Pulmonary vessel labeling method, device, nonvolatile storage medium, and electronic device

Abstract

This application discloses a method, apparatus, non-volatile storage medium, and electronic device for marking pulmonary vessels. The method for marking pulmonary vessels includes: determining a first overlapping region between a first pulmonary bronchus image and a first pulmonary vascular tree image, and removing vascular tree images located within the first overlapping region from the first pulmonary vascular tree image to obtain a second pulmonary vascular tree image; merging the second pulmonary vascular tree image and a first pulmonary arteriovenous trunk image to obtain a third pulmonary vascular tree image, wherein the first pulmonary arteriovenous trunk image includes a trunk vascular region with a diameter greater than a preset diameter threshold and the arteriovenous type of each voxel point belonging to the trunk vascular region; determining the boundary points of the trunk vascular region in the third pulmonary vascular tree image based on the first pulmonary arteriovenous trunk image, and using the boundary points as initial seed points for neighborhood search, adding target markers to each voxel point belonging to the peripheral vascular region in the third pulmonary vascular tree image.

Description

Technical Field

[0001] This application relates to the field of medical image data processing technology, and in particular to a method, apparatus, non-volatile storage medium, and electronic device for marking pulmonary blood vessels. Background Technology

[0002] When performing fine classification and annotation of pulmonary arteriovenous vessels, related technologies typically use threshold segmentation to obtain a complete pulmonary vascular tree, and then assign arteriovenous values ​​to the vascular tree through seed point search.

[0003] Currently, manual identification methods have a high accuracy rate for classifying the arteriovenous types of the pulmonary arteriovenous trunk. However, due to the presence of a lot of interference noise (e.g., noise from the bronchial walls) in pulmonary vascular tree images, manual labeling of the arteriovenous types of peripheral vessels is often inaccurate and labor-intensive, resulting in technical problems such as poor accuracy and low efficiency in vascular tree arteriovenous labeling.

[0004] There is currently no effective solution to the above problems. Summary of the Invention

[0005] This application provides a method, apparatus, non-volatile storage medium, and electronic device for marking pulmonary vessels, to at least solve the technical problem of poor accuracy in marking arteriovenous vessels in pulmonary vascular tree images due to noise interference when assigning arteriovenous values ​​to vascular trees through seed point search in related technologies.

[0006] According to one aspect of the embodiments of this application, a method for labeling pulmonary vessels is provided, comprising: determining a first overlapping region between a first pulmonary bronchus image and a first pulmonary vascular tree image, and removing the vascular tree images located within the first overlapping region in the first pulmonary vascular tree image to obtain a second pulmonary vascular tree image, wherein the second pulmonary vascular tree image includes a peripheral vascular region with a diameter not greater than a preset diameter threshold; merging the second pulmonary vascular tree image and a first pulmonary arteriovenous trunk image to obtain a third pulmonary vascular tree image, wherein the first pulmonary arteriovenous trunk image includes a trunk vascular region with a diameter greater than a preset diameter threshold and the arteriovenous type of each voxel point belonging to the trunk vascular region; determining the boundary points of the trunk vascular region in the third pulmonary vascular tree image based on the first pulmonary arteriovenous trunk image, and performing a neighborhood search using the boundary points as initial seed points to add target markers to each voxel point belonging to the peripheral vascular region in the third pulmonary vascular tree image, wherein the target markers are used to characterize the arteriovenous type.

[0007] Optionally, merging the second pulmonary vascular tree image and the first pulmonary arteriovenous trunk image to obtain the third pulmonary vascular tree image includes: determining the Hu value of each voxel point in the second pulmonary vascular tree image and deleting voxel points in the second pulmonary vascular tree image whose Hu value is greater than a preset threshold value to obtain the fourth pulmonary vascular tree image; merging the fourth pulmonary vascular tree image and the first pulmonary arteriovenous trunk image to obtain the third pulmonary vascular tree image, wherein, during the merging process, if there is a second overlapping region between the fourth pulmonary vascular tree image and the first pulmonary arteriovenous trunk image, the portion of the second overlapping region in the first pulmonary arteriovenous trunk image is retained in the third pulmonary vascular tree image.

[0008] Optionally, determining the boundary points of the main pulmonary artery and vein region in the third pulmonary vascular tree image based on the first pulmonary artery and vein main trunk image includes: performing an erosion operation on the first pulmonary artery and vein main trunk image to obtain a second pulmonary artery and vein main trunk image, wherein the erosion operation is used to eliminate voxel points of the artery and vein main trunk boundary in the first pulmonary artery and vein main trunk image; performing an intersection operation on the second pulmonary artery and vein main trunk image and the first pulmonary artery and vein main trunk image to obtain a target intersection image, wherein the target intersection image contains the common voxel points of the second pulmonary artery and vein main trunk image and the first pulmonary artery and vein main trunk image; performing an inversion operation on the target intersection image, and determining the voxel points obtained after the inversion operation as the boundary points of the main pulmonary vascular region.

[0009] Optionally, the target markers include: a first marker for characterizing the artery corresponding to the voxel point, a second marker for characterizing the vein corresponding to the voxel point, and a third marker for characterizing the voxel point that was not found. The boundary point is used as the initial seed point for neighborhood search. Adding target markers to each voxel point belonging to the peripheral vascular region in the third lung vascular tree image includes: determining the arteriovenous type corresponding to the initial seed point as the arteriovenous type of the initial seed point, wherein the arteriovenous type includes: artery and vein; obtaining a set of voxel points within a preset neighborhood of the initial seed point in the third lung vascular tree image; determining the target markers corresponding to each voxel point of the vascular tree within the preset neighborhood based on the distribution of voxel points in the voxel point set; determining voxel points within the preset neighborhood that have been marked with the first or second marker as new seed points and continuing the neighborhood search until no new seed points are found; after the neighborhood search is completed, adding the third marker to the voxel points of the vascular tree in the third lung vascular tree image that have not been marked with the target marker.

[0010] Optionally, the target marker further includes: a fourth marker for characterizing voxel points whose arterial / venous type cannot be determined; determining the target marker corresponding to each voxel point of the vascular tree within a preset neighborhood based on the distribution of voxel points in the voxel point set includes: adding a first marker to voxel points within the preset neighborhood whose arterial / venous type is undetermined when the seed point's arterial / venous type is arterial and the preset neighborhood does not contain voxel points whose arterial / venous type is vein; adding a second marker to voxel points within the preset neighborhood whose arterial / venous type is undetermined when the seed point's arterial / venous type is vein and the preset neighborhood does not contain voxel points whose arterial / venous type is arterial; adding a fourth marker to voxel points within the preset neighborhood whose arterial / venous type is undetermined when the seed point's arterial / venous type is arterial and both voxel points whose arterial / venous type is vein and voxel points whose arterial / venous type is undetermined exist within the preset neighborhood; adding a fourth marker to voxel points within the preset neighborhood whose arterial / venous type is undetermined when the seed point's arterial / venous type is vein and both voxel points whose arterial / venous type is arterial and voxel points whose arterial / venous type is undetermined exist within the preset neighborhood.

[0011] Optionally, the method further includes: sending the third pulmonary vascular tree image with added target markers to the front-end interactive interface for display; in response to the operation instructions of the front-end interactive interface, determining the arterial and venous types corresponding to the voxel points of the third marker and / or the fourth marker, and obtaining the target pulmonary vascular tree image, wherein the operation instructions are used to delete or correct the target markers corresponding to the voxel points of the third pulmonary vascular tree image.

[0012] Optionally, before determining the overlapping region between the first bronchial image and the first pulmonary vascular tree image, the method further includes: using a deep learning segmentation network to divide the computed tomography image of the lungs to obtain the original bronchial image, the original pulmonary vascular tree image, and the first pulmonary arteriovenous trunk image; using a preset dilation template to dilate the original bronchial image to obtain the first bronchial image, wherein the preset dilation template is a convolutional kernel of a preset size, and the dilation operation is used to expand the bronchial tract in the original bronchial image to the airway wall; and using a preset erosion template to erode the original pulmonary vascular tree image to obtain the first pulmonary vascular tree image, wherein the preset erosion template is a convolutional kernel of a preset size, and the erosion operation is used to eliminate noise points in the original pulmonary vascular tree image.

[0013] According to another aspect of the embodiments of this application, a pulmonary vascular marking device is also provided, comprising: a tracheal noise elimination module, configured to determine a first overlapping region between a first pulmonary bronchial image and a first pulmonary vascular tree image, and remove the vascular tree image located within the first overlapping region in the first pulmonary vascular tree image to obtain a second pulmonary vascular tree image, wherein the second pulmonary vascular tree image includes a peripheral vascular region with a diameter not greater than a preset diameter threshold; an image merging module, configured to merge the second pulmonary vascular tree image and a first pulmonary arteriovenous trunk image to obtain a third pulmonary vascular tree image, wherein the first pulmonary arteriovenous trunk image includes a trunk vascular region with a diameter greater than a preset diameter threshold and the arteriovenous type of each voxel point belonging to the trunk vascular region; and a neighborhood search module, configured to determine the boundary points of the trunk vascular region in the third pulmonary vascular tree image based on the first pulmonary arteriovenous trunk image, and perform a neighborhood search using the boundary points as initial seed points to add target markers to each voxel point belonging to the peripheral vascular region in the third pulmonary vascular tree image, wherein the target markers are used to characterize the arteriovenous type.

[0014] According to another aspect of the embodiments of this application, a non-volatile storage medium is also provided, which stores a computer program that, when executed by a processor, implements a method for marking pulmonary vessels.

[0015] According to another aspect of the embodiments of this application, an electronic device is also provided, including 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 a method for marking pulmonary vessels.

[0016] Based on the above scheme, this application involves determining the first overlapping region between the first pulmonary bronchial image and the first pulmonary vascular tree image, and removing the vascular tree images located within the first overlapping region from the first pulmonary vascular tree image to obtain a second pulmonary vascular tree image. The second pulmonary vascular tree image includes peripheral vascular regions with a diameter not greater than a preset diameter threshold. The second pulmonary vascular tree image and the first pulmonary arteriovenous trunk image are then merged to obtain a third pulmonary vascular tree image. The first pulmonary arteriovenous trunk image includes trunk vascular regions with a diameter greater than a preset diameter threshold and the arteriovenous types of each voxel point belonging to the trunk vascular region. Based on the first pulmonary arteriovenous trunk image, the boundary points of the trunk vascular regions in the third pulmonary vascular tree image are determined, and the boundary points are marked as... A neighborhood search is performed on the initial seed points, and target labels are added to each voxel point belonging to the peripheral vascular region in the third pulmonary vascular tree image. The target labels are used to characterize the arteriovenous type. Airway wall noise in the pulmonary vascular tree image is removed by pulmonary bronchial images, and the boundary points of the pulmonary arteriovenous trunk in the noise-removed pulmonary vascular tree image are determined as seed points for neighborhood search using the information of the pulmonary arteriovenous trunk. The pulmonary vascular tree is automatically assigned arteriovenous labels, which improves the efficiency and accuracy of arteriovenous classification and labeling of vascular trees. This solves the technical problem of poor accuracy of arteriovenous labeling of vascular trees caused by noise interference in the pulmonary vascular tree image when assigning arteriovenous labels to vascular trees through seed point search. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments or related technologies of this application, the accompanying drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a schematic diagram of a method for marking pulmonary vessels according to an embodiment of this application;

[0019] Figure 2a This is a schematic diagram of a first pulmonary arteriovenous trunk image provided according to an embodiment of this application;

[0020] Figure 2b This is a schematic diagram of a raw lung bronchial image provided according to an embodiment of this application;

[0021] Figure 2c This is a schematic diagram of a raw pulmonary vascular tree image provided according to an embodiment of this application;

[0022] Figure 3 This is a schematic diagram of a method for peripheral arteriovenous annotation based on a pulmonary vascular tree according to an embodiment of this application;

[0023] Figure 4 This is a schematic diagram of airway wall noise provided according to an embodiment of this application;

[0024] Figure 5 This is a schematic diagram of airway wall noise removal according to an embodiment of this application;

[0025] Figure 6 This is a schematic diagram of a Hu value filtering process provided according to an embodiment of this application;

[0026] Figure 7 This is a schematic diagram of a process for extracting blood vessel boundaries according to an embodiment of this application;

[0027] Figure 8 This is a schematic diagram of an arteriovenous neighborhood search according to an embodiment of this application;

[0028] Figure 9 This is a schematic diagram illustrating the marking of manual review locations according to an embodiment of this application;

[0029] Figure 10 This is a schematic diagram illustrating the marking of manual review locations according to an embodiment of this application;

[0030] Figure 11 This is a schematic diagram of a pulmonary vascular marking device according to an embodiment of this application;

[0031] Figure 12 This is a hardware structure block diagram of a computer terminal (or electronic device) for implementing a lung vascular marking method, according to an embodiment of this application. Detailed Implementation

[0032] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0033] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0034] To facilitate a better understanding of the embodiments of this application by those skilled in the art, some technical terms or nouns involved in the embodiments of this application are explained as follows:

[0035] Computed Tomography (CT) uses precisely collimated X-ray beams, gamma rays, ultrasound waves, etc., along with highly sensitive detectors, to scan a specific part of the human body one section after another. It features fast scanning time and clear images and can be used to examine a variety of diseases.

[0036] Hu value: This refers to the tissue density unit in a CT scan, also known as the CT value. It represents the density of tissue in the scan and is determined by measuring the degree of X-ray absorption in the tissue. It is used for disease diagnosis and evaluation of treatment effectiveness. The technical solution of this application is described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments.

[0037] In the task of segmenting pulmonary arteries and veins, a pulmonary artery and vein dataset needs to be constructed. For the classification of arteriovenous types in the main pulmonary arteries and veins, manual identification achieves high accuracy. However, for the peripheral lung region (vessel diameter <= 3mm), due to the thinner vessels, manual annotation is insufficient to cover this area. Therefore, the quality of the dataset is often improved by iteratively updating the pulmonary artery and vein annotations. Specifically, a deep learning model can be trained by manually annotating the main pulmonary artery and vein regions. After obtaining the deep learning pulmonary artery and vein segmentation model, the model can be updated by further extending the vessel length based on the existing segmentation results, thereby optimizing the vessel annotation.

[0038] Currently, pulmonary arteriovenous vascular trees with relatively complete pulmonary vascular information can be obtained through vascular extraction algorithms. However, these trees do not contain arteriovenous category information. They need to be merged with the arteriovenous segmentation results to complete the arteriovenous vascular classification, thereby further optimizing the arteriovenous vascular dataset and enabling the updating of the deep learning pulmonary arteriovenous vascular segmentation model.

[0039] However, since there is often a lot of interference noise in the pulmonary arteriovenous vascular tree (e.g., noise from the bronchial walls), the process of assigning arteriovenous values ​​to the vascular tree based on seed point search is prone to misclassification of arteriovenous values, resulting in technical problems such as poor accuracy of arteriovenous labeling.

[0040] To address the aforementioned problems, relevant solutions are provided in the embodiments of this application. The technical solutions provided in each embodiment of this specification are described in detail below with reference to the accompanying drawings.

[0041] This application provides a method for labeling pulmonary vessels, which can accurately and efficiently assign arterial and venous values ​​to the pulmonary vascular tree, optimize peripheral pulmonary arterial and venous information, and obtain fully labeled pulmonary vessels. Figure 1 This is a schematic diagram of a method for lung vascular marking according to an embodiment of this application, as shown below. Figure 1 As shown, the method includes the following steps:

[0042] Step S102: Determine the first overlapping region between the first lung bronchus image and the first lung vascular tree image, and remove the images of the vascular trees located in the first overlapping region in the first lung vascular tree image to obtain the second lung vascular tree image, wherein the second lung vascular tree image includes a peripheral vascular region with a diameter not greater than a preset diameter threshold.

[0043] In this embodiment, before performing step S102, a deep learning segmentation network can be used to divide the computed tomography (CT) images (enhanced CT or non-enhanced CT) of the lungs to extract the original bronchial image corresponding to the bronchial tree, the original pulmonary vascular tree image corresponding to the pulmonary vascular tree, and the first pulmonary arteriovenous trunk image corresponding to the pulmonary arteriovenous trunk. The original bronchial image is used to characterize the bronchial wall information, and the first pulmonary arteriovenous trunk image is used to characterize the trunk blood vessel information.

[0044] Specifically, a trained pulmonary arteriovenous trunk segmentation network (e.g., a segmentation network with a V-net network structure) can be used to extract images of the first pulmonary arteriovenous trunk on enhanced or non-enhanced CT scans, such as... Figure 2a As shown, the main arteriovenous trunks are blood vessels in the lung region with a diameter greater than a preset diameter threshold; in this embodiment, the preset diameter is 3 mm. The original lung and bronchial images are obtained using a lung and bronchial segmentation network (e.g., a segmentation network with a V-net network structure), as shown below. Figure 2b As shown; and, the original pulmonary vascular tree image is obtained using a deep learning segmentation network (or, the original pulmonary vascular tree image can also be obtained from the CT image based on the HU value using a vessel extraction algorithm), such as Figure 2c As shown, the pulmonary vascular tree consists of blood vessels in the left and right lobes of the lungs, and overlaps to some extent with the main pulmonary arteriovenous trunk area, but does not include the complete main pulmonary arteriovenous trunk.

[0045] During the training process of the above-mentioned pulmonary arteriovenous trunk segmentation network, the blood vessels (corresponding to the main trunk blood vessels with a diameter greater than 3mm) to which the voxel points with CT values ​​(i.e., Henle units) Hu > -400 can be marked on the sample CT images of the training set. The above-mentioned pulmonary arteriovenous trunk segmentation network is then trained based on the marked training set, so that the trained pulmonary arteriovenous trunk segmentation network can segment the pulmonary arteriovenous trunk.

[0046] After obtaining the images of the first pulmonary arteriovenous trunk, the original pulmonary bronchus, and the original pulmonary vascular tree, these three images can be used as pre-acquired raw data for subsequent processing steps. Figure 3 This is a schematic diagram of a method for peripheral arteriovenous annotation based on a pulmonary vascular tree, according to an embodiment of this application. Figure 3 As shown, Figure 3 The pulmonary bronchus, pulmonary vascular tree, and pulmonary arteriovenous trunk shown correspond to the original pulmonary bronchus image, the original pulmonary vascular tree image, and the first pulmonary arteriovenous trunk image, respectively.

[0047] The following section provides further explanation of the subsequent processing procedures for the images of the first pulmonary arteriovenous trunk, the original pulmonary bronchus, and the original pulmonary vascular tree.

[0048] First, the original lung bronchial images and the original lung vascular tree images are preprocessed to remove airway wall noise and some isolated noise points. The specific steps are as follows.

[0049] In some embodiments of this application, before determining the overlapping region between the first bronchial image and the first pulmonary vascular tree image, the method further includes the following steps: using a deep learning segmentation network to divide the computed tomography image to obtain the original bronchial image, the original pulmonary vascular tree image, and the first pulmonary arteriovenous trunk image; using a preset dilation template to dilate the original bronchial image to obtain the first bronchial image, wherein the preset dilation template is a convolution kernel of a preset size, and the dilation operation is used to expand the bronchial tubes in the original bronchial image to the airway wall; and using a preset erosion template to erode the original pulmonary vascular tree image to obtain the first pulmonary vascular tree image, wherein the preset erosion template is a convolution kernel of a preset size, and the erosion operation is used to eliminate noise points in the original pulmonary vascular tree image.

[0050] For the original lung and bronchial image, the lung and bronchial need to be expanded to the airway wall. Specifically, a preset expansion template of a preset size can be used to traverse all pixel positions of the lung and bronchial to obtain the traversed lung and bronchial region. In this embodiment, the preset expansion template is illustrated by a 3x3 convolution kernel, which corresponds to a matrix with a value of 1. The convolution kernel is logically ANDed with the image region it covers. If the region covered by the convolution kernel is all 0 after the AND operation, the result is 0; otherwise, the result is 1, thus expanding the image. After preprocessing, the first lung and bronchial image is finally obtained.

[0051] For the original pulmonary vascular tree image, the purpose of preprocessing is to remove airway wall noise and isolated noise points in the vascular tree. Specifically, a preset erosion template of a preset size can be used to traverse all pixel positions of the pulmonary vascular tree to obtain the traversed pulmonary vascular tree region. In this embodiment, the preset erosion template is illustrated by a 3x3 convolution kernel. The convolution kernel corresponds to a matrix with a value of 1. A logical "AND" operation is performed with the covered area. If all pixels in the area covered by the convolution kernel are 1 after the "AND" operation, then the pixel is 1; otherwise, it is 0. After preprocessing, the first pulmonary vascular tree image is finally obtained.

[0052] The first pulmonary vascular tree image mentioned above may contain some airway wall noise, i.e., blood vessels that overlap with the airway wall, such as... Figure 4 As shown, it is therefore necessary to determine the first overlapping region between the first bronchial image and the first pulmonary vascular tree image, and then remove the blood vessels in the first pulmonary vascular tree image that overlap with the airway wall and are located within the first overlapping region, such as... Figure 5 As shown, the square labeled 1 represents the voxel point of the blood vessel in the first lung vascular tree image, and the square labeled 2 represents the voxel point of the bronchus in the first lung bronchus image. Figure 5 The area within the dashed circle represents the first overlapping region. After the blood vessels in the overlapping region of the airway wall in the first pulmonary vascular tree image and the first pulmonary bronchial image are cleared, the second pulmonary vascular tree image is obtained.

[0053] Step S104: Merge the second pulmonary vascular tree image and the first pulmonary arteriovenous trunk image to obtain a third pulmonary vascular tree image. The first pulmonary arteriovenous trunk image includes a trunk vascular region with a diameter greater than the preset diameter threshold and the arteriovenous types of each voxel point belonging to the trunk vascular region.

[0054] After obtaining the second pulmonary vascular tree image, the first pulmonary arteriovenous trunk image and the processed second pulmonary vascular tree image can be merged for region search. The specific steps are as follows.

[0055] In some embodiments of this application, merging a second pulmonary vascular tree image and a first pulmonary arteriovenous trunk image to obtain a third pulmonary vascular tree image includes the following steps: determining the Hu value of each voxel point in the second pulmonary vascular tree image and deleting voxel points in the second pulmonary vascular tree image whose Hu value is greater than a preset numerical threshold to obtain a fourth pulmonary vascular tree image, wherein the Hu value is used to characterize the degree of tissue absorption of X-rays; merging the fourth pulmonary vascular tree image and the first pulmonary arteriovenous trunk image to obtain a third pulmonary vascular tree image, wherein, during the merging process, if there is a second overlapping region between the fourth pulmonary vascular tree image and the first pulmonary arteriovenous trunk image, the portion of the second overlapping region in the first pulmonary arteriovenous trunk image is retained in the third pulmonary vascular tree image.

[0056] Specifically, the portion of the main pulmonary arteriovenous vessels that is easily labeled manually in the second pulmonary vascular tree image (i.e., the arteriovenous vessels in the region corresponding to CT value Hu>-400 that can be effectively identified using the trained pulmonary arteriovenous trunk segmentation network) is first removed. Specifically, for example... Figure 6 As shown, the Hu value (i.e., CT value) of each voxel in the second pulmonary vascular tree image is determined, and the area with Hu value greater than a preset threshold (e.g., -400) is cleared. Thus, the fourth pulmonary vascular tree image is obtained after clearing airway wall noise, isolated noise points and easily labeled vascular areas in the vascular tree.

[0057] It should be noted that, in the embodiments of this application, Figure 3 The order in which the two steps, “delete the position of Hu>-400 in the vascular tree” and “remove the blood vessels in the region overlapping with the lung bronchi”, are executed is not limited. In short, after these two steps are completed, the fourth lung vascular tree image, which has cleared the airway wall noise, isolated noise points and easily labeled blood vessel regions in the vascular tree, can be obtained.

[0058] After obtaining the fourth pulmonary vascular tree image, it is necessary to merge the fourth pulmonary vascular tree image with the first pulmonary arteriovenous trunk image. If there is a second overlapping region between the fourth pulmonary vascular tree image and the first pulmonary arteriovenous trunk image during the merging process, the second overlapping region is removed from the fourth pulmonary vascular tree image based on the first pulmonary arteriovenous trunk image to obtain the third pulmonary vascular tree image.

[0059] After obtaining the third pulmonary vascular tree image, the subsequent neighborhood search steps can be performed, which will be further introduced below.

[0060] Step S106: Based on the first pulmonary arteriovenous trunk image, determine the boundary points of the trunk vessel region in the third pulmonary vascular tree image, and use the boundary points as initial seed points for neighborhood search. Add target labels to each voxel point in the third pulmonary vascular tree image that belongs to the peripheral vessel region. The target labels are used to characterize the arteriovenous type.

[0061] Before performing a neighborhood search, seed points need to be widely distributed in the third pulmonary vascular tree image. In this embodiment, the initial seed points for the neighborhood search are set as the boundary points of the arteriovenous trunk (trunk vessel region), which increases the number of deployed seed points, resulting in faster computation speed and fewer search operations. The process of determining the boundary points of the pulmonary arteriovenous trunk (trunk vessel region) will be further described below.

[0062] In this embodiment, the boundary points of the pulmonary arteriovenous trunk can be determined by eroding the known arteriovenous trunk in the first pulmonary arteriovenous trunk image and then intersecting and inverting the result after erosion with the first pulmonary arteriovenous trunk image before erosion. The specific steps are as follows.

[0063] In some embodiments of this application, determining the boundary points of the main pulmonary artery and vein region in the third pulmonary vascular tree image based on the first pulmonary artery and vein main trunk image includes the following steps: performing an erosion operation on the first pulmonary artery and vein main trunk image to obtain a second pulmonary artery and vein main trunk image, wherein the erosion operation is used to eliminate voxel points of the artery and vein main trunk boundary in the first pulmonary artery and vein main trunk image; performing an intersection operation on the second pulmonary artery and vein main trunk image and the first pulmonary artery and vein main trunk image to obtain a target intersection image, wherein the target intersection image contains the common voxel points of the second pulmonary artery and vein main trunk image and the first pulmonary artery and vein main trunk image; performing an inversion operation on the target intersection image, and determining the voxel points obtained after the inversion operation as the boundary points of the main pulmonary vascular region.

[0064] Specifically, morphological operations are used to erode the main arterial and venous vessels in the first pulmonary arterial and venous trunk image to obtain the second pulmonary arterial and venous trunk image. The eroded portion of the main arterial and venous vessels reduces the number of pixels at the boundaries. Then, the eroded portion of the main arterial and venous vessels in the second pulmonary arterial and venous trunk image is subtracted from the first image (i.e., an intersection operation followed by an inversion operation) to obtain the boundary points of the main arterial and venous trunk. Figure 7 As shown.

[0065] Then, all the boundary points of the arteriovenous trunks in the third pulmonary vascular tree image are used as initial seed points. This method can obtain a large number of initial seed points, thereby speeding up the calculation. After determining the initial seed points, the category (i.e. the corresponding target label) of each initial seed point is known. The initial seed points are used to search outwards for adjacent vascular trees to add target labels to the voxel points of the blood vessels at the category positions. If it is adjacent to an artery point, the vascular tree is assigned the value of artery; if it is adjacent to a vein, it is assigned the value of vein. The specific steps are as follows.

[0066] In some embodiments of this application, the target markers include: a first marker for characterizing the artery corresponding to the voxel point, a second marker for characterizing the vein corresponding to the voxel point, and a third marker for characterizing the voxel point not being searched; the boundary point is used as the initial seed point for neighborhood search, and adding target markers to each voxel point in the third lung vascular tree image belonging to the peripheral vascular region includes the following steps: determining the arteriovenous type of the arteriovenous trunk corresponding to the initial seed point as the arteriovenous type of the initial seed point, wherein the arteriovenous type includes: artery and vein; obtaining a set of voxel points in the third lung vascular tree image located within a preset neighborhood of the initial seed point; determining the target markers corresponding to each voxel point of the vascular tree within the preset neighborhood based on the distribution of voxel points in the voxel point set; determining the voxel points within the preset neighborhood that have been marked with the first or second marker as new seed points and continuing the neighborhood search until no new seed points are found; after the neighborhood search is completed, adding the third marker to the voxel points of the vascular tree in the third lung vascular tree image that have not been marked with the target marker.

[0067] In some embodiments of this application, the target marker further includes: a fourth marker for characterizing voxel points whose arterial / venous type cannot be determined; determining the target marker corresponding to each voxel point of the vascular tree within a preset neighborhood based on the distribution of voxel points in the voxel point set includes the following steps: when the arterial / venous type of the seed point is artery, and the preset neighborhood does not contain voxel points whose arterial / venous type is vein, adding a first marker to the voxel points within the preset neighborhood whose arterial / venous type is not determined; when the arterial / venous type of the seed point is vein, and the preset neighborhood does not contain voxel points whose arterial / venous type is artery. In the case where the seed point's arterial / venous type is arterial, and there are voxel points with both arterial / venous type and voxel points with undetermined arterial / venous type within the preset neighborhood, a fourth label is added to the voxel points with undetermined arterial / venous type within the preset neighborhood; in the case where the seed point's arterial / venous type is vein, and there are voxel points with both arterial / venous type and voxel points with undetermined arterial / venous type within the preset neighborhood, a fourth label is added to the voxel points with undetermined arterial / venous type within the preset neighborhood.

[0068] Specifically, during the neighborhood search process, the target labels corresponding to each voxel point of the vascular tree in the third lung vascular tree image that has not been labeled with the target are determined. Figure 8 This is a schematic diagram of an arteriovenous neighborhood search according to an embodiment of this application, such as... Figure 8 As shown, when there are only voxel points of vascular trees with undetermined arterial and venous types within the preset neighborhood of the seed point where the arterial and venous type is artery, the vascular tree is assigned the value of artery, that is, a first label is added to the voxel point (the first label is represented by 2 in this embodiment); when there are only voxel points of vascular trees with undetermined arterial and venous types within the preset neighborhood of the seed point where the arterial and venous type is vein, the vascular tree is assigned the value of vein, that is, a second label is added to the voxel point (the second label is represented by 3 in this embodiment).

[0069] Meanwhile, for locations identified as having classification disputes during neighborhood search, these areas can be marked as prompts during manual annotation, allowing for correction of the marked points from the previous stage during the manual correction phase. Figure 9 This is a schematic diagram illustrating the marking of manual review locations according to an embodiment of this application, such as... Figure 9 As shown, when the artery / vein type of the seed point is artery, and there are voxel points with artery / vein type of vein and voxel points with undetermined artery / vein type within the preset neighborhood, it is determined that the position needs to be manually marked, and a fourth mark is added to the voxel points with undetermined artery / vein type within the preset neighborhood (the fourth mark is represented by 5 in this embodiment); and when the artery / vein type of the seed point is vein, and there are voxel points with artery type of artery and voxel points with undetermined artery / vein type within the preset neighborhood, it is determined that the position needs to be manually marked, and a fourth mark is added to the voxel points with undetermined artery / vein type within the preset neighborhood.

[0070] In addition, since blood vessels in the vascular tree may not be adjacent to arteries and veins, there may be parts that cannot be found during neighborhood search. Although these parts exist in the vascular tree, they may not necessarily be blood vessels, but may be continuous noise. Therefore, they need to be independently labeled and manually reviewed. Specifically, after the neighborhood search in the third lung vascular tree image is completed, for the voxel points of the vascular tree that were not searched, that is, the voxel points of the vascular tree that have not been marked with the target, a third label is added to them (in this embodiment, the third label is represented by 6).

[0071] It should be noted that the embodiments of this application do not limit the preset neighborhood range for neighborhood search. For example, a 26-neighbor search algorithm can be used to assign arterial and venous labels to the vascular trees in the segmented arterial and venous vascular neighborhood. In 3D spatial search, the arterial and venous classification results obtained by using the 26-neighbor search algorithm are more accurate. Alternatively, a 4-neighbor search algorithm or an 8-neighbor search algorithm can be used.

[0072] To further improve the accuracy of pulmonary vascular classification and labeling, after the above labeling process, the third pulmonary vascular tree image with added target labels can be manually reviewed. The specific steps are as follows.

[0073] In some embodiments of this application, the method further includes the following steps: sending the third pulmonary vascular tree image with added target markers to a front-end interactive interface for display; in response to the operation instructions of the front-end interactive interface, determining the arterial and venous types corresponding to the voxel points of the third marker and / or the fourth marker, and obtaining the target pulmonary vascular tree image, wherein the operation instructions are used to delete or correct the target markers corresponding to the voxel points of the third pulmonary vascular tree image.

[0074] Specifically, manual correction only requires reviewing the labeled areas, significantly reducing the time and workload of peripheral vascular labeling. During manual correction, misclassifications near the labeled points (i.e., the voxel points of the third and / or fourth labels mentioned above) are corrected, and unlabeled vascular trees are reviewed and manually classified or deleted. After manual correction, the vascular trees are post-processed to delete unclassified vascular trees and labeled points that do not require correction. Ultimately, relatively complete pulmonary arterial and venous vascular information with labeled peripheral pulmonary vessels is obtained, such as... Figure 10 As shown.

[0075] According to an embodiment of this application, an embodiment of a lung vascular marking device is also provided. Figure 11 This is a schematic diagram of a pulmonary vascular marking device according to an embodiment of this application. Figure 11 As shown, the device includes:

[0076] The tracheal noise elimination module 110 is used to determine the first overlapping region between the first lung bronchus image and the first lung vascular tree image, and remove the images of the vascular trees located in the first overlapping region in the first lung vascular tree image to obtain a second lung vascular tree image. The second lung vascular tree image includes a peripheral vascular region with a diameter not greater than a preset diameter threshold.

[0077] Optionally, before determining the overlapping region between the first bronchial image and the first pulmonary vascular tree image, the method further includes: using a deep learning segmentation network to divide the computed tomography image of the lungs to obtain the original bronchial image, the original pulmonary vascular tree image, and the first pulmonary arteriovenous trunk image; using a preset dilation template to dilate the original bronchial image to obtain the first bronchial image, wherein the preset dilation template is a convolutional kernel of a preset size, and the dilation operation is used to expand the bronchial tract in the original bronchial image to the airway wall; and using a preset erosion template to erode the original pulmonary vascular tree image to obtain the first pulmonary vascular tree image, wherein the preset erosion template is a convolutional kernel of a preset size, and the erosion operation is used to eliminate noise points in the original pulmonary vascular tree image.

[0078] Image merging module 112 is used to merge the second pulmonary vascular tree image and the first pulmonary arteriovenous trunk image to obtain a third pulmonary vascular tree image. The first pulmonary arteriovenous trunk image includes a trunk vascular region with a diameter greater than a preset diameter threshold and the arteriovenous types of each voxel point belonging to the trunk vascular region.

[0079] Optionally, merging the second pulmonary vascular tree image and the first pulmonary arteriovenous trunk image to obtain the third pulmonary vascular tree image includes: determining the Hu value of each voxel point in the second pulmonary vascular tree image and deleting voxel points in the second pulmonary vascular tree image whose Hu value is greater than a preset threshold value to obtain the fourth pulmonary vascular tree image; merging the fourth pulmonary vascular tree image and the first pulmonary arteriovenous trunk image to obtain the third pulmonary vascular tree image, wherein, during the merging process, if there is a second overlapping region between the fourth pulmonary vascular tree image and the first pulmonary arteriovenous trunk image, the portion of the second overlapping region in the first pulmonary arteriovenous trunk image is retained in the third pulmonary vascular tree image.

[0080] The neighborhood search module 114 is used to determine the boundary points of the main vascular region in the third pulmonary vascular tree image based on the first pulmonary arteriovenous trunk image, and use the boundary points as initial seed points for neighborhood search, and add target labels to each voxel point in the third pulmonary vascular tree image that belongs to the peripheral vascular region, wherein the target labels are used to characterize the arteriovenous type.

[0081] Optionally, determining the boundary points of the main pulmonary artery and vein region in the third pulmonary vascular tree image based on the first pulmonary artery and vein main trunk image includes: performing an erosion operation on the first pulmonary artery and vein main trunk image to obtain a second pulmonary artery and vein main trunk image, wherein the erosion operation is used to eliminate voxel points of the artery and vein main trunk boundary in the first pulmonary artery and vein main trunk image; performing an intersection operation on the second pulmonary artery and vein main trunk image and the first pulmonary artery and vein main trunk image to obtain a target intersection image, wherein the target intersection image contains the common voxel points of the second pulmonary artery and vein main trunk image and the first pulmonary artery and vein main trunk image; performing an inversion operation on the target intersection image, and determining the voxel points obtained after the inversion operation as the boundary points of the main pulmonary vascular region.

[0082] Optionally, the target markers include: a first marker for characterizing the artery corresponding to the voxel point, a second marker for characterizing the vein corresponding to the voxel point, and a third marker for characterizing the voxel point that was not found. The boundary point is used as the initial seed point for neighborhood search. Adding target markers to each voxel point belonging to the peripheral vascular region in the third lung vascular tree image includes: determining the arteriovenous type corresponding to the initial seed point as the arteriovenous type of the initial seed point, wherein the arteriovenous type includes: artery and vein; obtaining a set of voxel points within a preset neighborhood of the initial seed point in the third lung vascular tree image; determining the target markers corresponding to each voxel point of the vascular tree within the preset neighborhood based on the distribution of voxel points in the voxel point set; determining voxel points within the preset neighborhood that have been marked with the first or second marker as new seed points and continuing the neighborhood search until no new seed points are found; after the neighborhood search is completed, adding the third marker to the voxel points of the vascular tree in the third lung vascular tree image that have not been marked with the target marker.

[0083] Optionally, the target marker further includes: a fourth marker for characterizing voxel points whose arterial / venous type cannot be determined; determining the target marker corresponding to each voxel point of the vascular tree within a preset neighborhood based on the distribution of voxel points in the voxel point set includes: adding a first marker to voxel points within the preset neighborhood whose arterial / venous type is undetermined when the seed point's arterial / venous type is arterial and the preset neighborhood does not contain voxel points whose arterial / venous type is vein; adding a second marker to voxel points within the preset neighborhood whose arterial / venous type is undetermined when the seed point's arterial / venous type is vein and the preset neighborhood does not contain voxel points whose arterial / venous type is arterial; adding a fourth marker to voxel points within the preset neighborhood whose arterial / venous type is undetermined when the seed point's arterial / venous type is arterial and both voxel points whose arterial / venous type is vein and voxel points whose arterial / venous type is undetermined exist within the preset neighborhood; adding a fourth marker to voxel points within the preset neighborhood whose arterial / venous type is undetermined when the seed point's arterial / venous type is vein and both voxel points whose arterial / venous type is arterial and voxel points whose arterial / venous type is undetermined exist within the preset neighborhood.

[0084] Optionally, the neighborhood search module 114 is further configured to: send the third pulmonary vascular tree image with added target markers to the front-end interactive interface for display; and, in response to the operation instructions of the front-end interactive interface, determine the arterial and venous types corresponding to the voxel points of the third marker and / or the fourth marker to obtain the target pulmonary vascular tree image, wherein the operation instructions are used to delete or correct the target markers corresponding to the voxel points of the third pulmonary vascular tree image.

[0085] It should be noted that each module in the above-mentioned pulmonary vascular marking device can be a program module (e.g., a set of program instructions to implement a certain function) or a hardware module. For the latter, it can be manifested in the following forms, but is not limited to them: each of the above modules is manifested as a processor, or the functions of each of the above modules are implemented by a processor.

[0086] It should be noted that the pulmonary vascular marking device provided in this embodiment can be used to perform... Figure 1 The lung vascular marking method shown above is also applicable to the embodiments of this application, and will not be repeated here.

[0087] This application embodiment also provides a non-volatile storage medium storing a computer program. When the computer program is executed by a processor, it implements the following lung vascular labeling method: determining a first overlapping region between a first pulmonary bronchus image and a first pulmonary vascular tree image, and removing the vascular tree images located within the first overlapping region from the first pulmonary vascular tree image to obtain a second pulmonary vascular tree image, the second pulmonary vascular tree image including peripheral vascular regions with a diameter not greater than a preset diameter threshold; merging the second pulmonary vascular tree image and the first pulmonary arteriovenous trunk image to obtain a third pulmonary vascular tree image, wherein the first pulmonary arteriovenous trunk image includes trunk vascular regions with a diameter greater than a preset diameter threshold and arteriovenous types of each voxel point belonging to the trunk vascular region; determining the boundary points of the trunk vascular region in the third pulmonary vascular tree image based on the first pulmonary arteriovenous trunk image, and using the boundary points as initial seed points for neighborhood search, adding target labels to each voxel point belonging to the peripheral vascular region in the third pulmonary vascular tree image, wherein the target labels are used to characterize the arteriovenous type.

[0088] This application also provides a computer terminal (or electronic device) in its embodiments. Figure 12 A hardware block diagram of a computer terminal (or electronic device) for implementing a method for labeling pulmonary vessels is shown. Figure 12As shown, the computer terminal 120 (or electronic device) may include one or more processors 1202 (shown as 1202a, 1202b, ..., 1202n in the figure) (processor 1202 may include, but is not limited to, a microprocessor MCU or a programmable logic device FPGA, etc.), a memory 1204 for storing data, and a transmission device 1206 for communication functions. In addition, it may also include: a display, an input / output interface (I / O interface), a universal serial bus (USB) port (which may be included as one of the ports of a BUS bus), a network interface, a power supply, and / or a camera. Those skilled in the art will understand that... Figure 12 The structure shown is for illustrative purposes only and does not limit the structure of the aforementioned electronic device. For example, the computer terminal 120 may also include... Figure 12 The more or fewer components shown, or having the same Figure 12 The different configurations shown.

[0089] It should be noted that the aforementioned one or more processors 1202 and / or other data processing circuits are generally referred to herein as "data processing circuits". These data processing circuits may be embodied, in whole or in part, in software, hardware, firmware, or any other combination thereof. Furthermore, the data processing circuits may be a single, independent processing module, or may be wholly or partially integrated into any other element within the computer terminal 120 (or electronic device). As involved in the embodiments of this application, the data processing circuits serve as a processor control mechanism (e.g., selection of a variable resistor termination path connected to an interface).

[0090] The memory 1204 can be used to store software programs and modules of application software, such as the program instructions / data storage device corresponding to the lung vascular marking method in this embodiment. The processor 1202 executes various functional applications and data processing by running the software programs and modules stored in the memory 1204, thereby implementing the aforementioned lung vascular marking method. The memory 1204 may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 1204 may further include memory remotely located relative to the processor 1202, and these remote memories can be connected to the computer terminal 120 via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0091] The transmission device 1206 is used to receive or send data via a network. Specific examples of the network described above may include a wireless network provided by the communication provider of the computer terminal 120. In one example, the transmission device 1206 includes a Network Interface Controller (NIC), which can connect to other network devices via a base station to communicate with the Internet. In another example, the transmission device 1206 may be a Radio Frequency (RF) module, used for wireless communication with the Internet.

[0092] The display may be, for example, a touchscreen liquid crystal display (LCD) that allows the user to interact with the user interface of the computer terminal 120 (or electronic device).

[0093] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or modules may be electrical, mechanical, or other forms.

[0094] The modules described as separate components may or may not be physically separate. Similarly, the components shown as modules may or may not be physical modules; they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0095] Furthermore, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module. The integrated modules described above can be implemented in hardware or as software functional modules.

[0096] If the integrated module is implemented as a software functional module and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to related technologies, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a readable storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned readable storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, ROM, RAM, magnetic disks, or optical disks.

[0097] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0098] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0099] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0100] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

Claims

1. A method for marking pulmonary blood vessels, characterized in that, include: A first overlapping region is determined between a first lung bronchus image and a first lung vascular tree image, and images of vascular trees located within the first overlapping region in the first lung vascular tree image are removed to obtain a second lung vascular tree image, wherein the second lung vascular tree image includes a peripheral vascular region with a diameter not greater than a preset diameter threshold. The second pulmonary vascular tree image and the first pulmonary arteriovenous trunk image are merged to obtain a third pulmonary vascular tree image, wherein the first pulmonary arteriovenous trunk image includes a trunk vascular region with a diameter greater than the preset diameter threshold and the arteriovenous types of each voxel point belonging to the trunk vascular region; Based on the first pulmonary arteriovenous trunk image, the boundary points of the trunk vessel region in the third pulmonary vascular tree image are determined, and the boundary points are used as initial seed points for neighborhood search. Target markers are added to each voxel point in the third pulmonary vascular tree image that belongs to the peripheral vessel region, wherein the target markers are used to characterize the arteriovenous type.

2. The method of claim 1, wherein, By merging the second pulmonary vascular tree image and the first pulmonary arteriovenous trunk image, the third pulmonary vascular tree image is obtained, including: Determine the Hu value of each voxel in the second pulmonary vascular tree image, and delete voxel points in the second pulmonary vascular tree image whose Hu value is greater than a preset threshold value to obtain the fourth pulmonary vascular tree image; The fourth pulmonary vascular tree image and the first pulmonary arteriovenous trunk image are merged to obtain the third pulmonary vascular tree image. During the merging process, if there is a second overlapping region between the fourth pulmonary vascular tree image and the first pulmonary arteriovenous trunk image, the portion of the second overlapping region in the first pulmonary arteriovenous trunk image is retained in the third pulmonary vascular tree image.

3. The method of claim 1, wherein the step of identifying the pulmonary vessel is performed by a computer. Based on the first pulmonary arteriovenous trunk image, determining the boundary points of the trunk vessel region in the third pulmonary vascular tree image includes: An erosion operation is performed on the first pulmonary arteriovenous trunk image to obtain a second pulmonary arteriovenous trunk image, wherein the erosion operation is used to eliminate voxel points at the boundaries of the arteriovenous trunk in the first pulmonary arteriovenous trunk image; An intersection operation is performed on the second pulmonary arteriovenous trunk image and the first pulmonary arteriovenous trunk image to obtain a target intersection image, wherein the target intersection image contains common pixels of the second pulmonary arteriovenous trunk image and the first pulmonary arteriovenous trunk image; The target intersection image is inverted, and the voxel points obtained after the inversion operation are determined as the boundary points of the main blood vessel region.

4. The method of claim 3, wherein, The target markers include: a first marker for characterizing the artery corresponding to the voxel point, a second marker for characterizing the vein corresponding to the voxel point, and a third marker for characterizing the voxel point not being searched; the boundary points are used as initial seed points for neighborhood search, and adding target markers to each voxel point belonging to the peripheral vascular region in the third lung vascular tree image includes: The arterial and venous type corresponding to the initial seed point is determined as the arterial and venous type of the initial seed point, wherein the arterial and venous type includes: artery and vein; Obtain the set of voxel points located within a preset neighborhood of the initial seed point in the third lung vascular tree image; Based on the distribution of voxels in the voxel set, the target markers corresponding to each voxel of the vascular tree within the preset neighborhood are determined; The voxel points within the preset neighborhood range that have been marked with the first or second label are identified as new seed points and the neighborhood search continues until no new seed points can be found. After the neighborhood search is completed, the third label is added to the voxel points of the vascular tree in the third lung vascular tree image that have not been labeled with the target label.

5. The method for marking pulmonary vessels according to claim 4, characterized in that, The target marker further includes: a fourth marker for characterizing that the arterial / venous type of the voxel point cannot be determined; the target marker corresponding to each voxel point of the vascular tree within the preset neighborhood range is determined based on the distribution of voxel points in the voxel point set, including: If the artery / vein type of the seed point is artery, and there are no voxel points of the artery / vein type being vein within the preset neighborhood, the first marker is added to the voxel points within the preset neighborhood where the artery / vein type is not determined. If the artery / vein type of the seed point is vein, and there are no voxel points of the artery / vein type being artery within the preset neighborhood, add the second marker to the voxel points within the preset neighborhood where the artery / vein type is not determined; If the artery / vein type of the seed point is artery, and there are voxel points with the artery / vein type of vein and voxel points with the artery / vein type not determined within the preset neighborhood, the fourth marker is added to the voxel points with the artery / vein type not determined within the preset neighborhood. If the artery / vein type of the seed point is vein, and there are voxel points with the artery / vein type being artery and voxel points with the artery / vein type not determined within the preset neighborhood, the fourth marker is added to the voxel points with the artery / vein type not determined within the preset neighborhood.

6. The method of claim 5, wherein the step of identifying the pulmonary blood vessels is performed by a computer. The method further includes: The third pulmonary vascular tree image with the target marker added is sent to the front-end interactive interface for display. In response to the operation instructions of the front-end interactive interface, the arterial and venous types corresponding to the voxel points of the third marker and / or the fourth marker are determined to obtain a target pulmonary vascular tree image, wherein the operation instructions are used to delete or correct the target markers corresponding to the voxel points of the third pulmonary vascular tree image.

7. The method of claim 1, wherein, Before determining the overlapping region between the first bronchial image and the first pulmonary vascular tree image, the method further includes: A deep learning segmentation network is used to segment the computed tomography images of the lungs to obtain the original lung bronchial images, the original lung vascular tree images, and the first pulmonary arteriovenous trunk images. A pre-defined dilation template is used to dilate the original lung bronchial image to obtain the first lung bronchial image. The pre-defined dilation template is a convolutional kernel of a pre-defined size. The dilation operation is used to expand the lung bronchi in the original lung bronchial image to the airway walls. The original pulmonary vascular tree image is subjected to an erosion operation using a preset erosion template to obtain the first pulmonary vascular tree image. The preset erosion template is a convolution kernel of a preset size, and the erosion operation is used to eliminate noise points in the original pulmonary vascular tree image.

8. A lung vessel labeling apparatus, characterized by comprising: include: The tracheal noise elimination module is used to determine the first overlapping region between the first lung bronchus image and the first lung vascular tree image, and remove the images of the vascular trees located in the first overlapping region in the first lung vascular tree image to obtain a second lung vascular tree image. The second lung vascular tree image includes a peripheral vascular region with a diameter not greater than a preset diameter threshold. The image merging module is used to merge the second pulmonary vascular tree image and the first pulmonary arteriovenous trunk image to obtain a third pulmonary vascular tree image. The first pulmonary arteriovenous trunk image includes a trunk vascular region with a diameter greater than the preset diameter threshold and the arteriovenous types of each voxel point belonging to the trunk vascular region. The neighborhood search module is used to determine the boundary points of the main vascular region in the third pulmonary vascular tree image based on the first pulmonary arteriovenous trunk image, and to use the boundary points as initial seed points for neighborhood search, and to add target labels to each voxel point in the third pulmonary vascular tree image that belongs to the peripheral vascular region, wherein the target labels are used to characterize the arteriovenous type.

9. A non-volatile storage medium, comprising: The non-volatile storage medium stores a computer program that, when executed by a processor, implements the pulmonary vascular marking method according to any one of claims 1 to 7.

10. An electronic device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the lung vascular marking method as described in any one of claims 1 to 7.

Images