Vessel centerline extraction method, system, device and readable storage medium

Abstract

The application provides a blood vessel center line extraction method, comprising: obtaining an initial ridge point domain of a binary image; determining an initial starting point and an initial ending point according to the initial ridge point domain, and extracting a target center line through a preset value and a minimum cost path according to the initial starting point and the initial ending point; in a ridge point domain without the target center line, determining a reference starting point and a reference ending point according to all target center lines, and extracting a next target center line through a preset value and a minimum cost path according to the reference starting point and the reference ending point; repeating the extraction of the next target center line in the ridge point domain without the target center line until there is no ridge point domain without the target center line; and outputting all extracted target center lines. The application can automatically and quickly extract an accurate center line through the above method. The application also discloses a blood vessel center line extraction system, equipment and a readable storage medium.

Description

Technical Field

[0001] This invention relates to the field of medical image processing, and specifically to a method, system, device, and readable storage medium for extracting the centerline of blood vessels. Background Technology

[0002] Currently, cardiovascular and cerebrovascular diseases pose a serious threat to human health, especially to the health of middle-aged and elderly people over 50 years of age. These diseases are characterized by high morbidity, high disability rate, and high mortality rate. Even with the most advanced and comprehensive treatment methods available today, more than 50% of stroke survivors are still unable to live independently. Worldwide, up to 15 million people die from cardiovascular and cerebrovascular diseases each year, making it the leading cause of death.

[0003] With the rapid development of medical imaging technologies such as CT angiography and magnetic resonance angiography (MRA), obtaining information on vascular ridges and centerlines using image post-processing techniques is of great significance for describing the morphology and structure of blood vessels and achieving vascular target reconstruction.

[0004] Currently, most centerline extraction schemes require interactive operations or prior knowledge of specific domains to determine the start and end points. Although centerline extraction schemes based on refinement algorithms can extract centerlines without interactive operations, they suffer from problems such as the extracted centerlines not being close enough to the center, difficulty in determining the endpoints and connection points of the centerlines during extraction, which is not conducive to subsequent centerline editing, the presence of many noisy branches requiring pruning, and computationally intensive processes. Summary of the Invention

[0005] The purpose of this invention is to automatically and quickly extract accurate and minimally disturbed centerlines from blood vessels using a centerline extraction method, saving time on centerline extraction and subsequent centerline editing, and improving efficiency. It addresses existing centerline extraction schemes based on thinning algorithms, which suffer from problems such as centerlines not being close enough to the center, difficulty in determining the endpoints and connection points of the centerline during extraction, hindering subsequent centerline editing, the presence of numerous noisy branches requiring pruning, and computationally intensive processes.

[0006] In a first aspect, embodiments of the present invention provide a method for extracting the centerline of a blood vessel, the method comprising:

[0007] Obtain the initial ridge point region of the binary image;

[0008] The initial start point and initial end point are determined based on the initial ridge point domain, and the target centerline is extracted based on the initial start point and initial end point through preset values ​​and the minimum cost path;

[0009] In the ridge point region where no target centerline has been extracted, the reference start point and reference end point are determined based on all extracted target centerlines, and the next target centerline is extracted based on the reference start point and reference end point through preset values ​​and minimum cost paths;

[0010] Repeat the process of extracting the next target centerline in the ridge point region where no target centerline has been extracted, until there are no more ridge point regions where no target centerline has been extracted;

[0011] Output all extracted target centerlines.

[0012] The above technical solution achieves fully automated centerline extraction without any interactive operations, saving extraction time and improving efficiency. By constraining the centerline's centrality through preset values, the extracted centerline is made closer to the center. During the extraction of the target centerline, reference start and end points are simultaneously acquired. These reference start and end points are clearly defined, corresponding to the endpoints and connection points of vascular branches, respectively, enabling accurate extraction of the target centerline. Furthermore, the reference start, end, and target centerline form a graph, facilitating interactive operations on the extracted centerline. This method extracts a smoother centerline after traversing the ridge region where the target centerline was not extracted, eliminating the need for pruning operations.

[0013] In some embodiments, the preset value is set by a function that calculates the minimum cost path.

[0014] In some embodiments, the function for calculating the minimum cost path is:

[0015]

[0016] Where d(v1, v2) is the Euclidean distance between ridge points v1 and v2; ξ(v1) and ξ(v2) are the weights that constrain the centrality of ridge points v1 and v2, respectively, and ε is a parameter to avoid division by zero.

[0017] In some embodiments, the initial starting point is the ridge point farthest from any ridge point in the initial ridge point domain.

[0018] In some embodiments, obtaining the initial ridge region of the binary image includes:

[0019] Calculate the Euclidean distance transform of the binary image to obtain the distance transform domain;

[0020] Based on the ridge points in the distance transformation domain, the initial ridge point domain is obtained. The ridge points are the maxima and neighborhoods of the distance transformation domain.

[0021] In some embodiments, after obtaining the initial ridge point domain based on the ridge points in the distance transform domain, the process includes:

[0022] Construct a graph where vertices are the pixels corresponding to ridges, edges are the lines connecting adjacent pixels, and edge weights are the Euclidean distances between vertices.

[0023] In some embodiments, determining an initial start point and an initial end point based on an initial ridge point domain, and extracting a target centerline based on the initial start point and initial end point using preset values ​​and a minimum cost path, includes:

[0024] Choose any vertex as the seed point, and use Dijkstra's algorithm to calculate the point farthest from the seed point as the initial starting point;

[0025] The point farthest from the initial starting point is calculated using Dijkstra's algorithm and used as the initial ending point.

[0026] In some embodiments, after calculating the point farthest from the initial starting point using Dijkstra's algorithm as the initial ending point, the method further includes:

[0027] Dijkstra's algorithm is used to calculate the minimum cost path between the initial starting point and the initial ending point as the target centerline by using a function that computes the minimum cost path.

[0028] In some embodiments, in the ridge point region where no target centerline has been extracted, a reference start point and a reference end point are determined based on all extracted target centerlines, and the next target centerline is extracted based on the reference start point and the reference end point using a preset value and a minimum cost path, including:

[0029] Mark the ridge point region where the target centerline was not extracted as the unextracted ridge point region;

[0030] Traverse and determine if the unextracted ridge region exists:

[0031] If it exists, determine the reference start point and reference end point based on all extracted target center lines, and extract the next target center line based on the reference start point and reference end point through preset values ​​and minimum cost paths;

[0032] If it does not exist, output all extracted target centerlines.

[0033] In some embodiments, the reference endpoint is: performing connected component processing on the unextracted ridge point domain, calculating the farthest geodesic distance from the connected component to all extracted target centerlines and the corresponding ridge points, and selecting the ridge point corresponding to the largest farthest geodesic distance as the reference endpoint.

[0034] In some embodiments, the reference starting point is: the shortest geodesic distance from the reference endpoint to all extracted target centerlines is calculated, and the first intersection of the path corresponding to the shortest geodesic distance and all extracted target centerlines is used as the reference starting point.

[0035] In some embodiments, before obtaining the initial ridge point domain of the binary image, the method further includes: inputting a binary image; determining whether the binary image is empty; if so, re-inputting the binary image; if not, obtaining the initial ridge point domain of the binary image.

[0036] In some embodiments, after outputting all extracted target centerlines, the method further includes: performing mean filtering on all extracted target centerlines.

[0037] Secondly, the present invention provides a system for extracting the centerline of a blood vessel, characterized in that it comprises:

[0038] The data receiving module is used to receive binary images input by the user.

[0039] The data processing module, connected to the data receiving module, is used for...

[0040] Obtain the initial ridge point region of the binary image;

[0041] The initial start point and initial end point are determined based on the initial ridge point domain, and the target centerline is extracted based on the initial start point and initial end point through preset values ​​and the minimum cost path;

[0042] In the ridge point region where no target centerline has been extracted, the reference start point and reference end point are determined based on all extracted target centerlines, and the next target centerline is extracted based on the reference start point and reference end point through preset values ​​and minimum cost paths;

[0043] Repeat the process of extracting the next target centerline in the ridge point region where no target centerline has been extracted, until there are no more ridge point regions where no target centerline has been extracted;

[0044] The data output module, connected to the data processing module, is used to output all extracted target centerlines.

[0045] By employing the above technical solution, through the coordinated action of the data receiving module, data processing module, and data output module, the blood vessel centerline extraction system automatically and quickly extracts accurate and smooth centerlines, saving extraction time and improving extraction efficiency. This addresses the problems of existing centerline extraction schemes based on thinning algorithms, such as the extracted centerlines not being close enough to the center, difficulty in determining the endpoints and connection points of the centerlines during extraction, facilitating subsequent centerline editing, the presence of numerous noisy branches requiring pruning, and computationally intensive processes.

[0046] Thirdly, the present invention provides an electronic device including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method for extracting the centerline of a blood vessel provided in any embodiment of the first aspect.

[0047] By adopting the above technical solution, electronic devices can automatically and quickly extract accurate and smooth center lines, improving efficiency and saving time.

[0048] Fourthly, the present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method for extracting the centerline of blood vessels provided in any embodiment of the first aspect. Attached Figure Description

[0049] Figure 1 The flowchart of the method for extracting the centerline of blood vessels in an embodiment of the present invention is shown. Figure 1 ;

[0050] Figure 2 The flowchart of the method for extracting the centerline of blood vessels in an embodiment of the present invention is shown. Figure 2 ;

[0051] Figure 3 The flowchart of the method for extracting the centerline of blood vessels in an embodiment of the present invention is shown. Figure 3 ;

[0052] Figure 4 The flowchart of the method for extracting the centerline of blood vessels in an embodiment of the present invention is shown. Figure 4 ;

[0053] Figure 5 The flowchart of the method for extracting the centerline of blood vessels in an embodiment of the present invention is shown. Figure 5 ;

[0054] Figure 6 A schematic diagram of a binary image domain is shown in an embodiment of the present invention;

[0055] Figure 7 A schematic diagram of the distance transformation domain in an embodiment of the present invention is shown;

[0056] Figure 8 A schematic diagram of the ridge point region in an embodiment of the present invention is shown;

[0057] Figure 9 A schematic diagram of the structure of the blood vessel centerline extraction system in an embodiment of the present invention is shown;

[0058] Figure 10 A schematic diagram of the electronic device for extracting the centerline of blood vessels in an embodiment of the present invention is shown. Detailed Implementation

[0059] The following specific embodiments illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. Although the description of the present invention is presented in conjunction with preferred embodiments, this does not mean that the features of the invention are limited to these embodiments. On the contrary, the purpose of describing the invention in conjunction with embodiments is to cover other options or modifications that may be derived based on the claims of the present invention. To provide a deep understanding of the invention, many specific details will be included in the following description. The invention may also be implemented without using these details. Furthermore, to avoid confusion or obscuring the focus of the invention, some specific details will be omitted in the description. It should be noted that, unless otherwise specified, the embodiments and features in the embodiments of the present invention can be combined with each other.

[0060] It should be noted that in this specification, similar reference numerals and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0061] The terms “first”, “second”, etc., are used only to distinguish descriptions and should not be interpreted as indicating or implying relative importance.

[0062] To make the objectives, technical solutions, and advantages of the present invention clearer, the embodiments of the present invention will be described in further detail below with reference to the accompanying drawings.

[0063] Firstly, reference Figure 1 An embodiment of the present invention provides a method for extracting the centerline of a blood vessel, comprising:

[0064] S1: Obtain the initial ridge point domain of the binary image.

[0065] Since ridge points only occupy a portion of the binary image, extracting the center line from the ridge points can not only reduce the computational load but also constrain the range of the center line, improving the accuracy and centrality of the extracted center line.

[0066] S2: Determine the initial start point and initial end point based on the initial ridge point domain, and extract the target centerline based on the initial start point and initial end point through preset values ​​and minimum cost paths.

[0067] By extracting the target centerline using preset values, the obtained target centerline is more accurate and closer to the center. This is especially true for larger and more tortuous blood vessels, where preset values ​​enable the extraction of a more accurate and central target centerline.

[0068] S3: In the ridge point region where the target centerline has not been extracted, determine the reference start point and reference end point based on all extracted target centerlines, and extract the next target centerline based on the reference start point and reference end point through preset values ​​and minimum cost paths.

[0069] Through step S3, the present invention achieves the simultaneous extraction of the next target centerline and the acquisition of the reference start point and reference end point. All extracted target centerlines, reference start points and reference end points constitute a graph, which facilitates interactive operations on the target centerlines.

[0070] S4: Repeat the process of extracting the next target centerline from the ridge point regions where no target centerline has been extracted, until there are no more ridge point regions where no target centerline has been extracted.

[0071] Execute step S5: Output all extracted target centerlines.

[0072] The above technical solution achieves fully automated centerline extraction without any interactive operations, saving extraction time and improving efficiency. By constraining the centerline's centrality through preset values, the extracted centerline is made closer to the center. During the extraction of the target centerline, reference start and end points are simultaneously acquired. These reference start and end points are clearly defined, corresponding to the endpoints and connection points of vascular branches, respectively, enabling accurate extraction of the target centerline. Furthermore, the reference start, end, and target centerline form a graph, facilitating interactive operations on the extracted centerline. This method extracts a smoother centerline after traversing the ridge region where the target centerline was not extracted, eliminating the need for pruning operations.

[0073] In some possible embodiments provided by the present invention, the preset value is set by a function that calculates the minimum cost path.

[0074] In this embodiment, the function for calculating the minimum cost path is:

[0075]

[0076] Where d(v1, v2) is the Euclidean distance between ridge points v1 and v2; ξ(v1) and ξ(v2) are the weights that constrain the centrality of ridge points v1 and v2, respectively. The closer ridge points v1 and v2 are to the center, the larger the weights ξ(v1) and ξ(v2) are, and the farther ridge points v1 and v2 are from the center, the smaller the weights ξ(v1) and ξ(v2) are; ε is a parameter to avoid division by zero.

[0077] Extracting the target centerline using a function with the minimum cost path results in a more accurate and closer-to-center centerline. This is especially beneficial for larger and more tortuous blood vessels, where the minimum cost path function can extract a more accurate and central centerline.

[0078] In some possible embodiments provided by the present invention, the initial starting point is the ridge point farthest from any ridge point in the initial ridge point domain.

[0079] In some possible embodiments provided by the present invention, reference is made to Figure 2 and combined Figure 1 S1: Obtain the initial ridge point domain of the binary image, including:

[0080] S11: Calculate the Euclidean distance transform of the binary image to obtain the distance transform domain;

[0081] S12: Obtain the initial ridge point domain based on the ridge points in the distance transformation domain. Here, the ridge points are the maxima and neighborhoods of the distance transformation domain.

[0082] For 3D tubular objects, the maxima of the distance transform domain are the centerlines. Blood vessels approximate 3D tubular objects, therefore the centerlines of blood vessels are located at the maxima of the distance transform domain and in the vicinity of them. Using the above technical solution, the centerlines of blood vessels are extracted based on the distance transform domain and the ridge point domain, resulting in extracted centerlines that are closer to the center and eliminating the need for pruning operations.

[0083] In some possible embodiments provided by the present invention, S12: After obtaining the initial ridge point domain based on the ridge points of the distance transform domain, the method for extracting the center line of the blood vessel further includes: constructing a graph, wherein the vertices of the graph are the pixels corresponding to the ridge points, the edges of the graph are the lines connecting adjacent pixels, and the weight of the edges is the Euclidean distance between the vertices.

[0084] In some possible embodiments provided by the present invention, reference is made to Figure 3 and combined Figure 1 S2: Determine the initial start point and initial end point based on the initial ridge point domain, and extract the target centerline based on the initial start point and initial end point using preset values ​​and the minimum cost path, including:

[0085] S21: Choose any vertex as the seed point, and use Dijkstra's algorithm to calculate the point farthest from the seed point as the initial starting point.

[0086] S22: Calculate the point farthest from the initial starting point using Dijkstra's algorithm and use it as the initial ending point.

[0087] In some possible embodiments provided by the present invention, reference continues to be made. Figure 3 S22: After calculating the point farthest from the initial starting point using Dijkstra's algorithm as the initial ending point, the method for extracting the centerline of blood vessels also includes:

[0088] S23: Use Dijkstra's algorithm to calculate the minimum cost path between the initial starting point and the initial ending point as the target centerline by using a function that calculates the minimum cost path.

[0089] By adopting the above technical solution, the initial starting point and initial ending point are clearly defined and accurately determined, and the minimum cost path is accurately extracted as the target centerline, making the extracted target centerline more accurate and closer to the center.

[0090] In some possible embodiments provided by the present invention, reference is made to Figure 4 and combined Figure 1 S3: In the ridge point region where the target centerline has not been extracted, determine the reference start point and reference end point based on all extracted target centerlines, and extract the next target centerline based on the reference start point and reference end point through preset values ​​and the minimum cost path, including:

[0091] S31: Mark the ridge point region where the target centerline was not extracted as the unextracted ridge point region.

[0092] S32: Traverse and determine if the unextracted ridge point region exists.

[0093] By traversing and determining whether the unextracted ridge point region exists in step S32, we can prevent the repeated extraction of the center line in the ridge point region where the center line has already been extracted. This not only saves extraction time and improves extraction efficiency, but also improves the accuracy of center line extraction.

[0094] If it exists, proceed to step S33: determine the reference start point and reference end point based on all extracted target center lines, and extract the next target center line based on the reference start point and reference end point through preset values ​​and minimum cost paths.

[0095] If it does not exist, proceed to step S5: output all extracted target centerlines.

[0096] Repeat steps S31 to S32, S33, or S5, i.e., execute S31: mark the ridge point regions where no target centerline has been extracted as unextracted ridge point regions; S32: traverse and determine whether unextracted ridge point regions exist. Continue until S32: traverse and determine whether unextracted ridge point regions exist. If the result is that no unextracted ridge point regions exist, execute S5: output all target centerlines.

[0097] In this embodiment, the extracted centerline is smoother after traversing the unextracted ridge point domain, eliminating the need for pruning. During the extraction of the next target centerline, a reference start point and a reference end point are simultaneously acquired. These reference start and end points are clearly defined, corresponding to the endpoints and connection points of branches, respectively, enabling accurate centerline extraction. Furthermore, the reference start point, reference end point, and target centerline form a graph, facilitating interactive operations on the extracted centerline.

[0098] In some other possible embodiments provided by the present invention, the reference endpoint is: to perform connected component processing on the unextracted ridge point domain, calculate the farthest geodesic distance from the connected component to all extracted target centerlines and the corresponding ridge point, and select the ridge point corresponding to the largest farthest geodesic distance as the reference endpoint.

[0099] In some other possible embodiments provided by the present invention, the reference starting point is: the shortest geodesic distance from the reference endpoint to all extracted target centerlines is calculated, and the first intersection of the path corresponding to the shortest geodesic distance and all extracted target centerlines is used as the reference starting point.

[0100] By adopting the above technical solution, the reference starting point and reference ending point are clearly defined and accurately determined in order to accurately extract the next target centerline.

[0101] Other possible embodiments provided by the present invention, refer to Figure 5 Before S1: obtaining the initial ridge region of the binary image, the method further includes:

[0102] S0: Input binary image;

[0103] S01: Determine if the binary image is empty. If it is, execute S0: Input the binary image, i.e., re-input the binary image; otherwise, execute S1: Obtain the initial ridge point domain of the binary image.

[0104] In this embodiment, S1: Obtain the initial ridge point domain of the binary image, which facilitates the subsequent extraction of only the center line of the ridge points, thereby reducing the amount of computation, constraining the range of the center line, and improving the accuracy and centrality of the center line.

[0105] In some other possible embodiments provided by the present invention, after S5: outputting all extracted target centerlines, the method for extracting the centerline of blood vessels further includes:

[0106] All extracted target centerlines are subjected to mean filtering to smooth them out.

[0107] refer to Figures 1 to 5 The following describes an exemplary implementation of a method for extracting the centerline of blood vessels.

[0108] refer to Figure 5 S0: Input binary image.

[0109] refer to Figure 6 Binary images are generally the result of blood vessel segmentation, with blood vessels as the foreground and other parts as the background.

[0110] S01: Determine if the binary image is empty:

[0111] If so, return to step S0: Input a binary image.

[0112] If the binary image is empty, the ridge point region cannot be obtained from the binary image to extract the blood vessel centerline. Therefore, it is necessary to re-enter the binary image until the input binary image is not empty.

[0113] If not, proceed to step S1: obtain the initial ridge region of the binary image.

[0114] refer to Figure 1 and Figure 2 Step S1: Obtain the initial ridge point region of the binary image, including:

[0115] S11: Calculate the Euclidean distance transform of the binary image to obtain the distance transform domain.

[0116] In this embodiment, the distance from each foreground pixel to the background is calculated using a distance transformation (also known as a distance function or skewing algorithm) to obtain the distance transformation domain. As mentioned earlier, for a 3D tubular object, the maximum value of the distance transformation domain is the centerline. Blood vessels approximate 3D tubular objects, therefore the centerline of the blood vessels lies in the region around the maximum value of the distance transformation domain. Figure 7 A schematic diagram of the distance transformation domain is shown in this embodiment.

[0117] S12: Obtain the ridge point domain based on the ridge points in the distance transform domain.

[0118] The maxima and neighborhoods of the distance transform domain are ridge points, and calculating the ridge points yields the ridge point domain. In this embodiment, extracting the centerline of the ridge points can reduce the computational load and improve computational efficiency; on the other hand, it improves the accuracy and centrality of extracting the vessel centerline. Figure 8 A schematic diagram of the ridge point region is shown in this embodiment.

[0119] The graph is constructed by using the pixels corresponding to the ridge points as vertices, the lines connecting adjacent pixels as edges, and the Euclidean distance between vertices as edge weights.

[0120] refer to Figure 1 S2: Determine the initial start point and initial end point based on the initial ridge point domain, and extract the target centerline based on the initial start point and initial end point through preset values ​​and minimum cost paths.

[0121] In this embodiment, reference Figure 3 Step S2 includes:

[0122] S21: Choose any vertex as the seed point, and use Dijkstra's algorithm to calculate the point farthest from the seed point as the initial starting point.

[0123] S22: Calculate the point farthest from the initial starting point using Dijkstra's algorithm and use it as the initial ending point.

[0124] S23: Use Dijkstra's algorithm to calculate the minimum cost path between the initial starting point and the initial ending point as the target centerline by using a function that calculates the minimum cost path.

[0125] In this embodiment, the function for calculating the minimum cost path, i.e., the edge weight formula, is as follows:

[0126]

[0127] Where d(v1, v2) is the Euclidean distance between ridge points v1 and v2; ξ(v1) and ξ(v2) are the weights that constrain the centrality of ridge points v1 and v2, respectively. The closer ridge points v1 and v2 are to the center, the larger the weights ξ(v1) and ξ(v2) are, and the farther ridge points v1 and v2 are from the center, the smaller the weights ξ(v1) and ξ(v2) are; ε is a parameter to avoid division by zero.

[0128] In this embodiment, the target centerline extracted based on the initial starting point and initial ending point using the minimum cost path function is more accurate and closer to the center.

[0129] refer to Figure 1 Execute S3: In the ridge point region where no target centerline has been extracted, determine the reference start point and reference end point based on all extracted target centerlines, and extract the next target centerline based on the reference start point and reference end point through preset values ​​and minimum cost paths.

[0130] Specifically, refer to Figure 4 Execute S31: Mark the ridge point region where the target centerline was not extracted as the ridge point region.

[0131] In this embodiment, after determining the initial starting point and initial ending point, and extracting the target centerline by calculating the minimum cost path, the ridge point regions of the extracted target centerline must be marked as extracted ridge point regions, and the ridge point regions of the unextracted target centerline must be marked as unextracted ridge point regions.

[0132] For example, the input binary image includes a first blood vessel, a second blood vessel, a third blood vessel, and a fourth blood vessel. After extracting the centerline (i.e., the target centerline) of the first blood vessel using the initial starting point and initial ending point determined in the above steps and by calculating the minimum cost path function, the first blood vessel must be marked as the ridge point region where the blood vessel centerline has been extracted, i.e., the extracted ridge point region; the ridge point regions where the second, third, and fourth blood vessels are located are the ridge point regions where the blood vessel centerlines have not been extracted, i.e., the unextracted ridge point regions.

[0133] refer to Figure 4 S32: Traverse and determine if the unextracted ridge region exists:

[0134] If it does not exist, proceed to step S5: output all extracted target centerlines.

[0135] For example, the input binary image includes only the first blood vessel, and the centerline of the first blood vessel is determined through the steps described above. The first blood vessel is marked as an extracted ridge point region. In this embodiment, the binary image does not contain any unextracted ridge point regions, and the centerline of the first blood vessel is output.

[0136] If it exists, execute S33: determine the reference start point and reference end point based on all extracted target center lines, and extract the next target center line based on the reference start point and reference end point through the preset value and minimum cost path.

[0137] In this embodiment, the unextracted ridge point domain is processed as a connected component, and the farthest geodesic distance from the connected component to all extracted target centerlines and the corresponding ridge point are calculated. The ridge point corresponding to the largest farthest geodesic distance is selected as the reference endpoint.

[0138] Calculate the shortest geodesic distance from the reference endpoint to all extracted target centerlines. The first intersection of the path corresponding to the shortest geodesic distance and all extracted target centerlines is used as the reference starting point.

[0139] For example, as described above, the centerline of the second, third, or fourth blood vessel is extracted based on the centerline of the first blood vessel.

[0140] For the unextracted ridge point regions, connected component processing is performed. The connected component for the second vessel is ccl2, for the third vessel it is ccl3, and for the fourth vessel it is ccl4. The furthest geodesic distances d2, d3, and d4 from each connected component to the centerline of the first vessel are calculated, and the corresponding ridge points are v2, v3, and v4. Based on the calculation results, d2 > d3 > d4. Therefore, the ridge point v2 of the second vessel is used as the reference endpoint (also called the endpoint or source point) of the centerline of the second vessel (i.e., the next target centerline).

[0141] The shortest geodesic distance from the reference endpoint v2 of the second blood vessel to the centerline of the first blood vessel is calculated, and the intersection point v of the path corresponding to the shortest geodesic distance and the centerline of the first blood vessel is determined. 21 This serves as the reference starting point (also known as the connection point of the center lines) for the second vascular center line (i.e., the next target center line).

[0142] Based on the reference endpoint v2 and reference starting point v of the second blood vessel 21 The centerline of the second blood vessel is extracted by calculating the minimum cost path function and the minimum cost path.

[0143] Using the above technical solution, the reference start point and reference end point of the second blood vessel centerline are obtained simultaneously when extracting the centerline of the second blood vessel. The centerline of the second blood vessel, its reference start point and reference end point constitute a graph, which facilitates interactive operations on the centerline of the second blood vessel.

[0144] Next, execute S4: Repeat the process of extracting the next target centerline from the ridge point regions where no target centerline has been extracted, until there are no more ridge point regions where no target centerline has been extracted.

[0145] That is, repeat S31 to S33 until S32: traverse and determine whether the unextracted ridge point domain exists, and output the result that there is no unextracted ridge point domain.

[0146] Execute S5: Output all extracted target centerlines.

[0147] For example, as described above, in an embodiment where the input binary image includes a first blood vessel, a second blood vessel, a third blood vessel, and a fourth blood vessel, the ridge point regions containing the center lines of the first and second blood vessels are marked as extracted ridge point regions; the ridge point regions containing the third and fourth blood vessels are marked as unextracted ridge point regions. By traversing and determining that the unextracted ridge point regions are not empty, the reference endpoint and reference start point of the blood vessel center line of the third or fourth blood vessel are determined based on the center lines of the first and second blood vessels to extract the blood vessel center line of the third or fourth blood vessel.

[0148] After calculation, since d3 > d4, the ridge point v3 of the third blood vessel is used as the reference endpoint (also called the endpoint or source point) of the third blood vessel's centerline (i.e., the next target centerline). Based on the reference endpoint v3 of the third blood vessel, the shortest geodesic distances to the centerlines of the first and second blood vessels are calculated. The paths corresponding to the shortest geodesic distances intersect the centerlines of the first and second blood vessels respectively. The path corresponding to the shortest geodesic distance intersects the centerline of the first blood vessel first. Therefore, the intersection point v3 of the path corresponding to the shortest geodesic distance and the centerline of the first blood vessel is... 31 The first intersection point among all intersection points, intersection point v 31 This serves as the reference starting point (also known as the connection point of the center lines) for the third vascular center line (i.e., the next target center line).

[0149] Based on the reference endpoint v3 and reference starting point v of the third blood vessel 31 The centerline of the third blood vessel is extracted by calculating the minimum cost path function and the minimum cost path.

[0150] Mark the ridge point regions containing the centerlines of the first, second, and third blood vessels as extracted ridge point regions; mark the ridge point region containing the fourth blood vessel as unextracted ridge point regions. Traverse and determine if the unextracted ridge point regions are empty. Based on the centerlines of the first, second, and third blood vessels, determine the reference endpoint v4 and reference starting point v4 of the fourth blood vessel according to the steps described above. 41 And based on the reference endpoint v4 and the reference starting point v 41 Determine the center line of the fourth vessel.

[0151] Next, the ridge point regions containing the centerlines of the first, second, third, and fourth blood vessels are marked as extracted ridge point regions. If any unextracted ridge point regions are found to be empty, then S5 is executed: output all extracted target centerlines. That is, output the centerlines of the first, second, third, and fourth blood vessels.

[0152] Then, mean filtering is applied to all extracted target centerlines.

[0153] In this embodiment, by way of example, the center lines of the first blood vessel, the second blood vessel, the third blood vessel, and the fourth blood vessel are subjected to mean filtering to form smooth center lines.

[0154] In this embodiment, centerline extraction is fully automated without any interactive operations, saving extraction time. This embodiment constrains the centrality of the centerline by calculating the minimum cost path function, making the extracted target centerline closer to the center. The extracted centerline is smoother after traversing the unextracted ridge domain, eliminating the need for pruning. During the extraction of the next target centerline, a reference start point and a reference end point are simultaneously acquired. These reference start and end points are clearly defined, corresponding to the endpoints and connection points of branches, respectively, achieving accurate centerline extraction. Furthermore, the reference start and end points, along with all extracted target centerlines, form a graph, facilitating interactive operations on the extracted centerlines.

[0155] Secondly, refer to Figure 9 The present invention provides a blood vessel centerline extraction system 1, comprising:

[0156] The data receiving module 11 is used to receive the binary image input by the user.

[0157] Data processing module 12, connected to data receiving module 11, is used for...

[0158] Obtain the initial ridge point region of the binary image;

[0159] The initial start point and initial end point are determined based on the initial ridge point domain, and the target centerline is extracted based on the initial start point and initial end point through preset values ​​and the minimum cost path;

[0160] In the ridge point region where no target centerline has been extracted, the reference start point and reference end point are determined based on all extracted target centerlines, and the next target centerline is extracted based on the reference start point and reference end point through preset values ​​and minimum cost paths;

[0161] Repeat the process of extracting the next target centerline from the ridge point regions where no target centerline has been extracted, until there are no more ridge point regions where no target centerline has been extracted.

[0162] The data output module 13 is connected to the data processing module 12 and is used to output all extracted target centerlines.

[0163] By employing the above technical solution, under the synergistic action of the data receiving module 11, the data processing module 12, and the data output module 13, the blood vessel centerline extraction system 1 automatically and quickly extracts accurate and smooth centerlines, saving extraction time and improving extraction efficiency. This solves the problems of existing centerline extraction schemes based on thinning algorithms, such as difficulty in determining the endpoints and connection points of the centerline, hindering subsequent centerline editing, the extracted centerline not being close enough to the center, the presence of numerous noisy branches requiring pruning, and time-consuming calculations.

[0164] Thirdly, refer to Figure 10 This invention provides an electronic device 2, including a memory 21, a processor 22, and a computer program stored in the memory 21 and executable on the processor 22. When the processor 22 executes the computer program, it implements the method for extracting the centerline of blood vessels in any of the foregoing embodiments. The memory 21 may include, for example, a system memory, a fixed non-volatile storage medium, etc. The system memory may store, for example, an operating system, application programs, a bootloader, and other programs.

[0165] In this embodiment, the electronic device 2 can automatically and accurately locate the center line of the blood vessel, reduce noise interference of the center line of the blood vessel, eliminate the need for pruning, improve efficiency, and save time.

[0166] Fourthly, embodiments of the present invention also provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the aforementioned methods for extracting the centerline of a blood vessel.

[0167] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable non-transitory storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0168] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0169] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0170] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0171] While the present invention has been illustrated and described with reference to certain preferred embodiments, those skilled in the art should understand that the above description is a further detailed explanation of the invention in conjunction with specific embodiments, and should not be construed as limiting the specific implementation of the invention to these descriptions. Various changes in form and detail can be made by those skilled in the art, including several simple deductions or substitutions, without departing from the spirit and scope of the invention.

Claims

1. A method for extracting the centerline of a blood vessel, characterized in that, The method includes: Obtain the initial ridge point region of the binary image; The initial start point and initial end point are determined based on the initial ridge point domain, and the target centerline is extracted based on the initial start point and the initial end point through a preset value and a minimum cost path. The preset value is set by a function that calculates the minimum cost path. In the ridge point region where the target centerline has not been extracted, a reference start point and a reference end point are determined based on all extracted target centerlines. The next target centerline is then extracted based on the reference start point and the reference end point through the preset value and the minimum cost path. The reference end point is determined by performing connected component processing on the unextracted ridge point region, calculating the farthest geodesic distance from the connected component to all extracted target centerlines and the corresponding ridge point, and selecting the ridge point corresponding to the largest farthest geodesic distance as the reference end point. The reference start point is determined by calculating the shortest geodesic distance from the reference end point to all extracted target centerlines, and the first intersection point of the path corresponding to the shortest geodesic distance and all extracted target centerlines as the reference start point. Repeat the process of extracting the next target centerline in the ridge point region where the target centerline has not been extracted, until there are no more ridge point regions where the target centerline has not been extracted. Output all extracted target centerlines.

2. The method for extracting the centerline of a blood vessel as described in claim 1, characterized in that, The function for calculating the minimum cost path is: in, It is a ridge point The Euclidean distance between them; and These are the constrained ridge points. The centrality weight, This is to avoid dividing by zero.

3. The method for extracting the centerline of a blood vessel as described in claim 1, characterized in that, The initial starting point is the ridge point that is farthest from any ridge point in the initial ridge point domain.

4. The method for extracting the centerline of a blood vessel as described in claim 1, characterized in that, The process of obtaining the initial ridge point region of the binary image includes: Calculate the Euclidean distance transform of the binary image to obtain the distance transform domain; The initial ridge point domain is obtained based on the ridge points of the distance transform domain, where the ridge points are the maxima and neighborhoods of the distance transform domain.

5. The method for extracting the centerline of a blood vessel as described in claim 4, characterized in that, After obtaining the initial ridge point domain based on the ridge points in the distance transform domain, the process includes: Construct a graph where the vertices are the pixels corresponding to the ridge points, the edges are the lines connecting adjacent pixels, and the weight of each edge is the Euclidean distance between the vertices.

6. The method for extracting the centerline of a blood vessel as described in claim 5, characterized in that, The step of determining the initial start point and initial end point based on the initial ridge point domain, and extracting the target centerline based on the initial start point and initial end point through preset values ​​and the minimum cost path, includes: Choose any vertex as the seed point, and use Dijkstra's algorithm to calculate the point farthest from the seed point as the initial starting point; The point farthest from the initial starting point is calculated using Dijkstra's algorithm and used as the initial ending point.

7. The method for extracting the centerline of a blood vessel as described in claim 6, characterized in that, After calculating the point farthest from the initial starting point using Dijkstra's algorithm as the initial ending point, the method further includes: The Dijkstra algorithm is used to calculate the minimum cost path between the initial starting point and the initial ending point using the function that calculates the minimum cost path, which serves as the target centerline.

8. The method for extracting the centerline of a blood vessel as described in claim 1, characterized in that, In the ridge point region where the target centerline has not been extracted, a reference start point and a reference end point are determined based on all extracted target centerlines, and the next target centerline is extracted based on the reference start point and reference end point using a preset value and a minimum cost path, including: The ridge point region where the target centerline was not extracted is marked as the unextracted ridge point region; Traverse and determine if the unextracted ridge region exists: If it exists, determine the reference start point and reference end point based on all the extracted target center lines, and extract the next target center line based on the reference start point and reference end point through a preset value and the minimum cost path; If it does not exist, output all extracted target centerlines.

9. The method for extracting the centerline of a blood vessel as described in claim 1, characterized in that, Before obtaining the initial ridge point domain of the binary image, the method further includes: Input the binary image; Determine whether the binary image is empty. If it is, re-input the binary image; otherwise, obtain the initial ridge point domain of the binary image.

10. The method for extracting the centerline of a blood vessel as described in claim 1, characterized in that, After outputting all the extracted target centerlines, the method further includes performing mean filtering on all the extracted target centerlines.

11. A system for extracting the centerline of a blood vessel, characterized in that, include: The data receiving module is used to receive binary images input by the user. The data processing module, connected to the data receiving module, is used for... Obtain the initial ridge point region of the binary image; The initial start point and initial end point are determined based on the initial ridge point domain, and the target centerline is extracted based on the initial start point and the initial end point through a preset value and a minimum cost path. The preset value is set by a function that calculates the minimum cost path. In the ridge point region where the target centerline has not been extracted, a reference start point and a reference end point are determined based on all extracted target centerlines. The next target centerline is then extracted based on the reference start point and the reference end point through the preset value and the minimum cost path. The reference end point is determined by performing connected component processing on the unextracted ridge point region, calculating the farthest geodesic distance from the connected component to all extracted target centerlines and the corresponding ridge point, and selecting the ridge point corresponding to the largest farthest geodesic distance as the reference end point. The reference start point is determined by calculating the shortest geodesic distance from the reference end point to all extracted target centerlines, and the first intersection point of the path corresponding to the shortest geodesic distance and all extracted target centerlines as the reference start point. Repeat the process of extracting the next target centerline in the ridge point region where the target centerline has not been extracted, until there are no more ridge point regions where the target centerline has not been extracted. The data output module, connected to the data processing module, is used to output all extracted target centerlines.

12. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method for extracting the centerline of a blood vessel as described in any one of claims 1 to 10.

13. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the method for extracting the centerline of a blood vessel as described in any one of claims 1 to 10.

Images