Coronary artery center line segmentation naming method and device, electronic equipment and storage medium

Abstract

The invention discloses a coronary artery center line segment naming method and device, electronic equipment and a storage medium. The method comprises the following steps: segmenting a coronary artery image to obtain an aorta segmentation result and a coronary artery segmentation result; segmenting the coronary artery segmentation result through a point cloud segmentation network to obtain a coronary artery segmentation result corresponding to the coronary artery segmentation result; extracting a coronary artery center line from the coronary artery segmentation result, and determining a target segmentation result of the coronary artery center line according to the coronary artery segmentation result; according to the aorta segmentation result and the coronary artery segmentation result, determining a coronary artery left tree root node and a coronary artery right tree root node from the target segmentation result; and determining the name of each coronary artery sub-segment in the target segmentation result according to the coronary artery left root node and the coronary artery right root node. According to the method, the accuracy of segmented naming of the coronary artery center line can be improved.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of medical image processing, and particularly relates to a coronary centerline segmentation naming method and device, electronic equipment and a storage medium. BACKGROUND

[0002] Coronary arteries are arteries that branch around the heart. Coronary artery segmentation naming is to distinguish and name each segment of the blood vessel according to the origin, bifurcation, spatial distribution and trend of the arterial blood vessel around the heart. According to the standard of the International Cardiovascular CT Association (SCCT), the coronary blood vessel is divided into 18 segments. Automatic coronary artery segmentation naming has important value in an automatic heart coronary artery auxiliary diagnosis system, including providing support for automatic lesion positioning, supporting doctors to quickly locate lesions and navigate in multiple blood vessels, and assisting in the structuring and standardization of coronary artery disease diagnosis reports.

[0003] In the prior art, the coronary centerline segmentation naming method can adopt a rule-based naming method or a convolutional neural network-based method. The rule-based naming method first calculates various attributes of all vessel segments of the vessel tree, and then traverses each bifurcation point according to the attributes to calculate the naming score of each bifurcation point. This rule-based naming method has poor robustness because the coronary artery pattern varies greatly between different patients, and cannot accurately segment and name different patients. The convolutional neural network-based method uses a convolutional neural network to extract features, and then uses a fully connected layer to analyze the features. This application has the advantage of better extracting local features, but for blood vessel trees with strong topological structure, the convolutional neural network ignores the connectivity, resulting in insufficient accuracy of the segmentation naming result. SUMMARY

[0004] The purpose of the embodiments of the present application is to provide a coronary centerline segmentation naming method, device, electronic equipment and storage medium, which can solve the problem of insufficient accuracy of coronary centerline segmentation naming.

[0005] To solve the above technical problems, the present application is implemented as follows: In a first aspect, the embodiments of the present application provide a coronary centerline segmentation naming method, which comprises: segmenting a coronary image to obtain an aorta segmentation result and a coronary segmentation result; segmenting the coronary segmentation result by a point cloud segmentation network to obtain a coronary segmentation result corresponding to the coronary segmentation result; extracting a coronary centerline from the coronary segmentation result, and determining a target segmentation result of the coronary centerline according to the coronary segmentation result; determine, from the aorta segmentation result and the coronary segmentation result, a left coronary tree root node and a right coronary tree root node in the target segmentation result; determine, according to the left coronary tree root node and the right coronary tree root node, names of each coronary subsegment in the target segmentation result.

[0006] In a second aspect, an embodiment of the present application provides a device for naming coronary centerline segments, including: a segmentation module configured to segment a coronary image to obtain an aorta segmentation result and a coronary segmentation result; a coronary segmentation module configured to perform segmentation processing on the coronary segmentation result by using a point cloud segmentation network to obtain a coronary segmentation result corresponding to the coronary segmentation result; a centerline segmentation module configured to extract a coronary centerline from the coronary segmentation result and determine a target segmentation result of the coronary centerline according to the coronary segmentation result; a root node determination module configured to determine, from the aorta segmentation result and the coronary segmentation result, a left coronary tree root node and a right coronary tree root node in the target segmentation result; a segmentation naming module configured to determine, according to the left coronary tree root node and the right coronary tree root node, names of each coronary subsegment in the target segmentation result.

[0007] In a third aspect, an embodiment of the present application provides an electronic device, which includes a processor, a memory, and a program or instruction stored in the memory and executable on the processor, and the program or instruction is executed by the processor to implement the steps of the method according to the first aspect.

[0008] In a fourth aspect, an embodiment of the present application provides a readable storage medium, which stores a program or instruction, and the program or instruction is executed by a processor to implement the steps of the method according to the first aspect.

[0009] In a fifth aspect, an embodiment of the present application provides a chip, which includes a processor and a communication interface, the communication interface is coupled with the processor, and the processor is configured to run a program or instruction to implement the method according to the first aspect.

[0010] In the embodiment of the present application, the aorta segmentation result and the coronary artery segmentation result are obtained by segmenting the coronary image, the coronary artery segmentation result is segmented by the point cloud segmentation network to obtain the coronary artery segmentation result corresponding to the coronary artery segmentation result, the coronary artery centerline is extracted from the coronary artery segmentation result, the target segmentation result of the coronary artery centerline is determined according to the coronary artery segmentation result, the coronary artery left tree root node and the coronary artery right tree root node are determined from the target segmentation result according to the aorta segmentation result and the coronary artery segmentation result, and the names of each coronary artery subsegment in the target segmentation result are determined according to the coronary artery left tree root node and the coronary artery right tree root node. In this way, the point cloud segmentation network is used for segmentation processing, which can accurately identify the changes in the topological structure such as vessel rupture and redundant branches, thereby improving the accuracy of coronary artery centerline segmentation naming. BRIEF DESCRIPTION OF DRAWINGS

[0011] Figure 1 is a schematic diagram of the 18 segments of coronary arteries specified by SCCT; Figure 2 is a flowchart of a method for naming coronary artery centerline segmentation provided by an embodiment of the present application; Figure 3 is a structural schematic diagram of a device for naming coronary artery centerline segmentation provided by an embodiment of the present application; Figure 4 is a hardware structural schematic diagram of an electronic device for implementing an embodiment of the present application. DETAILED DESCRIPTION

[0012] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work fall within the scope of protection of the present application.

[0013] The terms "first", "second", and the like in the specification and claims of the present application are used to distinguish similar objects, and are not used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances, so that the embodiments of the present application can be implemented in an order other than those illustrated or described herein. In addition, "and / or" in the specification and claims indicates at least one of the connected objects, and the character " / ", generally indicates that the front and rear associated objects are in a "or" relationship.

[0014] The method for naming coronary artery centerline segmentation provided by the embodiments of the present application will be described in detail below in combination with the drawings and specific embodiments and their application scenarios.

[0015] Figure 1 is a schematic diagram of the 18 segments of coronary arteries specified by SCCT, asFigure 1 As shown, according to the 18-segment coronary artery segmentation system published by SCCT in 2014, in order, the coronary artery segments are proximal right coronary artery (pRCA), middle right coronary artery (mRCA), distal right coronary artery (dRCA), right coronary artery posterior descending branch (R PDA), left main (LM), proximal left anterior descending (pLAD), middle left anterior descending (mLAD), distal left anterior descending (dLAD), first diagonal branch (D1), second diagonal branch (D2), proximal left circumflex (pLCX), first obtuse marginal branch (OM1), middle and distal left circumflex (LCx), second obtuse marginal branch (OM2), left circumflex posterior descending branch (L PDA), right coronary artery posterior descending branch (R PLB), intermediate branch (RI), and left circumflex posterior descending branch (L PLB).

[0016] Embodiments of the present application are to identify the coronary centerline and each segment in the coronary image.

[0017] Figure 2 is a flowchart of a method for coronary centerline segmentation naming provided by an embodiment of the present application, which is applied to the scenario of coronary centerline identification and segmentation naming of a coronary image. As shown, the method comprises steps 210 to 250. Figure 2 As shown, the method comprises steps 210 to 250.

[0018] Step 210, segmenting the coronary image to obtain an aorta segmentation result and a coronary segmentation result.

[0019] The coronary image is a 3D medical image obtained by medical imaging means, which can be a CT image, a nuclear magnetic resonance image, etc.

[0020] The coronary image can be preprocessed based on a convolutional neural network to segment the aorta and the coronary in the coronary image to obtain the aorta segmentation result and the coronary segmentation result. The convolutional neural network model can include VNet, UNet, or ResUNet, etc. The features of the coronary image are extracted by the convolutional neural network, and the prediction result of the aorta and the coronary in the coronary image is obtained by the softmax function. If it is a coronary, it can be marked as 1, if it is an aorta, it can be marked as 2, and other regions are marked as 0 to obtain the mask image G_aorta_cor_two_label of the aorta and the coronary, i.e. the three-dimensional segmentation result of the aorta and the coronary.

[0021] Based on the three-dimensional segmentation results of the aorta and the coronary artery, the aorta segmentation result and the coronary artery segmentation result can be obtained. In the three-dimensional segmentation results of the aorta and the coronary artery, the value of the aorta can be set to 0, and only the mask image G_cor_one_label of the coronary artery is retained, that is, the coronary artery segmentation result is obtained, which is used as the basis for the next step of processing.

[0022] The aorta segmentation result and the coronary artery segmentation result can be different mask images in the same image, or independent mask images.

[0023] In step 220, the coronary artery segmentation result is segmented by the point cloud segmentation network to obtain a coronary artery segmentation result corresponding to the coronary artery segmentation result.

[0024] The point cloud segmentation network is a pre-trained neural network for segmenting the point cloud data of the coronary artery. The point cloud segmentation network can include PointNet, PointNet++, etc.

[0025] The point cloud data corresponding to the coronary artery segmentation result is determined, and the point cloud data is input into the point cloud segmentation network. The point cloud data is segmented by the point cloud segmentation network to identify the target number of coronary artery segments in the point cloud data, that is, each point is marked with a label corresponding to the segment. The segmented point cloud data is converted to an image coordinate to obtain a coronary artery segmentation result corresponding to the coronary artery segmentation result. The target number can be determined according to different coronary artery segmentation systems. If the segmentation system shown in FIG. 18 is used, the target number is 18. Figure 1

[0026] In some embodiments of the present application, the coronary artery segmentation result is segmented by the point cloud segmentation network to obtain a coronary artery segmentation result corresponding to the coronary artery segmentation result, including: converting the coronary artery segmentation result to point cloud data; segmenting the point cloud data by the point cloud segmentation network to obtain a point cloud segmentation result; and performing inverse point cloud processing on the point cloud segmentation result to obtain the coronary artery segmentation result.

[0027] The point cloud segmentation network includes a multi-layer perceptron and a maximum pooling operation.

[0028] The point cloud is a set of points representing a target space in the same spatial reference coordinate system. For the coronary artery segmentation result G_cor_one_label, the pixel coordinates can be converted to spatial coordinates by a pre-set sampling interval (spacing) to obtain the position coordinates of each sampling point (i.e., pixel point), and the point cloud data corresponding to the coronary artery segmentation result is obtained. The pre-set sampling interval is the sampling interval corresponding to the acquisition of the coronary image.

[0029] ​Point cloud neural networks can capture the implicit spatial relationships between points. Unlike traditional neural networks, which are input to ordered feature maps and can only process data sampled in a uniform grid, point cloud neural networks take random-ordered 3D point coordinates as input, enabling them to handle data sampled in a non-uniform grid. Point cloud segmentation networks can leverage the symmetry of multilayer perceptrons (MLPs) and max pooling to address the disorder problem of point cloud data and eliminate the influence of disorder on the final result.

[0030] The point cloud data corresponding to the coronary artery segmentation results is input into the point cloud segmentation network. The point cloud segmentation network is highly robust to changes in topology such as ruptured blood vessels and redundant branches. After processing by the point cloud segmentation network, the category of the coronary artery sub-segment to which each point belongs can be obtained, that is, the point cloud segmentation result is obtained.

[0031] Finally, the point cloud segmentation result is reversed, that is, the coordinates of the sampling points are converted into pixel coordinates to obtain the segmented mask image G_cor_multi_label of the coronary artery, which is the coronary artery segmentation result.

[0032] By converting the coronary artery segmentation results into point cloud data, and then segmenting the point cloud data using a point cloud segmentation network, the accuracy of the coronary artery segmentation results can be improved because the point cloud segmentation network can accurately capture the implicit spatial relationships between points, thereby accurately utilizing the connectivity relationships in the vascular tree.

[0033] Step 230: Extract the coronary artery centerline from the coronary artery segmentation results, and determine the target segmentation result of the coronary artery centerline based on the coronary artery segmentation results.

[0034] The coronary artery region in the coronary artery segmentation result is refined into a line as the coronary artery centerline. Each point in the coronary artery centerline is compared with the coronary artery segmentation result to determine the segment to which each point in the coronary artery centerline belongs, thus obtaining the target segmentation result of the coronary artery centerline.

[0035] In some embodiments of this application, extracting the coronary artery centerline from the coronary artery segmentation result includes: determining the coronary artery region from the coronary artery segmentation result; and refining the coronary artery region based on a three-dimensional refinement algorithm to obtain the coronary artery centerline.

[0036] From the coronary artery segmentation result G_cor_one_label, find the two largest connected regions, which are the left and right trees of the coronary artery. Remove some small noise areas to obtain the coronary artery region. Then, process the coronary artery region using a three-dimensional thinning algorithm (such as 3D Thinning) to obtain the skeleton line C_cor_one_label of the coronary artery region, which serves as the coronary artery centerline.

[0037] By identifying the coronary artery region in the coronary artery segmentation results, noise areas can be removed, and a three-dimensional thinning algorithm can be applied to the coronary artery region for thinning processing, avoiding the influence of noise areas and obtaining an accurate coronary artery centerline.

[0038] In some embodiments of this application, determining the target segmentation result of the coronary artery centerline based on the coronary artery segmentation result includes: for each pixel on the coronary artery centerline, determining the coronary artery sub-segment to which the pixel belongs based on the neighboring points of the pixel in the coronary artery segmentation result, and obtaining the initial segmentation result of the coronary artery centerline; determining the maximum connected component of each coronary artery sub-segment in the initial segmentation result, and merging the maximum connected components in which the number of pixels is less than a number threshold into adjacent maximum connected components, and obtaining the target segmentation result.

[0039] Based on the coronary artery segmentation result G_cor_multi_label, the coronary artery sub-segment to which each pixel on the coronary artery centerline belongs is determined. Based on the coronary artery sub-segments to which the neighboring points of a pixel on the coronary artery centerline belong in the coronary artery segmentation result, the coronary artery sub-segment to which the pixel belongs is determined. This process is performed on each pixel on the coronary artery centerline to obtain the initial segmentation result of the coronary artery centerline.

[0040] Based on the initial segmentation result C_cor_multi_label, the largest connected component of each coronary artery segment is found. From all the largest connected components, the largest connected component with a pixel count less than a threshold is identified. The coronary artery segment represented by this largest connected component is considered a noise segment and needs to be merged. This largest connected component is then designated as the connected component to be merged. If the largest connected components adjacent to each other on both sides of the connected component to be merged belong to the same coronary artery segment, then the connected component to be merged is merged with the adjacent largest connected components on both sides into one connected component. The coronary artery field to which the connected component to be merged belongs is set as the coronary artery segment to which the adjacent largest connected components belong. If the largest connected components adjacent to each other on both sides of the connected component to be merged belong to different coronary artery segments, then the connected component to be merged is merged into the previous largest connected component, i.e., the coronary artery segment to which the previous largest connected component belongs is taken as the coronary artery segment to which the connected component to be merged belongs. After merging all connected components to be merged into adjacent largest connected components, the final segmentation result, i.e., the target segmentation result, is obtained. The previous largest connected component refers to the largest connected component located before the connected component to be merged, originating from either the left or right root node of the coronary artery.

[0041] It should be noted that the operation of merging connected components can also be performed after the left root node and the right root node of the coronary vein have been determined.

[0042] By merging the largest connected regions with fewer pixels into adjacent largest connected regions after determining the initial segmentation results of the coronary artery centerline, the situation of individual pixels being misclassified on the coronary artery centerline can be avoided, thereby further improving the accuracy of the segmentation results.

[0043] In one optional implementation, determining the coronary artery sub-segment to which the pixel belongs based on its neighboring points in the coronary artery segmentation results includes: determining the coronary artery sub-segment to which the pixel belongs as the coronary artery sub-segment with the largest number of neighboring points belonging to the same coronary artery sub-segment.

[0044] Based on the coronary artery segmentation result G_cor_multi_label, each pixel on the coronary artery centerline undergoes a voting process among its surrounding neighbors. The pixel with the most votes is identified as the segmentation result for that pixel. Specifically, for the current pixel, the coronary artery segment to which each neighboring point belongs in the coronary artery segmentation result is determined. The number of neighboring points corresponding to each coronary artery segment is counted, and the coronary artery segment with the most neighboring points is identified as the coronary artery segment to which the pixel belongs. This process is iteratively applied to all pixels on the coronary artery centerline to obtain the segmentation mask image C_cor_multi_label, which is the initial segmentation result.

[0045] If the neighboring points around the current pixel belong to the same coronary artery segment, the pixel can be directly identified as belonging to that segment. If the neighboring points around the current pixel belong to different coronary artery segments, it means that the pixel is located near two or more segments, such as near the boundary between two adjacent segments. The more neighboring points a pixel has, the more likely it is to belong to the segment to which those points belong. By identifying the segment with the most neighboring points among the segments around the current pixel, the accuracy of determining the segment to which the pixel belongs can be improved.

[0046] Step 240: Based on the aortic segmentation results and the coronary artery segmentation results, determine the left root node and the right root node of the coronary artery from the target segmentation results.

[0047] The process involves identifying the boundary points between the aortic and coronary artery segmentation results. Based on these boundary points, the left and right root nodes of the coronary arteries are then determined from the target segmentation results; specifically, the left and right root nodes of the coronary arteries along the coronary centerline are identified. Alternatively, the initial segmentation results can be used as the processing object. This involves determining the left and right root nodes of the coronary arteries from the initial segmentation results, following a similar process to that used from the target segmentation results.

[0048] In some embodiments of this application, determining the left root node and right root node of the coronary artery from the target segmentation results based on the aortic segmentation results and the coronary artery segmentation results includes: performing dilation processing on the aortic segmentation results to obtain the aortic dilation region; determining two sets of intersection points between the aortic dilation region and the coronary artery segmentation results; and determining the left root node and right root node of the coronary artery from the target segmentation results based on the two sets of intersection points.

[0049] Based on the mask images G_aorta_cor_two_label of the aorta and coronary arteries, the junction between the aorta and coronary arteries can be found, namely the left and right root nodes of the coronary artery tree. First, morphological dilation is used to dilate the aorta segmentation results to obtain the aorta dilated region. Then, the intersection of the aorta dilated region and the coronary artery segmentation results is determined, resulting in two sets of intersection points.

[0050] The endpoints in the target segmentation result can be determined, and the endpoints located within the range of the two boundary point sets can be determined from all the endpoints to obtain the left root node and right root node of the coronary vein.

[0051] By dilating the aortic segmentation results, the dilated region will intersect with the coronary artery segmentation results, resulting in two sets of boundary points between the dilated region and the coronary artery segmentation results. These boundary point sets are the pixels near the root nodes of the left and right coronary artery tree. Based on these boundary point sets, pixels in the target segmentation result that are far from these boundary point sets can be excluded, thus accurately determining the root nodes of the left and right coronary artery tree in the target segmentation result.

[0052] In one optional implementation, determining the left root node and right root node of the coronary artery from the target segmentation result based on two boundary point sets includes: clustering each boundary point set to obtain the cluster center point corresponding to each boundary point set; determining the endpoints in the target segmentation result; and determining the endpoint with the smallest distance to each cluster center point from the endpoints as the left root node and right root node of the coronary artery.

[0053] Perform KMeans clustering on the two boundary point sets respectively to obtain the cluster center point corresponding to each boundary point set, that is, obtain the coordinates of the two cluster center points.

[0054] Based on the coronary artery centerline C_cor_one_label, each pixel on the centerline is classified. When a pixel has only one neighbor, it is designated as an endpoint, resulting in two endpoint sets. For each cluster center, the distances between each endpoint in the endpoint set on the same side of the cluster center and the cluster center are calculated. The endpoint with the smallest distance is designated as the root node, resulting in the left root node P_left and the right root node P_right of the coronary artery.

[0055] Clustering the boundary point set yields cluster centers. Then, by using the distance between the endpoints of the coronary artery centerline and the cluster centers, endpoints that are far from the cluster centers can be eliminated. Thus, the endpoints with the smallest distance from the cluster centers among the endpoints of the coronary artery centerline are identified as the left and right root nodes of the coronary artery tree, which can further improve the accuracy of determining the left and right root nodes of the coronary artery tree.

[0056] Step 250: Determine the name of each coronary sub-segment in the target segmentation result based on the left and right root nodes of the coronary artery tree.

[0057] The target segmentation results (i.e., the centerline segments after merging connected components, i.e., coronary artery sub-segments) are then subjected to anatomical structure correction. This involves starting from the root node of the left coronary tree and sequentially determining the names of each coronary artery sub-segment in the left tree, and starting from the root node of the right coronary tree and sequentially determining the names of each coronary artery sub-segment in the right tree. For example, starting from the root node P_right of the right coronary tree, the sub-segments are sequentially: proximal right coronary artery (pRCA), mid-right coronary artery (mRCA), distal right coronary artery (dRCA), and the posterior descending artery of the right coronary artery (R...). PDA) and the right coronary posterior branch (R A coronary artery right tree (PLB) is formed, and this parent-child node relationship is used to correct each merged coronary artery segment. After this anatomical correction, the true coronary artery segment names are obtained, which are the final segmentation results and names of the coronary artery centerline.

[0058] The method for naming coronary artery centerlines provided in this application involves segmenting a coronary artery image to obtain aortic and coronary artery segmentation results. A point cloud segmentation network is then used to segment the coronary artery segmentation results, resulting in corresponding coronary artery segmentation results. The coronary artery centerline is extracted from these segmentation results, and a target segmentation result is determined based on the segmentation. The left and right root nodes of the coronary artery tree are determined from the target segmentation result based on the aortic and coronary artery segmentation results. Based on these root nodes, the names of each coronary artery sub-segment in the target segmentation result are determined. This segmentation process using a point cloud segmentation network accurately identifies topological changes such as vascular ruptures and redundant branches, thereby improving the accuracy of coronary artery centerline segmentation naming.

[0059] It should be noted that the coronary artery centerline segmentation naming method provided in this application embodiment can be executed by a coronary artery centerline segmentation naming device, or a control module within that device for executing the method of loading coronary artery centerline segmentation naming. This application embodiment uses the execution of the method of loading coronary artery centerline segmentation naming by a coronary artery centerline segmentation naming device as an example to illustrate the coronary artery centerline segmentation naming method provided in this application embodiment.

[0060] Figure 3 This is a schematic diagram of a device for segmented naming of coronary artery centerlines provided in an embodiment of this application, as shown below. Figure 3 As shown, the device for segmented naming of the coronary artery centerline includes: The segmentation module 310 is used to segment the coronary artery image to obtain the aortic segmentation result and the coronary artery segmentation result; The coronary artery segmentation module 320 is used to segment the coronary artery segmentation result through a point cloud segmentation network to obtain the coronary artery segmentation result corresponding to the coronary artery segmentation result. The centerline segmentation module 330 is used to extract the coronary artery centerline from the coronary artery segmentation result and determine the target segmentation result of the coronary artery centerline based on the coronary artery segmentation result. The root node determination module 340 is used to determine the left root node and the right root node of the coronary artery from the target segmentation results based on the aortic segmentation results and the coronary artery segmentation results. The segmentation naming module 350 is used to determine the name of each coronary sub-segment in the target segmentation result based on the left root node and the right root node of the coronary artery.

[0061] Optionally, the coronary artery segmentation module is specifically used for: The coronary artery segmentation results are converted into point cloud data; The point cloud data is segmented using a point cloud segmentation network to obtain point cloud segmentation results. The point cloud segmentation results are subjected to inverse point cloudification processing to obtain the coronary artery segmentation results.

[0062] Optionally, the point cloud segmentation network includes a multilayer perceptron and a max pooling operation.

[0063] Optionally, the centerline segmentation module includes a centerline extraction unit, which is used for: The coronary artery region is determined from the coronary artery segmentation results; The coronary artery region is refined using a three-dimensional refinement algorithm to obtain the coronary artery centerline.

[0064] Optionally, the centerline segmentation module includes a centerline segmentation unit, which is used for: For each pixel on the coronary artery centerline, the coronary artery sub-segment to which the pixel belongs is determined based on the neighboring points of the pixel in the coronary artery segmentation result, thus obtaining the initial segmentation result of the coronary artery centerline; Determine the maximum connected component of each coronary artery segment in the initial segmentation result, and merge the maximum connected components in which the number of pixels is less than a threshold into adjacent maximum connected components to obtain the target segmentation result.

[0065] Optionally, determining the coronary artery sub-segment to which the pixel belongs based on its neighboring points in the coronary artery segmentation result includes: The coronary vein segment with the largest number of neighboring points belonging to the same coronary vein segment among the neighboring points is determined as the coronary vein segment to which the pixel belongs.

[0066] Optionally, the root node determination module includes: An expansion processing unit is used to expand the aortic segmentation result to obtain an expanded aortic region. The boundary point determination unit is used to determine two sets of boundary points between the aortic expansion region and the coronary artery segmentation result; The root node determination unit is used to determine the left root node of the coronary artery and the right root node of the coronary artery from the target segmentation result based on the two boundary point sets.

[0067] Optionally, the root node determination unit is specifically used for: Each set of boundary points is clustered to obtain the cluster center point corresponding to each set of boundary points. Determine the endpoints in the target segmentation result; The endpoint with the smallest distance to each of the cluster centers is determined from the endpoints and is used as the left root node and the right root node of the coronary artery.

[0068] The device for naming the coronary artery centerline segments in this application embodiment can be a device, or a component, integrated circuit, or chip in a terminal. The device can be a mobile electronic device or a non-mobile electronic device. For example, mobile electronic devices can be mobile phones, tablets, laptops, PDAs, in-vehicle electronic devices, wearable devices, ultra-mobile personal computers (UMPCs), netbooks, or personal digital assistants (PDAs), etc., while non-mobile electronic devices can be servers, network-attached storage (NAS), personal computers (PCs), televisions (TVs), ATMs, or self-service machines, etc. This application embodiment does not impose specific limitations.

[0069] The device for segmenting and naming the coronary artery centerline in this embodiment can be a device with an operating system. This operating system can be Android, iOS, or other possible operating systems; this embodiment does not specifically limit its use.

[0070] The coronary artery centerline segmentation naming provided in this application embodiment can achieve… Figure 2 The various processes implemented in the method implementation examples will not be described again here to avoid repetition.

[0071] The coronary artery centerline segmentation and naming device provided in this application segmentes a coronary artery image to obtain aortic segmentation results and coronary artery segmentation results. It then uses a point cloud segmentation network to segment the coronary artery segmentation results, obtaining corresponding coronary artery segmentation results. The coronary artery centerline is extracted from the segmentation results, and a target segmentation result is determined based on the segmentation results. The left and right root nodes of the coronary artery tree are determined from the target segmentation results based on the aortic and coronary artery segmentation results. Based on these nodes, the names of each coronary artery sub-segment in the target segmentation result are determined. This segmentation process using a point cloud segmentation network accurately identifies topological changes such as vascular ruptures and redundant branches, thereby improving the accuracy of coronary artery centerline segmentation and naming.

[0072] Optionally, embodiments of this application also provide an electronic device, including a processor, a memory, and a program or instructions stored in the memory and executable on the processor. When the program or instructions are executed by the processor, they implement the various processes of the above-described method embodiment for segmented naming of coronary artery centerlines and achieve the same technical effect. To avoid repetition, they will not be described again here.

[0073] It should be noted that the electronic devices in the embodiments of this application include the mobile electronic devices and non-mobile electronic devices described above.

[0074] Figure 4 A schematic diagram of the hardware structure of an electronic device to implement an embodiment of this application. The electronic device 400 includes, but is not limited to, components such as: a radio frequency unit 401, a network module 402, an audio output unit 403, an input unit 404, a sensor 405, a display unit 406, a user input unit 407, an interface unit 408, a memory 409, and a processor 410. The input unit 404 may include an image processor 4041 and a microphone 4042; the display unit 406 may include a display panel 4061; and the user input unit 407 may include a touch panel 4071 and other input devices (such as a keyboard) 4072.

[0075] Those skilled in the art will understand that the electronic device 400 may also include a power supply (such as a battery) for supplying power to various components. The power supply may be logically connected to the processor 410 through a power management system, thereby enabling functions such as managing charging, discharging, and power consumption through the power management system. Figure 4 The electronic device structure shown does not constitute a limitation on the electronic device. The electronic device may include more or fewer components than shown, or combine certain components, or have different component arrangements, which will not be elaborated here. The memory 409 stores a program or instruction that can run on the processor. When the program or instruction is executed by the processor 410, it implements the various processes of the above-described method embodiment for naming the coronary artery centerline segment and achieves the same technical effect. To avoid repetition, it will not be described again here.

[0076] This application also provides a readable storage medium storing a program or instructions. When the program or instructions are executed by a processor, they implement the various processes of the above-described method embodiment for naming segmented coronary artery centerlines and achieve the same technical effect. To avoid repetition, these will not be described again here.

[0077] The processor is the processor in the electronic device described in the above embodiments. The readable storage medium includes computer-readable storage media, such as computer read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk.

[0078] This application embodiment also provides a chip, which includes a processor and a communication interface. The communication interface is coupled to the processor. The processor is used to run programs or instructions to implement the various processes of the above-described method embodiment for segmented naming of coronary artery centerlines, and can achieve the same technical effect. To avoid repetition, it will not be described again here.

[0079] It should be understood that the chip mentioned in the embodiments of this application may also be referred to as a system-on-a-chip, system chip, chip system, or system-on-a-chip, etc.

[0080] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element. Furthermore, it should be noted that the scope of the methods and apparatuses in the embodiments of this application is not limited to performing functions in the order shown or discussed, but may also include performing functions substantially simultaneously or in the reverse order, depending on the functions involved. For example, the described methods may be performed in a different order than described, and various steps may be added, omitted, or combined. Additionally, features described with reference to certain examples may be combined in other examples.

[0081] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal (which may be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0082] The embodiments of this application have been described above with reference to the accompanying drawings. However, this application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit and scope of the claims, and all of these forms are within the protection scope of this application.

Claims

1. A method for segmented naming of coronary artery centerlines, characterized in that, include: The coronary artery image is segmented to obtain the aortic segmentation result and the coronary artery segmentation result; The coronary artery segmentation result is segmented using a point cloud segmentation network to obtain the coronary artery segmentation result corresponding to the coronary artery segmentation result. Extract the coronary artery centerline from the coronary artery segmentation results, and determine the target segmentation result of the coronary artery centerline based on the coronary artery segmentation results; Based on the aortic segmentation results and the coronary artery segmentation results, the left root node and the right root node of the coronary artery are determined from the target segmentation results; Based on the left root node and the right root node of the coronary artery, the name of each coronary artery sub-segment in the target segmentation result is determined.

2. The method according to claim 1, characterized in that, The step of segmenting the coronary artery segmentation result using a point cloud segmentation network to obtain the corresponding coronary artery segmentation result includes: The coronary artery segmentation results are converted into point cloud data; The point cloud data is segmented using a point cloud segmentation network to obtain point cloud segmentation results. The point cloud segmentation results are subjected to inverse point cloudification processing to obtain the coronary artery segmentation results.

3. The method according to claim 1, characterized in that, Extracting the coronary artery centerline from the coronary artery segmentation results includes: The coronary artery region is determined from the coronary artery segmentation results; The coronary artery region is refined using a three-dimensional refinement algorithm to obtain the coronary artery centerline.

4. The method according to any one of claims 1-3, characterized in that, The step of determining the target segmentation result of the coronary artery centerline based on the coronary artery segmentation result includes: For each pixel on the coronary artery centerline, the coronary artery sub-segment to which the pixel belongs is determined based on the neighboring points of the pixel in the coronary artery segmentation result, thus obtaining the initial segmentation result of the coronary artery centerline; Determine the maximum connected component of each coronary artery segment in the initial segmentation result, and merge the maximum connected components in which the number of pixels is less than a threshold into adjacent maximum connected components to obtain the target segmentation result.

5. The method according to claim 4, characterized in that, Determining the coronary artery sub-segment to which the pixel belongs based on its neighboring points in the coronary artery segmentation result includes: The coronary vein segment with the largest number of neighboring points belonging to the same coronary vein segment among the neighboring points is determined as the coronary vein segment to which the pixel belongs.

6. The method according to any one of claims 1-3, characterized in that, The step of determining the left root node and right root node of the coronary artery from the target segmentation results based on the aortic segmentation results and the coronary artery segmentation results includes: The aortic segmentation results are subjected to dilation processing to obtain the aortic dilation region; Determine two sets of intersection points between the aortic dilatation region and the coronary artery segmentation result; Based on the two sets of intersection points, the left root node of the coronary artery and the right root node of the coronary artery are determined from the target segmentation results.

7. The method according to claim 6, characterized in that, The step of determining the left root node and the right root node of the coronary artery from the target segmentation result based on the two boundary point sets includes: Each set of boundary points is clustered to obtain the cluster center point corresponding to each set of boundary points. Determine the endpoints in the target segmentation result; The endpoint with the smallest distance to each of the cluster centers is determined from the endpoints and is used as the left root node and the right root node of the coronary artery.

8. A device for segmented naming of coronary artery centerlines, characterized in that, include: The segmentation module is used to segment the coronary artery image to obtain the aortic segmentation result and the coronary artery segmentation result; The coronary artery segmentation module is used to segment the coronary artery segmentation result through a point cloud segmentation network to obtain the coronary artery segmentation result corresponding to the coronary artery segmentation result. The centerline segmentation module is used to extract the coronary artery centerline from the coronary artery segmentation results and determine the target segmentation result of the coronary artery centerline based on the coronary artery segmentation results. The root node determination module is used to determine the left root node and the right root node of the coronary artery from the target segmentation results based on the aortic segmentation results and the coronary artery segmentation results. The segmentation naming module is used to determine the name of each coronary sub-segment in the target segmentation result based on the left root node and the right root node of the coronary artery tree.

9. An electronic device, characterized in that, It includes a processor, a memory, and a program or instructions stored in the memory and executable on the processor, wherein the program or instructions, when executed by the processor, implement the steps of the method for segmenting and naming the coronary artery centerline as described in any one of claims 1-7.

10. A readable storage medium, characterized in that, The readable storage medium stores a program or instructions that, when executed by a processor, implement the steps of the method for naming coronary artery centerline segments as described in any one of claims 1-7.

Images