Methods, apparatus, equipment, media, and procedures for obtaining the aortic centerline

Abstract

This application provides a method, apparatus, device, medium, and program product for obtaining the aortic centerline, comprising: preprocessing a CTA image based on a pre-trained convolutional neural network algorithm to obtain target segmentation prediction results and trunk segmentation prediction results; preprocessing the CTA image based on a pre-trained convolutional neural network algorithm to obtain target segmentation prediction results and trunk segmentation prediction results for the target aorta; obtaining a set of vessel vertices and a first set of endpoints for vessel segments in the target aorta based on the vascular skeleton line of the target segmentation prediction results; extracting the aortic sinus endpoints from the first set of endpoints based on the trunk segmentation prediction results, and determining a second set of endpoints; extracting a second shortest path set from the first shortest path set based on the Dijkstra algorithm and the set of vessel vertices, and determining the second shortest path set as the aortic centerline to improve the efficiency of obtaining the aortic centerline.

Description

Technical Field

[0001] This application relates to the field of medical imaging, and more particularly to a method, apparatus, device, medium, and procedure for obtaining the aortic centerline. Background Technology

[0002] The aortic centerline refers to the characteristic curve of the aortic vessels obtained from aortic vascular imaging. Aortic vascular imaging specifically refers to three-dimensional vascular images obtained from a patient's aorta using computed tomography angiography (CTA) imaging equipment.

[0003] In related technologies, the system first receives manually marked vessel endpoints (e.g., start point, bifurcation point, and end point) and connects these endpoints with the vessel center points to obtain the aortic centerline. However, manually marking the vessel endpoints takes considerable time, resulting in low efficiency in obtaining the aortic centerline. Summary of the Invention

[0004] This application provides a method, apparatus, device, medium, and procedure for obtaining the aortic centerline, thereby improving the efficiency of obtaining the aortic centerline.

[0005] In a first aspect, a method for obtaining the aortic centerline is provided, comprising: First, preprocessing a computed tomography (CTA) angiography image based on a pre-trained convolutional neural network algorithm to obtain the target segmentation prediction result of the target aorta and the trunk segmentation prediction result of the target aorta. Next, based on the vascular skeleton line of the target segmentation prediction result, obtaining the set of vascular vertices and the first endpoint set of the vascular segments in the target aorta; then, based on the trunk segmentation prediction result, extracting the aortic sinus endpoints from the first endpoint set, and determining the first endpoint set after deleting the aortic sinus endpoints as the second endpoint set. Finally, calculating the first shortest path set based on Dijkstra's algorithm and the vascular vertex set, and extracting the second shortest path set from the first shortest path set, determining the second shortest path set as the aortic centerline.

[0006] The CTA image includes a slice image of the target aorta; the vessel vertices in the vessel vertex set include at least one vessel endpoint and at least one bifurcation point; the first endpoint set includes each vessel endpoint in at least one vessel endpoint arranged according to the preset image layering direction; the first path set includes the shortest path from the aortic sinus endpoint to each vessel vertex in the vessel vertex set; and the second shortest path set includes the shortest path from the aortic sinus endpoint to each vessel endpoint in the second endpoint set.

[0007] In a second aspect, a device for acquiring the aortic centerline is provided, comprising: a first acquisition module, a second acquisition module, a first processing module, and a second processing module;

[0008] The first acquisition module is used to preprocess the computed tomography angiography (CTA) images based on a pre-trained convolutional neural network algorithm to obtain the target segmentation prediction results of the target aorta and the trunk segmentation prediction results of the target aorta; the CTA images include slice images of the target aorta.

[0009] The second acquisition module is used to acquire the set of vascular vertices and the set of first endpoints of the vascular segment in the target aorta based on the vascular skeleton line of the target segmentation prediction result; the vascular vertices in the vascular vertices set include at least one vascular endpoint and at least one bifurcation point; the set of first endpoints includes each vascular endpoint in at least one vascular endpoint arranged according to the preset image layering direction.

[0010] The first processing module is used to extract the aortic sinus endpoints from the first endpoint set based on the trunk segmentation prediction results, and to determine the first endpoint set after deleting the aortic sinus endpoints as the second endpoint set.

[0011] The second processing module is used to calculate a first shortest path set based on Dijkstra's algorithm and the set of vessel vertices, and extract a second shortest path set from the first shortest path set, and determine the second shortest path set as the aortic centerline; the first path set includes the shortest path from the aortic sinus endpoint to each vessel vertex in the set of vessel vertices, and the second shortest path set includes the shortest path from the aortic sinus endpoint to each vessel endpoint in the second endpoint set.

[0012] Thirdly, an electronic device is provided, comprising: a processor and a memory for storing a computer program, the processor for calling and running the computer program stored in the memory, and performing the methods as described in the first aspect or its various implementations.

[0013] Fourthly, a computer-readable storage medium is provided for storing a computer program that causes a computer to perform the methods described in the first aspect or its various implementations.

[0014] Fifthly, a computer program product is provided, including computer program instructions that cause a computer to perform the methods as described in the first aspect or its various implementations.

[0015] Sixthly, a computer program is provided that causes a computer to perform the methods described in the first aspect or its various implementations.

[0016] The technical solution provided in this application allows an electronic device to first preprocess CTA images using a pre-trained convolutional neural network algorithm to obtain the target segmentation prediction results for the target aorta and the trunk segmentation prediction results for the target aorta. Next, based on the vascular skeleton lines from the target segmentation prediction results, a set of vascular vertices and a first set of endpoints for the vascular segments in the target aorta are obtained. Then, based on the trunk segmentation prediction results, the aortic sinus endpoints are extracted from the first set of endpoints, and the first set of endpoints with the aortic sinus endpoints removed is determined as the second set of endpoints. Finally, based on Dijkstra's algorithm and the set of vascular vertices, a first set of shortest paths is calculated, and a second set of shortest paths is extracted from the first set of shortest paths, which is then determined as the aortic centerline. In this process, the electronic device can automatically find the starting point (the aortic sinus endpoint) and the ending point (the vascular endpoints in the second set of endpoints), which improves the efficiency of obtaining the aortic centerline compared to marking vascular endpoints with a watch. In the above process, electronic devices can automatically acquire the aortic centerline using Dijkstra's algorithm. As the performance of electronic devices gradually improves, the speed of acquiring the aortic centerline gradually increases, enabling electronic devices to acquire the aortic centerline in batches. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only 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 An application scenario diagram provided for an embodiment of this application;

[0019] Figure 2 A flowchart illustrating a first method for obtaining the aortic centerline provided in this application embodiment;

[0020] Figure 3 A flowchart illustrating a second method for obtaining the aortic centerline provided in this application embodiment;

[0021] Figure 4 A schematic diagram of a blood vessel contour provided for an embodiment of this application;

[0022] Figure 5 A flowchart illustrating a third method for obtaining the aortic centerline provided in this application embodiment;

[0023] Figure 6 A flowchart illustrating the fourth method for obtaining the aortic centerline provided in this application embodiment;

[0024] Figure 7 A schematic diagram of a trunk starting point provided in an embodiment of this application;

[0025] Figure 8 A flowchart illustrating the fifth method for obtaining the aortic centerline provided in this application embodiment;

[0026] Figure 9 A schematic diagram of the aortic centerline provided in an embodiment of this application;

[0027] Figure 10 A schematic diagram of a device 1000 for obtaining the aortic centerline provided in an embodiment of this application;

[0028] Figure 11 This is a schematic block diagram of the electronic device 1100 provided in the embodiments of this application. Detailed Implementation

[0029] 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, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0030] 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 server 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 devices.

[0031] As described above, in related technologies, the vascular endpoints (e.g., start points, bifurcation points, and end points) manually marked by the marker are first received, and the various vascular endpoints and their center points are connected to obtain the aortic centerline. However, if the vascular endpoints are manually marked by the marker, it takes a considerable amount of time, resulting in low efficiency in obtaining the aortic centerline.

[0032] To address the aforementioned technical problems, the inventive concept of this application is as follows: An electronic device can map the continuity of blood vessels and the structural characteristics of the aorta (ascending and descending aorta) into image features, automatically identifying the starting point (the endpoint of the aortic sinus) and the ending point (the vessel endpoints in the second endpoint set). Compared to marking vessel endpoints on a watch, this improves the efficiency of acquiring the aortic centerline. The electronic device calculates the shortest path between the starting point and each ending point using Dijkstra's algorithm to acquire the aortic centerline. As the performance of the electronic device gradually improves, the speed of acquiring the aortic centerline gradually increases, enabling the electronic device to acquire aortic centerlines in batches.

[0033] It should be understood that the technical solution of this application can be applied to the following scenarios, but is not limited to:

[0034] In some possible ways, Figure 1 An application scenario diagram provided for an embodiment of this application, such as... Figure 1 As shown, this application scenario may include electronic device 110 and network device 120. Electronic device 110 can establish a connection with network device 120 through a wired network or a wireless network.

[0035] For example, electronic device 110 may be a desktop computer, laptop computer, tablet computer, etc., but is not limited thereto. Network device 120 may be a terminal device or a server, but is not limited thereto. In one embodiment of this application, electronic device 110 may send a request message to network device 120, which may be used to request the acquisition of CTA images. Further, electronic device 110 may receive a response message sent by network device 120, which includes CTA images. Electronic device 110 may also perform image processing on the CTA images to obtain the aortic centerline.

[0036] also, Figure 1 An electronic device and a network device are given as examples, but in practice, other numbers of electronic devices and network devices may be included, and this application does not limit this.

[0037] In other possible implementations, the technical solution of this application may also be executed by the aforementioned electronic device 110, or by the aforementioned network device 120, and this application does not impose any restrictions on this.

[0038] After introducing the application scenarios of the embodiments of this application, the technical solution of this application will be described in detail below:

[0039] Figure 2 A flowchart illustrating a method for obtaining the aortic centerline provided in this application embodiment, the method being as follows: Figure 1The electronic device 110 shown performs, but is not limited to, its functions. For example... Figure 2 As shown, the method may include the following steps:

[0040] S210: The electronic device preprocesses the computed tomography angiography (CTA) images based on a pre-trained convolutional neural network algorithm to obtain the target segmentation prediction results of the target aorta and the trunk segmentation prediction results of the target aorta.

[0041] In this embodiment, a CTA image refers to an X-ray scan of the location of the target aorta after intravenous injection of a contrast agent, which is then processed by a computer. Specifically, a CTA image includes multiple slice images. The size of a CTA image can be M*N*L, where M and N are the length and width of each slice image, for example, M is 512 and N is 512, and L is the number of slice images in the CTA image, that is, the number of slice layers. Typically, slice images are acquired layer by layer, following the direction of the human body from foot to head.

[0042] In this embodiment, the CTA image includes a slice image of the target aorta, and also includes human tissues such as viscera, bones, muscles, and fat in the same layer as the target aorta. The purpose of preprocessing the CTA image is to segment the blood vessel contour image from the CTA image.

[0043] It is understandable that electronic devices can send information retrieval requests and receive CTA images sent by other electronic devices in response to the information retrieval requests. Electronic devices can also directly receive CTA images.

[0044] In some implementations, such as Figure 3 As shown, the electronic device acquires the target segmentation prediction results of the target aorta and the trunk segmentation prediction results of the target aorta, which may include:

[0045] S310: Electronic device acquires slice images of each layer in the CTA image.

[0046] S320: The electronic device inputs the grayscale values ​​of each pixel in each slice image into a pre-trained convolutional neural network algorithm to obtain the grayscale feature values ​​of each pixel in each slice image.

[0047] S330: The electronic device obtains the main trunk mask and branch mask of the target aorta based on the preset main trunk threshold, the preset branch threshold, and the grayscale feature values ​​of each pixel in each slice image.

[0048] S340: The electronic device determines the image corresponding to the trunk mask as the trunk segmentation prediction result of the target aorta, and merges the trunk mask and branch mask to obtain the artery mask, and determines the image corresponding to the artery mask as the target segmentation prediction result of the target aorta.

[0049] In the embodiments of this application, the sliced ​​image can be represented by grayscale values, and each sliced ​​image can be recorded by a matrix, the number of rows and columns of the matrix being the same as the number of rows and columns of the CTA image.

[0050] For example, each slice image is represented by a matrix. Where a mn This represents the grayscale value in the m-th row and n-th column.

[0051] In this embodiment, the pre-trained convolutional neural network algorithm refers to a neural network algorithm trained with pre-set samples. It is a deep learning model commonly used to analyze visual images. Specifically, it can employ Unet, VGGnet, or CNN networks.

[0052] For example, after pre-training a convolutional neural network algorithm, the grayscale feature values ​​of each pixel in each slice image are obtained, and each slice image is represented by a matrix. Where f mn This represents the grayscale feature value of the m-th row and n-th column.

[0053] For example, the trunk mask can be represented by a matrix, the number of rows and columns of which is the same as the number of rows and columns corresponding to the gray-level feature values ​​of the sliced ​​image. Specifically, for each pixel's gray-level feature value, it is compared with a preset trunk threshold T1. If the gray-level feature value is greater than T1, the corresponding value for that pixel in the trunk mask matrix is ​​set to 1, and the corresponding values ​​for the remaining pixels are set to 0. Then, the image corresponding to the trunk mask is determined as the trunk segmentation prediction result of the target aorta. Figure 4 As shown in (a), the trunk segmentation prediction results include the vascular contour of the aortic trunk of the target aorta.

[0054] For example, similarly, the branch mask can also be represented by a matrix, with the number of rows and columns being the same as the number of rows and columns corresponding to the gray-level feature values ​​of the sliced ​​image. Specifically, for each pixel's gray-level feature value, it is compared with a preset branch threshold T2. If the gray-level feature value is greater than T2, the corresponding value for that pixel in the branch mask matrix is ​​set to 2, and the corresponding values ​​for the remaining pixels are set to 0. Then, the image corresponding to the branch mask is determined as the branch segmentation prediction result of the target aorta. It should be noted that the specific values ​​of T1 and T2 are not limited in this embodiment. Figure 4 As shown in (b), the branch segmentation prediction results include the vascular contours of the aortic branches of the target aorta.

[0055] For example, the main trunk mask and branch trunk masks are merged to obtain the artery mask. The image corresponding to the artery mask is then identified as the target segmentation prediction result for the aorta. The image corresponding to the artery mask is the mask image of the slice image. The mask image for each slice image layer... Where b mn This represents the pixel mask in row m and column n. The first identifier for the main aorta of the target aorta is 1, the second identifier for the aortic branches of the target aorta is 2, and the identifiers for other parts besides the main aorta and aortic branches are 0. For example... Figure 4 As shown in (c), the target segmentation prediction result includes the vascular contours of the aortic trunk and aortic branches of the target aorta.

[0056] Thus, by preprocessing CTA images using a pre-trained convolutional neural network to obtain a mask image of the CTA image, the mask image can clearly distinguish the main aortic trunk, aortic branches, and non-aortic parts in the image, so as to obtain a clear and accurate vascular contour image (i.e., the target segmentation prediction result of the target aorta and the main aortic trunk segmentation prediction result). This can improve the accuracy of the target segmentation prediction result and the trunk segmentation prediction result, thereby improving the accuracy of the obtained aortic centerline.

[0057] S220: The electronic device obtains the set of vascular vertices and the set of first endpoints of the vascular segment in the target aorta based on the vascular skeleton line of the target segmentation prediction result.

[0058] Understandably, the vascular skeleton line refers to the thin vascular curve that is consistent with the original vascular connectivity and topology.

[0059] The set of vessel vertices includes at least one vessel endpoint and at least one bifurcation point. The first set of endpoints includes each vessel endpoint from at least one vessel endpoint arranged according to a preset image layering direction. The ascending order of the layering direction values ​​in the preset image layering direction corresponds to the arrangement of the photographed object from foot to head in the slice image.

[0060] In some implementations, such as Figure 5 As shown, the electronic device obtains the set of vessel vertices and the set of first endpoints of the vessel segments in the target aorta based on the vessel skeleton lines from the target segmentation prediction results. This may include:

[0061] S510: The electronic device refines the target segmentation prediction results according to the preset refinement algorithm to obtain the vascular skeleton line of the target aorta.

[0062] The preset thinning algorithm can be a morphological thinning algorithm or a binary thinning algorithm. It can determine the position of the target pixel by gradually eroding the target segmentation prediction result from the outer layer, and obtain the vascular skeleton line.

[0063] S520: The electronic device obtains the vascular vertices of the vascular segment in the target aorta based on the number of neighboring points of each pixel on the vascular skeleton line, and obtains the vascular vertex set.

[0064] In this embodiment, both the endpoints and bifurcation points of blood vessels are considered vessel vertices. After determining whether all pixels on the vessel skeleton line are vessel vertices, a set of vessel vertices is obtained.

[0065] In this embodiment, the number of neighboring points of each pixel on the vascular skeleton line is obtained. The number of neighboring points can be 1, 2, 3 or more. If the number of neighboring points is 1, then the pixel is a vascular endpoint; if the number of neighboring points is 2, then the pixel is a normal connection point; if the number of neighboring points is 3 or more, then the pixel is a bifurcation point.

[0066] For example, the vascular skeleton line is segmented according to the vessel endpoints and bifurcation points. The starting and ending points of each segment are either endpoints or bifurcation points. At least one vessel endpoint is obtained, and the i-th vessel endpoint {e i =(x i ,y i ,z i Given |i = 1, 2, ..., k}, find at least one bifurcation point to obtain the j-th bifurcation point {b}. j =(x j ,y j ,z j Let |j=1,2,…,p}, and combine at least one vessel endpoint and at least one bifurcation point to form the vessel vertex set V0=(v1,v2,…,v m (m=k+p).

[0067] S530: The electronic device extracts at least one blood vessel endpoint with a neighborhood point count of 1 from the blood vessel vertex set, and generates a first endpoint set by dividing each blood vessel endpoint in the at least one blood vessel endpoint according to a preset image layering direction.

[0068] For example, following the ascending order of the preset image layering direction, the first endpoint set E = (e1, e2, ..., e) is obtained. k ), which is the set of all the endpoints of blood vessels.

[0069] Thus, based on the number of neighboring points of each pixel on the vascular skeleton line, the set of vascular vertices and the set of first endpoints are obtained respectively, so that no vascular endpoints on the aortic centerline are missed, and the shortest path can be obtained between each vascular endpoint and the aortic sinus endpoint, thereby improving the accuracy of obtaining the aortic centerline.

[0070] S230: The electronic device extracts the aortic sinus endpoints from the first endpoint set based on the trunk segmentation prediction results, and determines the first endpoint set with the aortic sinus endpoints deleted as the second endpoint set.

[0071] It should be noted that the trunk segmentation prediction result refers to the vascular outline of the aortic trunk. The aorta is the largest artery in the human body, originating from the left ventricle of the heart, extending upwards, to the right, and then downwards in a slightly arched shape, continuing down the spine, and branching into many smaller arteries within the thoracic and abdominal cavities. The aorta's anatomical structure presents an arched shape, first originating from the left ventricle, then extending upwards and to the right, and then downwards along the thoracic cavity to the abdominal cavity, exhibiting a complex morphological structure. Specifically, the aortic trunk includes the ascending aorta, the aortic arch, and the descending aorta.

[0072] In some implementations, such as Figure 6 As shown, the electronic device extracts the aortic sinus endpoints from the first endpoint set based on the trunk segmentation prediction results, and determines the first endpoint set after deleting the aortic sinus endpoints as the second endpoint set, which may include:

[0073] S610: The electronic device searches for the trunk starting point based on the trunk segmentation prediction result, starting from the minimum segmentation direction value corresponding to the blood vessel endpoint in the first endpoint set and ending at the maximum segmentation direction value, according to the preset image segmentation direction.

[0074] The number of connected components corresponding to the current layer slice image where the trunk starts is 1, and the number of connected components corresponding to the next layer slice image is 2; the trunk starts is the center point of the connected component corresponding to the current layer slice image.

[0075] For example, the first endpoint set E0 = (e1, e2, ..., e k ), obtain e k The Z value is Z max The Z value of e1 is Z min The Z-value gradually increases along the preset image layering direction. From Z... min To Z max Within a certain range, find the center point of the region where the current layer slice image has two connected regions and the next layer slice image has one connected region, such as... Figure 7 As shown, c0 = (x0, y0, z0).

[0076] S620: The electronic device extracts the aortic sinus endpoints from the first set of endpoints.

[0077] In this embodiment, the aortic sinus endpoint is the vessel endpoint in the first set of endpoints that is closest to the trunk starting point. Specifically, the distance between the trunk starting point and each vessel endpoint in the first set of endpoints is calculated, and the minimum distance value is found. The vessel endpoint corresponding to the minimum distance value is determined as the aortic sinus endpoint.

[0078] S630: The electronic device determines the first set of endpoints, from which the aortic sinus endpoints are removed, as the second set of endpoints.

[0079] It should be noted that the vascular endpoints in the second set of endpoints are the termination points on the central axis of the aorta.

[0080] Thus, in the above process, by determining the endpoint of the aortic sinus in the ascending aorta through the structural characteristics of the aortic trunk, the accuracy of the determined position of the aortic sinus endpoint can be improved, thereby improving the accuracy of the aortic centerline.

[0081] S240: The electronic device calculates a first shortest path set based on Dijkstra's algorithm and the set of vessel vertices, extracts a second shortest path set from the first shortest path set, and determines the aortic centerline from the second shortest path set.

[0082] The first path set includes the shortest path from the aortic sinus endpoint to each vessel vertex in the vessel vertex set, and the second shortest path set includes the shortest path from the aortic sinus endpoint to each vessel endpoint in the second endpoint set.

[0083] In some implementations, such as Figure 8 As shown, the electronic device extracts the aortic sinus endpoints from the first endpoint set based on the trunk segmentation prediction results, and determines the first endpoint set after deleting the aortic sinus endpoints as the second endpoint set, which may include:

[0084] S810: The electronic device moves the aortic sinus endpoint in the set of vessel apexes such that the first element in the set of vessel apexes is the aortic sinus endpoint.

[0085] It is understandable that the first element in the set of vessel vertices is the aortic sinus endpoint, which ensures that the aortic sinus endpoint is the starting point in the directed vertex graph.

[0086] For example, the set of blood vessel vertices V0 = (v1, v2, ..., v m Locate the aortic sinus from it. i Corresponding vertex v j(j≤m), move to the first element position in the set of vessel vertices to obtain the updated set of vessel vertices V=(v1,v2,…,v m ).

[0087] S820: The electronic device generates a directed vertex graph using the length of the segment between each blood vessel vertex in the set of blood vessel vertices as the weight.

[0088] Among them, the directed graph of vertices is used to identify the connection relationship and connection distance between each blood vessel vertex in the blood vessel vertex set.

[0089] In some implementations, after obtaining the set of vessel vertices, the electronic device can also record the length of the vessel segment between two adjacent vessel vertices. For example, the set of vessel vertices V0 = (v1, v2, ..., v...). m (m=k+p), from vertex v i To vertex v j The segmented length of the blood vessel w i,j The segment length of a blood vessel is determined based on the number of common connection points between two adjacent vertices of the blood vessel.

[0090] In this embodiment of the application, a directed vertex graph is generated based on the set of blood vessel vertices and the length of blood vessel segments between blood vessel vertices in the set of blood vessel vertices.

[0091] Specifically, the set of vessel vertices V = (v1, v2, ..., v m ), where e is the edge of each edge. i,j ={v i ,v j The weight of} is a non-negative real number w. i,j e i,j w represents two adjacent vessel vertices. i,j This indicates that from vertex v i To vertex v j The length of the blood vessel segments. Set the n-order weight matrix [w] of the directed graph with vertices. ij ] n*n , where w ii =∞, i = 1, 2, ..., n, and the remaining values ​​are set as follows:

[0092]

[0093] It should be noted that for two adjacent vessel vertices (v... i v j Electronic devices can also be connected via P i,j Record from vertex v i To vertex v j The set of pixels that have been passed through.

[0094] For example, the set of vessel vertices includes 5 vessel endpoints and 3 branching points. The adjacency order of each vessel vertex is as follows: vessel endpoint E0 is connected to branching point B0, branching point B0 is connected to branching point B1, branching point B0 is connected to branching point B2, branching point B1 is connected to vessel endpoint E1, branching point B1 is connected to vessel endpoint E2, branching point B2 is connected to vessel endpoint E3, and branching point B2 is connected to vessel endpoint E4. The set of vessel vertices includes 8 vessel vertices, therefore the directed vertex graph is an 8th order weight matrix [w]. ij ] n*n , where w ii =∞, i=1,2,…,8, the matrix element positions corresponding to two adjacent blood vessel vertices in the weight matrix are set to the corresponding blood vessel segment lengths.

[0095] S830: The electronic device computes the first set of shortest paths based on Dijkstra's algorithm and the directed graph of vertices.

[0096] Generally, the aortic centerline is a line formed by the centers of multiple axial interfaces of the target aorta. The aortic centerline can serve as the basis for subsequent quantitative analysis of lesions.

[0097] The first shortest path set includes the shortest path between the aortic sinus endpoint and each vessel vertex in the vessel vertex set.

[0098] In some implementations, calculating the first shortest path set includes: defining a set S, setting a single source point s∈V, where V is the set of vessel vertices. Finding the shortest path from the source point s (the endpoint of the aortic sinus) to all nodes in V-{s}. That is, the shortest path from the source point s to the vertex v. i A path ∈V can only pass through vertices already included in set S, and cannot pass through any vertices not yet in set S. The length of the shortest path relative to set S is denoted as dist[s,v]. i The set of points that find the shortest path relative to set S is denoted as dist_path[s,v]. i By continuously calculating dist[s,v] i Furthermore, the set S is continuously expanded. As the set S is continuously expanded, the shortest path relative to the set S will become shorter and shorter, until v... i Upon entering set S, we obtain dist[s,v] i ] and dist_path[s,v i ].

[0099] Specifically as follows:

[0100] 1) Initially, S = {v1};

[0101] 2) For v i =VS, which connects the source point s to the point v. i The weights between them are assigned to dist[s,v] i ], connect the source point s with point v i The set of points between them is assigned to dist_path[s,v] i ];

[0102] 3) Select the vertex v in set VS that has the shortest distance relative to set S in the shortest path. j , and this vertex v j Merge it into set S, thereby expanding S;

[0103] 4) Update v i = dist[s,v] of vertices in VS i Regarding the v selected in the previous step... j If dist[s,v] j ]+w ji <dist[s,v i If , then update the source vertex s to vertex v. i The shortest path dist[s,v] i ] = dist[s,v j ]+w ji At the same time, update the point set dist_path[s,v] corresponding to the shortest path. i ] = dist_path[s,v j ]+P ji ;

[0104] 5) Repeat steps 2), 3) and 4) above until S = V.

[0105] After the above calculations, the shortest path between the aortic sinus endpoint and each vessel vertex in the set of vessel vertices can be obtained, i.e., the first shortest path set. It can be understood that in dist_path[s,v]... i The shortest path is recorded in the set of pixels it passes through.

[0106] S840: The electronic device extracts a second shortest path set from the first shortest path set based on each vascular endpoint in the second endpoint set, and determines the second shortest path set as the aortic centerline.

[0107] In this embodiment of the application, the aortic sinus endpoint is taken as the starting point and any blood vessel endpoint in the second endpoint set is taken as the ending point, and the second shortest path set is searched from the first shortest path set.

[0108] Subsequently, the electronic device can also determine the centerline corresponding to each shortest path based on the point set corresponding to each shortest path in the second set of shortest paths. For example... Figure 9 As shown, the electronic device merges the centerlines corresponding to all shortest paths in the second shortest path set to obtain the aortic centerline (which can also be called the aortic centerline tree).

[0109] In summary, during the process of acquiring the aortic centerline of the target aorta, the electronic device can automatically locate the starting point (the endpoint of the aortic sinus) and the ending point (the vessel endpoint in the second endpoint set), which improves the efficiency of acquiring the aortic centerline compared to marking vessel endpoints with a watch. Furthermore, the electronic device can automatically acquire the aortic centerline using Dijkstra's algorithm. As the performance of the electronic device gradually improves, the speed of acquiring the aortic centerline gradually increases, enabling the electronic device to acquire aortic centerlines in batches.

[0110] Figure 10 This application provides a schematic diagram of a device 1000 for obtaining the aortic centerline. (See attached diagram.) Figure 10 As shown, the device 1000 includes:

[0111] The first acquisition module 1001 is used to preprocess the computed tomography angiography (CTA) image based on a pre-trained convolutional neural network algorithm to obtain the target segmentation prediction result of the target aorta and the trunk segmentation prediction result of the target aorta; the CTA image includes slice images of the target aorta.

[0112] The second acquisition module 1002 is used to acquire the set of vascular vertices and the set of first endpoints of the vascular segment in the target aorta based on the vascular skeleton line of the target segmentation prediction result; the vascular vertices in the vascular vertices set include at least one vascular endpoint and at least one bifurcation point; the set of first endpoints includes each vascular endpoint in at least one vascular endpoint arranged according to the preset image layering direction.

[0113] The first processing module 1003 is used to extract the aortic sinus endpoints from the first endpoint set based on the trunk segmentation prediction results, and to determine the first endpoint set after deleting the aortic sinus endpoints as the second endpoint set.

[0114] The second processing module 1004 is used to calculate a first shortest path set based on Dijkstra's algorithm and the set of vessel vertices, and extract a second shortest path set from the first shortest path set, and determine the second shortest path set as the aortic centerline; the first path set includes the shortest path from the aortic sinus endpoint to each vessel vertex in the set of vessel vertices, and the second shortest path set includes the shortest path from the aortic sinus endpoint to each vessel endpoint in the second endpoint set.

[0115] It should be understood that the apparatus embodiments for obtaining the aortic centerline and the method embodiments for obtaining the aortic centerline can correspond to each other, and similar descriptions can be found in the method embodiments for obtaining the aortic centerline. To avoid repetition, further details are omitted here. Specifically, Figure 10 The apparatus 1000 shown can execute the above-described method embodiment for obtaining the aortic centerline, and the aforementioned and other operations and / or functions of each module in the apparatus 1000 are respectively for implementing the corresponding process in the above-described method for obtaining the aortic centerline, which will not be described in detail here for the sake of brevity.

[0116] The apparatus 1000 of this application embodiment has been described above from the perspective of functional modules in conjunction with the accompanying drawings. It should be understood that this functional module can be implemented in hardware, in software instructions, or in a combination of hardware and software modules. Specifically, the steps of the method embodiment for obtaining the aortic centerline in this application embodiment can be completed by integrated logic circuits in the processor hardware and / or by software instructions. The steps of the method for obtaining the aortic centerline disclosed in this application embodiment can be directly manifested as execution by a hardware decoding processor, or by a combination of hardware and software modules in the decoding processor. Optionally, the software module can be located in a mature storage medium in the art, such as random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, etc. This storage medium is located in memory, and the processor reads information from the memory and, in conjunction with its hardware, completes the steps of the above-described method embodiment for obtaining the aortic centerline.

[0117] Figure 11 This is a schematic block diagram of the electronic device 1100 provided in the embodiments of this application.

[0118] like Figure 11 As shown, the electronic device 11000 may include:

[0119] The system includes a memory 1110 and a processor 1120. The memory 1110 stores computer programs and transfers the program code to the processor 1120. In other words, the processor 1120 can retrieve and run the computer program from the memory 1110 to implement the methods described in the embodiments of this application.

[0120] For example, the processor 1120 can be used to execute the above-described method embodiments according to instructions in the computer program.

[0121] In some embodiments of this application, the processor 1120 may include, but is not limited to:

[0122] General-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc.

[0123] In some embodiments of this application, the memory 1110 includes, but is not limited to:

[0124] Volatile memory and / or non-volatile memory. Non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. Volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of RAM are available, such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDR SDRAM), Enhanced Synchronous DRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), and Direct Rambus RAM (DR RAM).

[0125] In some embodiments of this application, the computer program may be divided into one or more modules, which are stored in the memory 1110 and executed by the processor 1120 to complete the method provided in this application. The one or more modules may be a series of computer program instruction segments capable of performing specific functions, which describe the execution process of the computer program in the electronic device.

[0126] like Figure 11 As shown, the electronic device may also include:

[0127] Transceiver 1130, which can be connected to processor 1120 or memory 1110.

[0128] The processor 1120 can control the transceiver 1130 to communicate with other devices; specifically, it can send information or data to other devices or receive information or data sent by other devices. The transceiver 1130 may include a transmitter and a receiver. The transceiver 1130 may further include antennas, and the number of antennas may be one or more.

[0129] It should be understood that the various components in the electronic device are connected through a bus system, which includes a data bus, a power bus, a control bus, and a status signal bus.

[0130] This application also provides a computer storage medium storing a computer program thereon, which, when executed by a computer, enables the computer to perform the methods of the above-described method embodiments. Alternatively, embodiments of this application also provide a computer program product containing instructions that, when executed by a computer, cause the computer to perform the methods of the above-described method embodiments.

[0131] When implemented using software, it can be implemented entirely or partially as a computer program product. This computer program product includes one or more computer instructions. When these computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., digital video disc (DVD)), or a semiconductor medium (e.g., solid-state disk (SSD)).

[0132] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0133] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, 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.

[0134] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. For example, the functional modules in the various embodiments of this application may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module.

[0135] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method of obtaining an aortic centerline, the method comprising: The method comprises the following steps: Pretreatment is performed on a computed tomography angiography (CTA) image based on a pre-trained convolutional neural network algorithm to obtain a target segmentation prediction result of a target aorta and a trunk segmentation prediction result of the target aorta; the CTA image comprises slice images of the target aorta; A blood vessel vertex set and a first endpoint set of a blood vessel segment in the target aorta are obtained according to a blood vessel skeleton line of the target segmentation prediction result; the blood vessel vertices in the blood vessel vertex set comprise at least one blood vessel endpoint and at least one bifurcation point; the first endpoint set comprises each blood vessel endpoint in the at least one blood vessel endpoint arranged according to a preset image layering direction; An aortic sinus endpoint is extracted from the first endpoint set according to the trunk segmentation prediction result, and the first endpoint set with the aortic sinus endpoint removed is determined as a second endpoint set; A first shortest path set is calculated according to a Dijkstra algorithm and the blood vessel vertex set, and a second shortest path set is extracted from the first shortest path set, and the second shortest path set is determined as the aortic center line; the first shortest path set comprises a shortest path from the aortic sinus endpoint to each blood vessel vertex in the blood vessel vertex set, and the second shortest path set comprises a shortest path from the aortic sinus endpoint to each blood vessel endpoint in the second endpoint set.

2. The method of claim 1, wherein, The method of pretreating a computed tomography angiography (CTA) image based on a pre-trained convolutional neural network algorithm to obtain a target segmentation prediction result of a target aorta and a trunk segmentation prediction result of the target aorta comprises the following steps: Each slice image in the CTA image is obtained; A gray value of each pixel point in each slice image is input into the pre-trained convolutional neural network algorithm to obtain a gray feature value of each pixel point in each slice image; A trunk mask and a branch mask of the target aorta are obtained according to a preset trunk threshold, a preset branch threshold and the gray feature value of each pixel point in each slice image; An image corresponding to the trunk mask is determined as the trunk segmentation prediction result of the target aorta, and the trunk mask and the branch mask are combined to obtain an artery mask, and an image corresponding to the artery mask is determined as the target segmentation prediction result of the target aorta.

3. The method according to claim 1 or 2, characterized in that, The method of obtaining a blood vessel vertex set and a first endpoint set of a blood vessel segment in the target aorta according to a blood vessel skeleton line of the target segmentation prediction result comprises the following steps: The target segmentation prediction result is refined according to a preset refinement algorithm to obtain a blood vessel skeleton line of the target aorta; A blood vessel vertex of a blood vessel segment in the target aorta is obtained according to a number of neighborhood points of each pixel point on the blood vessel skeleton line to obtain the blood vessel vertex set; At least one blood vessel endpoint with the number of neighborhood points being 1 is extracted from the blood vessel vertex set, and each blood vessel endpoint in the at least one blood vessel endpoint is generated into the first endpoint set according to a preset image layering direction.

4. The method of claim 2, wherein, The method comprises the following steps: According to the main stem segmentation prediction result, aortic sinus end point is extracted from the first end point set, and the first end point set in which the aortic sinus end point is deleted is determined as the second end point set, which comprises the following steps: According to the main stem segmentation prediction result, aortic sinus end point is extracted from the first end point set, and the first end point set in which the aortic sinus end point is deleted is determined as the second end point set, which comprises the following steps: According to the main stem segmentation prediction result, aortic sinus end point is extracted from the first end point set, and the first end point set in which the aortic sinus end point is deleted is determined as the second end point set, which comprises the following steps:

5. The method of claim 1, wherein, According to the main stem segmentation prediction result, aortic sinus end point is extracted from the first end point set, and the first end point set in which the aortic sinus end point is deleted is determined as the second end point set, which comprises the following steps: According to the Dijkstra algorithm and the blood vessel vertex set, a first shortest path set is calculated, and a second shortest path set is extracted from the first shortest path set, and the second shortest path set is determined as the aortic center line, which comprises the following steps: The aortic sinus end point is moved in the blood vessel vertex set, so that the first element of the blood vessel vertex set is the aortic sinus end point; The blood vessel segment length between each blood vessel vertex in the blood vessel vertex set is taken as the weight value to generate a vertex directed graph; According to the Dijkstra algorithm and the vertex directed graph, a first shortest path set is calculated; 6. An apparatus for obtaining an aortic centerline, the apparatus comprising: According to each blood vessel end point in the second end point set, a second shortest path set is extracted from the first shortest path set, and the second shortest path set is determined as the aortic center line. The method comprises the following steps: The first acquisition module is used for pre-processing the computed tomography angiography (CTA) image of the target aorta based on a pre-trained convolutional neural network algorithm to obtain a target segmentation prediction result of the target aorta and a main stem segmentation prediction result of the target aorta; the CTA image comprises a slice image of the target aorta; The second acquisition module is used for obtaining a blood vessel vertex set and a first end point set of a blood vessel segment in the target aorta according to a blood vessel skeleton line of the target segmentation prediction result; the blood vessel vertex in the blood vessel vertex set comprises at least one blood vessel end point and at least one bifurcation point; the first end point set comprises each blood vessel end point in the at least one blood vessel end point arranged according to a preset image layering direction; The first processing module is used for extracting aortic sinus end point from the first end point set according to the main stem segmentation prediction result, and determining the first end point set in which the aortic sinus end point is deleted as the second end point set. The second processing module is configured to calculate a first shortest path set according to the Dijkstra algorithm and the set of vessel vertices, and extract a second shortest path set from the first shortest path set, so as to determine the second shortest path set as the aortic center line; the first shortest path set comprises a shortest path from the aortic sinus end point to each vessel vertex in the set of vessel vertices, and the second shortest path set comprises a shortest path from the aortic sinus end point to each vessel end point in the second end point set.

7. The apparatus of claim 6, wherein, The first acquisition module is specifically configured to: acquire each slice image in the CTA image; input a gray value of each pixel point in the each slice image into the pre-trained convolutional neural network algorithm to obtain a gray feature value of each pixel point in the each slice image; obtain a main stem mask value and a branch mask value of the target aorta according to a preset main stem threshold, a preset branch threshold and the gray feature value of each pixel point in the each slice image; determine an image corresponding to the main stem mask value as a main stem segmentation prediction result of the target aorta, and combine the main stem mask value and the branch mask value to obtain an artery mask value, and determine an image corresponding to the artery mask value as a target segmentation prediction result of the target aorta.

8. An electronic device, comprising: comprise: a processor and a memory, the memory being configured to store a computer program, and the processor being configured to invoke and run the computer program stored in the memory to execute the method in any one of claims 1-5.

9. A computer-readable storage medium, characterized in that, A computer program for storing, the computer program enabling a computer to execute the method in any one of claims 1-5.

10. A computer program product comprising instructions, characterized in that, When the computer program product runs on an electronic device, the electronic device is enabled to execute the method in any one of claims 1-5.

Images