Vascular structure root node extraction method and device, computer equipment and storage medium

Abstract

The invention relates to a vessel structure root node extraction method and device, computer equipment and a storage medium. The method comprises the following steps: carrying out center line extraction on a bronchus segmentation mask of a target object to obtain a bronchus center line corresponding to the bronchus segmentation mask; according to the bronchial tube center line, determining a root bronchial tube position corresponding to the bronchial tube segmentation mask; according to the bronchial tube center line and the root bronchial tube position, generating a bronchial tube topological structure corresponding to the bronchial tube segmentation mask; and identifying at least one vascular structure root node in the bronchial segment segmentation mask according to the bronchial topological structure. By adopting the method, the segmentation accuracy of the lung image can be improved.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a method, apparatus, computer equipment, storage medium and computer program product for extracting root nodes of vascular structures. Background Art

[0002] There are usually two methods for the problem of segmenting lung images. The first method uses an end-to-end lung image segmentation algorithm based on deep learning. This method does not require a complex image processing process, but the results lack interpretability and are not convenient to optimize or adjust when the results are not ideal. The second method is to determine the lung image segmentation interface by combining the segmentation information of arterial or venous images, or by combining the segmentation information of bronchial images. However, there is a defect in both methods. The lung image segmentation interface may not pass through the root node of the vascular segmentation, resulting in low segmentation accuracy of the lung image. Summary of the invention

[0003] Based on this, it is necessary to provide a method, apparatus, computer device, computer-readable storage medium and computer program product for extracting root nodes of vascular structures that can improve the segmentation accuracy of lung images in response to the above technical problems.

[0004] In the first aspect, the present application provides a method for extracting a root node of a vascular structure. The method comprises: performing centerline extraction on a bronchial segmentation mask of a target object to obtain a bronchial centerline corresponding to the bronchial segmentation mask; determining the root bronchial position corresponding to the bronchial segmentation mask according to the bronchial centerline; generating a bronchial topological structure corresponding to the bronchial segmentation mask according to the bronchial centerline and the root bronchial position; and identifying at least one vascular structure root node in the bronchial segmentation mask according to the bronchial topological structure.

[0005] In the second aspect, the present application also provides a device for extracting the root node of a vascular structure. The device includes: a bronchial centerline extraction module, which is used to extract the centerline of the bronchial segmentation mask of the target object to obtain the bronchial centerline corresponding to the bronchial segmentation mask; a bronchial position determination module, which is used to determine the root bronchial position corresponding to the bronchial segmentation mask according to the bronchial centerline; a bronchial topology structure generation module, which is used to generate the bronchial topology corresponding to the bronchial segmentation mask according to the bronchial centerline and the root bronchial position; and a vascular structure root node identification module, which is used to identify at least one vascular structure root node in the bronchial segmentation mask according to the bronchial topology.

[0006] In a third aspect, the present application also provides a computer device. The computer device includes a memory and a processor, the memory stores a computer program, and the processor implements the following steps when executing the computer program: extracting the centerline of the bronchial segmentation mask of the target object to obtain the bronchial centerline corresponding to the bronchial segmentation mask; determining the root bronchial position corresponding to the bronchial segmentation mask according to the bronchial centerline; generating the bronchial topology corresponding to the bronchial segmentation mask according to the bronchial centerline and the root bronchial position; and identifying at least one vascular structure root node in the bronchial segmentation mask according to the bronchial topology.

[0007] In a fourth aspect, the present application also provides a computer-readable storage medium. The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the following steps are implemented: extracting the centerline of the bronchial segmentation mask of the target object to obtain the bronchial centerline corresponding to the bronchial segmentation mask; determining the root bronchial position corresponding to the bronchial segmentation mask according to the bronchial centerline; generating the bronchial topology corresponding to the bronchial segmentation mask according to the bronchial centerline and the root bronchial position; and identifying at least one vascular structure root node in the bronchial segmentation mask according to the bronchial topology.

[0008] In a fifth aspect, the present application also provides a computer program product. The computer program product includes a computer program, which implements the following steps when executed by a processor: extracting the centerline of the bronchial segmentation mask of the target object to obtain the bronchial centerline corresponding to the bronchial segmentation mask; determining the root bronchial position corresponding to the bronchial segmentation mask according to the bronchial centerline; generating the bronchial topology corresponding to the bronchial segmentation mask according to the bronchial centerline and the root bronchial position; and identifying at least one vascular structure root node in the bronchial segmentation mask according to the bronchial topology.

[0009] The above-mentioned method, device, computer equipment, storage medium and computer program product for extracting the root node of a vascular structure perform centerline extraction on the bronchial segmentation mask of the target object to obtain the bronchial centerline corresponding to the bronchial segmentation mask; determine the root bronchial position corresponding to the bronchial segmentation mask based on the bronchial centerline; generate the bronchial topological structure corresponding to the bronchial segmentation mask based on the bronchial centerline and the root bronchial position; and identify at least one vascular structure root node in the bronchial segmentation mask based on the bronchial topological structure.

[0010] By extracting the bronchial centerline corresponding to the bronchial segmentation mask of the target object and combining it with the topological structure generation algorithm, the bronchial topological structure corresponding to the bronchial segmentation mask is obtained; further, by searching the bronchial topological structure, the vascular structure root node is identified, and in the case of complex structures such as bronchi and blood vessels, at least one bronchial segmentation structure common vascular structure root node can be accurately extracted, which helps to improve the segmentation accuracy of lung images. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Figure 1 A diagram of an application environment of a method for extracting a root node of a vascular structure in one embodiment;

[0012] Figure 2 A schematic diagram of a flow chart of a method for extracting a root node of a vascular structure in an embodiment;

[0013] Figure 3 A schematic diagram of a process for obtaining a bronchial topological structure in one embodiment;

[0014] Figure 4 A schematic diagram of a flow chart of a method for determining a bronchial growth direction in one embodiment;

[0015] Figure 5 A schematic diagram of a flow chart of a method for obtaining a bronchial topological structure in another embodiment;

[0016] Figure 6 A schematic diagram of a flow chart of a method for obtaining a vascular structure root node in one embodiment;

[0017] Figure 7 It is a flowchart of a method for obtaining a bronchial segment root node branch in one embodiment;

[0018] Figure 8 A schematic diagram of a flow chart of a method for obtaining a bronchial segment root node branch in another embodiment;

[0019] Fig. 9 It is a flowchart of a method for obtaining a bronchial segment root node branch in another embodiment;

[0020] Fig.10 A schematic diagram of a flow chart of a method for obtaining a target common parent node in one embodiment;

[0021] Fig.11 A schematic diagram of a flow chart of a method for obtaining a target common parent node in another embodiment;

[0022] Fig.12 A schematic flow chart of a method for obtaining a vascular structure root node in another embodiment;

[0023] Fig.13A schematic diagram of a flow chart of a bronchial topology verification method in one embodiment;

[0024] Fig.14 is a schematic diagram of a bronchial segmentation mask in one embodiment;

[0025] Fig.15 A schematic diagram of the extraction result of the bronchial centerline in one embodiment;

[0026] Fig.16 A schematic diagram of various markings for points on a center line in one embodiment;

[0027] Fig.17 A schematic diagram of the extraction result of a root bronchus in one embodiment;

[0028] Fig.18 A schematic diagram of the bronchial topological structure in one embodiment;

[0029] Fig.19 is a schematic diagram of the recognition result of the root node of the vascular structure in one embodiment;

[0030] Fig. 20 A schematic diagram of a bronchial segment interface in one embodiment;

[0031] Fig.21 A schematic diagram of Strahler classification results in one embodiment;

[0032] Fig. 22 It is a structural block diagram of a vascular structure root node extraction device in one embodiment;

[0033] Fig.23 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0034] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0035] The present invention provides a method for extracting a vascular structure root node, which can be applied to Figure 1In the application environment shown. Among them, the terminal 102 communicates with the server 104 through the network. The data storage system can store the data that the server 104 needs to process. The data storage system can be integrated on the server 104, or it can be placed on the cloud or other network servers. The server 104 obtains the bronchial segmentation mask of the target object from the terminal 102, and further uses the server 104 to extract the centerline of the bronchial segmentation mask of the target object to obtain the bronchial centerline corresponding to the bronchial segmentation mask; according to the bronchial centerline, determine the root bronchial position corresponding to the bronchial segmentation mask; according to the bronchial centerline and the root bronchial position, generate the bronchial topology corresponding to the bronchial segmentation mask; according to the bronchial topology, identify at least one vascular structure root node in the bronchial segmentation mask. Among them, the terminal 102 can be but is not limited to various Internet of Things devices and portable wearable devices. The server 104 can be implemented with an independent server or a server cluster consisting of multiple servers.

[0036] In one embodiment, Figure 2 As shown, a method for extracting the root node of a vascular structure is provided, and the method is applied to Figure 1 The server in the example is used to illustrate the following steps:

[0037] Step 202 : extracting the centerline of the bronchus segmentation mask of the target object to obtain the bronchus centerline corresponding to the bronchus segmentation mask.

[0038] The target object may be a human or an animal, wherein the target object includes a region of interest, such as the liver, kidneys, lungs, etc. in the human body.

[0039] The bronchial segmentation mask may be an image obtained by segmenting the bronchial segmentation structure of the lungs, wherein the bronchial segmentation mask is a binary image composed of 0 and 1, such as Fig.13 The bronchial segmentation mask shown in FIG. 1 can define the image bronchial segmentation mask by specifying data values, data ranges, finite or infinite values, regions of interest, and annotation files. Fig.14 As shown, in some embodiments, Fig.14 The main part of the bronchial segmentation mask shown can be represented by color A (ie, any color, such as yellow, red, green, blue, black, etc.), while other bronchial segment structures can be represented by colors different from color A.

[0040] The bronchial centerline may be a set of center points obtained by traversing the axes of each bronchus in the bronchial segmentation mask.

[0041] Specifically, the server 104 obtains the bronchial segmentation mask of the target object from the terminal 102. The server 104 extracts the centerline of the bronchial segmentation mask by using an algorithm based on topological refinement to obtain the bronchial centerline corresponding to the bronchial segmentation mask. The extraction result of the bronchial centerline is as follows: Fig.15 The bronchial centerline shown, in some embodiments, as Fig.15 The bronchial centerline of the main trunk can be represented by color A (i.e., any color, such as yellow, red, green, blue, black, etc.), while the bronchial centerlines of other bronchial segment structures can be represented by colors different from color A; multiple center points on the axis of the bronchus, including the centerline endpoints, centerline bifurcation points, and centerline continuity points of the bronchial centerline. Fig.16 In the example of the bronchus shown, it can be seen that the axis of the bronchus has a centerline, the end of the centerline is the centerline endpoint, the point where the centerline bifurcates is the centerline bifurcation point, and the continuous points on the centerline are the centerline continuous points; further, as Fig.16 Various marks of the points on the center line are shown, the connection point of the center line endpoint is 1, the connection point of the center line bifurcation point is greater than or equal to 3, and the connection point of the center line continuous point is 2.

[0042] Step 204: determine the root bronchus position corresponding to the bronchus segmentation mask according to the bronchus centerline.

[0043] The root bronchus position may be the position information of the root bronchus of the lung, including the starting position, the end position, the curve point position, the radius and the average radius, etc. Fig.13 The bronchus segmentation mask shown, the black part in the middle is the root bronchus.

[0044] Specifically, all centerline branches are obtained through the information of the bronchial centerline, and the information of all bronchial branches is counted, including the starting point, end point, curve point, radius and average radius, etc.; that is, the root bronchus corresponding to the root node of the entire bronchus is extracted by searching the centerline endpoint, centerline bifurcation point and centerline continuous point of the bronchial centerline, as well as the information of the centerline endpoint, the information of the centerline bifurcation point and the information of the centerline continuous point. The root bronchus is used as the starting position for the subsequent extraction of the bronchial topological structure, and the branch with the largest weighted sum of the average radius and the length of the branch is used as the root node of the root bronchus. The extraction result of the root bronchus and the position of the root bronchus are shown in FIG. Fig.17 As shown, in some embodiments, Fig.17The extraction result of the root bronchus and the position of the root bronchus shown can be represented by color A (i.e., any color, such as yellow, red, green, blue, black, etc.), while other bronchial segment structures can be represented by colors different from color A.

[0045] Step 206: Generate a bronchial topological structure corresponding to the bronchial segmentation mask according to the bronchial centerline and the root bronchial position.

[0046] Among them, the bronchial topology structure can be a structure that studies the geometric figures of the pulmonary bronchi or some properties of the space that can remain unchanged after continuous changes in shape. It only considers the positional relationship of the pulmonary bronchi without considering their shape and size.

[0047] Specifically, the bronchial "growth direction" in the bronchial topological structure is confirmed by the position of the root bronchus and the centerline of the bronchus, that is, the edges of all bronchi should be directional. When extracting the topological structure, it is necessary to first determine the extension direction starting from the root node of the root bronchus; when the extension direction is determined, based on the centerline of the bronchus, a traversal algorithm is used to search for the bronchial segmentation mask of the target object to obtain the bronchial topological structure, where the bronchial topological structure is as follows: Fig.18 As shown, in some embodiments, as Fig.18 Different branches of the shown bronchial topology may be represented by different colors.

[0048] Step 208: Identify at least one vascular structure root node in the bronchial segmentation mask according to the bronchial topology.

[0049] The vascular structure root node may be a bifurcation point at the boundary between two associated bronchial segments in the lung structure.

[0050] Specifically, when the bronchial topology structure corresponding to the bronchial segmentation mask is determined, all target common parent nodes in the bronchial topology structure are searched, that is, the target common parent nodes corresponding to the boundaries of the two associated bronchial segments in the lung structure, and further the endpoints with a large number of connections are selected from the target common parent nodes as the root nodes of each vascular structure corresponding to the bronchial segmentation mask. Fig.19 is the recognition result of each vascular structure root node in the bronchial segmentation mask; wherein, Fig.19 The center of each white circle is a different root node of the vascular structure.

[0051] In the above-mentioned method for extracting the root node of the vascular structure, the centerline of the bronchial segmentation mask of the target object is extracted to obtain the bronchial centerline corresponding to the bronchial segmentation mask; based on the bronchial centerline, the root bronchial position corresponding to the bronchial segmentation mask is determined; based on the bronchial centerline and the root bronchial position, the bronchial topological structure corresponding to the bronchial segmentation mask is generated; based on the bronchial topological structure, at least one vascular structure root node in the bronchial segmentation mask is identified.

[0052] By extracting the bronchial centerline corresponding to the bronchial segmentation mask of the target object and combining it with the topological structure generation algorithm, the bronchial topological structure corresponding to the bronchial segmentation mask is obtained; further, by searching the bronchial topological structure, the vascular structure root node is identified, and in the case of complex structures such as bronchi and blood vessels, at least one bronchial segmentation structure common vascular structure root node can be accurately extracted, which helps to improve the segmentation accuracy of lung images.

[0053] In one embodiment, Figure 3 As shown, according to the bronchial centerline and the root bronchial position, the bronchial topological structure corresponding to the bronchial segmentation mask is generated, including:

[0054] Step 302, determining the bronchial growth direction according to the root bronchial position and the bronchial centerline.

[0055] The bronchial growth direction may be an extension direction starting from the root node of the root bronchus.

[0056] Specifically, since the set of each bronchi corresponds to the bronchial centerline, each bronchus has a corresponding bronchial centerline. Then, the centerline corresponding to the root bronchus is extracted from the bronchial centerline according to the position of the root bronchus, that is, the root bronchial centerline, wherein the root bronchial centerline has two endpoints, and the number of connection points of the two endpoints is not the same. Further, the bronchial growth direction is determined according to the size relationship of the number of connection points of the two endpoints of the root bronchial centerline. Fig.17 As shown, Fig.17 The direction indicated by the arrow in the figure is the growth direction of the bronchus, wherein the number of connection points in the positive direction indicated by the arrow is greater than that in the negative direction indicated by the arrow.

[0057] Step 304 , taking the bronchial root node of the root bronchial centerline as the starting point, searching for the bronchial topological structure according to the bronchial growth direction.

[0058] The bronchial root node may be one of the two centerline endpoints of the root bronchus, and the specific decision needs to be made according to the number of connection points of the centerline endpoints and the growth direction of the bronchus.

[0059] Specifically, the endpoint with more connection points among the two endpoints of the bronchial centerline is selected as the bronchial root node (starting point) for generating the bronchial topology structure, and the bronchus connected to the root bronchus is searched according to the bronchial growth direction until the other end bifurcation point of the first bronchus connected is found (such as Fig.17 Similarly, the bifurcation point at the other end of the connected first bronchus is taken as the next starting point, and according to the bronchial growth direction, the search for the second bronchus connected to the next starting point is performed again until the bifurcation point at the other end of the connected second bronchus is found, and so on, until no bifurcation point appears at the other end of the Nth bronchus, then the search is stopped to obtain the bronchial topology structure.

[0060] In this embodiment, the bronchial growth direction is determined by the root bronchial position and the bronchial centerline, and the bronchial root node is further used as the starting point to search for the bronchial topological structure according to the bronchial growth direction. This can provide an accurate growth starting point and growth direction of the bronchial topological structure, ensure the smooth generation of the bronchial topological structure, and improve the growth efficiency of the bronchial topological structure.

[0061] In one embodiment, Figure 4 As shown in the figure, the direction of bronchial growth is determined according to the position of the root bronchus and the bronchial centerline, including:

[0062] Step 402, selecting the root bronchus centerline from the bronchus centerlines according to the root bronchus position information.

[0063] The root bronchus centerline may be a portion of the bronchial centerline corresponding to the root bronchus.

[0064] Specifically, according to the position information of the bronchus, the partial centerline of the bronchial centerline corresponding to the root bronchus is located; since the bronchial centerline has a set of multiple center points, a subset of the center points of the partial centerline corresponding to the root bronchus is further extracted as the root bronchial centerline corresponding to the root bronchus.

[0065] Step 404, determining the bronchial growth direction according to the centerline endpoint information of the root bronchial centerline.

[0066] The centerline endpoint information may be the number of connection points between two endpoints of the root bronchus centerline.

[0067] Specifically, the first and last points of the root bronchial centerline of the root bronchus are usually the end point and the multi-branch intersection point, that is, the connection point of the starting point is equal to 1, and the connection point of the end point is greater than or equal to 3. The total number of connection points is as follows Fig.16As shown. At this time, the growth direction of the root bronchus should be from the point with connection point 1 to the point with connection point 3. However, sometimes due to the presence of a small protrusion in the root bronchus, the protrusion becomes a separate branch when the center line is extracted, and the root bronchus is divided into multiple branches by the protrusion. When the bronchial root node is determined by methods such as the average radius, the middle parts of multiple branches of the bronchial root node may be extracted. At this time, the connection points of the starting point and the end point are greater than or equal to 3, and the starting point and the end point cannot be determined by the value of the connection point. Therefore, the bronchial growth direction is determined by calculating the total number of points connected to the two end points of the root bronchial center line of the root bronchus, where the end point with a smaller total number of connection points is the starting point of the bronchial growth direction, and the end point with a larger total number of connection points is the end point of the bronchial growth direction.

[0068] In this embodiment, by selecting the root bronchial centerline from the bronchial centerlines based on the root bronchial position information, and further determining the bronchial growth direction based on the centerline endpoint information of the root bronchial centerline, the difference between the two different centerline endpoint information can be clearly identified, and the bronchial growth direction for generating the bronchial topological structure can be accurately obtained.

[0069] In one embodiment, Figure 5 As shown in the figure, the bronchial root node of the root bronchial centerline is used as the starting point, and the bronchial topology structure is searched according to the bronchial growth direction, including:

[0070] Step 502: Use the bronchial root node as the branch parent node.

[0071] The branch parent node may be a common node of the edges of all nodes of any bronchial branch structure.

[0072] Specifically, the end point of the root bronchus centerline with the majority of connection points is taken as the bronchus root node, and combined with the bronchial growth direction, the bronchus root node is used as the starting point for searching the topological structure of each bronchi adjacent to the root bronchus, that is, as the branch parent node for searching the topological structure of each bronchi adjacent to the root bronchus.

[0073] Step 504 , starting from the branch parent node, extending in the bronchial growth direction to the bifurcation point, and obtaining the branch child node.

[0074] The branch child node may be a node corresponding to an outgoing edge of any common node of the bronchial branch structure.

[0075] Specifically, the parent node of the branch is used as the starting point to determine the initial search node for the search of the bronchial branch structure, and the bronchial growth direction is used as the search direction. The current bronchial branch structure is searched by breadth-first traversal or depth-first traversal. That is, the initial search node for the search of the bronchial branch structure is extended in the bronchial growth direction until one of the bifurcation points of the bronchial branch structure is searched. The bifurcation point is used as the branch child node of the bronchial branch structure. The branch child node is as follows: Fig.18 As shown, it also serves as the end search node for the search of bronchial branch structures.

[0076] Step 506, taking the branch child node as a new branch parent node, returning to execute the step of taking the branch parent node as the starting point, extending to the bifurcation point in the bronchial growth direction, and obtaining the branch child node, until the bronchial topology structure is obtained.

[0077] Specifically, the branch child node of the last search of the bronchial branch structure is used as the branch parent node of the next search of the bronchial branch structure, that is, the end search node of the last search of the bronchial branch structure is used as the initial search node of the next search of the bronchial branch structure, and the step of "taking the bronchial growth direction as the search direction, searching the current bronchial branch structure by breadth-first traversal or depth-first traversal, that is, taking the initial search node of the search for the bronchial branch structure, extending along the bronchial growth direction until one of the bifurcation points of the bronchial branch structure is searched, and then taking the bifurcation point as the branch child node of the bronchial branch structure and also as the end search node of the search for the bronchial branch structure" is repeatedly executed, and multiple cycles of searching are performed until the number of connection points of the end search node of the search for the bronchial branch structure is 1, then the search is stopped, and the results of multiple searches are integrated to obtain the bronchial topological structure.

[0078] In this embodiment, by using a circular search method, taking the branch parent node as the starting point and extending to the bifurcation point in the bronchial growth direction, the bronchial topology structure is finally obtained. In the process of generating the bronchial topology structure, it can be accurate to the order of magnitude of the bifurcation point, which is beneficial to improving the data expression accuracy of the bronchial topology structure.

[0079] In one embodiment, Figure 6 As shown, according to the bronchial topology, identifying at least one vascular structure root node in the bronchial segmentation mask includes:

[0080] Step 602, searching for the structural information of each bronchus in the bronchial topology structure, and obtaining the root node branches of each bronchial segment.

[0081] The bronchial structure information may be the topological structure information of each bronchial branch structure in the bronchial topological structure.

[0082] The bronchial segment root node branch may be a common parent node of each bronchial branch structure, that is, a parent node of the bronchial segment structure.

[0083] Specifically, by searching the bronchial topology, the bronchial structure information of each bronchial branch structure in the bronchial topology is obtained. For different bronchial segment structures, each bronchial structure information or each bronchial branch structure has the same common parent node; further, based on different bronchial segment structures, by searching each bronchial structure information or each bronchial branch structure in the opposite direction of the bronchial growth direction, the common parent node of each bronchial structure information or each bronchial branch structure is obtained, which is one of the bronchial segment root node branches. For multiple bronchial segment structures in the entire bronchial topology, each bronchial segment root node branch is obtained.

[0084] Step 604 : Based on the preset bronchial segment association relationship, at least one bronchial segment root node branch is searched from the opposite direction of the bronchial growth direction to obtain a target common parent node.

[0085] The preset bronchial segment association relationship may be an association relationship between different bronchial segment structures in clinical data, wherein there is a relationship between the bronchial segment association relationship and the lung segment interface of the lung image, that is, two bronchial segments associated with each lung segment interface, and the lung segment interface is determined by decomposing a large number of lungs; wherein, if Fig. 20 Shown are the interfaces between the various lung segments in the lung image.

[0086] The target common parent node may be a plurality of common parent nodes branching from the root nodes of each bronchial segment in the same bronchial segment interface in the direction opposite to the bronchial growth direction.

[0087] Specifically, according to clinical experience, the lung image can be segmented into 13 lung segments, and the segmentation result of each lung segment corresponds to a bronchial segment interface. On the basis of the same bronchial segment interface, according to the opposite direction of the bronchial growth direction, with the root node branch of each bronchial segment as the starting point, the public parent node with the lowest node hierarchy relationship in the root node branch of each bronchial segment in the bronchial segment interface is searched as the target public parent node in the bronchial segment interface. For each bronchial segment interface, the same search method is used to obtain the target public parent node corresponding to each bronchial segment interface.

[0088] Step 606: Determine the vascular structure root node in the bronchial segmentation mask according to the target common parent node.

[0089] Specifically, based on the same bronchial segmentation interface, since the target common parent node has two endpoints and the number of connection points of the two endpoints is not the same, the number of connection points of the two endpoints of the target common parent node is obtained by traversing the target common parent node; further, the number of connection points of the two endpoints is compared, and the endpoint with more connection points is selected as the vascular structure root node in the bronchial segmentation mask. For each bronchial segmentation interface, the same method is used to find the vascular structure root node, and each vascular structure root node in the bronchial segmentation mask is obtained.

[0090] In this embodiment, by searching for the common target public parent node of each bronchial segment root node branch under the bronchial segment interface as a constraint condition, so as to determine the vascular structure root node in the target public parent node, the vascular structure root node can be better extracted, which helps to improve the quality of lung segmentation.

[0091] In one embodiment, Figure 7 As shown, the bronchial structure information of each bronchial topology is searched to obtain the root node branches of each bronchial segment, including:

[0092] Step 702: Calculate the label of each bronchial branch structure in the bronchial topological structure to obtain a label value of each branch.

[0093] The bronchial branch structure may be a structural representation of each bronchial branch in the lungs in the bronchial topological structure.

[0094] The branch label value may be a segmentation type describing each bronchial branch structure, that is, describing different locations of the current bronchial branch structure in lung anatomy (anatomical position, anatomical knowledge).

[0095] Specifically, in practice, the label values ​​of all points on the center line corresponding to each bronchial branch structure are generally the same, but different label values ​​may exist at special points on the center line. Therefore, the center lines corresponding to each bronchial branch structure in the bronchial topological structure are traversed to obtain the label values ​​of each point on the center line of each bronchial branch structure. The label value of the point with the highest proportion of the label values ​​of each point on the center line of each bronchial branch structure is used as the branch label value of the corresponding bronchial branch structure, where the branch label value represents the type of the bronchial branch structure.

[0096] Step 704: classify the corresponding bronchial branch structures according to the branch label values ​​to determine the segmented structures of each bronchus.

[0097] The bronchial segment structure may be a bronchial segment composed of at least one bronchial branch structure.

[0098] Specifically, because the branch label value of each bronchial branch structure is used to describe which bronchial segment structure each bronchial branch structure belongs to, each bronchial branch structure is classified according to each branch label value to obtain each bronchial segment structure in the bronchial topological structure, wherein each branch label value has a corresponding relationship with the bronchial branch structure.

[0099] Step 706, searching each bronchial segment structure to obtain the root node branch of each bronchial segment.

[0100] Specifically, for any bronchial segmentation structure, each parent node corresponding to each bronchial branch structure is searched in the opposite direction of the bronchial growth direction, and the overlapping parent node with the highest level is selected from each parent node as the bronchial segment root node branch of the bronchial segmentation structure. The same search method is used for each bronchial segmentation structure to obtain the bronchial segment root node branch corresponding to each bronchial segmentation structure.

[0101] In this embodiment, by calculating the branch label value of each bronchial branch structure in the bronchial topological structure, and further using the branch label value to classify each bronchial branch structure, so as to determine each bronchial segment root node branch based on the bronchial segment structure obtained after classification, each bronchial branch structure can be accurately classified to ensure the accuracy of obtaining the bronchial segment root node branch.

[0102] In one embodiment, Figure 8 As shown, each bronchial segment structure is searched to obtain the root node branches of each bronchial segment, including:

[0103] Step 802: Search each bronchial branch structure in any bronchial segment structure to obtain each ancestor branch node.

[0104] The ancestor branch node may be any multiple parent nodes of a bronchial branch structure, wherein a hierarchical relationship exists between the multiple parent nodes of the bronchial branch structure.

[0105] Specifically, based on any bronchial segment structure as the search unit, the parent node of each bronchial branch structure in the bronchial segment structure is searched in the opposite direction of the bronchial growth direction to obtain each ancestor branch node corresponding to each bronchial branch structure.

[0106] Step 804, traverse each ancestor branch node in the opposite direction of the bronchial growth direction to determine the bronchial segment root node branch.

[0107] Specifically, for any bronchial branch structure, since there is a corresponding relationship between the bronchial branch structure and at least one ancestor branch node, each target ancestor branch node corresponding to the bronchial branch structure is selected from each ancestor branch node. For example, the ancestor branch nodes of the bronchial branch structure with a label value of 1 are target ancestor branch node 1, target ancestor branch node 2, and target ancestor branch node 3; and the ancestor branch nodes of the bronchial branch structure with a label value of 2 are target ancestor branch node 1, target ancestor branch node 4, and target ancestor branch node 5. Further, each bronchial branch structure in the bronchial branch structure traverses each target ancestor branch node of the bronchial branch structure in the opposite direction of the bronchial growth direction, and the hierarchical relationship of each target ancestor branch node can be obtained. Finally, the parent node with the highest hierarchical relationship is selected from the overlapping target ancestor branch nodes as the bronchial segment root node branch of the bronchial branch structure. The above-mentioned search method is used for each bronchial branch structure to determine each bronchial segment root node branch.

[0108] In this embodiment, by traversing multiple ancestral branch nodes corresponding to each bronchial branch structure in the opposite direction of bronchial growth, the parent node with the highest seniority in the bronchial segment structure can be determined at each ancestral branch node, which helps to simplify the subsequent steps of obtaining the target common parent node of the bronchial segment structure.

[0109] In one embodiment, Fig. 9 As shown, each ancestor branch node is traversed in the opposite direction of the bronchial growth direction to determine the bronchial segment root node branch, including:

[0110] Step 902: determining at least one target ancestor branch node of each bronchial branch structure from each ancestor branch node.

[0111] The target ancestor branch node may be an ancestor branch node corresponding to the selected bronchial branch structure.

[0112] Specifically, since each bronchial branch structure has at least one ancestor branch node that has a corresponding relationship with the bronchial branch structure during the generation process of the bronchial topological structure, for each bronchial branch structure, the above-mentioned corresponding relationship is used to select each ancestor branch node corresponding to each bronchial branch structure from each ancestor branch node as each target ancestor branch node.

[0113] Step 904: traverse each target ancestor branch node in the opposite direction of the bronchial growth direction to determine the overlapping ancestor branch nodes of each bronchial branch structure.

[0114] The overlapping ancestor branch nodes may be parent nodes of the same part of each target ancestor branch node in different bronchial branch structures.

[0115] Specifically, for any bronchial branch structure, the search direction is opposite to the bronchial growth direction, and each target ancestor branch node in the bronchial branch structure is searched to obtain the relative generation relationship of each target ancestor branch node in the bronchial branch structure in the bronchial topological structure. For example, the ancestor branch node of the bronchial branch structure with a label value of 1 is the target ancestor branch node 1, the target ancestor branch node 2, and the target ancestor branch node 3. For each bronchial branch structure, the above-mentioned search method for the relative generation relationship in the bronchial topological structure is used to determine the relative generation relationship of each target ancestor branch node in each bronchial branch structure in the bronchial topological structure. Since the relative generation relationship of each target ancestor branch node of different bronchial branch structures in the bronchial topological structure may overlap, each target ancestor branch node with overlapping relative relationship is regarded as each overlapping ancestor branch node.

[0116] Step 906: determine the bronchial segment root node branch according to the node hierarchy relationship of each overlapping ancestor branch node.

[0117] The node hierarchical relationship may be a relationship between a parent node and a child node between respective overlapping ancestor branch nodes.

[0118] Specifically, the hierarchical relationship of each overlapping ancestral branch node is traversed to determine the node hierarchical relationship between different overlapping ancestral branch nodes. Since the node hierarchical relationship includes the generation of each overlapping ancestral branch node, the overlapping ancestral branch node with the highest generation among each overlapping ancestral branch node is determined through the node hierarchical relationship to serve as the bronchial segment root node branch.

[0119] In this embodiment, by traversing each target ancestor branch node in the opposite direction of the bronchial growth direction, the highest generation among the common overlapping ancestor branch nodes is obtained as the bronchial segment root node branch. The highest generation branch can be accurately located in the current bronchial segment structure, avoiding the main trunk branch.

[0120] In one embodiment, Fig.10 As shown, based on the preset bronchial segment association relationship, at least one bronchial segment root node branch is searched from the opposite direction of the bronchial growth direction to obtain the target common parent node, including:

[0121] Step 1002: Select at least one target bronchial segment root node branch in the bronchial segment interface from each bronchial segment root node branch.

[0122] The target bronchial segment root node branch may be each bronchial segment root node branch under the constraint of the bronchial segment interface.

[0123] Specifically, based on clinical data, 13 bronchial segment interfaces were defined in the lung images, such as Fig. 20 As shown, two bronchial segment structure label values ​​associated with each bronchial segment interface are set and saved in a predetermined order. In each bronchial segment interface, at least one bronchial segment root node branch with the bronchial segment interface as a constraint condition is selected from each bronchial segment root node branch as each target bronchial segment root node branch.

[0124] Step 1004, traverse the root node branches of each target bronchial segment in the opposite direction of the bronchial growth direction to obtain each overlapping common parent node.

[0125] The overlapping common parent node may be a common parent node shared by the root node branches of each target bronchial segment in the same bronchial segment interface.

[0126] Specifically, the search direction is opposite to the bronchial growth direction, and each target bronchial segment root node branch in the bronchial segment interface is searched to obtain the relative generation relationship of each target bronchial segment root node branch in the bronchial segment interface in the bronchial topology structure. Since the relative generation relationship of different target bronchial segment root node branches in the bronchial topology structure may overlap, each target bronchial segment root node branch with overlapping relative relationship is used as each overlapping common parent node. For each bronchial segment interface, the above-mentioned overlapping common parent node search method is used to determine each overlapping common parent node in each bronchial segment interface.

[0127] Step 1006: Determine the target common parent node according to the node hierarchy relationship of each overlapping common parent node.

[0128] The node hierarchical relationship may be a relationship between parent nodes and child nodes between respective overlapping common parent nodes.

[0129] Specifically, since each overlapping common parent node is a common parent node with different node hierarchical relationships in the same bronchial segment interface, for example: the overlapping common parent node 1 and the overlapping common parent node 2 are both overlapping common parent nodes of the root node branches of each bronchial segment, but the node hierarchy relationship of the overlapping common parent node 1 is different from the node hierarchy relationship of the overlapping common parent node 2. Therefore, among the overlapping common parent nodes obtained by the search, the overlapping common parent node with the lowest node hierarchy relationship is found as the target common parent node, that is, the overlapping common parent node with the lowest generation of each overlapping common parent node is found as the target common parent node. For each bronchial segment interface, the above-mentioned target common parent node search method is used to determine the target common parent node in each bronchial segment interface.

[0130] In this embodiment, by traversing each target bronchial segment root node branch in the opposite direction of the bronchial growth direction, the lowest generation among the common parent nodes of the common part is obtained as the target common parent node. The lowest generation branch can be accurately located in the current bronchial segment interface to avoid locating the main trunk branch.

[0131] In one embodiment, Fig.11 As shown, according to the node hierarchical relationship of each overlapping common parent node, the target common parent node is determined, including:

[0132] Step 1102, determining the hierarchical relationship value of each overlapping common parent node according to the mapping relationship between the hierarchical relationship of each node and the corresponding overlapping common parent node.

[0133] The hierarchical relationship value may be a digital representation of the hierarchical relationship of nodes with overlapping common parent nodes.

[0134] Specifically, according to the one-to-one correspondence between each node hierarchical relationship and each overlapping common parent node, that is, according to the mapping relationship between each node hierarchical relationship and each overlapping common parent node, the node hierarchical relationship of each overlapping common parent node is assigned from high to low, that is, the hierarchical relationship value of the node hierarchical relationship of the overlapping common parent node with a higher level is greater than the hierarchical relationship value of the node hierarchical relationship of the overlapping common parent node with a lower level, so as to obtain the hierarchical relationship value of each overlapping common parent node.

[0135] Step 1104 , selecting the overlapping common parent node with the smallest hierarchical relationship value as the target common parent node.

[0136] Specifically, each hierarchical relationship value is traversed to determine the minimum hierarchical relationship value, and the coincident common parent node corresponding to the minimum hierarchical relationship value is used as the target common parent node.

[0137] In this embodiment, by digitizing the generations of different overlapping common parent nodes in the node hierarchy relationship, the target common parent node can be accurately located using the hierarchy relationship value, thereby improving the accuracy of obtaining the target common parent node.

[0138] In one embodiment, Fig.12 As shown, according to the target common parent node, the vascular structure root node in the bronchial segmentation mask is determined, including:

[0139] Step 1202: for any target common parent node, traverse the target common parent node to obtain the number of first endpoint connections and the number of second endpoint connections.

[0140] The first endpoint connection number and the second endpoint connection number may both be the number of connection points of two endpoints of the target common parent node.

[0141] Specifically, for any target common parent node, traverse the center line of the target common parent node corresponding to the target common parent node in the bronchial topology structure to obtain the two endpoints of the target common parent node, and further traverse the number of connection points between the two endpoints of the target common parent node to obtain the first endpoint connection number and the second endpoint connection number.

[0142] Step 1204: When the number of first endpoint connections is greater than the number of second endpoint connections, the endpoints corresponding to the first endpoint connection number are used as the root nodes of the vascular structure.

[0143] Specifically, if the first endpoint connection number is greater than the second endpoint connection number, that is, the first endpoint connection number minus the second endpoint connection number is a positive number, the endpoint of the target common parent node corresponding to the first endpoint connection number is used as the vascular structure root node.

[0144] Alternatively, in step 1206, when the number of first endpoint connections is less than the number of second endpoint connections, the endpoint corresponding to the second endpoint connection number is used as the root node of the vascular structure.

[0145] Specifically, if the first endpoint connection number is less than the second endpoint connection number, that is, the first endpoint connection number minus the second endpoint connection number is a negative number, the endpoint of the target common parent node corresponding to the second endpoint connection number is used as the vascular structure root node.

[0146] In this embodiment, by comparing the number of connections of the first endpoint and the number of connections of the second endpoint in the target common parent node, the endpoint with a larger number of connections in the target common parent node can be obtained and used as the root node of the vascular structure, which helps to avoid using the endpoint connected to the main branch as the root node of the vascular structure.

[0147] In one embodiment, Fig.13 As shown, after the step of generating the bronchial topological structure corresponding to the bronchial segmentation mask according to the bronchial centerline and the root bronchial position information, the method includes:

[0148] Step 1302: classify the bronchial topological structures to obtain classification information of each topological structure.

[0149] The topological structure classification information may be information obtained after classifying the bronchial topological structure, and is used to verify whether the bronchial topological structure is correct.

[0150] Specifically, the Strahler classification method is used to classify the bronchial topological structure to obtain the classification information of each topological structure. Among them, the Strahler classification rules are as follows: (1) Each leaf node is marked as level 1; (2) When the nodes of the next level have the same level, the level of the bifurcation point increases by 1 based on the nodes of the next level; (3) When the levels of the nodes of the next level are inconsistent, the level of the bifurcation point is the higher level of the nodes of the next level. The connectivity of the blood vessel endpoints is 1. The Strahler classification results are as follows: Fig.21 shown.

[0151] Step 1304: verify the bronchial topological structure according to the classification information of each topological structure.

[0152] Specifically, the generated bronchial topology is verified according to each topology classification information to obtain a topology verification result; wherein the topology verification result includes that the bronchial topology is correct and that the bronchial topology is abnormal.

[0153] Step 1306, when the topology verification result indicates that the bronchial topology is correct, obtain at least one vascular structure root node of the bronchial segmentation mask according to the bronchial segmentation interface and the bronchial topology of the target object.

[0154] Specifically, if the topology verification result indicates that the bronchial topology is correct, it means that the bronchial topology can be used to identify at least one vascular structure root node of the bronchial segmentation mask; further, the step of "obtaining at least one vascular structure root node of the bronchial segmentation mask based on the bronchial segmentation interface and the bronchial topology of the target object" is executed to identify each vascular structure root node.

[0155] Alternatively, in step 1308, when the topology verification result indicates that the bronchial topology is abnormal, the process returns to executing the process of generating a bronchial topology corresponding to a bronchial segmentation mask according to the bronchial centerline and the root bronchial position information.

[0156] Specifically, if the topology verification result indicates that the bronchial topology is abnormal, it means that the bronchial topology cannot be used to identify at least one vascular structure root node of the bronchial segmentation mask; further, return to execute the step of "generating the bronchial topology corresponding to the bronchial segmentation mask according to the bronchial centerline and the root bronchial position information" to re-acquire the bronchial topology corresponding to the bronchial segmentation mask.

[0157] In this embodiment, the bronchial topology is verified by using the Strahler classification method to determine whether the bronchial topology is reasonable. Different operation methods are performed under different verification results, so as to check whether there are abnormalities in the bronchial topology and avoid deviations in the subsequent identification of the root node of the vascular structure.

[0158] It should be understood that, although the steps in the flowcharts involved in the above embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps is not strictly limited in order, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above embodiments may include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0159] Based on the same inventive concept, the embodiment of the present application also provides a vascular structure root node extraction device for implementing the above-mentioned vascular structure root node extraction method. The implementation solution provided by the device is similar to the implementation solution recorded in the above-mentioned method, so the specific limitations in the embodiments of one or more vascular structure root node extraction devices provided below can refer to the limitations of a vascular structure root node extraction method above, and will not be repeated here. 。

[0160] In one embodiment, Fig. 22 As shown, a vascular structure root node extraction device is provided, comprising: a bronchial centerline extraction module 2202, a bronchial position determination module 2204, a bronchial topology structure generation module 2206 and a vascular structure root node identification module 2208, wherein:

[0161] The bronchial centerline extraction module 2202 is used to extract the centerline of the bronchial segmentation mask of the target object to obtain the bronchial centerline corresponding to the bronchial segmentation mask;

[0162] A bronchial position determination module 2204 is used to determine the root bronchial position corresponding to the bronchial segmentation mask according to the bronchial centerline;

[0163] A bronchial topology structure generating module 2206 is used to generate a bronchial topology structure corresponding to a bronchial segmentation mask according to the bronchial centerline and the root bronchial position;

[0164] The vascular structure root node identification module 2208 is used to identify at least one vascular structure root node in the bronchial segmentation mask according to the bronchial topology.

[0165] In one embodiment, the bronchial topology structure generating module 2206 is also used to determine the bronchial growth direction according to the root bronchial position and the bronchial centerline; taking the bronchial root node of the root bronchial centerline as the starting point, the bronchial topology structure is searched according to the bronchial growth direction.

[0166] In one embodiment, the bronchial topology generation module 2206 is further used to select a root bronchial centerline from the bronchial centerlines according to the root bronchial position information; and determine the bronchial growth direction according to the centerline endpoint information of the root bronchial centerline.

[0167] In one embodiment, the bronchial topology structure generation module 2206 is also used to use the bronchial root node as the branch parent node; use the branch parent node as the starting point, extend to the bifurcation point in the bronchial growth direction, and obtain a branch child node; use the branch child node as the new branch parent node, return to execute the step of using the branch parent node as the starting point, extending to the bifurcation point in the bronchial growth direction, and obtaining a branch child node, until the bronchial topology structure is obtained.

[0168] In one embodiment, the vascular structure root node identification module 2208 is also used to search for the bronchial structure information of each bronchial topology structure to obtain the root node branch of each bronchial segment; based on the preset bronchial segment association relationship, search for at least one bronchial segment root node branch from the opposite direction of the bronchial growth direction to obtain the target common parent node; according to the target common parent node, determine the vascular structure root node in the bronchial segmentation mask.

[0169] In one embodiment, the vascular structure root node identification module 2208 is also used to calculate the labels of each bronchial branch structure in the bronchial topological structure to obtain the label value of each branch; the branch label value represents the type of the bronchial branch structure; the corresponding bronchial branch structure is classified according to each branch label value to determine each bronchial segment structure; and each bronchial segment structure is searched to obtain the root node branch of each bronchial segment.

[0170] In one embodiment, the vascular structure root node identification module 2208 is also used to search for each bronchial branch structure in any bronchial segment structure to obtain each ancestral branch node; traverse each ancestral branch node in the opposite direction of the bronchial growth direction to determine the bronchial segment root node branch.

[0171] In one embodiment, the vascular structure root node identification module 2208 is also used to determine at least one target ancestor branch node of each bronchial branch structure from each ancestor branch node; traverse each target ancestor branch node in the opposite direction of the bronchial growth direction to determine the overlapping ancestor branch nodes of each bronchial branch structure; and determine the bronchial segment root node branch based on the node hierarchical relationship of each overlapping ancestor branch node.

[0172] In one embodiment, the vascular structure root node identification module 2208 is also used to select at least one target bronchial segment root node branch in the bronchial segment interface from the root node branches of each bronchial segment; traverse each target bronchial segment root node branch in the opposite direction of the bronchial growth direction to obtain each overlapping common parent node; determine the target common parent node according to the node hierarchical relationship of each overlapping common parent node.

[0173] In one embodiment, the vascular structure root node identification module 2208 is also used to determine the hierarchical relationship value of each overlapping common parent node based on the mapping relationship between the hierarchical relationship of each node and the corresponding overlapping common parent node; and select the overlapping common parent node with the smallest hierarchical relationship value as the target common parent node.

[0174] In one embodiment, the vascular structure root node identification module 2208 is also used to traverse the target common parent node for any target common parent node to obtain the first endpoint connection number and the second endpoint connection number; when the first endpoint connection number is greater than the second endpoint connection number, the endpoint corresponding to the first endpoint connection number is used as the vascular structure root node; or, when the first endpoint connection number is less than the second endpoint connection number, the endpoint corresponding to the second endpoint connection number is used as the vascular structure root node.

[0175] In one embodiment, the bronchial topology generation module 2206 is also used to classify the bronchial topology to obtain the classification information of each topology; verify the bronchial topology according to the classification information of each topology; when the topology verification result indicates that the bronchial topology is correct, execute at least one vascular structure root node of the bronchial segmentation mask according to the bronchial segmentation interface and the bronchial topology of the target object; or, when the topology verification result indicates that the bronchial topology is abnormal, return to execute the bronchial topology corresponding to the bronchial segmentation mask according to the bronchial centerline and the root bronchial position information.

[0176] Each module in the above-mentioned vascular structure root node extraction device can be implemented in whole or in part by software, hardware and a combination thereof. Each of the above-mentioned modules can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute the operations corresponding to each of the above modules.

[0177] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Fig.23 As shown. The computer device includes a processor, a memory and a network interface connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store server data. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a method for extracting a root node of a vascular structure is implemented.

[0178] Those skilled in the art will understand that Fig.23 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0179] In one embodiment, a computer device is further provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above method embodiments when executing the computer program.

[0180] In one embodiment, a computer-readable storage medium is provided, storing a computer program, which implements the steps in the above method embodiments when executed by a processor.

[0181] In one embodiment, a computer program product or computer program is provided, the computer program product or computer program includes computer instructions, the computer instructions are stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device performs the steps in the above-mentioned method embodiments.

[0182] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties.

[0183] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.

[0184] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0185] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A method for extracting root nodes of vascular structures, characterized in that: The method comprises: Performing centerline extraction on the bronchial segmentation mask of the target object to obtain the bronchial centerline corresponding to the bronchial segmentation mask; Determining the root bronchus position corresponding to the bronchus segmentation mask according to the bronchus centerline; Generating a bronchial topological structure corresponding to the bronchial segmentation mask according to the bronchial centerline and the root bronchial position; At least one vascular structure root node in the bronchial segmentation mask is identified based on the bronchial topology.

2. The method according to claim 1, characterized in that The step of generating a bronchial topological structure corresponding to the bronchial segmentation mask according to the bronchial centerline and the root bronchial position includes: Determining the direction of bronchial growth according to the root bronchial position and the bronchial centerline; The bronchial root node of the root bronchial centerline is used as a starting point, and the bronchial topological structure is searched out according to the bronchial growth direction.

3. The method according to claim 2, characterized in that Determining the bronchial growth direction according to the root bronchial position and the bronchial centerline includes: According to the root bronchus position information, selecting a root bronchus centerline from the bronchus centerlines; The bronchial growth direction is determined according to the centerline endpoint information of the root bronchial centerline.

4. The method according to claim 2, characterized in that: The step of taking the bronchial root node of the root bronchial centerline as a starting point and searching for the bronchial topological structure according to the bronchial growth direction includes: Taking the bronchial root node as the branch parent node; Taking the branch parent node as the starting point, extending in the bronchial growth direction to the bifurcation point, to obtain a branch child node; The branch child node is used as the new branch parent node, and the step of taking the branch parent node as the starting point, extending to the bifurcation point in the bronchial growth direction, and obtaining the branch child node is returned to be executed until the bronchial topology structure is obtained.

5. The method according to claim 2, characterized in that: The step of identifying at least one vascular structure root node in the bronchial segmentation mask according to the bronchial topology structure includes: Searching for each bronchial structure information in the bronchial topological structure to obtain the root node branch of each bronchial segment; Based on the preset bronchial segment association relationship, searching for at least one bronchial segment root node branch from the opposite direction of the bronchial growth direction to obtain a target common parent node; According to the target common parent node, a vascular structure root node in the bronchial segmentation mask is determined.

6. The method according to claim 5, characterized in that The step of searching for the structural information of each bronchus in the bronchial topology structure to obtain the root node branches of each bronchial segment includes: Calculating labels for each bronchial branch structure in the bronchial topological structure to obtain label values ​​for each branch; the branch label values ​​represent the type of the bronchial branch structure; Classifying the corresponding bronchial branch structures according to each branch label value to determine each bronchial segment structure; The bronchial segment structures are searched to obtain the root node branches of the bronchial segments.

7. The method according to claim 6, characterized in that The searching of each bronchial segment structure to obtain each bronchial segment root node branch includes: Searching each of the bronchial branch structures in any of the bronchial segment structures to obtain each ancestor branch node; Each of the ancestor branch nodes is traversed in the opposite direction of the bronchial growth direction to determine the bronchial segment root node branch.

8. The method according to claim 7, characterized in that The traversing each of the ancestor branch nodes in the reverse direction of the bronchial growth direction to determine the bronchial segment root node branch includes: Determine at least one target ancestor branch node of each of the bronchial branch structures from each of the ancestor branch nodes; Traversing each of the target ancestor branch nodes in the opposite direction of the bronchial growth direction to determine the overlapping ancestor branch nodes of each of the bronchial branch structures; The bronchial segment root node branch is determined according to the node hierarchical relationship of each of the overlapping ancestral branch nodes.

9. The method according to claim 5, characterized in that The step of searching at least one bronchial segment root node branch from the opposite direction of the bronchial growth direction based on the preset bronchial segment association relationship to obtain a target common parent node includes: Selecting at least one target bronchial segment root node branch in the bronchial segment interface from each of the bronchial segment root node branches; Traversing the root node branches of each target bronchial segment along the direction opposite to the bronchial growth direction to obtain each overlapping common parent node; The target common parent node is determined according to the node hierarchical relationship of each overlapping common parent node.

10. The method according to claim 9, characterized in that The step of determining the target common parent node according to the node hierarchical relationship of the overlapping common parent nodes includes: Determine the hierarchical relationship value of each of the overlapping common parent nodes according to the mapping relationship between the hierarchical relationship of each of the nodes and the corresponding overlapping common parent nodes; The overlapping common parent node with the smallest hierarchical relationship value is selected as the target common parent node.

11. The method according to claim 5, characterized in that The step of determining the vascular structure root node in the bronchial segmentation mask according to the target common parent node includes: For any of the target common parent nodes, traverse the target common parent nodes to obtain the number of first endpoint connections and the number of second endpoint connections; When the number of the first endpoint connections is greater than the number of the second endpoint connections, the endpoints corresponding to the first endpoint connections are used as the root nodes of the vascular structure; or, When the number of the first endpoint connections is smaller than the number of the second endpoint connections, the endpoints corresponding to the second endpoint connections are used as the root nodes of the vascular structure.

12. The method according to claim 1, characterized in that After the step of generating the bronchial topological structure corresponding to the bronchial segmentation mask according to the bronchial centerline and the root bronchial position information, the method includes: Classifying the bronchial topological structures to obtain classification information of each topological structure; Verifying the bronchial topology according to each of the topological structure classification information; When the topology verification result indicates that the bronchial topology is correct, performing the step of obtaining at least one vascular structure root node of the bronchial segmentation mask according to the bronchial segmentation interface of the target object and the bronchial topology; or, When the topology verification result indicates that the bronchial topology is abnormal, the process returns to executing the step of generating the bronchial topology corresponding to the bronchial segmentation mask according to the bronchial centerline and the root bronchial position information.

13. A vascular structure root node extraction device, characterized in that: The device comprises: A bronchial centerline extraction module is used to extract the centerline of the bronchial segmentation mask of the target object to obtain the bronchial centerline corresponding to the bronchial segmentation mask; A bronchial position determination module, used to determine the root bronchial position corresponding to the bronchial segmentation mask according to the bronchial centerline; A bronchial topology structure generating module, used for generating a bronchial topology structure corresponding to the bronchial segmentation mask according to the bronchial centerline and the root bronchial position; The vascular structure root node identification module is used to identify at least one vascular structure root node in the bronchial segmentation mask according to the bronchial topology.

14. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 12 are implemented.

15. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 12 are implemented.

16. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 12 are implemented.

Images