Matching method and device for coronary branch points, electronic equipment and readable storage medium

Abstract

The application provides a coronary branch point matching method and device, electronic equipment and a readable storage medium. A coronary skeleton image of each two-dimensional coronary angiography image is determined. Based on the starting point information of the coronary starting point in each coronary skeleton image, a plurality of initial transfer relationship matrices between the plurality of coronary skeleton images are constructed. For each coronary skeleton image, by traversing each blood vessel node on the coronary center line in the coronary skeleton image, a plurality of blood vessel branch points presented by the coronary to be reconstructed in the coronary skeleton image are determined. The second blood vessel branch point matched with the first blood vessel branch point in the first skeleton image is matched to obtain a plurality of matched groups of matching branch point pairs between the plurality of coronary skeleton images. The target transfer relationship matrix for three-dimensional reconstruction of the coronary to be reconstructed is determined. In this way, the matching of the blood vessel branch points between the plurality of coronary skeleton images can be accurately realized, and the accuracy of the matching result is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of image processing, in particular to a coronary artery branch point matching method and device, electronic equipment and readable storage medium. BACKGROUND

[0002] Cardiovascular and cerebrovascular diseases are known as the first killer threatening human life and health. With the increasing improvement of people's living standards, the material goods of life are also more and more sufficient, and the incidence of cardiovascular and cerebrovascular diseases gradually increases. With the continuous development of science and technology, medical diagnosis technology is also constantly innovating; at the same time, the vigorous development of medical imaging technology brings convenient measures and effective means for diagnosing various diseases.

[0003] By observing the coronary angiography image, the structure of the coronary artery and the blood vessel stenosis can be objectively diagnosed, and if necessary, the reconstruction of the three-dimensional coronary angiography image based on the two-dimensional coronary angiography image is needed. However, during the shooting process of the two-dimensional coronary angiography image, the shooting object's heartbeat, breathing and other reasons will cause a certain deviation in the presentation position of the coronary artery in the two-dimensional coronary angiography image at different shooting angles. At this time, if the reconstruction of the three-dimensional coronary angiography image is accurately implemented, it is necessary to match the blood vessel branch nodes in the two-dimensional coronary angiography images at different shooting angles to determine the relative positions between different two-dimensional coronary angiography images. At present, the matching points need to be selected manually through a man-machine interface. In this case, whether the matching result is accurate is limited by the operator's work experience, and if there are many blood vessel branch points to be matched, the operator cannot complete the matching of the blood vessel branch points in a short time. SUMMARY

[0004] Therefore, the purpose of the present application is to provide a coronary artery branch point matching method, device, electronic equipment and readable storage medium, which can more accurately match the blood vessel branch points with the help of the transfer relationship matrix, and improve the accuracy of the matching result.

[0005] The present application provides a coronary artery branch point matching method, which comprises:

[0006] The two-dimensional coronary angiography images of the coronary artery to be reconstructed at different collection angles are preprocessed respectively, and the coronary center line of the coronary artery to be reconstructed at the collection angle of each two-dimensional coronary angiography image is determined to obtain a plurality of coronary skeleton images;

[0007] Based on the starting point information of the coronary starting point of the coronary artery to be reconstructed in each coronary skeleton image, a plurality of initial transfer relationship matrices between the plurality of coronary skeleton images are constructed;

[0008] For each coronary skeleton image, a depth-first search method is adopted to determine a plurality of vessel branch points existing in the coronary to be reconstructed under the acquisition angle of the coronary skeleton image by traversing each vessel node on the coronary centerline in the coronary skeleton image;

[0009] A first vessel branch point is determined from a plurality of vessel branch points of a first skeleton image, and a second vessel branch point matched with the first vessel branch point in a second skeleton image is matched by establishing a polar plane; wherein the first skeleton image has the least number of vessel branch nodes among the plurality of coronary skeleton images; and the second skeleton image has the most number of vessel branch nodes among the plurality of coronary skeleton images;

[0010] Based on the node information of the first vessel branch point, the node information of the second vessel branch point, and the starting point information of the coronary starting point, the plurality of initial transition relationship matrices are updated to obtain a plurality of updated transition relationship matrices;

[0011] According to a predetermined matching order, the matching of all vessel branch points in the first skeleton image is completed one by one, a plurality of matched branch point pairs between the plurality of coronary skeleton images are determined, and the plurality of transition relationship matrices are updated in real time after the matching process of each vessel branch point is completed;

[0012] Based on the node information carried by the plurality of matched branch point pairs, the plurality of transition relationship matrices updated in real time are updated to obtain a plurality of target transition relationship matrices between the plurality of coronary skeleton images, so that the three-dimensional reconstruction of the coronary to be reconstructed is completed by using the plurality of target transition relationship matrices.

[0013] In a possible implementation, the first vessel branch point is determined from a plurality of vessel branch points of a first skeleton image, including:

[0014] A first vessel branch point is randomly determined from a plurality of vessel branch points of the first skeleton image; or,

[0015] The vessel branch point with the highest branch level of the coronary branch vessel among the plurality of vessel branch points of the first skeleton image is determined as the first vessel branch point.

[0016] In a possible implementation, the second vessel branch point matched with the first vessel branch point in the second skeleton image is matched by establishing a polar plane, including:

[0017] By means of the polar plane, a first polar line corresponding to the first vessel branch point in the second skeleton image is established to determine a plurality of candidate branch points corresponding to the first vessel branch point in the second skeleton image;

[0018] corresponding to each candidate branch point in the first skeleton image to determine a second blood vessel branch point matched with the first blood vessel branch point.

[0019] In a possible implementation, the determining, by means of the epipolar plane, a plurality of candidate branch points corresponding to the first blood vessel branch point in the second skeleton image, comprises:

[0020] The first relative distance between each blood vessel branch point in the second skeleton image and the first epipolar line is determined by means of the epipolar plane.

[0021] The first tree structure depth of the coronary artery to be reconstructed at each blood vessel branch point in the second skeleton image and the second tree structure depth of the coronary artery to be reconstructed at the first blood vessel branch point are respectively determined.

[0022] The depth difference between the first blood vessel branch point and each blood vessel branch point in the second skeleton image is respectively determined based on the determined plurality of first tree structure depths and the second tree structure depth.

[0023] The blood vessel branch point in the second skeleton image with the first relative distance less than a preset distance threshold and the depth difference less than or equal to a preset depth threshold is determined as the candidate branch point corresponding to the first blood vessel branch point.

[0024] In a possible implementation, the determining, by means of the epipolar plane, a plurality of candidate branch points corresponding to the first blood vessel branch point in the second skeleton image, comprises:

[0025] The second epipolar line corresponding to each candidate branch point in the first skeleton image is determined by means of the epipolar plane.

[0026] The second relative distance between the first blood vessel branch point and each second epipolar line is determined.

[0027] The candidate branch point corresponding to the second epipolar line with the smallest second relative distance is determined as the second blood vessel branch point matched with the first blood vessel branch point.

[0028] In a possible implementation, the matching of all blood vessel branch points in the first skeleton image is completed one by one in the predetermined matching order to determine a plurality of matched branch point pairs between the first skeleton image and the second skeleton image, comprising:

[0029] After the matching of the first blood vessel branch point is completed, a first blood vessel branch point to be matched is determined again from other blood vessel branch points in the first skeleton image according to a predetermined matching order, wherein the other blood vessel branch points are blood vessel branch points in the first skeleton image except for the first blood vessel branch node for which the matching has been completed;

[0030] A second blood vessel branch point matched with the first blood vessel branch point to be matched in the second skeleton image is matched by means of the polar plane;

[0031] The node information of each first blood vessel branch point for which the matching has been completed and the node information of the second blood vessel branch point matched with each first blood vessel branch point are used to update the plurality of transition relation matrices until the matching of all blood vessel branch points in the first skeleton image is completed, and a plurality of matched branch point pairs are obtained.

[0032] In a possible implementation, for each coronary skeleton image, a plurality of blood vessel branch points existing in the coronary to be reconstructed at the acquisition angle of the coronary skeleton image are determined by traversing each blood vessel node on the coronary centerline in the coronary skeleton image in a depth-first search manner, including:

[0033] For each coronary skeleton image, a plurality of blood vessel branch points existing in the coronary to be reconstructed at the acquisition angle of the coronary skeleton image are determined by traversing each blood vessel node on the coronary centerline in the coronary skeleton image in a depth-first search manner, taking the coronary starting point in the coronary skeleton image as a traversal starting point, establishing a blood vessel tree structure of the coronary to be reconstructed at the acquisition angle of the coronary skeleton image, and determining the branch level of each blood vessel node.

[0034] Embodiments of the present application further provide a coronary branch point matching device, the matching device comprising:

[0035] An image preprocessing module is configured to pre-process two-dimensional coronary angiography images of the coronary to be reconstructed at different acquisition angles, to obtain a plurality of coronary skeleton images by determining a coronary centerline of the coronary to be reconstructed at the acquisition angle of each two-dimensional coronary angiography image;

[0036] A matrix construction module is configured to construct a plurality of initial transition relation matrices between the plurality of coronary skeleton images based on starting point information of a coronary starting point of the coronary to be reconstructed in each coronary skeleton image;

[0037] A branch point determination module is configured to, for each coronary skeleton image, determine a plurality of blood vessel branch points existing in the coronary to be reconstructed at the acquisition angle of the coronary skeleton image by traversing each blood vessel node on the coronary centerline in the coronary skeleton image in a depth-first search manner.

[0038] a branch point matching module, configured to determine a first blood vessel branch point from a plurality of blood vessel branch points of a first skeleton image, and match a second blood vessel branch point in a second skeleton image corresponding to the first blood vessel branch point by establishing a plane of poles; wherein the first skeleton image is the one with the least number of blood vessel branch points among the plurality of coronary skeleton images; and the second skeleton image is the one with the most number of blood vessel branch points among the plurality of coronary skeleton images;

[0039] a first matrix updating module, configured to update the plurality of initial transition relation matrices based on node information of the first blood vessel branch point, node information of the second blood vessel branch point, and starting point information of a starting point of a coronary artery, to obtain a plurality of updated transition relation matrices;

[0040] a second matrix updating module, configured to complete the matching of all blood vessel branch points in the first skeleton image one by one according to a predetermined matching order, to determine a plurality of matched branch point pairs between the plurality of coronary skeleton images, and to update the plurality of transition relation matrices in real time after completing the matching process of each blood vessel branch point;

[0041] a matrix determining module, configured to update the plurality of updated transition relation matrices based on node information carried by the plurality of matched branch point pairs, to obtain a plurality of target transition relation matrices between the plurality of coronary skeleton images, and to complete the three-dimensional reconstruction of the coronary artery to be reconstructed by using the plurality of target transition relation matrices.

[0042] In a possible implementation, when the branch point matching module is configured to determine a first blood vessel branch point from a plurality of blood vessel branch points of a first skeleton image, the branch point matching module is configured to:

[0043] randomly determine a first blood vessel branch point from the plurality of blood vessel branch points of the first skeleton image; or,

[0044] determine a blood vessel branch point with the highest branch level of the coronary branch blood vessel from the plurality of blood vessel branch points of the first skeleton image as the first blood vessel branch point.

[0045] In a possible implementation, when the branch point matching module is configured to match a second blood vessel branch point in a second skeleton image corresponding to the first blood vessel branch point by establishing a plane of poles, the branch point matching module is configured to:

[0046] determine a plurality of candidate branch points in the second skeleton image corresponding to the first blood vessel branch point by establishing a first polar line corresponding to the first blood vessel branch point in the second skeleton image by means of the plane of poles;

[0047] corresponding to the first blood vessel branch point in the second skeleton image, to determine a second blood vessel branch point matched with the first blood vessel branch point.

[0048] In a possible implementation, when the branch point matching module is configured to determine a plurality of candidate branch points corresponding to the first blood vessel branch point in the second skeleton image by establishing a first epipolar line corresponding to the first blood vessel branch point in the second skeleton image by means of the epipolar plane, the branch point matching module is configured to:

[0049] establish, by means of the epipolar plane, the first epipolar line corresponding to the first blood vessel branch point in the second skeleton image, to determine a first relative distance between each blood vessel branch point in the second skeleton image and the first epipolar line, respectively;

[0050] determine a first tree structure depth of the coronary artery to be reconstructed at each blood vessel branch point in the second skeleton image and a second tree structure depth of the coronary artery to be reconstructed at the first blood vessel branch point, respectively;

[0051] determine a depth difference between the first blood vessel branch point and each blood vessel branch point in the second skeleton image based on the determined plurality of first tree structure depths and the second tree structure depth, respectively;

[0052] determine, as a candidate branch point corresponding to the first blood vessel branch point, a blood vessel branch point in the second skeleton image with a first relative distance less than a preset distance threshold and a depth difference less than or equal to a preset depth threshold.

[0053] In a possible implementation, when the branch point matching module is configured to determine a second blood vessel branch point matched with the first blood vessel branch point by establishing a second epipolar line corresponding to each candidate branch point in the first skeleton image, the branch point matching module is configured to:

[0054] establish, by means of the epipolar plane, the second epipolar line corresponding to each candidate branch point in the first skeleton image;

[0055] determine a second relative distance between the first blood vessel branch point and each second epipolar line;

[0056] determine, as the second blood vessel branch point matched with the first blood vessel branch point, a candidate branch point corresponding to a second epipolar line with a smallest second relative distance.

[0057] In a possible implementation, the second matrix updating module is configured to:

[0058] After the matching of the first blood vessel branch point is completed, a first blood vessel branch point to be matched is determined again from other blood vessel branch points in the first skeleton image according to the predetermined matching order, where the other blood vessel branch points are blood vessel branch points in the first skeleton image except the first blood vessel branch node for which the matching is completed.

[0059] The second blood vessel branch point matched with the first blood vessel branch point to be matched in the second skeleton image is matched by means of the polar plane.

[0060] The node information of each first blood vessel branch point for which the matching is completed and the node information of the second blood vessel branch point matched with each first blood vessel branch point are used to update the plurality of transition relation matrices until the matching of all blood vessel branch points in the first skeleton image is completed, and a plurality of matched branch point pairs are obtained.

[0061] In a possible implementation, the branch point determining module is configured to:

[0062] For each coronary skeleton image, the blood vessel tree structure of the coronary artery to be reconstructed under the acquisition angle of the coronary skeleton image is established by using the depth-first search manner to traverse each blood vessel node on the coronary center line in the coronary skeleton image with the coronary starting point in the coronary skeleton image as the traversal starting point, and the plurality of blood vessel branch points existing in the coronary artery to be reconstructed under the acquisition angle of the coronary skeleton image and the branch level of each blood vessel node are determined.

[0063] The application further provides an electronic device, which comprises a processor, a memory and a bus, the memory stores machine readable instructions executable by the processor, when the electronic device is running, the processor and the memory communicate through the bus, and the machine readable instructions are executed by the processor to perform the steps of the coronary branch point matching method as described above.

[0064] The embodiment of the present application further provides a computer readable storage medium, and the computer readable storage medium stores a computer program, and the computer program performs the steps of the matching method of the coronary artery branch point when the computer program is run by a processor.

[0065] The matching method, device, electronic equipment and readable storage medium of the coronary artery branch point provided by the embodiment of the present application are used for respectively pre-processing a plurality of two-dimensional coronary angiography images, determining a coronary skeleton image of each two-dimensional coronary angiography image, constructing a plurality of initial transition relationship matrices between the plurality of coronary skeleton images based on the starting point information of the coronary starting point in each coronary skeleton image, determining a plurality of vessel branch points of the coronary to be reconstructed in the coronary skeleton image by traversing each vessel node on the coronary center line in the coronary skeleton image for each coronary skeleton image by using a depth-first search method, matching the second vessel branch point matched with the first vessel branch point in the first skeleton image by establishing a polar plane, updating the plurality of initial transition relationship matrices in the matching process, completing the matching of each vessel branch point in the first skeleton image in a predetermined matching order, obtaining a plurality of matched branch point pairs between the plurality of coronary skeleton images, and updating the plurality of transition relationship matrices based on the node information carried by the plurality of matched branch point pairs, and obtaining a target transition relationship matrix used for three-dimensional reconstruction of the coronary to be reconstructed. In this way, the matching of the vessel branch points between the plurality of coronary skeleton images can be accurately realized, the accuracy of the matching result is improved, the matching of the vessel branch points is quickly realized without human intervention, and the matching efficiency of the vessel branch points is improved.

[0066] In order to make the above objectives, characteristics and advantages of the present application more apparent, the following will describe a preferred embodiment in detail, and the accompanying drawings will be described as follows. BRIEF DESCRIPTION OF DRAWINGS

[0067] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the embodiments. It should be understood that the following drawings only show some embodiments of the present application, and therefore should not be regarded as a limitation to the scope. For those skilled in the art, other related drawings can also be obtained without creative labor on the basis of these drawings.

[0068] Figure 1 A flowchart of a matching method of a coronary artery branch point provided by the embodiment of the present application;

[0069] Figure 2 A schematic diagram of a branch point matching process provided by the embodiment of the present application;

[0070] Figure 3A structure schematic diagram of a coronary artery branch point matching device provided by an embodiment of the present application;

[0071] Figure 4 A structure schematic diagram of an electronic device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0072] To make the objectives, technical solutions and advantages of the embodiments of the present application clearer, the following will be combined with the accompanying drawings for the embodiments of the present application to make a clear and complete description of the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. The components of the embodiments of the present application described and shown in the accompanying drawings can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the accompanying drawings is not intended to limit the scope of the claimed present application, but only represents selected embodiments of the present application. Based on the embodiments of the present application, every other embodiment obtained by a person skilled in the art without creative work belongs to the scope of protection of the present application.

[0073] It is found through research that the structure of the coronary artery and the blood vessel stenosis can be objectively diagnosed by observing the coronary angiography image, and if necessary, the reconstruction of the three-dimensional coronary angiography image based on the two-dimensional coronary angiography image is needed. However, in the shooting process of the two-dimensional coronary angiography image, the shooting object's heartbeat, breathing and other reasons will cause a certain deviation in the presentation position of the coronary artery in the two-dimensional coronary angiography images at different shooting angles. At this time, if the reconstruction of the three-dimensional coronary angiography image is accurately implemented, the blood vessel branch nodes in the two-dimensional coronary angiography images at different shooting angles need to be matched to determine the relative positions between different two-dimensional coronary angiography images. At present, the matching points need to be selected manually through the man-machine interface. In this case, whether the matching result is accurate is limited by the operator's work experience. If there are many blood vessel branch points to be matched, the operator cannot complete the matching of the blood vessel branch points in a short time.

[0074] Based on this, the embodiments of the present application provide a coronary artery branch point matching method, which can realize the accurate matching of the blood vessel branch points in the coronary artery without manual matching, and improve the matching efficiency of the blood vessel branch points.

[0075] Please refer to Figure 1 , Figure 1 A flowchart of a coronary artery branch point matching method provided by an embodiment of the present application. As shown in Figure 1 , the coronary artery branch point matching method provided by the embodiments of the present application comprises the following steps.

[0076] S101, respectively, the two-dimensional coronary angiography image of the to-be-reconstructed coronary artery under different collection angles is preprocessed, the coronary center line of the to-be-reconstructed coronary artery under the collection angle of each two-dimensional coronary angiography image is determined, and a plurality of coronary skeleton images are obtained.

[0077] S102, based on the starting point information of the coronary starting point of the to-be-reconstructed coronary artery in each coronary skeleton image, a plurality of initial transfer relationship matrices between the plurality of coronary skeleton images are constructed.

[0078] S103, for each coronary skeleton image, a depth-first search method is adopted, each blood vessel node on the coronary center line in the coronary skeleton image is traversed, and a plurality of blood vessel branch points existing in the to-be-reconstructed coronary artery under the collection angle of the coronary skeleton image are determined.

[0079] S104, a first blood vessel branch point is determined from the plurality of blood vessel branch points of the first skeleton image, and a second blood vessel branch point matched with the first blood vessel branch point in the second skeleton image is matched by establishing a polar plane.

[0080] S105, based on the node information of the first blood vessel branch point, the node information of the second blood vessel branch point, and the starting point information of the coronary starting point, the plurality of initial transfer relationship matrices are updated to obtain a plurality of updated transfer relationship matrices.

[0081] S106, according to a predetermined matching order, the matching of all blood vessel branch points in the first skeleton image is completed one by one, a plurality of groups of matched branch point pairs between the plurality of coronary skeleton images are determined, and the plurality of updated transfer relationship matrices are updated in real time after each matching process is completed.

[0082] S107, based on the node information carried by the plurality of groups of matched branch point pairs, the plurality of real-time updated transfer relationship matrices are updated again to obtain a plurality of target transfer relationship matrices between the plurality of coronary skeleton images, so that the three-dimensional reconstruction of the to-be-reconstructed coronary artery is completed by using the plurality of target transfer relationship matrices.

[0083] The method for matching coronary branch points provided by the embodiments of the present application comprises the following steps: the coronary skeleton images of each two-dimensional coronary angiography image are determined by respectively pre-processing a plurality of two-dimensional coronary angiography images; a plurality of initial transition relationship matrices between the coronary skeleton images are constructed based on the starting point information of the coronary starting points in each coronary skeleton image; for each coronary skeleton image, the plurality of blood vessel branch points of the coronary to be reconstructed presented in the coronary skeleton image are determined by traversing each blood vessel node on the coronary centerline in the coronary skeleton image using a depth-first search method; the second blood vessel branch point matched with the first blood vessel branch point in the first skeleton image is matched by establishing a polar plane; the updating of the plurality of initial transition relationship matrices is realized in the matching process; the matching of each blood vessel branch point in the first skeleton image is completed according to a predetermined matching order, and a plurality of matched groups of branch point pairs between the coronary skeleton images are obtained; finally, the plurality of transition relationship matrices are updated based on the node information carried by the plurality of matched groups of branch point pairs, and a target transition relationship matrix used for the three-dimensional reconstruction of the coronary to be reconstructed is obtained. In this way, the matching of the blood vessel branch points between the coronary skeleton images can be accurately realized, the accuracy of the matching result can be improved, the matching of the blood vessel branch points can be quickly realized without human intervention, and the matching efficiency of the blood vessel branch points can be improved.

[0084] In step S101, before the matching of the blood vessel branch points is performed, the blood vessel branch points presented by the coronary to be reconstructed in each two-dimensional coronary angiography image need to be determined. At this time, the two-dimensional coronary angiography images need to be skeletonized by pre-processing the two-dimensional coronary angiography images, so as to determine the plurality of blood vessel branch points included in each two-dimensional coronary angiography image.

[0085] Specifically, the two-dimensional coronary angiography images of the coronary to be reconstructed under different acquisition angles are pre-processed respectively, and the coronary centerline of the coronary to be reconstructed under the acquisition angle of each two-dimensional coronary angiography image is determined, so as to obtain the coronary skeleton image corresponding to each two-dimensional coronary angiography image.

[0086] In an embodiment, step S101 comprises:

[0087] S1011, for each two-dimensional coronary angiography image, the two-dimensional coronary angiography image is converted into a binary coronary image.

[0088] In this step, for each two-dimensional coronary angiography image, the two-dimensional coronary angiography image is converted into a binary coronary image by performing binary segmentation processing on the two-dimensional coronary angiography image; in this way, the coronary blood vessels are distinguished from the background image as the foreground image.

[0089] Specifically, pixels in the acquired two-dimensional coronary angiography image are marked as 1 if they are located at positions of coronary vessels, and marked as 0 if they are located at other positions. In this way, binary segmentation processing of the two-dimensional coronary angiography image is completed, and a binary coronary image corresponding to the two-dimensional coronary angiography image is obtained.

[0090] Here, the binary segmentation processing of the two-dimensional coronary angiography image can be implemented in a manner of deep learning semantic segmentation. For example, a pre-trained binary segmentation model can be used to implement the binary segmentation processing of the two-dimensional coronary angiography image.

[0091] The binary segmentation model can be obtained by training a pre-constructed U-Net segmentation model using a plurality of sample two-dimensional angiography images and label tags of each sample two-dimensional angiography image. In the label tags, coronary vessels and backgrounds in the sample two-dimensional angiography images are distinguished.

[0092] Here, since multiple vessel connected domains can be determined in the binary coronary image obtained after the binary processing, only the largest vessel connected domain is retained as a target connected domain for subsequent skeletonization processing. In this way, over-attention to small vessels that are not necessary to focus on can be reduced, the data processing efficiency is improved, and the accuracy of the matching result is ensured.

[0093] S1012, performing skeletonization processing on the target connected domain in the binary coronary image to determine a coronary centerline of the coronary artery to be reconstructed under the acquisition angle of the two-dimensional coronary angiography image, and obtaining a coronary skeleton image corresponding to the two-dimensional coronary angiography image.

[0094] In this step, based on the binary coronary image obtained after the binary segmentation processing, skeletonization processing is performed on the target connected domain in the binary coronary image to determine a coronary centerline of the coronary artery to be reconstructed under the acquisition angle of the two-dimensional coronary angiography image. In this way, the conversion of the two-dimensional coronary angiography image skeletonization is completed, and a coronary skeleton image corresponding to the two-dimensional coronary angiography image is obtained.

[0095] Here, in order to accurately implement the matching of the vessel branch points between different coronary skeleton images, a transfer relationship matrix representing the transfer relationship between a plurality of coronary skeleton images is used to implement the matching of the vessel branch points between the plurality of coronary skeleton images during the matching of the vessel branch points.

[0096] In step S102, first, the starting point information of the starting point (i.e., the coronary sinus of the coronary vessel) of the coronary artery to be reconstructed in each coronary skeleton image is determined; then, based on the starting point information of the plurality of coronary starting points, a plurality of initial transfer relationship matrices between the plurality of coronary skeleton images are constructed.

[0097] Here, at least two two-dimensional coronary angiography images (two-dimensional coronary angiography image DSA-0 and two-dimensional coronary angiography image DSA-1) are included, and therefore, at least two coronary skeleton images are included, and at least two transfer relationship matrices are included between the at least two coronary skeleton images, one of which is used to represent the transfer relationship from DSA-0 to DSA-1, and the other is used to represent the transfer relationship from DSA-1 to DSA-0.

[0098] In an implementation, the starting point information of the coronary starting point of the coronary to be reconstructed in each coronary skeleton image is determined by the following steps:

[0099] In step S201, for each coronary skeleton image, a pre-trained starting point prediction model is used to determine the starting point information of the coronary starting point of the coronary to be reconstructed at the acquisition angle of the coronary skeleton image.

[0100] In this step, for each coronary skeleton image, a pre-trained starting point prediction model is used to predict the location of the coronary starting point in the coronary skeleton image, determine the coronary starting point of the coronary to be reconstructed at the acquisition angle of the coronary skeleton image, and determine the starting point information of the coronary starting point according to the location of the coronary starting point.

[0101] The starting point information includes the position coordinates of the location of the coronary starting point, the vessel radius of the location of the coronary starting point, and the direction of the location of the coronary starting point.

[0102] The starting point prediction model can be obtained by training a pre-constructed U-Net segmentation model using a plurality of sample two-dimensional angiography images and sample labels of each sample two-dimensional angiography image, wherein the sample labels have labeled the location of the coronary starting point in the sample two-dimensional angiography image.

[0103] In an implementation, step S201 includes:

[0104] S2011, for each coronary skeleton image, a pre-trained starting point prediction model is used to determine an initial segmentation region located around the coronary starting point of the coronary to be reconstructed in the coronary skeleton image.

[0105] In this step, for each coronary skeleton image, a pre-trained starting point prediction model is used to predict the approximate location of the coronary starting point, and determine an initial segmentation region located around the coronary starting point of the coronary to be reconstructed in the coronary skeleton image.

[0106] S2012, dilate the initial segmentation region in the coronary skeleton image to determine the intersection between the initial segmentation region and the coronary centerline displayed in the coronary skeleton image.

[0107] In this step, the initial segmentation region is dilated in the coronary skeleton image until an intersection point between the initial segmentation region and the coronary centerline shown in the coronary skeleton image is found, and the dilation of the initial segmentation region is ended.

[0108] S2013, determining a third relative distance between each end point of the coronary centerline and the intersection point.

[0109] In this step, considering that the coronary starting point should be an end point of the coronary centerline and the predicted initial segmentation region is located around the coronary starting point, the end point closest to the intersection point between the initial segmentation region and the coronary centerline can be considered as the coronary starting point of the coronary to be reconstructed; at this time, the third relative distance between each end point of the coronary centerline and the intersection point needs to be determined respectively.

[0110] S2014, determining the end point with the smallest third relative distance as the coronary starting point of the coronary to be reconstructed at the acquisition angle of the coronary skeleton image.

[0111] In this step, the end point with the smallest third relative distance to the intersection point is determined as the coronary starting point of the coronary to be reconstructed at the acquisition angle of the coronary skeleton image.

[0112] In step S103, for each coronary skeleton image, a depth-first search method is used to traverse each blood vessel node on the coronary centerline shown in the coronary skeleton image, complete the search at the position of each blood vessel node, determine whether each blood vessel node is a blood vessel branch point, and further determine the multiple blood vessel branch points in the coronary to be reconstructed at the acquisition angle of the coronary skeleton image.

[0113] Here, the depth-first search (DFS) is a traversal method of generating tree structure, in which a tree structure of the coronary blood vessel is generated in the process of traversing each blood vessel node, so that whether each blood vessel node is a blood vessel branch point can be determined.

[0114] In an embodiment, step S103 comprises:

[0115] S1031, for each coronary skeleton image, a depth-first search method is used to traverse each blood vessel node on the coronary centerline in the coronary skeleton image, taking the coronary starting point (i.e. the coronary sinus of the coronary blood vessel) in the coronary skeleton image as the traversal starting point, establish the blood vessel tree structure of the coronary to be reconstructed at the acquisition angle of the coronary skeleton image, determine the multiple blood vessel branch points in the coronary to be reconstructed at the acquisition angle of the coronary skeleton image and the branch level of each blood vessel node.

[0116] In this step, for each coronary skeleton image, a traversal starting point is set as the coronary starting point of the coronary to be reconstructed, and a depth-first search is performed from the traversal starting point to complete the traversal of each blood vessel node on the coronary centerline in the coronary skeleton image, to determine the blood vessel branch points from the blood vessel nodes on the coronary centerline and to determine the branch level of each blood vessel node while establishing the blood vessel tree structure of the blood vessel tree structure of the coronary to be reconstructed at the angle of the coronary skeleton image.

[0117] The branch level of the blood vessel node represents the branch level of the blood vessel branch to which the blood vessel node belongs. For example, for a coronary to be reconstructed, the main blood vessel and the branch blood vessel can be distinguished according to the thickness of the blood vessel during the generation of the blood vessel tree structure of the coronary to be reconstructed by using the DFS method. In general, the starting point of the main blood vessel is the location of the coronary starting point (coronary sinus) of the coronary to be reconstructed. If a blood vessel node is located on the main blood vessel, the branch level of the blood vessel node is 0. If the branch blood vessel connected to the main blood vessel is a first-level branch blood vessel, the branch level of the blood vessel node located on the first-level branch blood vessel is 1.

[0118] Therefore, the branch level of each blood vessel node can be determined by analyzing the branch level of the branch blood vessel on which the blood vessel node is located during the traversal and generation of the blood vessel tree structure.

[0119] In an embodiment, step S1031 includes:

[0120] S10311, selecting a current blood vessel node to be searched from the blood vessel nodes on the coronary centerline in the coronary skeleton image during the traversal process, with the coronary starting point as the traversal starting point.

[0121] In this step, each blood vessel node on the coronary centerline in the coronary skeleton image is traversed one by one, with the coronary starting point as the traversal starting point, and a current blood vessel node to be searched is selected from the blood vessel nodes on the coronary centerline in the coronary skeleton image during the traversal process.

[0122] S10312, using a depth-first search method to determine the blood vessel radius at the current blood vessel node and whether the current blood vessel node is a blood vessel branch point by using a plurality of inflation shapes to inflate at the current blood vessel node, and establishing the blood vessel tree structure at the current blood vessel node.

[0123] In this step, a deep-first search method is used to expand the shape by expanding at the current blood vessel node, and when the expanded shape intersects with the blood vessel wall of the coronary blood vessel, the expansion of the expanded shape is stopped, the area of the expanded region is determined, the average of the area of the expanded shape is determined, and the radius of the virtual circle with the same area as the average is determined as the blood vessel radius at the current blood vessel node. At the same time, it is determined whether the current blood vessel node is a blood vessel branch point during the traversal process.

[0124] During the traversal process of the current blood vessel node, the blood vessel tree structure of the coronary artery to be reconstructed at the current blood vessel node is established in real time.

[0125] S10313, based on the positional relationship between the traversal initial point and the current blood vessel node, a direction vector of the current blood vessel node is determined.

[0126] In this step, the direction vector at the current blood vessel node position is determined based on the positional relationship between the current blood vessel node and the traversal initial point, wherein the positional relationship includes the relative direction and relative distance between the current blood vessel node and the traversal initial point.

[0127] S10314, based on the blood vessel radius of the current blood vessel node, the direction vector of the current blood vessel node, and the blood vessel length of the branch blood vessel to which the current blood vessel node belongs, the branch level of the current blood vessel node is determined.

[0128] In this step, the blood vessel radius of the current blood vessel node, the direction vector of the current blood vessel node, and the blood vessel length of the branch blood vessel to which the current blood vessel node belongs are comprehensively considered to sort the current blood vessel node and determine the branch level of the current blood vessel node.

[0129] In step S104, the plurality of coronary skeleton images includes at least two coronary skeleton images, a first skeleton image and a second skeleton image; from the plurality of blood vessel branch points of the first skeleton image, a first blood vessel branch point to be matched is determined;

[0130] Using the initial transfer relationship matrix between the first skeleton image and the second skeleton image, a second blood vessel branch point matched with the first blood vessel branch point in the second skeleton image is matched.

[0131] In order to ensure that each blood vessel branch point in the first skeleton image can find a matched second blood vessel branch point in the second skeleton image, an image with the least blood vessel branch nodes in the plurality of coronary skeleton images is taken as the first skeleton image, that is, the first skeleton image has the least blood vessel branch nodes in the plurality of skeleton images; correspondingly, an image with the most blood vessel branch nodes in the plurality of coronary skeleton images is taken as the second skeleton image, that is, the second skeleton image has the most blood vessel branch nodes in the plurality of skeleton images.

[0132] In an embodiment, the first blood vessel branch point is determined from the plurality of blood vessel branch points of the first skeleton image, comprising:

[0133] The first blood vessel branch point is randomly determined from the plurality of blood vessel branch points of the first skeleton image.

[0134] In this step, a blood vessel node is randomly selected from the plurality of blood vessel branch points included in the first skeleton image as the first blood vessel branch point.

[0135] Or, the blood vessel branch point with the highest branch level of the coronary branch vessel in the plurality of blood vessel branch points of the first skeleton image is determined as the first blood vessel branch point.

[0136] In this step, the first blood vessel branch point that should be preferentially matched can be determined according to the branch level of each blood vessel branch point; specifically, the blood vessel branch point with the highest branch level of the coronary branch vessel in the plurality of blood vessel branch points of the first skeleton image is determined as the first blood vessel branch point.

[0137] In an embodiment, please refer to Figure 2 , Figure 2 A branch point matching process provided by the embodiment of the application. As shown in Figure 2 , a second blood vessel branch point matched with the first blood vessel branch point in the second skeleton image is matched by establishing an epipolar plane, comprising:

[0138] Step 1, by means of the epipolar plane, a first epipolar line corresponding to the first blood vessel branch point in the second skeleton image is established to determine a plurality of candidate branch points corresponding to the first blood vessel branch point in the second skeleton image.

[0139] In this step, the first epipolar line corresponding to the first blood vessel branch point in the second skeleton image is established by using the initial transition relationship matrix between the first skeleton image and the second skeleton image, and the blood vessel branch points in the second skeleton image that meet the preset requirements are determined as the plurality of candidate branch points corresponding to the first blood vessel branch point in the second skeleton image.

[0140] In an embodiment, step 1 comprises:

[0141] S11, establishing a first polar line corresponding to the first blood vessel branch point in the second skeleton image by means of the polar plane, and determining a first relative distance between each blood vessel branch point in the second skeleton image and the first polar line, respectively.

[0142] S12, determining a first tree structure depth of the coronary artery to be reconstructed at each blood vessel branch point in the second skeleton image, and a second tree structure depth of the coronary artery to be reconstructed at the first blood vessel branch point, respectively.

[0143] In this step, in order to accurately determine the second blood vessel branch point matched with the first blood vessel branch point, the tree structure depth at the position of each blood vessel branch point needs to be referred to when matching the blood vessel branch points; at this time, the first tree structure depth of the coronary artery to be reconstructed at each blood vessel branch point in the second skeleton image and the second tree structure depth of the coronary artery to be reconstructed at the first blood vessel branch point need to be determined respectively.

[0144] S13, determining a depth difference value between the first blood vessel branch point and each blood vessel branch point in the second skeleton image based on the determined multiple first tree structure depths and the second tree structure depth.

[0145] S14, determining the blood vessel branch point in the second skeleton image with a first relative distance less than a preset distance threshold and a depth difference value less than or equal to a preset depth threshold as a candidate branch point corresponding to the first blood vessel branch point.

[0146] Step 2, determining the second blood vessel branch point matched with the first blood vessel branch point by establishing a second polar line corresponding to each candidate branch point in the first skeleton image.

[0147] In this step, the second blood vessel branch point matched with the first blood vessel branch point is determined from the multiple candidate branch points by establishing a second polar line corresponding to each candidate branch point in the first skeleton image.

[0148] In an embodiment, step 2 comprises:

[0149] Step 21, establishing a second polar line corresponding to each candidate branch point in the first skeleton image by means of the polar plane.

[0150] Step 22, determining a second relative distance between the first blood vessel branch point and each second polar line.

[0151] In this step, the second relative distance between the first blood vessel branch point and each second polar line in the first skeleton image is determined respectively.

[0152] Step 23, determining the candidate branch point corresponding to the second polar line with the second smallest relative distance as the second vessel branch point matched with the first vessel branch point.

[0153] In this step, the candidate branch point corresponding to the second polar line with the second smallest relative distance between the first vessel branch point is determined as the second vessel branch point matched with the first vessel branch point in the second skeleton image.

[0154] Here, the first vessel branch point and the second vessel branch point are the same vessel branch point in the coronary artery to be reconstructed, and only differ in the acquisition angle.

[0155] In step S105, after completing the matching of the first vessel branch point in the first skeleton image, in order to gradually improve the accuracy of the matching results of each subsequent vessel node, the initial transition relationship matrix needs to be updated; specifically, the node information of the first vessel branch point, the node information of the second vessel branch point, and the starting point information of the starting point of the coronary artery are used to update the established plurality of initial transition relationship matrices to obtain a plurality of updated transition relationship matrices.

[0156] Among them, the node information of the first vessel branch point includes the position coordinates of the position of the first vessel branch point, the vessel radius of the position of the first vessel branch point, and the direction of the position of the first vessel branch point, etc. The node information of the second vessel branch point includes the position coordinates of the position of the second vessel branch point, the vessel radius of the position of the second vessel branch point, and the direction of the position of the second vessel branch point, etc.

[0157] In step S106, according to the matching method of the first vessel branch point described above, the matching of each vessel branch point in the first skeleton image is completed one by one according to the predetermined matching order of each vessel branch node in the first skeleton image; that is, the matched vessel branch point of each vessel branch point in the first skeleton image in the second skeleton image is determined; and further, a plurality of matched branch point pairs between a plurality of coronary skeleton images can be determined; Specifically, for each vessel branch point in the first skeleton image, the vessel branch point and the vessel branch point matched with the vessel branch point in the second skeleton image form a matched branch point pair, and in this way, a plurality of matched branch point pairs can be obtained.

[0158] Here, after completing the matching process of each vessel branch point, the updated plurality of transition relationship matrices obtained above are updated in real time.

[0159] In one embodiment, the matching of all vessel branch points in the first skeleton image one by one according to the predetermined matching order, and determining a plurality of matched branch point pairs between the first skeleton image and the second skeleton image, comprises:

[0160] Step a, after the matching of the first blood vessel branch point is completed, a first blood vessel branch point to be matched is determined again from other blood vessel branch points in the first skeleton image according to a predetermined matching order.

[0161] In this step, after the matching of the first blood vessel branch point of the first skeleton image is completed, a first blood vessel branch point to be matched is determined again from other blood vessel branch points of the first skeleton image which have not been matched according to a predetermined matching order.

[0162] The other blood vessel branch points are blood vessel branch points in the first skeleton image except the first blood vessel branch node which has completed matching.

[0163] Step b, the second blood vessel branch point matched with the first blood vessel branch point to be matched in the second skeleton image is matched by means of the polar plane.

[0164] In this step, the second blood vessel branch point matched with the first blood vessel branch point to be matched determined in step a is matched again from the second skeleton image by means of the polar plane.

[0165] Step c, the node information of each first blood vessel branch point which has completed matching and the node information of the second blood vessel branch point matched with each first blood vessel branch point are used to update the plurality of transition relationship matrices until the matching of all blood vessel branch points in the first skeleton image is completed, and a plurality of matched branch point pairs are obtained.

[0166] In this step, the node information of each first blood vessel branch point which has completed matching and the node information of the second blood vessel branch point matched with each first blood vessel branch point are used to update the plurality of transition relationship matrices, and the plurality of transition relationship matrices obtained after the update are used again to perform the matching of blood vessel branch points in the next round, and so on, until the matching of all blood vessel branch points in the first skeleton image is completed, and a plurality of matched branch point pairs are obtained.

[0167] In step S107, based on the node information carried by the plurality of matched branch point pairs, the plurality of transition relationship matrices obtained in real time are updated again, and thus a plurality of target transition relationship matrices between the plurality of coronary skeleton images are obtained; further, the three-dimensional reconstruction of the coronary artery to be reconstructed can be completed by using the plurality of target transition relationship matrices.

[0168] Here, the node information carried by each matched branch point pair includes the node information of the first blood vessel branch point in the matched branch point pair and the node information of the second blood vessel branch point matched with the first blood vessel branch point.

[0169] The method for matching coronary artery branch points provided by the embodiments of the present application comprises the following steps: the two-dimensional coronary angiography images are preprocessed respectively to determine the coronary skeleton images of the two-dimensional coronary angiography images; the initial transition relationship matrices between the coronary skeleton images are constructed based on the starting point information of the coronary starting points in each coronary skeleton image; for each coronary skeleton image, the deep-first-search method is adopted to determine the multiple vessel branch points of the coronary artery to be reconstructed by traversing each vessel node on the coronary center line in the coronary skeleton image; the second vessel branch point matched with the first vessel branch point in the first skeleton image is matched by establishing the polar plane; the multiple initial transition relationship matrices are updated in the matching process; the matching of each vessel branch point in the first skeleton image is completed according to the predetermined matching order, and the matched multiple sets of matching branch point pairs between the multiple coronary skeleton images are obtained; finally, the multiple transition relationship matrices are updated based on the node information carried by the multiple sets of matching branch point pairs, and the target transition relationship matrix used for the three-dimensional reconstruction of the coronary artery to be reconstructed is obtained. In this way, the matching of the vessel branch points between the multiple coronary skeleton images can be accurately realized, the accuracy of the matching result is improved, the matching of the vessel branch points can be quickly realized without human intervention, and the matching efficiency of the vessel branch points is improved.

[0170] Please refer to Figure 3 , Figure 3 The structure diagram of the matching device for coronary artery branch points provided by the embodiments of the present application is shown in FIG. 3. Figure 3 As shown in FIG. 3, the matching device 300 comprises:

[0171] The image preprocessing module 310 is configured to pre-process the two-dimensional coronary angiography images of the coronary artery to be reconstructed under different acquisition angles, and obtain the multiple coronary skeleton images by determining the coronary center line of the coronary artery to be reconstructed under the acquisition angle of each two-dimensional coronary angiography image.

[0172] The matrix construction module 320 is configured to construct the multiple initial transition relationship matrices between the multiple coronary skeleton images based on the starting point information of the coronary starting points in each coronary skeleton image.

[0173] The branch point determination module 330 is configured to, for each coronary skeleton image, adopt the deep-first-search method to determine the multiple vessel branch points in the coronary artery to be reconstructed by traversing each vessel node on the coronary center line in the coronary skeleton image.

[0174] The branch point matching module 340 is configured to determine a first blood vessel branch point from a plurality of blood vessel branch points of a first skeleton image, and match a second blood vessel branch point in a second skeleton image that matches the first blood vessel branch point by establishing a polar plane; wherein the first skeleton image is the one with the least number of blood vessel branch points among the plurality of coronary skeleton images; and the second skeleton image is the one with the most number of blood vessel branch points among the plurality of coronary skeleton images.

[0175] The first matrix updating module 350 is configured to update the plurality of initial transition relationship matrices based on node information of the first blood vessel branch point, node information of the second blood vessel branch point, and starting point information of a coronary starting point, to obtain updated plurality of transition relationship matrices.

[0176] The second matrix updating module 360 is configured to complete the matching of all blood vessel branch points in the first skeleton image one by one according to a predetermined matching order, determine a plurality of matched branch point pairs between the plurality of coronary skeleton images, and update the plurality of transition relationship matrices in real time after completing the matching process of each blood vessel branch point.

[0177] The matrix determining module 370 is configured to update the plurality of transition relationship matrices updated in real time based on node information carried by the plurality of matched branch point pairs, to obtain a plurality of target transition relationship matrices between the plurality of coronary skeleton images, so as to complete the three-dimensional reconstruction of the coronary to be reconstructed by using the plurality of target transition relationship matrices.

[0178] Further, when the branch point matching module 340 is configured to determine a first blood vessel branch point from a plurality of blood vessel branch points of a first skeleton image, the branch point matching module 340 is configured to:

[0179] randomly determine a first blood vessel branch point from the plurality of blood vessel branch points of the first skeleton image; or,

[0180] determine a blood vessel branch point with the highest branch level of the coronary branch blood vessel from the plurality of blood vessel branch points of the first skeleton image as the first blood vessel branch point.

[0181] Further, when the branch point matching module 340 is configured to match a second blood vessel branch point in a second skeleton image that matches the first blood vessel branch point by establishing a polar plane, the branch point matching module 340 is configured to:

[0182] determine a plurality of candidate branch points in the second skeleton image that correspond to the first blood vessel branch point by establishing a first polar line corresponding to the first blood vessel branch point in the second skeleton image by means of the polar plane;

[0183] determine a second blood vessel branch point matched with the first blood vessel branch point by establishing a second polar line corresponding to each candidate branch point in the first skeleton image.

[0184] Further, the branch point matching module 340 is configured to, when determining the candidate branch points corresponding to the first blood vessel branch point in the second skeleton image by establishing the first polar line corresponding to the first blood vessel branch point in the second skeleton image, determine the first relative distance between each blood vessel branch point in the second skeleton image and the first polar line by means of the polar plane.

[0185] establish the first polar line corresponding to the first blood vessel branch point in the second skeleton image by means of the polar plane, and determine the first relative distance between each blood vessel branch point in the second skeleton image and the first polar line, respectively.

[0186] determine the first tree structure depth of the blood vessel at each blood vessel branch point in the second skeleton image and the second tree structure depth of the blood vessel at the first blood vessel branch point, respectively;

[0187] determine the depth difference between the first blood vessel branch point and each blood vessel branch point in the second skeleton image based on the determined first tree structure depths and the second tree structure depth, respectively;

[0188] determine the blood vessel branch point in the second skeleton image as the candidate branch point corresponding to the first blood vessel branch point when the first relative distance is less than a preset distance threshold and the depth difference is less than or equal to a preset depth threshold.

[0189] Further, the branch point matching module 340 is configured to, when determining the second blood vessel branch point matched with the first blood vessel branch point by establishing the second polar line corresponding to each candidate branch point in the first skeleton image, determine the second relative distance between the first blood vessel branch point and each second polar line.

[0190] establish the second polar line corresponding to each candidate branch point in the first skeleton image by means of the polar plane.

[0191] determine the second relative distance between the first blood vessel branch point and each second polar line.

[0192] determine the candidate branch point corresponding to the second polar line with the smallest second relative distance as the second blood vessel branch point matched with the first blood vessel branch point.

[0193] Further, the second matrix updating module 360 is configured to, when determining the matched groups of matching branch points between the first skeleton image and the second skeleton image by sequentially completing the matching of all blood vessel branch points in the first skeleton image according to a predetermined matching order, determine the second matrix updating module 360 is configured to:

[0194] After the matching of the first blood vessel branch point is completed, a first blood vessel branch point to be matched is determined again from other blood vessel branch points in the first skeleton image according to a predetermined matching order, wherein the other blood vessel branch points are blood vessel branch points in the first skeleton image except for the first blood vessel branch node for which the matching has been completed;

[0195] A second blood vessel branch point matched with the first blood vessel branch point to be matched in the second skeleton image is matched by means of the polar plane;

[0196] The plurality of transition relation matrices are updated by using the node information of each first blood vessel branch point for which the matching has been completed and the node information of the second blood vessel branch point matched with each first blood vessel branch point, until the matching of all blood vessel branch points in the first skeleton image is completed, so as to obtain a plurality of matched branch point pairs.

[0197] Further, the branch point determination module 330 is configured to, for each coronary skeleton image, determine the plurality of blood vessel branch points existing in the coronary artery to be reconstructed under the acquisition angle of the coronary skeleton image by traversing each blood vessel node on the coronary centerline in the coronary skeleton image in a depth-first search manner.

[0198] For each coronary skeleton image, the plurality of blood vessel branch points existing in the coronary artery to be reconstructed under the acquisition angle of the coronary skeleton image are determined by traversing each blood vessel node on the coronary centerline in the coronary skeleton image in a depth-first search manner with the coronary starting point in the coronary skeleton image as a traversal starting point, and a blood vessel tree structure of the coronary artery to be reconstructed under the acquisition angle of the coronary skeleton image is established, and the branch level of each blood vessel node is determined.

[0199] The matching device for coronary branch points provided in the embodiment of the present application determines the coronary skeleton image of each two-dimensional coronary angiography image by respectively pre-processing a plurality of two-dimensional coronary angiography images; and constructs a plurality of initial transition relationship matrices among the plurality of coronary skeleton images based on the starting point information of the coronary starting points in each coronary skeleton image; for each coronary skeleton image, the plurality of blood vessel branch points of the coronary to be reconstructed presented in the coronary skeleton image are determined by traversing each blood vessel node on the coronary centerline in the coronary skeleton image using the depth-first search method; the second blood vessel branch point matched with the first blood vessel branch point in the first skeleton image is matched by establishing a polar plane; and the plurality of initial transition relationship matrices are updated in the matching process; the matching of each blood vessel branch point in the first skeleton image is completed according to the predetermined matching order, and a plurality of matched groups of branch point pairs among the plurality of coronary skeleton images are obtained; finally, the plurality of transition relationship matrices are updated based on the node information carried by the plurality of matched groups of branch point pairs, and a target transition relationship matrix used for three-dimensional reconstruction of the coronary to be reconstructed is obtained. In this way, the matching of the blood vessel branch points among the plurality of coronary skeleton images can be accurately realized, the accuracy of the matching result is improved, the matching of the blood vessel branch points can be quickly realized without human intervention, and the matching efficiency of the blood vessel branch points can be improved.

[0200] Please refer to Figure 4 , Figure 4 A structure schematic diagram of an electronic device provided by the embodiment of the present application is shown in FIG. 4. Figure 4 As shown in FIG. 4, the electronic device 400 includes a processor 410, a memory 420 and a bus 430.

[0201] The memory 420 stores machine readable instructions executable by the processor 410, when the electronic device 400 is running, the processor 410 and the memory 420 communicate through the bus 430, and the machine readable instructions executed by the processor 410 can execute the steps of the matching method for coronary branch points in the method embodiment shown in the above Figure 1 The specific implementation can be referred to the method embodiment, which will not be described here.

[0202] The embodiment of the present application further provides a computer readable storage medium, the computer readable storage medium stores a computer program, and the computer program can execute the steps of the matching method for coronary branch points in the method embodiment shown in the above Figure 1 The specific implementation can be referred to the method embodiment, which will not be described here.

[0203] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the system, device and unit described above can refer to the corresponding process in the foregoing method embodiments, which will not be repeated here.

[0204] In several embodiments provided in the present application, it should be understood that the disclosed system, device and method can be implemented in other ways. The device embodiments described above are only schematic, for example, the division of the units is only a logical function division, and actual implementation can have another division manner, for example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the units shown or discussed can be indirect coupling or communication connection through some communication interfaces, devices or units, and can be electrical, mechanical or other forms.

[0205] The units described as separate components can or can not be physically separate, and the components shown as units can or can not be physical units, that is, they can be located in one place, or can be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the embodiment scheme.

[0206] In addition, the functional units in each embodiment of the present application can be integrated in one processing unit, or each unit can be physically present separately, or two or more units can be integrated in one unit.

[0207] If the functions are realized in the form of software function units and sold or used as independent products, they can be stored in a nonvolatile computer readable storage medium executable by a processor. Based on this understanding, the technical solutions of the present application or the essential part or part of the technical solutions that make contributions to the prior art can be embodied in the form of a software product. The computer software product is stored in a storage medium, including a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The foregoing storage medium includes: U disk, mobile hard disk, read-only memory (Read-Only Memory, ROM), random access memory (Random Access Memory, RAM), magnetic disk or optical disk and various program code storage media.

[0208] Finally, it should be noted that the above-described embodiments are merely specific embodiments of the present application, which are used to illustrate the technical solutions of the present application, but not to limit the same. The protection scope of the present application is not limited thereto. Although the present application has been described in detail with reference to the foregoing embodiments, it should be understood by those skilled in the art that any skilled person in the art can still modify or easily think of changes to the technical solutions recorded in the foregoing embodiments, or make equivalent replacements to some of the technical features, within the technical scope disclosed by the present application. The modifications, changes or replacements do not make the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A method for matching coronary artery branch points, characterized in that, The matching method includes: Two-dimensional coronary angiography images of the coronary artery to be reconstructed under different acquisition angles are preprocessed. By determining the coronary centerline of the coronary artery to be reconstructed under the acquisition angle of each two-dimensional coronary angiography image, multiple coronary skeleton images are obtained. Based on the starting point information of the coronary artery in each coronary artery skeleton image, multiple initial transition relationship matrices are constructed between the multiple coronary artery skeleton images; For each coronary artery skeleton image, a depth-first search method is used to determine the multiple vascular branch points in the coronary artery to be reconstructed at the acquisition angle of the coronary artery skeleton image by traversing each vascular node on the coronary artery centerline in the coronary artery skeleton image. The first vascular branch point is determined from multiple vascular branch points in the first skeleton image. By establishing an polar plane, the second vascular branch point in the second skeleton image that matches the first vascular branch point is found. Wherein, the first skeleton image is the one with the fewest vascular branch nodes among the multiple coronary skeleton images; the second skeleton image is the one with the most vascular branch nodes among the multiple coronary skeleton images. Based on the node information of the first blood vessel branch point, the node information of the second blood vessel branch point, and the starting point information of the coronary artery origin, the plurality of initial transition relationship matrices are updated to obtain the plurality of updated transition relationship matrices. According to the predetermined matching order, the matching of all vascular branch points in the first skeleton image is completed one by one, and multiple sets of matching branch point pairs that match each other among the multiple coronary skeleton images are determined. After the matching process of each vascular branch point is completed, the multiple transition relationship matrices are updated in real time. Based on the node information carried by the multiple sets of matching branch point pairs, the multiple transfer relationship matrices obtained in real time are updated to obtain multiple target transfer relationship matrices between the multiple coronary artery skeleton images, so as to complete the three-dimensional reconstruction of the coronary artery to be reconstructed using the multiple target transfer relationship matrices.

2. The matching method according to claim 1, characterized in that, Determining the first blood vessel branch point from multiple blood vessel branch points in the first skeleton image includes: The first blood vessel branch point is randomly determined from multiple blood vessel branch points in the first skeleton image; or, Among the multiple vascular branch points in the first skeleton image, the vascular branch point with the highest branch level of the coronary artery branch is determined as the first vascular branch point.

3. The matching method according to claim 1, characterized in that, The step of establishing a polar plane and matching the second blood vessel branch point in the second skeleton image that matches the first blood vessel branch point includes: With the aid of the polar plane, by establishing the first polar line corresponding to the first blood vessel branch point in the second skeleton image, multiple candidate branch points corresponding to the first blood vessel branch point in the second skeleton image are determined. By establishing a second polar line corresponding to each candidate branch point in the first skeleton image, a second blood vessel branch point that matches the first blood vessel branch point is determined.

4. The matching method according to claim 3, characterized in that, The step of determining multiple candidate branch points corresponding to the first blood vessel branch point in the second skeleton image by establishing a first polar line corresponding to the first blood vessel branch point in the second skeleton image using the polar plane includes: Using the polar plane, a first polar line corresponding to the first blood vessel branch point is established in the second skeleton image, and a first relative distance between each blood vessel branch point in the second skeleton image and the first polar line is determined respectively. The first tree structure depth of the coronary artery to be reconstructed at each vessel branch point in the second skeleton image and the second tree structure depth of the coronary artery to be reconstructed at the first vessel branch point are determined respectively. Based on the determined depths of the first tree structure and the second tree structure, the depth difference between the first blood vessel branch point and each blood vessel branch point in the second skeleton image is determined respectively. The blood vessel branch points in the second skeleton image whose first relative distance is less than a preset distance threshold and whose depth difference is less than or equal to a preset depth threshold are determined as candidate branch points corresponding to the first blood vessel branch points.

5. The matching method according to claim 3, characterized in that, The step of determining the second blood vessel branch point that matches the first blood vessel branch point by establishing a second polar line corresponding to each candidate branch point in the first skeleton image includes: Using the polar plane, a second polar line corresponding to each candidate branch point is established in the first skeleton image; Determine the second relative distance between the first blood vessel branch point and each second polar line; The candidate branch point corresponding to the second polar line with the smallest relative distance is determined as the second vascular branch point that matches the first vascular branch point.

6. The matching method according to claim 1, characterized in that, The process of matching all blood vessel branch points in the first skeleton image one by one according to a predetermined matching order, and determining multiple sets of matching branch point pairs between the first skeleton image and the second skeleton image, includes: After the first blood vessel branch point is matched, the first blood vessel branch point to be matched is determined again from other blood vessel branch points in the first skeleton image according to a predetermined matching order; wherein, the other blood vessel branch points are blood vessel branch points in the first skeleton image other than the first blood vessel branch nodes that have been matched. Using the polar plane, a second vascular branch point in the second skeleton image that matches the first vascular branch point to be matched is identified; Using the node information of each first blood vessel branch point that has been matched and the node information of the second blood vessel branch point that matches each first blood vessel branch point, the multiple transition relationship matrices are updated until the matching of all blood vessel branch points in the first skeleton image is completed, resulting in multiple sets of matched branch point pairs.

7. The matching method according to claim 1, characterized in that, For each coronary artery skeleton image, a depth-first search method is used to determine multiple vascular branch points in the coronary artery to be reconstructed at the acquisition angle of the coronary artery skeleton image by traversing each vessel node on the coronary artery centerline of the image. These branch points include: For each coronary artery skeleton image, a depth-first search method is used, taking the coronary artery origin point in the coronary artery skeleton image as the traversal starting point, and traversing each vessel node on the coronary artery centerline in the coronary artery skeleton image one by one, to establish the vessel tree structure of the coronary artery to be reconstructed under the acquisition angle of the coronary artery skeleton image, and to determine the multiple vessel branch points in the coronary artery to be reconstructed under the acquisition angle of the coronary artery skeleton image and the branch level of each vessel node.

8. A device for matching coronary artery branch points, characterized in that, The matching device includes: The image preprocessing module is used to preprocess the two-dimensional coronary angiography images of the coronary artery to be reconstructed at different acquisition angles. By determining the coronary centerline of the coronary artery to be reconstructed at the acquisition angle of each two-dimensional coronary angiography image, multiple coronary skeleton images are obtained. The matrix construction module is used to construct multiple initial transition relationship matrices between the multiple coronary artery skeleton images based on the starting point information of the coronary artery starting point in each coronary artery skeleton image to be reconstructed; The branch point determination module is used to determine multiple vascular branch points in the coronary artery to be reconstructed at the acquisition angle of each coronary artery skeleton image by using a depth-first search method and traversing each vascular node on the coronary artery centerline in the coronary artery skeleton image. The branch point matching module is used to determine the first blood vessel branch point from multiple blood vessel branch points in the first skeleton image, and to match the second blood vessel branch point in the second skeleton image that matches the first blood vessel branch point by establishing an polar plane; wherein, the first skeleton image is the one with the fewest blood vessel branch nodes among the multiple coronary skeleton images; and the second skeleton image is the one with the most blood vessel branch nodes among the multiple coronary skeleton images. The first matrix update module is used to update the multiple initial transition relationship matrices based on the node information of the first blood vessel branch point, the node information of the second blood vessel branch point, and the starting point information of the coronary artery origin point, so as to obtain the multiple updated transition relationship matrices. The second matrix update module is used to complete the matching of all vascular branch points in the first skeleton image one by one according to a predetermined matching order, determine multiple sets of matching branch point pairs that match each other among the multiple coronary skeleton images, and update the multiple transition relationship matrices in real time after completing the matching process of each vascular branch point. The matrix determination module is used to update multiple transition relationship matrices obtained in real time based on the node information carried by the multiple sets of matching branch point pairs, so as to obtain multiple target transition relationship matrices between the multiple coronary artery skeleton images, and to use the multiple target transition relationship matrices to complete the three-dimensional reconstruction of the coronary artery to be reconstructed.

9. An electronic device, characterized in that, include: The device includes a processor, a memory, and a bus. The memory stores machine-readable instructions executable by the processor. When the electronic device is running, the processor communicates with the memory via the bus. The machine-readable instructions are executed by the processor to perform the steps of the coronary artery branch point matching method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, performs the steps of the coronary artery branch point matching method as described in any one of claims 1 to 7.

Images