A blood vessel centerline optimization method, optimization device, electronic equipment and medium

Abstract

The application provides a blood vessel center line optimization method, an optimization device, an electronic device and a medium, comprising: extracting an initial blood vessel center line of a target blood vessel from an angiogram image of the target blood vessel; determining a blood vessel segmentation point of the initial blood vessel center line from a plurality of blood vessel center points included in the initial blood vessel center line based on a number of adjacent blood vessel center points of each blood vessel center point in the plurality of blood vessel center points; dividing the initial blood vessel center line into a plurality of initial blood vessel center segments according to the blood vessel segmentation point; determining a connection relationship of each initial blood vessel center segment based on a blood vessel structure feature and an image texture feature between each initial blood vessel center segment; and reconnecting the initial blood vessel center segments according to the connection relationship to obtain an optimized blood vessel center line. In this way, the deviation conditions such as false positives, fractures and adhesions can be effectively optimized and corrected, so as to generate a blood vessel center line structure with rich topological structure and reasonable blood vessel direction.

Description

Technical Field

[0001] This application relates to the field of medical image processing technology, and in particular to a method, device, electronic device, and medium for optimizing the centerline of blood vessels. Background Technology

[0002] The demand for automated analysis of coronary angiography vessels is increasingly recognized, and various automated technologies are emerging, such as automated vessel segmentation and automated extraction of vessel centerlines. However, vessel segmentation techniques are generally based on pixel-level segmentation, primarily referencing image grayscale, making it difficult to comprehensively consider multiple factors, such as spatial location and topological structure. These limitations in vessel segmentation methods lead to significant deviations in the subsequently extracted vessel centerlines, resulting in frequent false positives, breaks, and adhesions. Summary of the Invention

[0003] In view of this, the purpose of this application is to provide a method, device, electronic device and medium for optimizing the vascular centerline. By segmenting the initial vascular centerline and reconnecting the vascular segments by integrating vascular structural features and image texture features, it is possible to effectively optimize and correct deviations such as false positives, breaks and adhesions, thereby generating a vascular centerline structure with rich topology and reasonable vascular orientation.

[0004] This application provides a method for optimizing the centerline of a blood vessel, the optimization method comprising:

[0005] Extract the initial vascular centerline of the target vessel from the angiographic image of the target vessel;

[0006] Based on the number of adjacent vessel center points of each vessel center point among the plurality of vessel center points included in the initial vessel center line, the vessel segmentation points of the initial vessel center line are determined from the plurality of vessel center points.

[0007] The initial vascular centerline is divided into multiple initial vascular center segments based on the vascular segmentation points.

[0008] Based on the vascular structural features and image texture features between each initial vascular central segment, the connection relationship of each initial vascular central segment is determined.

[0009] The initial vascular center segment is reconnected according to the aforementioned connection relationship to obtain the optimized vascular centerline.

[0010] Furthermore, the step of extracting the initial vascular centerline of the target vessel from the angiographic image of the target vessel includes:

[0011] The vascular segmentation result image of the target blood vessel is determined based on the angiography image;

[0012] The blood vessel segmentation result image is iteratively filtered multiple times using a multi-scale Gaussian filter operator to obtain a filtered image of the target blood vessel.

[0013] For each blood vessel point in the blood vessel segmentation result image, determine whether the gray value of the corresponding pixel point in the filtered image is the largest in multiple directions around the pixel point;

[0014] If so, the vessel point is determined as the vessel center point to obtain the initial vessel centerline.

[0015] Furthermore, the number of adjacent vessel centers for each vessel center point is determined through the following steps:

[0016] The initial blood vessel centerline on the initial blood vessel centerline image is translated along multiple directions to obtain multiple translated blood vessel centerline images; wherein, the initial blood vessel centerline image is represented as a matrix, and the matrix element is a preset value indicating that the position of the matrix element is the center point of the blood vessel on the initial blood vessel centerline;

[0017] The initial blood vessel centerline image is multiplied by the multiple translated blood vessel centerline images to obtain multiple feature images;

[0018] The multiple feature images are added together to obtain adjacent feature images;

[0019] The feature value of each matrix element in the adjacent feature image is determined as the number of adjacent blood vessel center points of the corresponding blood vessel center point.

[0020] Furthermore, the vessel segmentation points include vessel endpoints and vessel bifurcation points; the step of dividing the initial vessel centerline into multiple initial vessel centerlines based on the vessel segmentation points includes: dividing the initial vessel centerline located between two vessel segmentation points into one initial vessel centerline segment.

[0021] Furthermore, determining the connectivity of each initial vascular central segment based on the vascular structural features and image texture features between each initial vascular central segment includes:

[0022] For the target initial vascular center segment, determine the candidate initial vascular center segments located within the search range of the target initial vascular center segment;

[0023] For each candidate initial vascular center segment, based on the gravitational field between the candidate initial vascular center segment and the target initial vascular center segment, the gravitational probability of the target initial vascular center segment connecting with the candidate initial vascular center segment is determined;

[0024] Based on the orientation angle of the candidate initial blood vessel center segment and the orientation angle of the target initial blood vessel center segment, the direction probability of the connection between the target initial blood vessel center segment and the candidate initial blood vessel center segment is determined;

[0025] Texture features are extracted from the candidate gray-level co-occurrence matrix of the candidate initial blood vessel center segment and the target gray-level co-occurrence matrix of the target initial blood vessel center segment, respectively. The texture probability of the target initial blood vessel center segment being connected to the candidate initial blood vessel center segment is determined based on the feature similarity between the texture features.

[0026] Based on the gravity probability, the orientation probability, and the texture probability, the combined probability of the target initial blood vessel center segment being connected to the candidate initial blood vessel center segment is determined.

[0027] The connection relationship of the target initial vascular central segment is determined based on the comprehensive probability of its connection with each candidate initial vascular central segment.

[0028] Furthermore, based on the comprehensive probability of the target initial vascular central segment connecting with each candidate initial vascular central segment, the connection relationship of the target initial vascular central segment is determined, including:

[0029] Determine whether the overall probability of the target initial vascular center segment connecting with each candidate initial vascular center segment is less than a preset probability threshold;

[0030] If so, then it is determined that the target initial vascular central segment should not be connected to other initial vascular central segments;

[0031] If not, then the target initial vascular segment should be connected to the candidate initial vascular segment with the highest overall probability.

[0032] Furthermore, the step of reconnecting the initial vascular center segment according to the aforementioned connection relationship to obtain the optimized vascular centerline includes:

[0033] When the connection relationship of any initial vascular central segment indicates that the initial vascular central segment should not be connected to other initial vascular central segments, the initial vascular central segment is discarded as noise.

[0034] When the connection relationship of any initial vascular central segment indicates that the initial vascular central segment should be connected to other initial vascular central segments, the vascular segmentation point connecting the initial vascular central segment to the vascular segmentation point of other initial vascular central segments is determined, and the vascular connection line between the two vascular segmentation points is determined to be the vascular central line.

[0035] This application embodiment also provides a device for optimizing the centerline of a blood vessel, the device comprising:

[0036] An extraction module is used to extract the initial vascular centerline of the target blood vessel from the angiographic image of the target blood vessel;

[0037] The first determining module is used to determine the vascular segmentation points of the initial vascular centerline from the plurality of vascular center points based on the number of adjacent vascular center points of each vascular center point among the plurality of vascular center points included in the initial vascular centerline.

[0038] A segmentation module is used to divide the initial blood vessel centerline into multiple initial blood vessel center segments based on the blood vessel segmentation points.

[0039] The second determining module is used to determine the connection relationship of each initial blood vessel central segment based on the vascular structure features and image texture features between each initial blood vessel central segment;

[0040] A connection module is used to reconnect the initial vascular center segment according to the connection relationship to obtain an optimized vascular centerline.

[0041] This application also provides an electronic device, including: 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. When the machine-readable instructions are executed by the processor, the steps of the above-described method for optimizing the vascular centerline are performed.

[0042] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, performs the steps of the above-described method for optimizing the centerline of a blood vessel.

[0043] This application provides a method, device, electronic device, and medium for optimizing the vascular centerline. By segmenting the initial vascular centerline and reconnecting the vascular segments by integrating vascular structural features and image texture features, it can effectively optimize and correct deviations such as false positives, breaks, and adhesions, thereby generating a vascular centerline structure with rich topology and reasonable vascular orientation.

[0044] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0045] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0046] Figure 1 A flowchart illustrating a method for optimizing the centerline of a blood vessel provided in an embodiment of this application is shown;

[0047] Figure 2 This illustration shows a calculation diagram for determining the number of adjacent vessel center points for each vessel center point, provided by an embodiment of this application.

[0048] Figure 3 This illustration shows a segmented schematic diagram of an initial vascular centerline provided in an embodiment of this application;

[0049] Figures 4(a) and 4(b) show schematic diagrams illustrating a method for determining the connection relationship of each initial vascular central segment according to an embodiment of this application;

[0050] Figure 5 This illustration shows a schematic diagram of an optimized vascular centerline provided by an embodiment of this application;

[0051] Figure 6 This illustration shows a structural schematic diagram of a device for optimizing the centerline of a blood vessel according to an embodiment of this application;

[0052] Figure 7 A schematic diagram of the structure of an electronic device provided in an embodiment of this application is shown. Detailed Implementation

[0053] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. The components of the embodiments of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely represents selected embodiments of this application. Based on the embodiments of this application, every other embodiment obtained by those skilled in the art without inventive effort falls within the scope of protection of this application.

[0054] Research has revealed a growing demand for automated analysis of coronary angiography vessels, with various automated techniques emerging, such as automated vessel segmentation and automated centerline extraction. However, vessel segmentation techniques are generally pixel-level, primarily relying on image grayscale and failing to comprehensively consider multiple factors, such as spatial location and topological structure. These limitations in vessel segmentation methods lead to significant deviations in the subsequently extracted vessel centerlines, resulting in frequent false positives, vessel breakage, and adhesions.

[0055] Based on this, embodiments of this application provide a method, device, electronic device, and medium for optimizing the vascular centerline. By segmenting the initial vascular centerline and reconnecting the vascular segments by integrating vascular structural features and image texture features, it is possible to effectively optimize and correct deviations such as false positives, breaks, and adhesions, thereby generating a vascular centerline structure with rich topology and reasonable vascular orientation.

[0056] Please see Figure 1 , Figure 1 A flowchart illustrating a method for optimizing the centerline of a blood vessel, as provided in an embodiment of this application. Figure 1 As shown in the embodiments of this application, the optimization method includes:

[0057] S101. Extract the initial vascular centerline of the target blood vessel from the angiography image of the target blood vessel.

[0058] Here, taking the coronary artery as the target vessel as an example, coronary angiography is a commonly used and effective method for diagnosing coronary atherosclerotic heart disease (CHD). It is a relatively safe and reliable invasive diagnostic technique that is now widely used in clinical practice and is considered the "gold standard" for diagnosing CHD. It involves injecting a contrast agent into the coronary arteries through a catheter, and then obtaining 2D coronary angiographic images via X-ray imaging.

[0059] In one possible implementation, step S101 may include:

[0060] S1011. Determine the vascular segmentation result image of the target blood vessel based on the angiography image.

[0061] In this step, any method in the existing technology can be used, such as using a blood vessel segmentation model to segment the angiography image to obtain the blood vessel segmentation result image of the target blood vessel. The blood vessel segmentation result image is a binary image with the same size as the original angiography image. The binary image has two types of pixel values, such as 0 and 1, where 0 represents the background and 1 represents the foreground, i.e., the blood vessel.

[0062] In the blood vessel segmentation image obtained in step S1011, the center position of the blood vessel and the gray value of the blood vessel boundary are the same, making it impossible to accurately locate the center line of the blood vessel. Therefore, in S1012 to S1014 below, this embodiment first uses a multi-scale iterative Gaussian filtering method to process the binary blood vessel segmentation result accordingly, making the pixel value of the blood vessel center different from that of the edge, i.e., achieving a peak effect (the center is higher than the two sides); secondly, a multi-directional non-maximum suppression method is used to extract the center point, i.e., to find the highest position (peak) in a local region, thereby obtaining the initial blood vessel center line. Specifically:

[0063] S1012. The blood vessel segmentation result image is iteratively filtered multiple times using a multi-scale Gaussian filter operator to obtain the filtered image of the target blood vessel.

[0064] Taking coronary arteries as an example, a prominent feature of the coronary artery tree is the inconsistent and highly variable diameter of the vessels. Therefore, when performing Gaussian filtering, using Gaussian filtering operators of different scales can cover vessels of varying diameters as much as possible. Furthermore, the embodiments of this application employ a multi-iteration method to maximize the difference between the vessel center and the vessel edge, thereby creating a peak effect.

[0065] Specifically, the formula for iteratively filtering the blood vessel segmentation result image using a multi-scale Gaussian filter operator can be expressed as:

[0066]

[0067] In the formula, N represents the total number of iterations; n represents the current iteration; I(n) represents the current filtered image obtained in the nth iteration; I(0) represents the blood vessel segmentation result image; G(σ) represents the convolution operation; n ) represents the standard deviation as σ n The Gaussian convolution kernel. For example, when N=3, σ1=9, σ2=6, σ3=3 can be selected.

[0068] S1013. For each blood vessel point in the blood vessel segmentation result image, determine whether the gray value of the corresponding pixel point in the filtered image is the largest in multiple directions around the pixel point.

[0069] S1014. If so, then the blood vessel point is determined as the blood vessel center point to obtain the initial blood vessel centerline.

[0070] For steps S1013 and S1014, the position of the vessel center point is calculated using the local directional maxima method. For example, a coordinate system can be established for each vessel point along the vessel's direction and perpendicular to it, selecting four directional angles: 0°, 45°, 90°, and 135°. For each foreground vessel point in the segmented image, based on the Gaussian-filtered image, it is calculated whether the vessel point is the pixel with the largest grayscale value among its four adjacent pixels. If so, the vessel point is retained as the vessel center point; otherwise, it is not retained. Thus, only the center point pixels on the "peaks" are retained, i.e., the vessel center points, thereby obtaining the vessel centerline image where the vessel centerline is located.

[0071] In another possible implementation, a vessel enhancement image can be obtained from the original angiography image based on the Hessian matrix filtering algorithm. Then, threshold segmentation is performed on the vessel enhancement image, and the segmentation result is skeletonized to obtain the initial centerline. The drawbacks of the above methods are that the segmentation result obtained from the enhancement image includes a lot of background, and the centerline extraction depends on the skeletonization algorithm. The obtained centerline may not be completely at the center of the vessel, which will also lead to significant deviations in the subsequently extracted vessel centerline, such as frequent false positives, breaks, and adhesions.

[0072] S102. Based on the number of adjacent blood vessel center points of each blood vessel center point among the multiple blood vessel center points included in the initial blood vessel center line, determine the blood vessel segmentation points of the initial blood vessel center line from the multiple blood vessel center points.

[0073] In one possible implementation, the number of adjacent vessel center points for each vessel center point can be determined through the following steps:

[0074] The initial vascular centerline image is translated along multiple directions to obtain multiple translated vascular centerline images. The initial vascular centerline image is represented as a matrix, where each matrix element represents a preset value indicating that the corresponding position is the vascular center point on the initial vascular centerline. The initial vascular centerline image is then multiplied by each of the multiple translated vascular centerline images to obtain multiple feature images. These feature images are then added to obtain adjacent feature images. The eigenvalue of each matrix element in the adjacent feature images is determined as the number of adjacent vascular center points corresponding to that matrix element.

[0075] Please refer to the following: Figure 2 , Figure 2 This is a schematic diagram illustrating a calculation method for determining the number of adjacent vessel center points for each vessel center point, provided as an embodiment of this application. Figure 2As shown, the initial vessel centerline image is a 4x4 binary image (which can be represented as a 4x4 matrix during calculation). A matrix element of "1" indicates that the corresponding position of the matrix element is the vessel center point on the initial vessel centerline. Assuming that the number of adjacent points of each vessel center point in the four neighborhood directions (up, down, left, and right) is calculated, the initial vessel centerline can be translated along the four directions to obtain four translated vessel centerline images. The four translated images are then multiplied by the initial vessel centerline image to obtain multiple feature images. Finally, the multiple feature images are added by matrix to obtain the adjacent feature images.

[0076] In practice, in order to more accurately determine the number of adjacent vessel center points in multiple directions, an 8-neighborhood neighbor point calculation can be performed on the center line, that is, the vessel center line is translated and calculated along eight different directions.

[0077] Thus, if the convolution calculation method in the existing technology is used, it is necessary to traverse each point to perform (8+1) x W x H calculations (W and H are the image size, usually 512x512), which results in a very large number of calculations; however, if the above image displacement plus matrix multiplication method is used, only 8+1=9 calculations are needed, which can greatly reduce the amount of calculation and improve the calculation speed.

[0078] After determining the number of adjacent vessel centers for each vessel center, the type of each vessel center can be determined. The types of vessel center centers include vessel endpoints, vessel bifurcation points, and vessel connection points, while vessel segmentation points include vessel endpoints and vessel bifurcation points.

[0079] Specifically, if the eigenvalue of a matrix element in the adjacency feature image is 1, it means that the center point of the blood vessel has 1 adjacent center point, and the center point of the blood vessel is an endpoint; if the eigenvalue of a matrix element in the adjacency feature image is 2, it means that the center point of the blood vessel has 2 adjacent center points, and the center point of the blood vessel is a connection point; if the eigenvalue of a matrix element in the adjacency feature image is greater than or equal to 3, it means that the center point of the blood vessel has greater than or equal to 3 adjacent center points, and the center point of the blood vessel is a bifurcation point.

[0080] S103. Divide the initial blood vessel centerline into multiple initial blood vessel center segments according to the blood vessel segmentation points.

[0081] The vessel segmentation points include vessel endpoints and vessel bifurcation points; in a specific implementation, step S103 may include: dividing the initial vessel centerline located between two vessel segmentation points into an initial vessel center segment.

[0082] Please see Figure 3 , Figure 3This is a segmented schematic diagram of an initial vascular centerline provided in an embodiment of this application. Figure 3 As shown, after segmenting the initial vascular centerline, the following appeared: Figure 3 The scenarios are "fracture" as shown in Case 1 and "adhesion" as shown in Case 2. However, for the "fracture" shown in Case 1, one possibility is that the two broken initial vascular central segments are genuine fractures, requiring connection; another possibility is that one of the "fractured" initial vascular central segments is "false positive" noise, meaning it doesn't actually belong to the vascular centralline structure and should be discarded. However, after segmenting the initial vascular centralline, it's impossible to determine which possibility the "fracture" shown in Case 1 belongs to or what optimization method should be adopted; further determination is needed based on subsequent steps.

[0083] S104. Based on the vascular structural features and image texture features between each initial vascular central segment, determine the connection relationship of each initial vascular central segment.

[0084] In this step, the matching probability between initial vascular central segments can be determined based on the vascular structural features and image texture features between each initial vascular central segment. The vascular structural features can include the distance and direction between vascular segments, and the image texture features can include the second moment of the angle and the contrast. Then, the connection relationship can be determined based on the matching probability to select initial vascular central segments for connection, thereby reconnecting broken segments, distinguishing adhesions, and discarding false positives.

[0085] Please refer to Figures 4(a) and 4(b), which are schematic diagrams illustrating the determination of the connection relationship of each initial vascular central segment according to an embodiment of this application. In one possible implementation, step S104 may include:

[0086] S1041. For the target initial vascular center segment, determine the candidate initial vascular center segments located within the search range of the target initial vascular center segment.

[0087] As shown in Figure 4(a), the target initial blood vessel center segment is p. When matching blood vessel center segments, only candidate initial blood vessel center segments within the search range are considered for calculation. For example, first, the direction of p is determined, then the direction of p is rotated 60° clockwise and 60° counterclockwise (a total of 120°) as the search angle, and the search radius is determined to be 25 pixels, thus obtaining the search range. Initial blood vessel center segments whose center points fall within this search range are determined as candidate initial blood vessel center segments. As shown in Figure 4(a), the candidate initial blood vessel center segments corresponding to the target initial blood vessel center segment p are q1 and q2.

[0088] S1042. For each candidate initial vascular center segment, based on the gravitational field between the candidate initial vascular center segment and the target initial vascular center segment, determine the gravitational probability that the target initial vascular center segment is connected to the candidate initial vascular center segment.

[0089] In this step, the gravitational field (tensor field) reflects the gravitational force between two points, and the greater the gravitational force, the higher the probability of a connected gravitational force. As shown in Figure 4(b), P and O represent the endpoints of the two initial blood vessel central segments p and q, respectively, where the gravitational force is to be calculated. Draw a circle passing through points O and P, with its center at C. v represents the direction vector of the initial blood vessel central segment where point O is located, and w represents the direction perpendicular to CP. Assuming that the gravitational force between points O and P needs to be calculated, the direction w is first determined using the following formula:

[0090] w = 2OP(OP·v) - v

[0091] Then, the gravitational probability of the initial target vessel center segment p connecting with the candidate initial vessel center segment q is calculated based on the direction w using the following formula:

[0092]

[0093] In the formula, p s (p,q) represents the gravitational probability of connecting the initial vascular center segment p with the candidate initial vascular center segment q; r represents the length of the radian OP; represents the curvature of radians OP; c represents the weighted ratio of length and curvature. And σ can be calculated using the following formula:

[0094]

[0095] In the formula, L represents the straight length of line segment CP; θ represents the angle between line segment CO and line segment CP.

[0096] As can be seen from the above formula, the closer the distance between two points and the smaller the arc, the greater the gravitational force, and vice versa; when the endpoints of the two initial blood vessel center segments coincide, the gravitational probability is set to 1.

[0097] S1043. Based on the direction angle of the candidate initial blood vessel center segment and the direction angle of the target initial blood vessel center segment, determine the direction probability of the target initial blood vessel center segment connecting with the candidate initial blood vessel center segment.

[0098] In practical implementation, the direction probability of connection between the target initial blood vessel central segment p and the candidate initial blood vessel central segment q can be determined by the following formula:

[0099]

[0100] In the formula, p d (p,q) represents the distance probability of the initial vascular center segment p connecting with the candidate initial vascular center segment q; d(p) and d(q) represent the directional angles of the initial vascular center segment p and the candidate initial vascular center segment q, respectively, such as 10° and 20°; c represents the preset weight coefficient.

[0101] S1044. Extract texture features from the candidate gray-level co-occurrence matrix of the candidate initial blood vessel center segment and the target gray-level co-occurrence matrix of the target initial blood vessel center segment, respectively, and determine the texture probability that the target initial blood vessel center segment is connected to the candidate initial blood vessel center segment based on the feature similarity between the texture features.

[0102] In this context, the matrix elements in the gray-level co-occurrence matrix represent the joint distribution of gray levels of two pixels with a certain spatial relationship. For example, the gray-level co-occurrence matrix of a 9×9 angiographic image centered on the point to be calculated (e.g., the segmentation point of the blood vessel) in the central segment can be calculated as a texture feature map. The calculation of the value of the matrix element at position (i, j) in the gray-level co-occurrence matrix P involves traversing all pixels with gray level i in the original image and calculating the number of pixels with a spacing of d pixels, a direction of θ, and a gray level of j. The formula can be expressed as:

[0103] P(i,j|d,θ)={(x,y)|f(x,y)=i,f(x+dx,y+dy)=j;x,y=0,1,…,N-1}

[0104] In the formula, d represents the spacing distance between pixels to be calculated in the original image, θ is the direction of the spacing pixels, and (x,y) is the coordinate position in the original image.

[0105] Based on the gray-level co-occurrence matrix P, the following two important statistical parameters are further calculated as texture features:

[0106] Second moment of angle: ASM=∑ i ∑ j P(i,j) 2

[0107] Contrast ratio: Con = ∑ i ∑ j (i,j) 2 P(i,j)

[0108] It should be noted that the stronger the connectivity between two points, the greater their texture feature similarity should be. Therefore, the texture probability of the target initial blood vessel center segment p being connected to the candidate initial blood vessel center segment q can be determined by the following formula:

[0109]

[0110] In the formula, pg (p,q) represents the texture probability that the initial blood vessel center segment p is connected to the candidate initial blood vessel center segment q; c = 0.001 represents the preset Gaussian scaling factor.

[0111] S1045. Based on the gravity probability, the direction probability, and the texture probability, determine the comprehensive probability that the target initial blood vessel center segment is connected to the candidate initial blood vessel center segment.

[0112] In practical implementation, the combined probability of connection between the target initial vascular central segment and the candidate initial vascular central segment can be determined using the following formula:

[0113] Prob(p,q)=w1p s (p,q)+w2p g (p,q)+w3p d (p,q)

[0114] In the formula, Prob(p,q) represents the overall probability; w1, w2 and w3 represent the weight coefficients of each item in the overall probability.

[0115] S1046. Based on the comprehensive probability of the target initial vascular central segment being connected to each candidate initial vascular central segment, determine the connection relationship of the target initial vascular central segment.

[0116] In a specific implementation, step S1046 may include: determining whether the overall probability of the target initial vascular center segment being connected to each candidate initial vascular center segment is less than a preset probability threshold; if so, determining that the target initial vascular center segment should not be connected to other initial vascular center segments; if not, determining that the target initial vascular center segment should be connected to the candidate initial vascular center segment with the highest overall probability.

[0117] S105. Reconnect the initial vascular center segment according to the connection relationship to obtain the optimized vascular centerline.

[0118] In one possible implementation, step S105 may include: when the connection relationship of any initial vascular central segment indicates that the initial vascular central segment should not be connected to other initial vascular central segments, discarding the initial vascular central segment as noise; when the connection relationship of any initial vascular central segment indicates that the initial vascular central segment should be connected to other initial vascular central segments, connecting the vascular segmentation point of the initial vascular central segment to the vascular segmentation point of other initial vascular central segments, and determining that the vascular connection line between the two vascular segmentation points belongs to the vascular central line.

[0119] Here, if the combined probability of a certain initial vascular central segment and other initial vascular central segments within the search range is less than a preset probability threshold, such as 0.3, then it is considered that the initial vascular central segment should not be connected to other initial vascular central segments. In this case, the initial vascular central segment is considered a "false positive" vascular central segment and should be discarded as noise.

[0120] Please compare and refer to the following: Figure 3 and Figure 5 , Figure 5 This is a schematic diagram illustrating an optimized blood vessel centerline provided in an embodiment of this application. As shown in the figure, for... Figure 3 In scenario 1, the "fracture" can be confirmed, based on the steps outlined above, as the two initial central segments of the vessel are indeed fractured. Figure 5 The dashed lines in box 1 connect the two closest segment points of the initial vessel central segments according to their connection relationships. For the "adhesion" shown in case 2, the above steps determine that initial vessel central segment a should be connected to initial vessel central segment c, and initial vessel central segment b should be connected to initial vessel central segment d. Therefore, although the morphology of the "adhesive" vessel central segments is not changed, the original four initial vessel central segments can be relabeled as optimized vessel central segments A and B. This makes the vessel central segments visually appear adhered due to their own vascular topology, but distinguishable to the computer. In this way, both broken and overlapping adhered vessels are correctly connected, and false positive vessels are discarded, providing an important prerequisite for subsequent coronary tree analysis, such as topology acquisition and vessel-by-vessel modeling.

[0121] This application provides a method for optimizing the vascular centerline, comprising: extracting an initial vascular centerline of the target vascular vessel from an angiographic image of the target vascular vessel; determining vascular segmentation points of the initial vascular centerline from the plurality of vascular center points included in the initial vascular centerline based on the number of adjacent vascular center points of each vascular center point; dividing the initial vascular centerline into multiple initial vascular center segments according to the vascular segmentation points; determining the connection relationship of each initial vascular center segment based on the vascular structural features and image texture features between each initial vascular center segment; and reconnecting the initial vascular center segments according to the connection relationship to obtain an optimized vascular centerline.

[0122] In this way, by segmenting the initial vascular centerline and reconnecting the vascular segments by integrating vascular structural features and image texture features, it is possible to effectively optimize and correct deviations such as false positives, breaks, and adhesions, thereby generating a vascular centerline structure with rich topology and reasonable vascular orientation.

[0123] Please see Figure 6 , Figure 6 This is a schematic diagram of a device for optimizing the centerline of a blood vessel, provided in an embodiment of this application. Figure 6 As shown, the optimization device 600 includes:

[0124] Extraction module 610 is used to extract the initial vascular centerline of the target blood vessel from the angiography image of the target blood vessel;

[0125] The first determining module 620 is used to determine the vascular segmentation points of the initial vascular centerline from the plurality of vascular center points based on the number of adjacent vascular center points of each vascular center point among the plurality of vascular center points included in the initial vascular centerline.

[0126] The segmentation module 630 is used to divide the initial blood vessel centerline into multiple initial blood vessel center segments based on the blood vessel segmentation points.

[0127] The second determining module 640 is used to determine the connection relationship of each initial blood vessel central segment based on the vascular structure features and image texture features between each initial blood vessel central segment.

[0128] The connection module 650 is used to reconnect the initial vascular center segment according to the connection relationship to obtain an optimized vascular centerline.

[0129] Furthermore, when extracting the initial vascular centerline of the target blood vessel from the angiographic image of the target blood vessel, the extraction module 610 is used to:

[0130] The vascular segmentation result image of the target blood vessel is determined based on the angiography image;

[0131] The blood vessel segmentation result image is iteratively filtered multiple times using a multi-scale Gaussian filter operator to obtain a filtered image of the target blood vessel.

[0132] For each blood vessel point in the blood vessel segmentation result image, determine whether the gray value of the corresponding pixel point in the filtered image is the largest in multiple directions around the pixel point;

[0133] If so, the vessel point is determined as the vessel center point to obtain the initial vessel centerline.

[0134] Furthermore, the optimization device 600 also includes a third determining module; the third determining module is used to determine the number of adjacent vessel center points for each vessel center point through the following steps:

[0135] The initial blood vessel centerline on the initial blood vessel centerline image is translated along multiple directions to obtain multiple translated blood vessel centerline images; wherein, the initial blood vessel centerline image is represented as a matrix, and the matrix element is a preset value indicating that the position of the matrix element is the center point of the blood vessel on the initial blood vessel centerline;

[0136] The initial blood vessel centerline image is multiplied by the multiple translated blood vessel centerline images to obtain multiple feature images;

[0137] The multiple feature images are added together to obtain adjacent feature images;

[0138] The feature value of each matrix element in the adjacent feature image is determined as the number of adjacent blood vessel center points of the corresponding blood vessel center point.

[0139] Furthermore, the vessel segmentation points include vessel endpoints and vessel bifurcation points; when the segmentation module 630 divides the initial vessel centerline into multiple initial vessel center segments based on the vessel segmentation points, the segmentation module 630 is used for:

[0140] The initial vascular centerline located between two vascular segmentation points is divided into an initial vascular center segment.

[0141] Furthermore, when determining the connection relationship of each initial vascular central segment based on the vascular structural features and image texture features between each initial vascular central segment, the second determining module 640 is used to:

[0142] For the target initial vascular center segment, determine the candidate initial vascular center segments located within the search range of the target initial vascular center segment;

[0143] For each candidate initial vascular center segment, based on the gravitational field between the candidate initial vascular center segment and the target initial vascular center segment, the gravitational probability of the target initial vascular center segment connecting with the candidate initial vascular center segment is determined;

[0144] Based on the orientation angle of the candidate initial blood vessel center segment and the orientation angle of the target initial blood vessel center segment, the direction probability of the connection between the target initial blood vessel center segment and the candidate initial blood vessel center segment is determined;

[0145] Texture features are extracted from the candidate gray-level co-occurrence matrix of the candidate initial blood vessel center segment and the target gray-level co-occurrence matrix of the target initial blood vessel center segment, respectively. The texture probability of the target initial blood vessel center segment being connected to the candidate initial blood vessel center segment is determined based on the feature similarity between the texture features.

[0146] Based on the gravity probability, the orientation probability, and the texture probability, the combined probability of the target initial blood vessel center segment being connected to the candidate initial blood vessel center segment is determined.

[0147] The connection relationship of the target initial vascular central segment is determined based on the comprehensive probability of its connection with each candidate initial vascular central segment.

[0148] Furthermore, when determining the connection relationship of the target initial vascular central segment based on the comprehensive probability of connection between the target initial vascular central segment and each candidate initial vascular central segment, the second determining module 640 is used to:

[0149] Determine whether the overall probability of the target initial vascular center segment connecting with each candidate initial vascular center segment is less than a preset probability threshold;

[0150] If so, then it is determined that the target initial vascular central segment should not be connected to other initial vascular central segments;

[0151] If not, then the target initial vascular segment should be connected to the candidate initial vascular segment with the highest overall probability.

[0152] Furthermore, when the connection module 650 is used to reconnect the initial vascular center segment according to the connection relationship to obtain an optimized vascular centerline, the connection module 650 is used to:

[0153] When the connection relationship of any initial vascular central segment indicates that the initial vascular central segment should not be connected to other initial vascular central segments, the initial vascular central segment is discarded as noise.

[0154] When the connection relationship of any initial vascular central segment indicates that the initial vascular central segment should be connected to other initial vascular central segments, the vascular segmentation point connecting the initial vascular central segment to the vascular segmentation point of other initial vascular central segments is determined, and the vascular connection line between the two vascular segmentation points is determined to be the vascular central line.

[0155] Please see Figure 7 , Figure 7 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Figure 7 As shown, the electronic device 700 includes a processor 710, a memory 720, and a bus 730.

[0156] The memory 720 stores machine-readable instructions executable by the processor 710. When the electronic device 700 is running, the processor 710 communicates with the memory 720 via the bus 730. When the machine-readable instructions are executed by the processor 710, they can perform the operations described above. Figure 1 So much so Figure 5 The steps of the method embodiment shown in the illustration are for optimizing the centerline of a blood vessel. For specific implementation details, please refer to the method embodiment, which will not be repeated here.

[0157] This application also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, can perform the above-described actions. Figures 1 to 5 The steps of the method embodiment shown in the illustration are for optimizing the centerline of a blood vessel. For specific implementation details, please refer to the method embodiment, which will not be repeated here.

[0158] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0159] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. The apparatus embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. Furthermore, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Additionally, the shown or discussed mutual couplings, direct couplings, or communication connections may be through some communication interfaces; indirect couplings or communication connections between devices or units may be electrical, mechanical, or other forms.

[0160] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0161] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0162] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a processor-executable, non-volatile, computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0163] Finally, it should be noted that the above-described embodiments are merely specific implementations of this application, used to illustrate the technical solutions of this application, and not to limit them. The scope of protection of this application is not limited thereto. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that any person skilled in the art can still modify or easily conceive of changes to the technical solutions described in the foregoing embodiments, or make equivalent substitutions for some of the technical features, within the scope of the technology disclosed in this application. Such modifications, changes, or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be covered within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method of optimizing a blood vessel centerline, characterized by, The optimization method includes: Extract the initial vascular centerline of the target vessel from the angiographic image of the target vessel; Based on the number of adjacent vessel center points of each vessel center point among the plurality of vessel center points included in the initial vessel center line, the vessel segmentation points of the initial vessel center line are determined from the plurality of vessel center points. The initial vascular centerline is divided into multiple initial vascular center segments based on the vascular segmentation points. Based on the vascular structural features and image texture features between each initial vascular central segment, the connection relationship of each initial vascular central segment is determined. The initial vascular center segment is reconnected according to the aforementioned connection relationship to obtain the optimized vascular centerline; Based on the vascular structural features and image texture features between each initial vascular central segment, the connectivity relationships of each initial vascular central segment are determined, including: For the target initial vascular center segment, determine the candidate initial vascular center segments located within the search range of the target initial vascular center segment; For each candidate initial vascular center segment, based on the gravitational field between the candidate initial vascular center segment and the target initial vascular center segment, the gravitational probability of the target initial vascular center segment connecting with the candidate initial vascular center segment is determined; Based on the orientation angle of the candidate initial blood vessel center segment and the orientation angle of the target initial blood vessel center segment, the direction probability of the connection between the target initial blood vessel center segment and the candidate initial blood vessel center segment is determined; Texture features are extracted from the candidate gray-level co-occurrence matrix of the candidate initial blood vessel center segment and the target gray-level co-occurrence matrix of the target initial blood vessel center segment, respectively. The texture probability of the target initial blood vessel center segment being connected to the candidate initial blood vessel center segment is determined based on the feature similarity between the texture features. Based on the gravity probability, the orientation probability, and the texture probability, the combined probability of the target initial blood vessel center segment being connected to the candidate initial blood vessel center segment is determined. Determine whether the overall probability of the target initial vascular center segment connecting with each candidate initial vascular center segment is less than a preset probability threshold; If so, then it is determined that the target initial vascular central segment should not be connected to other initial vascular central segments; The step of reconnecting the initial vascular center segment according to the aforementioned connection relationship to obtain the optimized vascular centerline includes: When the connection relationship of any initial vascular central segment indicates that the initial vascular central segment should not be connected to other initial vascular central segments, the initial vascular central segment is discarded as noise.

2. The optimization method according to claim 1, characterized in that, The step of extracting the initial vascular centerline of the target blood vessel from the angiographic image of the target blood vessel includes: The vascular segmentation result image of the target blood vessel is determined based on the angiography image; The blood vessel segmentation result image is iteratively filtered multiple times using a multi-scale Gaussian filter operator to obtain a filtered image of the target blood vessel. For each blood vessel point in the blood vessel segmentation result image, determine whether the gray value of the corresponding pixel point in the filtered image is the largest in multiple directions around the pixel point; If so, the vessel point is determined as the vessel center point to obtain the initial vessel centerline.

3. The optimization method according to claim 1, characterized in that, The number of adjacent vessel center points for each vessel center point is determined by the following steps: The initial blood vessel centerline on the initial blood vessel centerline image is translated along multiple directions to obtain multiple translated blood vessel centerline images; wherein, the initial blood vessel centerline image is represented as a matrix, and the matrix element is a preset value indicating that the position of the matrix element is the center point of the blood vessel on the initial blood vessel centerline; The initial blood vessel centerline image is multiplied by the multiple translated blood vessel centerline images to obtain multiple feature images; The multiple feature images are added together to obtain adjacent feature images; The feature value of each matrix element in the adjacent feature image is determined as the number of adjacent blood vessel center points of the corresponding blood vessel center point.

4. The optimization method according to claim 1, characterized in that, The vessel segmentation points include vessel endpoints and vessel bifurcation points; the step of dividing the initial vessel centerline into multiple initial vessel centerlines based on the vessel segmentation points includes: dividing the initial vessel centerline located between two vessel segmentation points into one initial vessel centerline segment.

5. The optimization method according to claim 1, characterized in that, Based on the vascular structural features and image texture features between each initial vascular central segment, the connectivity relationships of each initial vascular central segment are determined, including: If the overall probability of the target initial vascular center segment being connected to each candidate initial vascular center segment is not less than a preset probability threshold, then it is determined that the target initial vascular center segment should be connected to the candidate initial vascular center segment with the highest overall probability.

6. The optimization method according to claim 5, characterized in that, The step of reconnecting the initial vascular center segment according to the connection relationship to obtain the optimized vascular centerline further includes: When the connection relationship of any initial vascular central segment indicates that the initial vascular central segment should be connected to other initial vascular central segments, the vascular segmentation point connecting the initial vascular central segment to the vascular segmentation point of other initial vascular central segments is determined, and the vascular connection line between the two vascular segmentation points is determined to be the vascular central line.

7. A device for optimizing the centerline of a blood vessel, characterized in that, The optimization device includes: An extraction module is used to extract the initial vascular centerline of the target blood vessel from the angiographic image of the target blood vessel; The first determining module is used to determine the vascular segmentation points of the initial vascular centerline from the plurality of vascular center points based on the number of adjacent vascular center points of each vascular center point among the plurality of vascular center points included in the initial vascular centerline. A segmentation module is used to divide the initial blood vessel centerline into multiple initial blood vessel center segments based on the blood vessel segmentation points. The second determining module is used to determine the connection relationship of each initial blood vessel central segment based on the vascular structure features and image texture features between each initial blood vessel central segment; A connection module is used to reconnect the initial vascular center segment according to the connection relationship to obtain an optimized vascular centerline; When the second determining module determines the connectivity of each initial vascular central segment based on the vascular structural features and image texture features between each initial vascular central segment, the second determining module is used to: For the target initial vascular center segment, determine the candidate initial vascular center segments located within the search range of the target initial vascular center segment; For each candidate initial vascular center segment, based on the gravitational field between the candidate initial vascular center segment and the target initial vascular center segment, the gravitational probability of the target initial vascular center segment connecting with the candidate initial vascular center segment is determined; Based on the orientation angle of the candidate initial blood vessel center segment and the orientation angle of the target initial blood vessel center segment, the direction probability of the connection between the target initial blood vessel center segment and the candidate initial blood vessel center segment is determined; Texture features are extracted from the candidate gray-level co-occurrence matrix of the candidate initial blood vessel center segment and the target gray-level co-occurrence matrix of the target initial blood vessel center segment, respectively. The texture probability of the target initial blood vessel center segment being connected to the candidate initial blood vessel center segment is determined based on the feature similarity between the texture features. Based on the gravity probability, the orientation probability, and the texture probability, the combined probability of the target initial blood vessel center segment being connected to the candidate initial blood vessel center segment is determined. Determine whether the overall probability of the target initial vascular center segment connecting with each candidate initial vascular center segment is less than a preset probability threshold; If so, then it is determined that the target initial vascular central segment should not be connected to other initial vascular central segments; When the connection module is used to reconnect the initial vascular center segment according to the connection relationship to obtain the optimized vascular centerline, the connection module is used to: When the connection relationship of any initial vascular central segment indicates that the initial vascular central segment should not be connected to other initial vascular central segments, the initial vascular central segment is discarded as noise.

8. 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 a method for optimizing the vascular centerline as described in any one of claims 1 to 6.

9. 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 a method for optimizing the vascular centerline as described in any one of claims 1 to 6.

Images