some patients; and some patients may have allergic reactions to iodine contrast agents, which can be severe and life-threatening, and are therefore contraindicated for the examination; some patients with renal insufficiency and hyperthyroidism will also face many risks after the injection of iodine contrast agents, and iodine-containing contrast agents should be used with caution; in addition, the radiation exposure dose of CTPA scanning is much higher than that of ordinary chest CT plain scan examinations. Summary of the invention

[0004] The purpose of the present application is to provide a method, device, equipment and medium for generating pulmonary artery CT angiography images, which use chest CT plain scan images to generate CTPA images without the use of iodine-containing contrast agents, and are low-cost, safe and have a wide range of applications.

[0005] To achieve the above objectives, this application provides the following solutions.

[0006] In a first aspect, the present application provides a method for generating a pulmonary artery CT angiography image, comprising: obtaining a chest CT plain scan image of a target object; inputting the chest CT plain scan image of the target object into an image generation model to obtain a pulmonary artery CT angiography image of the target object; wherein the image generation model is constructed based on a deep learning method.

[0007] In a second aspect, the present application provides a pulmonary artery CT angiography image generating device, comprising: an image acquisition module, used to acquire a chest CT plain scan image of a target object; an image generation module, used to input the chest CT plain scan image of the target object into an image generation model to obtain a pulmonary artery CT angiography image of the target object; wherein the image generation model is constructed based on a deep learning method.

[0008] 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 any of the above-described methods for generating pulmonary artery CT angiography images.

[0009] In a fourth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the above-described methods for generating pulmonary artery CT angiography images.

[0010] According to the specific embodiments provided by the present application, the present application has the following technical effects: the present application provides a method, device, equipment and medium for generating pulmonary artery CT angiography images, constructs an image generation model based on a deep learning method, and uses the image generation model to generate a pulmonary artery CT angiography image of the target object based on a chest CT plain scan image of the target object. The present application directly uses a chest CT plain scan image based on a deep learning method to generate a pulmonary artery CT angiography image, without the use of iodine-containing contrast agents, and is low-cost, safe and has a wide range of applications. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0012] Figure 1 This is a diagram of the application environment of a method for generating a pulmonary artery CT angiography image in one embodiment of the present application.

[0013] Figure 2 A flowchart of a method for generating a pulmonary artery CT angiography image provided in one embodiment of the present application.

[0014] Figure 3 A flowchart of a method for determining an image generation model provided in one embodiment of the present application.

[0015] Figure 4A schematic diagram of the structure of a deep learning network model provided in one embodiment of the present application.

[0016] Figure 5 A schematic diagram of the structure of a generator provided in one embodiment of the present application.

[0017] Figure 6 A schematic diagram of the structure of a register provided in one embodiment of the present application.

[0018] Figure 7 A schematic diagram of the structure of a discriminator provided in one embodiment of the present application.

[0019] Figure 8 A schematic diagram of functional modules of a pulmonary artery CT angiography image generating device provided in another embodiment of the present application.

[0020] Fig. 9 A schematic diagram of the structure of a computer device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0021] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0022] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below with reference to the accompanying drawings and specific implementation methods.

[0023] In this application, all actions to obtain signals, information or data are carried out in compliance with the relevant data protection laws and policies of the country where they are located and with the authorization given by the owner of the corresponding device.

[0024] Chest CT scan is one of the most common routine chest examinations in clinical practice. This technology uses the natural differences in the human body's internal structure to display contrast. It can clearly show the status of multiple key organs such as the lungs, trachea, mediastinum, pleura, ribs and thoracic vertebrae, but it still cannot distinguish lesions such as blood vessel walls and vascular cavities. If pulmonary embolism is suspected, it is necessary to clearly display the lesions in the blood vessels. The most common solution is to perform pulmonary artery CT angiography with high-pressure injection of iodine-containing contrast agent. In particular, if the patient is suspected of pulmonary embolism in the emergency department at night, the success of pulmonary artery CT angiography depends on the hospital's equipment, the experience of the operating technician, the patient's cooperation, etc. Therefore, it is crucial to develop a solution that can clearly display vascular lesions in a non-invasive, simple, accurate and stable manner.

[0025] In order to solve the problem of the need for additional use of iodine-containing contrast agents and CT radiation in the current diagnostic process, combined with the needs of clinical practical work, the embodiment of the present application provides a method for generating pulmonary artery CT angiography images, which generates virtual pulmonary artery CT angiography images only with the help of chest CT plain scan images and deep learning methods, and can clearly outline anatomical structures such as pulmonary blood vessels. This solution has a wide range of applicability, and the most important thing is that it does not require the use of any iodine-containing contrast agents, reducing the use of iodine-containing contrast agents and potential risks to patients. In addition, patients do not need to undergo additional CT scans, which also reduces the risk of patients receiving ionizing radiation.

[0026] The method for generating a pulmonary artery CT angiography image provided in the embodiment of the present application can be applied to Figure 1 In the application environment shown. Among them, the terminal 102 communicates with the server 104 through the network. The data storage system can store the data that the server 104 needs to process. The data storage system can be set up separately, integrated on the server 104, or placed on the cloud or other servers. The terminal 102 can send the chest CT plain scan image of the target object to the server 104. After the server 104 receives the chest CT plain scan image of the target object, for the chest CT plain scan image of the target object, the server 104 obtains the chest CT plain scan image of the target object; the chest CT plain scan image of the target object is input into the image generation model to obtain the pulmonary artery CT angiography image (CTPA image) of the target object, wherein the image generation model is constructed based on a deep learning method. The server 104 can feed back the obtained pulmonary artery CT angiography image of the target object to the terminal 102. In addition, in some embodiments, the pulmonary artery CT angiography image generation method can also be implemented independently by the server 104 or the terminal 102. For example, the terminal 102 can directly process the chest CT plain scan image of the target object, or the server 104 can obtain the chest CT plain scan image of the target object from the data storage system and process the chest CT plain scan image of the target object.

[0027] The terminal 102 may be, but is not limited to, various desktop computers, laptop computers, smart phones, tablet computers, IoT devices, and portable wearable devices. The IoT devices may be smart speakers, smart TVs, smart air conditioners, smart vehicle-mounted devices, etc. The portable wearable devices may be smart watches, smart bracelets, head-mounted devices, etc. The server 104 may be implemented as an independent server or a server cluster consisting of multiple servers, or may be a cloud server.

[0028] In an exemplary embodiment, Figure 2As shown, a method for generating a pulmonary artery CT angiography image is provided. The method is executed by a computer device, and specifically can be executed by a computer device such as a terminal or a server alone, or can be executed by a terminal and a server together. In the embodiment of the present application, the method is applied to Figure 1 The server 104 in the example is used for explanation, and the steps include the following steps 201 to 202.

[0029] Step 201: Obtain a chest CT plain scan image of a target object.

[0030] Step 202: Input the chest CT plain scan image of the target object into the image generation model to obtain the pulmonary artery CT angiography image of the target object.

[0031] Among them, the image generation model is constructed based on the deep learning method.

[0032] By implementing the above-mentioned steps 201 to 202, chest CT plain scan images are used to generate pulmonary artery CT angiography images, without the need to use iodine-containing contrast agents, thereby reducing the economic burden on patients, having low costs, reducing potential risks to patients, being safe and having a wide range of applications, and improving diagnostic accuracy and work efficiency, thus bringing significant advantages to the field of medical imaging.

[0033] In another exemplary embodiment of the present application, see Figure 3 The method for determining the image generation model used in step 202 includes the following steps 301 to 303.

[0034] Step 301, obtaining training data.

[0035] The training data includes: chest CT plain scan images of the training subject and real pulmonary artery CT angiography images (real CTPA images) of the training subject.

[0036] Step 302, construct a deep learning network model.

[0037] The deep learning network model includes a generator, a aligner and a discriminator; the generator and the aligner are both connected to the discriminator. The structure of the deep learning network model is as follows: Figure 4 shown.

[0038] Step 303, using the chest CT plain scan image in the training data as the input of the generator, using the training data as the input of the aligner, training is performed with the goal of minimizing the total loss function, and using the trained generator as the image generation model.

[0039] Among them, please see Figure 4, the generator is used to perform feature encoding, deep feature extraction and decoding processing on the chest CT plain scan image in the training data to generate the pulmonary artery CT angiography image of the training object. The aligner is used to perform joint feature encoding, deep feature extraction and decoding processing on the chest CT plain scan image of the training object and the real pulmonary artery CT angiography image of the training object to obtain the correlation matrix between the chest CT plain scan image of the training object and the real pulmonary artery CT angiography image of the training object, and obtain the corrected pulmonary artery CT angiography image based on the correlation matrix and the real pulmonary artery CT angiography image of the training object; the correlation matrix is ​​used to describe the difference between the chest CT plain scan image and the real pulmonary artery CT angiography image. The discriminator is used to output a discrimination result based on the pulmonary artery CT angiography image generated by the generator and the corrected pulmonary artery CT angiography image obtained by the aligner.

[0040] The total loss function is determined based on the pulmonary artery CT angiography image generated by the generator, the corrected pulmonary artery CT angiography image output by the aligner, and the discrimination result output by the discriminator.

[0041] In another exemplary embodiment of the present application, see Figure 5 , the generator includes: a first encoding module, a first deep feature extraction module and a first decoding module connected in sequence. The first encoding module is used to perform down-sampling feature encoding on the chest CT plain scan image in the training data to obtain a first encoding result representing the characteristics of the chest CT plain scan image, and the first encoding result is used as a shallow feature map sequence extracted by the generator. The first deep feature extraction module is used to perform deep feature extraction on the first encoding result to obtain a first feature extraction result, and the first feature extraction result is used as a deep feature map sequence extracted by the generator. The first decoding module is used to decode the first feature extraction result to generate a pulmonary artery CT angiography image of the training object.

[0042] See also Figure 6, the aligner includes: a second encoding module, a second deep feature extraction module and a second decoding module connected in sequence. The second encoding module is used to perform down-sampling joint feature encoding on the chest CT plain scan image of the training object and the real pulmonary artery CT angiography image of the training object, and obtain a joint encoding result representing the correlation between the chest CT plain scan image and the real pulmonary artery CT angiography image, and the joint encoding result is used as a shallow joint feature map sequence extracted by the aligner. The second deep feature extraction module is used to perform deep feature extraction on the joint encoding results of the two types of images to obtain a second feature extraction result, and the second feature extraction result is used as a deep joint feature map sequence extracted by the aligner. The second decoding module is used to decode the second feature extraction result to obtain a correlation matrix between the chest CT plain scan image of the training object and the real pulmonary artery CT angiography image of the training object; the correlation matrix is ​​pixel-by-pixel multiplied with the input real pulmonary artery CT angiography image of the training object, and the real pulmonary artery CT angiography image is spatially registered with the chest CT plain scan image through the multiplication operation to obtain a corrected pulmonary artery CT angiography image.

[0043] See also Figure 7 , the discriminator includes a third encoding module; the third encoding module is used to perform feature encoding on the pulmonary artery CT angiography image generated by the generator and the corrected pulmonary artery CT angiography image obtained by the aligner to obtain a second encoding result, and output a discrimination result according to the second encoding result to guide model training. Specifically, the third encoding module performs feature encoding on the pulmonary artery CT angiography image generated by the generator and the corrected pulmonary artery CT angiography image obtained by the aligner, respectively, to obtain a first encoding feature map representing the generated pulmonary artery CT angiography image and a second encoding feature map representing the corrected pulmonary artery CT angiography image; the second encoding result includes a first encoding feature map and a second encoding feature map.

[0044] In another exemplary embodiment of the present application, please refer to Figure 6 , the first encoding module, the second encoding module and the third encoding module have the same structure; the first depth feature extraction module and the second depth feature extraction module have the same structure; the first decoding module and the second decoding module have the same structure.

[0045] The first encoding module includes: a padding layer, a convolution layer, a normalization layer and an activation layer connected in sequence. The padding layer is used to align the edge of the convolution kernel with the input image to achieve data dimension alignment; the convolution layer is used to extract image features; the normalization layer is used to improve the convergence speed of the model; and the activation layer is used to perform nonlinear transformation.

[0046] The first deep feature extraction module includes: a first basic feature extraction structure and a second basic feature extraction structure connected in sequence. The output of the second basic feature extraction structure is also connected to the input of the first basic feature extraction structure to form a residual feature extraction structure to extract deep features; the output of the second basic feature extraction structure is added to the input of the first basic feature extraction structure as the output of the first deep feature extraction module. The first basic feature extraction structure and the second basic feature extraction structure have the same structure; the first basic feature extraction structure includes: a padding layer, a convolution layer, a normalization layer, and an activation layer connected in sequence.

[0047] The first decoding module includes: an upsampling layer, a normalization layer and an output layer connected in sequence. The upsampling layer is used to increase the size of the feature map; the normalization layer is used to increase the convergence speed of the model; and the output layer is used to perform nonlinear transformation to achieve image reconstruction.

[0048] The structures of the second encoding module, the third encoding module, the second depth feature extraction module and the second decoding module are not described in detail here. Figure 6 and Figure 7 .

[0049] In another exemplary embodiment of the present application, the expression for correcting the pulmonary artery CT angiography image is as follows.

[0050] .

[0051] in, represents the corrected pulmonary artery CT angiography image; Represents the chest CT plain scan image in the training data; represents a real pulmonary artery CT angiography image of a training subject; express and The correlation matrix of Represents dot product.

[0052] In another exemplary embodiment of the present application, please refer to Figure 4 , the method for determining the total loss function comprises the following steps.

[0053] (1) Constructing a correlation loss function, a pixel loss function and a structural similarity loss function according to the pulmonary artery CT angiography image generated by the generator and the corrected pulmonary artery CT angiography image output by the aligner.

[0054] Among them, the expression of the correlation loss function is as follows.

[0055] .

[0056] L corr represents the correlation loss function; represents a pulmonary artery CT angiography image generated by the generator; express expectations; represents the corrected pulmonary artery CT angiography image; Represents the chest CT plain scan image in the training data; Represents a real pulmonary artery CT angiography image of a training subject.

[0057] The expression of the pixel loss function is as follows.

[0058] .

[0059] L pixel represents the pixel loss function; represents the mean square error.

[0060] The expression of the structural similarity loss function is as follows.

[0061] .

[0062] L ssim represents the structural similarity loss function; express The mean of express The mean of express The variance of express The variance of express and The covariance between and are constants to avoid the denominator being zero.

[0063] (2) Constructing an adversarial loss function based on the discrimination result output by the discriminator.

[0064] Among them, the expression of the adversarial loss function is as follows.

[0065] .

[0066] .

[0067] .

[0068] L adv represents the adversarial loss function; L G represents the generator loss function; L Drepresents the discriminator loss function; Represents the discriminator for The judgment result of Represents the discriminator for The discrimination result of .

[0069] (3) Constructing the total loss function according to the correlation loss function, the pixel loss function, the structural similarity loss function and the adversarial loss function.

[0070] Among them, the expression of the total loss function is as follows.

[0071] .

[0072] L represents the total loss function; Indicates L corr The weight coefficient of Indicates L pixel The weight coefficient of Indicates L ssim The weight coefficient of Indicates L adv The weight coefficient of .

[0073] The pulmonary artery CT angiography image generation method of the above embodiment enables the subject or patient to obtain a virtual pulmonary artery CT angiography image by relying only on chest CT plain scan images without using iodine-containing contrast agents, and can clearly outline anatomical structures such as pulmonary blood vessels.

[0074] The method for generating pulmonary artery CT angiography images in the above-mentioned embodiment has wide applicability, is truly non-invasive, simple and easy to implement, and is also applicable to patients who are allergic to iodine-containing contrast agents and other patients who are not suitable for the use of contrast agents.

[0075] The pulmonary artery CT angiography image generation method of the above embodiment uses chest CT plain scan data to generate a virtual pulmonary artery CT angiography image, so that the patient does not need to undergo additional scanning and the risk of the patient being exposed to ionizing radiation is greatly reduced.

[0076] The method for generating pulmonary artery CT angiography images in the above-mentioned embodiment can be implemented by using an intelligent system, thereby reducing the reliance on the experience of medical personnel operating the computer and improving the safety and effectiveness of examinations in clinical practice.

[0077] In practical applications, an implementation process of the above-mentioned pulmonary artery CT angiography image generation method can be described as follows: (1) Obtain training data. The training data includes chest CT plain scan images and real CTPA images obtained by pulmonary artery CT angiography of the same training object. (2) Input the training data into a deep learning network model for training, adjust the network parameters according to the error value of the total loss function, and obtain an image generation model for generating CTPA images. The deep learning network model includes a generator, a aligner, and a discriminator. (3) Input the chest CT plain scan image of the target object (i.e., the chest CT plain scan image to be enhanced) into the CTPA image generation model to obtain the target generated pulmonary artery CTA image.

[0078] The above method can generate CTPA images from chest CT plain scan images without using iodine-containing contrast agents, reducing the potential risks of iodine-containing contrast agents to patients and the risk of patients receiving radiation. It can be used as a non-invasive, simple, accurate, and stable method to clearly display vascular lesions and is applied in the field of pulmonary artery CT imaging.

[0079] Based on the same inventive concept, the embodiment of the present application also provides a pulmonary artery CT angiography image generation device for implementing the above-mentioned pulmonary artery CT angiography image generation method. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above-mentioned method, so the specific limitations in one or more pulmonary artery CT angiography image generation device embodiments provided below can refer to the limitations of the pulmonary artery CT angiography image generation method above, and will not be repeated here.

[0080] In an exemplary embodiment, Figure 8 As shown, a pulmonary artery CT angiography image generation device is provided, including: an image acquisition module 801, used to acquire a chest CT plain scan image of a target object. An image generation module 802, used to input the chest CT plain scan image of the target object into an image generation model to obtain a pulmonary artery CT angiography image of the target object. The image generation model is constructed based on a deep learning method.

[0081] The pulmonary artery CT angiography image generation device of this embodiment uses chest CT plain scan images combined with a deep learning method to generate pulmonary artery CT angiography images, which can improve the accuracy and efficiency of diagnosis while reducing the economic burden on patients. It brings significant clinical application advantages to the field of medical imaging and facilitates doctors to formulate more scientific and safe examination and treatment plans.

[0082] In an exemplary embodiment, a computer device is provided. The computer device may be a server or a terminal. The internal structure diagram thereof may be as follows: Fig. 9As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as 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 used 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 operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store chest CT plain scan images of the target object. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a method for generating a pulmonary artery CT angiography image is implemented.

[0083] Those skilled in the art will understand that Fig. 9 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0084] In an exemplary embodiment, a computer device is further provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above-mentioned method embodiments when executing the computer program.

[0085] In an exemplary embodiment, a computer-readable storage medium is provided, storing a computer program, and when the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0086] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, 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 embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM may be in various forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM).

[0087] The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. The non-relational database may include a distributed database based on blockchain, etc., but is not limited thereto. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, etc., but is not limited thereto.

[0088] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0089] This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the method and core ideas of this application. At the same time, for those skilled in the art, according to the ideas of this application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.

Claims

1. A method for generating a pulmonary artery CT angiography image, characterized in that: The method for generating a pulmonary artery CT angiography image comprises: Acquire a chest CT plain scan image of the target object; Inputting the chest CT plain scan image of the target object into the image generation model to obtain the pulmonary artery CT angiography image of the target object; Among them, the image generation model is constructed based on the deep learning method.

2. The method for generating pulmonary artery CT angiography images according to claim 1, characterized in that: The method for determining the image generation model comprises: Acquire training data; the training data includes: chest CT plain scan images of the training subject and real pulmonary artery CT angiography images of the training subject; Constructing a deep learning network model; the deep learning network model includes a generator, a aligner and a discriminator; the generator and the aligner are both connected to the discriminator; Using the chest CT plain scan image in the training data as the input of the generator, using the training data as the input of the aligner, training with the minimum total loss function as the goal, and using the trained generator as the image generation model; The generator is used to perform feature encoding, deep feature extraction and decoding processing on the chest CT plain scan image in the training data to generate a pulmonary artery CT angiography image of the training object; The aligner is used to perform joint feature encoding, deep feature extraction and decoding processing on the chest CT plain scan image of the training object and the real pulmonary artery CT angiography image of the training object, to obtain a correlation matrix between the chest CT plain scan image of the training object and the real pulmonary artery CT angiography image of the training object, and to obtain a corrected pulmonary artery CT angiography image based on the correlation matrix and the real pulmonary artery CT angiography image of the training object; The discriminator is used to output a discrimination result according to the pulmonary artery CT angiography image generated by the generator and the corrected pulmonary artery CT angiography image obtained by the registerer; The total loss function is determined based on the pulmonary artery CT angiography image generated by the generator, the corrected pulmonary artery CT angiography image output by the aligner, and the discrimination result output by the discriminator.

3. The method for generating pulmonary artery CT angiography images according to claim 2, characterized in that: The generator comprises: a first encoding module, a first deep feature extraction module and a first decoding module connected in sequence; the first encoding module is used to perform feature encoding on the chest CT plain scan image in the training data to obtain a first encoding result; the first deep feature extraction module is used to perform deep feature extraction on the first encoding result to obtain a first feature extraction result; the first decoding module is used to perform decoding processing on the first feature extraction result to generate a pulmonary artery CT angiography image of the training object; The aligner includes: a second encoding module, a second deep feature extraction module and a second decoding module connected in sequence; the second encoding module is used to perform joint feature encoding on the chest CT plain scan image of the training object and the real pulmonary artery CT angiography image of the training object to obtain a joint encoding result; the second deep feature extraction module is used to perform deep feature extraction on the joint encoding result to obtain a second feature extraction result; the second decoding module is used to perform decoding processing on the second feature extraction result to obtain a correlation matrix between the chest CT plain scan image of the training object and the real pulmonary artery CT angiography image of the training object; the correlation matrix is ​​multiplied pixel by pixel with the real pulmonary artery CT angiography image of the training object to obtain a corrected pulmonary artery CT angiography image; The discriminator includes a third encoding module; the third encoding module is used to perform feature encoding on the pulmonary artery CT angiography image generated by the generator and the corrected pulmonary artery CT angiography image obtained by the aligner to obtain a second encoding result, and output a discrimination result based on the second encoding result.

4. The method for generating pulmonary artery CT angiography images according to claim 3, characterized in that: The first encoding module, the second encoding module and the third encoding module have the same structure; the first depth feature extraction module and the second depth feature extraction module have the same structure; the first decoding module and the second decoding module have the same structure.

5. The method for generating pulmonary artery CT angiography images according to claim 4, characterized in that: The first encoding module includes: a padding layer, a convolution layer, a normalization layer and an activation layer connected in sequence; The first deep feature extraction module comprises: a first basic feature extraction structure and a second basic feature extraction structure connected in sequence; the output of the second basic feature extraction structure is also connected to the input of the first basic feature extraction structure; the output of the second basic feature extraction structure is added to the input of the first basic feature extraction structure as the output of the first deep feature extraction module; the first basic feature extraction structure and the second basic feature extraction structure have the same structure; the first basic feature extraction structure comprises: a padding layer, a convolution layer, a normalization layer and an activation layer connected in sequence; The first decoding module includes: an upsampling layer, a normalization layer and an output layer connected in sequence.

6. The method for generating pulmonary artery CT angiography images according to claim 2, characterized in that: The method for determining the total loss function comprises: constructing a correlation loss function, a pixel loss function and a structural similarity loss function according to the pulmonary artery CT angiography image generated by the generator and the corrected pulmonary artery CT angiography image output by the register; Among them, the expression of the correlation loss function is: ； L corr represents the correlation loss function; represents a pulmonary artery CT angiography image generated by the generator; express expectations; represents the corrected pulmonary artery CT angiography image; Represents the chest CT plain scan image in the training data; represents a real pulmonary artery CT angiography image of a training subject; The expression of the pixel loss function is: ； L pixel represents the pixel loss function; represents mean square error; The expression of the structural similarity loss function is: ； L ssim represents the structural similarity loss function; express The mean of express The mean of express The variance of express The variance of express and The covariance between and They are all constants to avoid the denominator being zero; Constructing an adversarial loss function according to the discrimination result output by the discriminator; Among them, the expression of the adversarial loss function is: ； ； ； L adv represents the adversarial loss function; L G represents the generator loss function; L D represents the discriminator loss function; Represents the discriminator for The judgment result of Represents the discriminator for The judgment result of Constructing the total loss function according to the correlation loss function, the pixel loss function, the structural similarity loss function and the adversarial loss function; Among them, the expression of the total loss function is: ； L represents the total loss function; Indicates L corr The weight coefficient of Indicates L pixel The weight coefficient of Indicates L ssim The weight coefficient of Indicates L adv The weight coefficient of .

7. The method for generating pulmonary artery CT angiography images according to claim 2, characterized in that: The expression for correcting the pulmonary artery CT angiography image is: ； in, represents the corrected pulmonary artery CT angiography image; Represents the chest CT plain scan image in the training data; represents a real pulmonary artery CT angiography image of a training subject; express and The correlation matrix of Represents dot product.

8. A pulmonary artery CT angiography image generating device, characterized in that: The pulmonary artery CT angiography image generating device comprises: An image acquisition module is used to acquire a chest CT plain scan image of a target object; An image generation module, used for inputting a chest CT plain scan image of a target object into an image generation model to obtain a pulmonary artery CT angiography image of the target object; Among them, the image generation model is constructed based on the deep learning method.

9. 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 method for generating pulmonary artery CT angiography images according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for generating a pulmonary artery CT angiography image according to any one of claims 1 to 7 is implemented.

Images