Rapid annotation methods, computer equipment, and computer program products for coronary CTA vessels

Abstract

This application relates to a rapid annotation method, computer equipment, and computer program product for coronary CTA vessels. The rapid annotation method includes: constructing a three-dimensional coronary artery model based on coronary CTA sequence images; extracting the coronary artery centerline from the three-dimensional model; extracting the centerline of a single coronary artery from the centerline; straightening and reconstructing the corresponding single vessel in the three-dimensional coronary artery model along the single coronary artery centerline to obtain a straightened single vessel model; performing cross-sectional sampling on the single vessel model according to the cutting plane passing through the single coronary artery centerline to obtain a cross-sectional image; performing image segmentation on the cross-sectional image to obtain the edge lines of the coronary vessels; and mapping the edge lines of the coronary vessels to the single vessel model to obtain a single vessel model with annotated coronary vessel boundary information. Compared to the scheme of directly segmenting the coronary vessel boundaries in the three-dimensional single vessel model, this application reduces the operational difficulty and saves computational costs.

Description

Technical Field

[0001] This application relates to the field of medical image processing technology, and in particular to a rapid annotation method for coronary CTA vessels, computer equipment, and computer program products. Background Technology

[0002] Coronary artery disease (CAD) is the leading cause of death worldwide, and its incidence continues to rise, posing a serious threat to people's health. As an effective and non-invasive method for diagnosing CAD, multi-slice spiral CT angiography (CTA) has developed rapidly. CTA image segmentation is a crucial foundation for reconstructing the three-dimensional geometric model of the coronary arteries, which provides cardiologists with a more accurate, faster, and more repeatable diagnostic method. Therefore, accurately extracting the coronary artery contour based on coronary CTA sequence images is an important clinical auxiliary analysis method for coronary artery stenosis, and can provide quantitative analysis of calcification, plaque burden, and stenosis degree. Consequently, vessel segmentation based on CTA images has become a hot research topic in the field of medical image processing.

[0003] In existing technologies, deep learning-based techniques are widely used for coronary CTA sequence image segmentation. Rapidly obtaining high-quality coronary CTA sequence image annotations is a crucial part of the segmentation task. Summary of the Invention

[0004] Therefore, it is necessary to provide a rapid annotation method for coronary CTA vessels to address the aforementioned technical problems.

[0005] The rapid annotation method for coronary CTA vessels in this application includes:

[0006] A three-dimensional model of the coronary artery is constructed based on the coronary CTA sequence image. The coronary artery centerline is extracted based on the three-dimensional model of the coronary artery. The centerline of a single coronary artery is extracted based on the coronary artery centerline.

[0007] Along the center line of the single coronary artery, straighten and reconstruct the corresponding single vessel in the three-dimensional model of the coronary artery to obtain the straightened single vessel model;

[0008] Based on the cutting plane passing through the center line of the single coronary artery, cross-sectional sampling is performed on the single blood vessel model to obtain a cross-sectional image;

[0009] Image segmentation is performed on the cross-sectional image to obtain the coronary artery edge line, and the coronary artery edge line is mapped to the single vessel model to obtain a single vessel model with coronary artery boundary information annotation.

[0010] Optionally, a single coronary artery centerline is extracted based on the coronary artery centerline, specifically including: the vein of the coronary artery centerline has a coronary artery origin and a number of corresponding coronary artery endpoints, each of the coronary artery endpoints and the corresponding coronary artery origin forms a single coronary artery centerline, and a single coronary artery centerline is extracted from the vein of the coronary artery centerline.

[0011] Optionally, the single blood vessel model is cylindrical, and the centerline of the single coronary artery is located at the axis of the cylinder.

[0012] Optionally, the corresponding single vessel in the reconstructed coronary artery 3D model is straightened to obtain a straightened single vessel model, specifically including:

[0013] Straighten and reconstruct the corresponding single blood vessel in the three-dimensional model of the coronary artery to obtain a straightened cubic single blood vessel model;

[0014] For a cubic single blood vessel model, a cylindrical single blood vessel model is obtained by cutting along the center line of the single coronary artery.

[0015] Optionally, the cubic single blood vessel model has a length, width, and height of X*X*H, where X and H represent X and H unit lengths, respectively.

[0016] The cylindrical single blood vessel model has a diameter and height of X*H, where X and H represent X and H unit lengths, respectively.

[0017] Optionally, the intercepting plane includes a plurality of planes evenly distributed at circumferential intervals, and the rapid annotation method includes:

[0018] Based on the cutting plane passing through the center line of the single coronary artery, cross-sectional sampling is performed on the single blood vessel model to obtain a cross-sectional image;

[0019] After obtaining the cross-sectional image, the cross-sectional image is uniformly processed so that the center line of the single coronary artery in the cross-sectional image is in the center of the cross-sectional image, and the specifications of each cross-sectional image are the same.

[0020] Optionally, image segmentation is performed on the cross-sectional image to obtain the coronary artery edge line, specifically including:

[0021] The cross-sectional image is coarsely segmented to obtain a coarse segmentation result;

[0022] The coarse segmentation result is then finely segmented using a trained Unet neural network to obtain the coronary artery edge line.

[0023] Optionally, the edge line of the coronary artery is mapped to the single vessel model to obtain a single vessel model with annotated coronary artery boundary information, specifically including:

[0024] The edge lines of the coronary arteries are mapped onto the single vessel model to obtain multiple edge lines of the coronary arteries on the single vessel model.

[0025] Based on the connected domain formed by connecting the edge lines of the multiple coronary arteries, the coronary artery contour on the single vessel model is obtained, thereby obtaining a single vessel model with annotated coronary artery boundary information.

[0026] This application also provides a computer device, including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the rapid labeling method for coronary CTA vessels described in this application.

[0027] This application also provides a computer program product, including computer instructions that, when executed by a processor, implement the steps of the rapid annotation method for coronary CTA vessels described in this application.

[0028] The rapid annotation method for coronary CTA vessels proposed in this application has at least the following effects:

[0029] This application obtains cross-sectional images by dimensionality reduction of a single blood vessel model. Different cross-sectional images can be used as samples to train a network model that performs image segmentation, thus alleviating the problem of insufficient data.

[0030] Compared to the approach of directly segmenting the coronary artery boundary in a 3D single-vessel model, this application reduces the operational difficulty and saves computational costs. Attached Figure Description

[0031] Figure 1 This is a flowchart illustrating a rapid annotation method for coronary CTA vessels in one embodiment of this application;

[0032] Figure 2 This is a flowchart of a rapid annotation method for coronary CTA vessels in one embodiment of this application;

[0033] Figure 3 for Figure 1 A schematic diagram of the straightened single blood vessel model (three-dimensional straightening CPR) in step S200;

[0034] Figure 4 for Figure 1 A schematic diagram of the cross-sectional image (two-dimensional CPR) in step S300;

[0035] Figure 5 for Figure 1 A schematic diagram of the coronary artery edge line (two-dimensional CPR segmentation result) in step S400;

[0036] Figure 6 for Figure 1 A schematic diagram of a single vessel model (3D straightened CPR segmentation result) with coronary vessel boundary information annotated in step S500;

[0037] Figure 7 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0038] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0039] See Figure 1 , Figures 3-6 One embodiment of this application provides a rapid annotation method for coronary CTA vessels, comprising:

[0040] Step S100: Construct a three-dimensional model of the coronary artery based on the coronary CTA sequence image, extract the coronary artery centerline based on the three-dimensional model of the coronary artery, and extract the centerline of a single coronary artery based on the coronary artery centerline.

[0041] Step S200: Straighten the corresponding single blood vessel in the three-dimensional model of the reconstructed coronary artery along the center line of the single coronary artery to obtain the straightened single blood vessel model.

[0042] Step S300: Based on the interception plane passing through the center line of a single coronary artery, perform cross-sectional sampling on the single vessel model to obtain a cross-sectional image;

[0043] Step S400: Perform image segmentation on the cross-sectional image to obtain the edge lines of the coronary arteries;

[0044] Step S500: Map the edge line of the coronary artery to the single vessel model to obtain the single vessel model with the boundary information of the coronary artery.

[0045] In this embodiment, a single coronary artery centerline refers to a single coronary artery centerline that is continuous from beginning to end and has no branching. Since the coronary artery centerlines are generally in a branching network state, a single coronary artery centerline can be obtained by traversing the network.

[0046] The straightening and reconstruction procedure, or CPR (Curved Planar Reformation), is based on the single coronary artery model obtained by straightening and reconstructing the centerline of a single coronary artery. Therefore, it involves cutting out the centerline of a single coronary artery that has a straight segment in the plane.

[0047] In this embodiment, the intercepting planes include multiple planes distributed circumferentially, such as those evenly spaced circumferentially. Correspondingly, multiple intercepted images are obtained. In this embodiment, cross-sectional images are obtained by dimensionality reduction of the single vessel model. Different cross-sectional images can be used as samples to train the network model performing image segmentation, improving the problem of insufficient data. Compared to the approach of directly segmenting the coronary artery boundary in a three-dimensional single vessel model, this embodiment reduces operational difficulty and saves computational resources.

[0048] See Figures 1-6 This application provides a method for rapid annotation of coronary CTA vessels in one embodiment, including steps (1) to (5), which are used to explain and illustrate steps S100 to S500 respectively. The method includes: (1) generating a centerline, i.e., obtaining a centerline from the CTA image; (2) generating a straightened CPR, i.e., generating a straightened three-dimensional CPR along the centerline direction; (3) generating a multi-angle two-dimensional CPR, i.e., generating a two-dimensional CPR by straightening the three-dimensional CPR; (4) segmenting the two-dimensional CPR, i.e., segmenting the two-dimensional CPR vessel boundary using a Unet network; and (5) converting the two-dimensional segmentation result into a three-dimensional segmentation result, i.e., converting the two-dimensional CPR vessel boundary into a three-dimensional CPR vessel boundary.

[0049] (1) Generate centerline, corresponding to step 100. This includes merging multi-slice DICOM format CTA images to construct a three-dimensional coronary artery model, inputting the three-dimensional coronary artery model into a pre-trained centerline extraction module, and obtaining the centerline of the coronary artery.

[0050] In step S100, a single coronary artery centerline is extracted based on the coronary artery centerline. Specifically, the coronary artery centerline has a coronary artery origin (left coronary artery origin and right coronary artery origin) and multiple corresponding coronary artery endpoints. Each coronary artery endpoint and its corresponding coronary artery origin form a single coronary artery centerline. The single coronary artery centerline is extracted from the coronary artery centerline.

[0051] (2) Generate straightening CPR, corresponding to step 200.

[0052] Specifically, this involves obtaining a straightened planar reformation (CPR) for each individual vessel along the extension direction of the coronary vessel centerline. The straightened CPR includes the vessel outline and information about the surrounding area (such as background, tissue, etc.). The straightened CPR is spatially represented as a cube, denoted as X*X*H, where H is the length after straightening along the centerline and X*X is the size of the cross-section.

[0053] (3) Generate multi-angle two-dimensional CPR, corresponding to the explanation step 300.

[0054] For a straightened CPR, in its X*X cross-section, with the center point on the centerline as the origin, generate a circle with a diameter of X. Divide the cross-section into n equal parts along the diameter of the circle. Divide all cross-sections into n equal parts along the direction after straightening the centerline, resulting in n two-dimensional CPR sheets of size X*H.

[0055] For steps S200 and S300, the following two implementation methods, which have substantially the same requirements for the cross-sectional image, can be adopted accordingly.

[0056] The first implementation involves straightening the corresponding individual blood vessels in the reconstructed three-dimensional coronary artery model to obtain a straightened individual blood vessel model (straightened CPR). Specifically, this includes: straightening the corresponding individual blood vessels in the reconstructed three-dimensional coronary artery model to obtain a straightened cubical individual blood vessel model; for the cubical individual blood vessel model, using the centerline of the individual coronary artery as an axis, cutting to obtain a cylindrical individual blood vessel model. This ensures that the individual blood vessel model is cylindrical, and the centerline of the individual coronary artery is located at the axis of the cylinder.

[0057] For a cylindrical single blood vessel model, a cross-sectional image of the same size can be directly obtained by using the cutting plane, with the centerline of the single coronary artery located at the center of the cross-sectional image. Specifically, for a cubic single blood vessel model, the dimensions of length, width, and height are X*X*H, where X and H represent X and H unit sampling lengths, respectively. For a cylindrical single blood vessel model, the dimensions of diameter and height are X*H, where X and H represent X and H unit lengths, respectively.

[0058] In the second implementation method, a cross-sectional sample is taken from a cubic single blood vessel model based on the cutting plane passing through the center line of a single coronary artery to obtain a cross-sectional image. After obtaining the cross-sectional image, the cross-sectional image is uniformly processed so that the center line of the single coronary artery in the cross-sectional image is in the center of the cross-sectional image, and the specifications of each cross-sectional image are the same.

[0059] In this embodiment, the cutting planes for obtaining the cross-sectional images all pass through the centerline of a single coronary artery, which has been straightened in step S200. This can be understood as the cutting planes comprising multiple planes formed by rotating at equal angles around the centerline of a single coronary artery.

[0060] Because different cross-sectional images all have a central region representing the blood vessel outline and an edge region representing the background or tissue, this method achieves clustered sampling of the blood vessel outline and divergent sampling of the edge regions with a limited number of cross-sectional images. Specifically, taking a cylindrical single blood vessel model as an example, the central region of the cross-sectional image is always the blood vessel outline. At the same rotation angle, the sampling arc length span of the cutting plane differs for the inner circle of the blood vessel outline and the outer circle of the background or tissue; the sampling arc length span of the inner circle of the blood vessel outline is smaller than that of the outer circle. Therefore, compared to the cutting method where the cutting plane is perpendicular to the direction of the centerline extension, this embodiment greatly reduces the requirement for the number of cutting planes, further reducing computational overhead.

[0061] (4) Two-dimensional CPR segmentation, corresponding to explanation step 400.

[0062] Step S400 involves performing image segmentation on the cross-sectional image to obtain the coronary artery edge line. Specifically, this includes: performing coarse segmentation on the cross-sectional image to obtain a coarse segmentation result; and performing fine segmentation on the coarse segmentation result using a trained Unet neural network to obtain the coronary artery edge line.

[0063] It can be understood that the coronary artery vessel edge line is the vessel contour on a two-dimensional cross-sectional image. Compared to segmenting vessels in three dimensions to obtain their contours, segmenting in two-dimensional CPR is significantly less difficult, and the maturity of segmentation algorithms for two-dimensional data is far greater than that for three-dimensional data. Specifically, the segmentation algorithm built into OpenCV can be used to perform coarse segmentation on the two-dimensional CPR, resulting in labeled dataset 1. Labeled dataset 1 is then refined using a professional annotation tool to obtain accurate vessel boundary annotations, resulting in labeled dataset 2. Using labeled dataset 2, a Unet neural network is trained to obtain a two-dimensional CPR segmentation model. This two-dimensional CPR segmentation model, when input to a cross-sectional image, outputs n images of the vessel boundaries (coronary artery vessel edge lines) in two-dimensional CPR.

[0064] (5) Convert the two-dimensional segmentation result into a three-dimensional segmentation result, corresponding to the explanation step S500.

[0065] Step S500 involves mapping the coronary artery edge lines to a single vessel model to obtain a single vessel model with annotated coronary artery boundary information. Specifically, this includes: mapping the coronary artery edge lines to a single vessel model to obtain multiple coronary artery edge lines on the single vessel model; obtaining the coronary artery contour on the single vessel model based on the connected domain formed by connecting the multiple coronary artery edge lines, thereby obtaining a single vessel model with annotated coronary artery boundary information.

[0066] Each cross-sectional image contains the segmented central region of the coronary vessels, and correspondingly, the two coronary vessel edge lines that make up the coronary vessel contour are obtained. The boundaries on both sides of the 2D CPR segmentation are the boundaries of the 3D straightened CPR along the diameter direction. In this embodiment, the obtained 3D straightened CPR generates n 2D CPR vessels. According to the generation order, the coronary vessel edge lines in the cropping plane are mapped to the single vessel model, and the 3D straightened CPR is updated and constructed. The vessel boundary information annotation of the 3D straightened CPR can then be obtained. This vessel boundary information annotation can be used for binarization processing of the 3D straightened CPR to highlight the vessel contour edges, or for deep learning model training.

[0067] Traditional annotation methods for coronary CTA vessels along the centerline in 3D straightened CPR involve directly annotating vessel boundaries on the cross-section of each layer, which is time-consuming, labor-intensive, and generates a large workload. The rapid annotation method for coronary CTA vessels provided in the embodiments of this application transforms the task of 3D straightened CPR into n 2D CPR images, converting the complex 3D segmentation task into a simple and mature 2D segmentation task. By converting to 2D CPR and performing coarse segmentation, the boundary information of the vessels can be obtained quickly.

[0068] This embodiment converts 3D straightening CPR into 2D CPR, obtains coarse segmentation results, and then checks and modifies them to obtain correct vessel boundaries before training the 2D segmentation model. This process is rapid and efficient. The trained 2D segmentation model can generate a large number of 3D straightening CPR annotations, significantly improving the problem of insufficient data annotation.

[0069] It should be understood that, although Figures 1-2 The steps in the flowchart are shown sequentially as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order in which these steps are executed, and they can be performed in other orders. Figure 1-2 At least some of the steps in the process may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least some of the sub-steps or stages of other steps.

[0070] In one embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 7As shown, the computer device includes a processor, memory, network interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The network interface is used to communicate with external terminals via a network connection. When the computer program is executed by the processor, it implements a rapid labeling method for coronary CTA vessels. The display screen can be an LCD screen or an e-ink display screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device casing, or an external keyboard, touchpad, or mouse.

[0071] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:

[0072] Step S100: Construct a three-dimensional model of the coronary artery based on the coronary CTA sequence image, extract the coronary artery centerline based on the three-dimensional model of the coronary artery, and extract the centerline of a single coronary artery based on the coronary artery centerline.

[0073] Step S200: Straighten the corresponding single blood vessel in the three-dimensional model of the reconstructed coronary artery along the center line of the single coronary artery to obtain the straightened single blood vessel model.

[0074] Step S300: Based on the interception plane passing through the center line of a single coronary artery, perform cross-sectional sampling on the single vessel model to obtain a cross-sectional image;

[0075] Step S400: Perform image segmentation on the cross-sectional image to obtain the edge lines of the coronary arteries;

[0076] Step S500: Map the edge line of the coronary artery to the single vessel model to obtain the single vessel model with the boundary information of the coronary artery.

[0077] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, the computer program performing the following steps when executed by a processor:

[0078] Step S100: Construct a three-dimensional model of the coronary artery based on the coronary CTA sequence image, extract the coronary artery centerline based on the three-dimensional model of the coronary artery, and extract the centerline of a single coronary artery based on the coronary artery centerline.

[0079] Step S200: Straighten the corresponding single blood vessel in the three-dimensional model of the reconstructed coronary artery along the center line of the single coronary artery to obtain the straightened single blood vessel model.

[0080] Step S300: Based on the interception plane passing through the center line of a single coronary artery, perform cross-sectional sampling on the single vessel model to obtain a cross-sectional image;

[0081] Step S400: Perform image segmentation on the cross-sectional image to obtain the edge lines of the coronary arteries;

[0082] Step S500: Map the edge line of the coronary artery to the single vessel model to obtain the single vessel model with the boundary information of the coronary artery.

[0083] In one embodiment, a computer program product is provided, including computer instructions that, when executed by a processor, perform the following steps:

[0084] Step S100: Construct a three-dimensional model of the coronary artery based on the coronary CTA sequence image, extract the coronary artery centerline based on the three-dimensional model of the coronary artery, and extract the centerline of a single coronary artery based on the coronary artery centerline.

[0085] Step S200: Straighten the corresponding single blood vessel in the three-dimensional model of the reconstructed coronary artery along the center line of the single coronary artery to obtain the straightened single blood vessel model.

[0086] Step S300: Based on the interception plane passing through the center line of a single coronary artery, perform cross-sectional sampling on the single vessel model to obtain a cross-sectional image;

[0087] Step S400: Perform image segmentation on the cross-sectional image to obtain the edge lines of the coronary arteries;

[0088] Step S500: Map the edge line of the coronary artery to the single vessel model to obtain the single vessel model with the boundary information of the coronary artery.

[0089] In this embodiment, the computer program product includes program code portions for performing the steps of the rapid annotation method for coronary CTA vessels in the various embodiments of this application when the computer program product is executed by one or more computing devices. The computer program product may be stored on a computer-readable recording medium. It may also be provided for download via a data network (e.g., via RAN, via the Internet, and / or via RBS). Alternatively or additionally, the method may be encoded in a field-programmable gate array (FPGA) and / or an application-specific integrated circuit (ASIC), or its functionality may be provided for download by means of a hardware description language.

[0090] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0091] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as the combination of these technical features does not contradict each other, it should be considered to be within the scope of this specification. When technical features of different embodiments are embodied in the same drawing, it can be regarded as the drawing also disclosing examples of combinations of the various embodiments involved.

[0092] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A method for fast annotation of coronary CTA vessels, characterized in that, The method comprises the following steps: a coronary artery three-dimensional model is constructed based on a coronary CTA sequence image, a coronary centerline is extracted based on the coronary artery three-dimensional model, and a single coronary centerline is extracted based on the coronary centerline; a single blood vessel model is obtained by straightening and reconstructing a corresponding single blood vessel in the coronary artery three-dimensional model along the single coronary centerline; a cross-section image is obtained by cross-section sampling of the single blood vessel model according to a cross-section plane intersecting the single coronary centerline; coronary vessel edge lines are obtained by performing image segmentation on the cross-section image, and the coronary vessel edge lines are mapped to the single blood vessel model to obtain a single blood vessel model labeled with coronary vessel boundary information.

2. The method of claim 1, wherein, The single coronary centerline is extracted based on the coronary centerline, and specifically, the coronary centerline has a coronary starting point and a plurality of coronary ending points, each coronary ending point and the corresponding coronary starting point form a single coronary centerline, and the single coronary centerline is obtained by cutting the coronary centerline.

3. The method of claim 1, wherein, The single blood vessel model is a cylinder, and the single coronary centerline is located at the axis of the cylinder.

4. The method of claim 3, wherein, The single blood vessel model is obtained by straightening and reconstructing the corresponding single blood vessel in the coronary artery three-dimensional model, and specifically: the single blood vessel model is obtained by straightening and reconstructing the corresponding single blood vessel in the coronary artery three-dimensional model to obtain a cuboid single blood vessel model; for the cuboid single blood vessel model, the single coronary centerline is taken as the axis to obtain a cylindrical single blood vessel model.

5. The method of claim 4, wherein, The cuboid single blood vessel model has a length, a width, and a height, and the specifications of the length, the width, and the height are X*X*H, where X and H represent X and H unit lengths, respectively. The cylindrical single blood vessel model has a diameter and a height, and the specifications of the diameter and the height are X*H, where X and H represent X and H unit lengths, respectively.

6. The method of claim 1, wherein, The cross-section plane includes a plurality of circumferentially uniformly distributed cross-section planes, and the method comprises the following steps: the cross-section image is obtained by cross-section sampling of the single blood vessel model according to the cross-section plane intersecting the single coronary centerline; after obtaining the cross-section image, the single coronary centerline in the cross-section image is located at the center of the cross-section image, and the specifications of each cross-section image are the same by uniform processing of the cross-section image.

7. The method of claim 1, wherein, The method comprises the following steps: coarse segmentation is performed on the cross-section image to obtain a coarse segmentation result; fine segmentation is performed on the coarse segmentation result by using a trained Unet neural network to obtain coronary vessel edge lines.

8. The method of claim 1, wherein, The method comprises the following steps: the coronary vessel edge lines are mapped to the single blood vessel model to obtain a plurality of coronary vessel edge lines on the single blood vessel model; a coronary vessel contour on the single blood vessel model is obtained based on a connected domain formed by the plurality of coronary vessel edge lines, and a single blood vessel model labeled with coronary vessel boundary information is further obtained.

9. Computer device comprising a memory, a processor and a computer program stored on the memory, characterized in that, The processor executes the computer program to implement the steps of the coronary CTA blood vessel rapid labeling method according to any one of claims 1-8.

10. A computer program product comprising computer instructions, characterized in that, The computer instructions, when executed by the processor, implement the steps of the method for fast labeling of coronary CTA vessels according to any one of claims 1-8.

Images