Method, system, device and storage medium for extracting centerline of cerebral artery

Abstract

The present invention discloses a method for extracting the centerline of a cerebral artery, comprising: obtaining a target mask image of the cerebral artery; performing refinement processing on the target mask image to preliminarily extract the centerline of the cerebral artery to obtain a refined image; determining the starting point of the internal carotid artery based on the refined image, and constructing a connectivity graph based on the refined image; calculating all minimum cost paths as target centerlines based on the starting point of the internal carotid artery and the connectivity graph; and outputting all target centerlines. The present invention automatically and rapidly extracts accurate, near-center centerlines through the above method, while increasing the robustness of the centerline extraction algorithm. The present invention also discloses a system, device, and storage medium for extracting the centerline of a cerebral artery.

Description

Technical Field

[0001] The present invention relates to the field of medical image processing, and in particular to a method, system, device and storage medium for extracting the centerline of a cerebral artery. Background Art

[0002] With the development and progress of medical imaging technology, there are more and more types of medical images. Common medical images include X-ray (X-ray, which is an electromagnetic wave with extremely high frequency, extremely short wavelength and high energy), DSA (short for Digital subtraction angiography), CT (short for Computed Tomography), MRI (short for Magnetic Resonance Imaging), etc.

[0003] Typically, doctors rely solely on clinical experience and visual interpretation of medical images to provide an objective diagnosis and treatment plan. With the advancement of computer technology, the use of computers for efficient and accurate analysis of medical images is becoming increasingly mainstream. To diagnose and develop treatment plans for cerebral arterial vascular lesions, clinicians need to provide quantitative analysis results and plan treatment plans based on these results. These results can include determining whether a cerebral arterial tumor exists, its specific location, the extent of the lesion, and the risk of rupture. This quantitative analysis is premised on the extraction of precise vascular centerlines. The extracted cerebral arterial centerlines can be used to diagnose vascular diseases, calculate lesion location and extent, and provide the trajectory and position of interventional devices for surgical navigation planning, thereby planning the surgical path.

[0004] Common blood vessel centerline extraction methods include manual calibration, topological refinement, and distance transformation. The manual calibration method requires the operator to specify the center points based on experience and connect these points into a centerline. It can accurately extract the centerline of a two-dimensional image, but it is very difficult to operate on a three-dimensional image. Therefore, the manual calibration method has the problems of low efficiency and dependence on the doctor's experience. The topological refinement method is a method of deleting boundary points from the outside to the inside to obtain the centerline, which is often achieved through template iteration. The topological refinement method can maintain the topological structure of the blood vessels, but the solution is complex and computationally intensive. The distance transformation method converts a binary image into a grayscale image. The grayscale value is the distance between the pixel point and the image boundary point. This method has high accuracy, but there may be discontinuities in the center point.

[0005] The existing methods for extracting the centerline of cerebral arteries have the problems of relying on manual experience, taking a long time to calculate, having low computational efficiency, and the extracted centerline being discontinuous and not close enough to the center.

[0006] The purpose of the present invention is to solve the problems of existing cerebral artery centerline extraction methods, such as reliance on manual experience, long calculation time, low computational efficiency, discontinuous extracted centerlines, and insufficient proximity to the center.

[0007] In a first aspect, the cerebral artery centerline extraction method provided by the present invention automatically and rapidly extracts an accurate centerline close to the center, while increasing the robustness of the centerline extraction algorithm.

[0008] To solve the above technical problems, an embodiment of the present invention discloses a method for extracting the centerline of a cerebral artery, comprising: obtaining a target mask image of the cerebral artery; performing refinement processing on the target mask image to preliminarily extract the centerline of the cerebral artery to obtain a refined image; determining the starting point of the internal carotid artery based on the refined image, and constructing a connectivity graph based on the refined image; calculating all minimum cost paths as target centerlines based on the starting point of the internal carotid artery and the connectivity graph; and outputting all target centerlines.

[0009] The above technical solution realizes the fully automatic extraction of the centerline of the cerebral artery, reduces the amount of calculation and improves the calculation speed. The starting point of the internal carotid artery is determined according to the refined image, and a connectivity map is constructed based on the refined image. This not only facilitates the subsequent extraction of the centerline path and improves the calculation speed, but also ensures that the centerline of the extracted cerebral artery is closer to the center.

[0010] According to another specific embodiment of the present invention, determining the starting point of the internal carotid artery based on the refined image and constructing a connectivity graph based on the refined image include: determining all end points through 26-neighborhood analysis based on the refined image; and determining the starting point of the internal carotid artery based on all the end points, where the starting point of the internal carotid artery is the end point farthest from the center of all the end points.

[0011] According to another specific embodiment of the present invention, the starting point of the internal carotid artery is determined based on all the end points, and the formula is:

[0012]

[0013] Among them, Max{} means taking the maximum value, || means Euclidean distance, Ei means the i-th end point, i is a positive integer, Represents the mean of all endpoints.

[0014] According to another specific embodiment of the present invention, the starting point of the internal carotid artery is determined based on the refined image, and a connectivity graph is constructed based on the refined image, including: constructing a weighted connectivity graph based on the refined image using a distance transformation algorithm; wherein the farther away from the edge of the refined image, the greater the weight of the connectivity graph and the smaller the weight value; the closer to the edge of the refined image, the smaller the weight of the connectivity graph and the larger the weight value.

[0015] According to another specific embodiment of the present invention, the weight calculation formula of the connectivity graph is as follows:

[0016] W i =-W i

[0017] Among them, W represents the weight, the actual value is the distance transformation grayscale image value DF(i), i is the position of all mask pixel values ​​in the refined image, W i Represents the weight value of the i-th pixel.

[0018] According to another specific embodiment of the present invention, all minimum cost paths are calculated as target center lines based on the starting point and connectivity graph of the internal carotid artery, including: using the Dijkstra algorithm to calculate the minimum cost path between the starting point of the internal carotid artery and each end point as the target center line; until the calculation of the minimum cost path between the starting point of the internal carotid artery and all end points is completed, all target center lines are obtained.

[0019] According to another specific embodiment of the present invention, before obtaining the target mask image of the cerebral artery, the method further includes: obtaining an initial mask image of the three-dimensional image of the cerebral artery; extracting a region of interest in the initial mask image; and filling holes in the cerebral blood vessel image in the region of interest to form a target mask image of the cerebral artery.

[0020] According to another specific embodiment of the present invention, obtaining an initial mask image of a three-dimensional image of a cerebral artery includes: inputting a three-dimensional image of the cerebral artery; obtaining an initial mask image of the cerebral aneurysm and surrounding blood vessels through a segmentation algorithm based on the three-dimensional image of the cerebral artery; wherein the starting position of the initial mask image is the starting point of the internal carotid artery proximal to the parent artery, and the ending position is a preset multiple of the diameter of the parent artery distal to the parent artery that is not shorter than the parent artery at the location of the aneurysm, the diameter of the cerebral artery branch in the initial mask image is greater than a first preset value, and the length of the cerebral artery branch is not less than a second preset value.

[0021] According to another specific embodiment of the present invention, the tumor-bearing artery is located at the origin of the internal carotid artery.

[0022] According to another specific embodiment of the present invention, extracting the region of interest in the initial mask image includes: extracting the starting point and the ending point of the region of interest; wherein the starting point of the region of interest is the first point in the initial mask image in each three-dimensional direction that is a mask pixel value, and the ending point is the first point in the three-dimensional direction that is a non-mask pixel value.

[0023] According to another specific embodiment of the present invention, the target mask image is refined and the centerline of the cerebral artery is preliminarily extracted to obtain a refined image, including: using a distance transform algorithm to refine the target mask image to obtain a distance transform grayscale image; performing a point deletion operation based on the distance transform grayscale image through 26-neighborhood analysis, including: determining whether each pixel point is connected to the 26-neighborhood: if so, retaining the pixel point; if not, deleting the pixel point.

[0024] In a second aspect, an embodiment of the present invention discloses a system for extracting the centerline of a cerebral artery, comprising: a data receiving module for receiving a three-dimensional image of a cerebral artery input by a user; a data processing module connected to the data receiving module for acquiring a target mask image of the cerebral artery; performing refinement processing on the target mask image to preliminarily extract the centerline of the cerebral artery to obtain a refined image; determining the starting point of the internal carotid artery based on the refined image, and constructing a connectivity graph based on the refined image; calculating all minimum cost paths as target centerlines based on the starting point of the internal carotid artery and the connectivity graph; and a data output module connected to the data processing module for outputting all target centerlines.

[0025] By adopting the above technical solution, under the coordinated action of the data receiving module, the data processing module and the data output module, the centerline of the cerebral artery is extracted fully automatically through the centerline extraction system of the cerebral artery, which reduces the amount of calculation, improves the calculation speed, and the extracted centerline of the cerebral artery is closer to the center.

[0026] In a third aspect, an embodiment of the present invention discloses an electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the method for extracting the centerline of the cerebral artery in any embodiment of the first aspect is implemented.

[0027] By adopting the above technical solution, the electronic device automatically and quickly extracts the centerline of the cerebral artery that is accurate and close to the center, thereby improving efficiency and saving time.

[0028] In a fourth aspect, an embodiment of the present invention discloses a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the method for extracting the centerline of the cerebral artery in any embodiment of the first aspect. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 The process of the method for extracting the centerline of the cerebral artery in the embodiment of the present invention is shown as follows: Figure 1 ;

[0030] Figure 2The process of the method for extracting the centerline of the cerebral artery in the embodiment of the present invention is shown as follows: Figure 2 ;

[0031] Figure 3 The process of the method for extracting the centerline of the cerebral artery in the embodiment of the present invention is shown as follows: Figure 3 ;

[0032] Figure 4 The process of the method for extracting the centerline of the cerebral artery in the embodiment of the present invention is shown as follows: Figure 4 ;

[0033] Figure 5 The process of the method for extracting the centerline of the cerebral artery in the embodiment of the present invention is shown as follows: Figure 5 ;

[0034] Figure 6 The process of the method for extracting the centerline of the cerebral artery in the embodiment of the present invention is shown as follows: Figure 6 ;

[0035] Figure 7 The process of the method for extracting the centerline of the cerebral artery in the embodiment of the present invention is shown as follows: Figure 7 ;

[0036] Figure 8 A schematic diagram showing a region of interest of a cerebral artery extracted in an embodiment of the present invention;

[0037] Figure 9 A schematic diagram showing a hole in a mask image of a cerebral artery according to an embodiment of the present invention;

[0038] Figure 10 A schematic structural diagram of a cerebral artery centerline extraction system according to an embodiment of the present invention is shown;

[0039] Figure 11 A schematic structural diagram of an electronic device for extracting the centerline of a cerebral artery in an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0040] The following is an explanation of the embodiments of the present invention by specific specific examples. Those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. Although the description of the present invention will be introduced in conjunction with the preferred embodiment, this does not mean that the features of this invention are limited to this embodiment. On the contrary, the purpose of introducing the invention in conjunction with the embodiment is to cover other options or modifications that may be extended based on the claims of the present invention. In order to provide a deep understanding of the present invention, the following description will include many specific details. The present invention can also be implemented without using these details. In addition, in order to avoid confusion or blurring the focus of the present invention, some specific details will be omitted in the description. It should be noted that the embodiments of the present invention and the features in the embodiments can be combined with each other without conflict.

[0041] It should be noted that in this specification, similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, it does not need to be further defined or explained in subsequent drawings.

[0042] The terms “first”, “second”, etc. are only used for distinguishing descriptions and should not be understood as indicating or implying relative importance.

[0043] In the description of this embodiment, it should be noted that, unless otherwise specified or limited, the terms "disposed," "connected," and "connected" should be understood broadly. For example, they can refer to fixed connections, detachable connections, or integral connections; mechanical connections or electrical connections; direct connections or indirect connections through an intermediate medium; and internal connections between two components. Those skilled in the art will understand the specific meanings of the above terms in this embodiment based on specific circumstances.

[0044] To make the objectives, technical solutions and advantages of the present invention more clear, the embodiments of the present invention will be described in further detail below with reference to the accompanying drawings.

[0045] First, reference Figure 1 In an embodiment of the present invention, a method for extracting the centerline of a cerebral artery is provided, the method comprising:

[0046] S1: Acquire a target mask image of the cerebral arteries.

[0047] Since the target mask image of the cerebral artery only occupies a portion of the initial mask image of the cerebral artery, namely 1 / 4 to 1 / 3, the amount of calculation in the subsequent processing flow is greatly reduced, and the efficiency of extracting the centerline of the cerebral artery is improved; the target mask image removes noise, making the extracted centerline of the cerebral artery more accurate.

[0048] S2: Performing refinement processing on the target mask image and preliminarily extracting the center line of the cerebral artery to obtain a refined image.

[0049] The thinning process ensures that the center lines of the initially extracted cerebral arteries are continuous center lines and ensures the connectivity of the thinned images.

[0050] S3: Determine the starting point of the internal carotid artery according to the refined image, and construct a connectivity map based on the refined image.

[0051] The refined image allows for fully automatic determination of the internal carotid artery's starting point, eliminating the need for any interaction. Furthermore, compared to existing techniques that require human interaction and traversal of the refined image's endpoints before determining the internal carotid artery's starting point, fully automatic determination of the internal carotid artery's starting point based on the refined image reduces computational effort and time.

[0052] A connectivity graph is constructed based on the refined image to facilitate the subsequent extraction of the centerline path of the cerebral artery and to ensure the centrality of the centerline path of the cerebral artery, that is, to make the centerline of the extracted cerebral artery more centered.

[0053] S4: Calculate all minimum cost paths as the target centerline based on the starting point and connectivity graph of the internal carotid artery.

[0054] The minimum cost path calculation is based on the connected thinned image and the weighted connected graph based on the distance transform algorithm, which ensures the centrality of the centerline of the extracted cerebral artery while maintaining the connectivity.

[0055] S5: Output all target center lines.

[0056] Using the above technical solution, in step S2, the target mask image of the cerebral artery is directly refined, which reduces the amount of calculation, improves the calculation speed, and increases the robustness of the algorithm. After the target mask image is refined, the centerline of the cerebral artery is extracted as a continuous centerline. Step S3: Determine the starting point of the internal carotid artery based on the refined image, and construct a connectivity graph based on the refined image. This not only facilitates the subsequent extraction of the centerline path and improves the calculation speed, but also ensures that the centerline of the extracted cerebral artery is closer to the center. In step S4, the target centerline is extracted using the minimum cost path calculation method in combination with the weighted refinement graph, which increases the accuracy of the centerline while improving the calculation speed.

[0057] The method for extracting the centerline of a cerebral artery provided by the present invention is used to optimize the extraction of the centerline of a cerebral artery in cerebral aneurysm planning software. For cerebral aneurysm planning software, the specific location of the aneurysm can be determined based on the centerline of the vascular blood vessel, the degree of aneurysm lesion can be calculated, and the risk of rupture can be estimated. At the same time, the interventional surgery path can also be planned.

[0058] In some possible embodiments provided by the present invention, reference is made to Figure 2 Combined with Figure 1 S1: Before acquiring the target mask image of the cerebral artery, the method further includes:

[0059] S01: Acquire an initial mask image of a 3D image of cerebral arteries.

[0060] The initial mask image is an image of the cerebral arteries with redundant bifurcated blood vessels removed, which is convenient for reducing the amount of subsequent calculations and improving calculation efficiency.

[0061] S02: Extracting the region of interest in the initial mask image.

[0062] Typically, the region of interest of a cerebral arterial aneurysm only occupies 1 / 4 to 1 / 3 of the initial mask image. After extracting the region of interest, the computational complexity of subsequent processing steps can be greatly reduced, thereby improving computational efficiency.

[0063] S03: Filling holes in the cerebral blood vessel image in the region of interest to form a target mask image of the cerebral artery.

[0064] In the region of interest of the extracted initial mask image, holes appear in the partially filled part of the cerebral artery, that is, holes appear in the cerebral artery in the region of interest (such as Figure 9 The presence of holes (shown as a hole at B) prevents accurate extraction of the cerebral artery centerline. Specifically, according to the definition of a centerline, the extracted centerline typically maintains the original image structure. Therefore, in the region of interest, to preserve the hole structure, the centerline near the hole is extracted as a circle surrounding the hole. This circle does not correspond to the original topological structure of the cerebral artery centerline, resulting in an incorrectly extracted centerline. In step S03, the holes in the cerebral vascular image in the region of interest are filled to prepare for subsequent accurate cerebral artery centerline extraction.

[0065] Exemplarily, the hole filling algorithm is a flood filling method. In some other possible embodiments, the present invention does not limit the hole filling algorithm.

[0066] In some possible embodiments provided by the present invention, reference is made to Figure 3 Combined with Figure 2 S01: Acquire an initial mask image of a 3D image of cerebral arteries, including:

[0067] S011: Input a three-dimensional image of cerebral arteries.

[0068] S012: Obtain an initial mask image of the cerebral aneurysm blood vessels and surrounding blood vessels using a segmentation algorithm based on the three-dimensional image of the cerebral artery blood vessels.

[0069] In step S012, redundant bifurcated blood vessels are removed from the three-dimensional image of the cerebral arteries by a segmentation algorithm, and an initial mask image is obtained, thereby reducing the amount of subsequent calculations and improving calculation efficiency.

[0070] In this embodiment, the starting position of the initial mask image is the starting point of the internal carotid artery proximal to the parent artery, and the ending position is a preset multiple of the diameter of the parent artery at the distal end of the parent artery that is not shorter than the tumor position. The diameter of the cerebral artery branch in the initial mask image is greater than the first preset value, and the length of the cerebral artery branch is not less than the second preset value. Exemplarily, the preset multiple is 3 times, or the preset multiple is 5 times. Exemplarily, the first preset value can range from 1mm to 3mm. If the first preset value is 1mm, the diameter of the cerebral artery branch in the initial mask image is greater than 1mm. Exemplarily, the second preset value can range from 3mm to 5mm. If the second preset value is 3mm, the length of the cerebral artery branch in the initial mask image is not less than 3mm.

[0071] For example, the segmentation algorithm adopts AI technology; for another example, the segmentation algorithm adopts traditional image processing technology. The present invention does not limit the adopted segmentation algorithm.

[0072] In some possible embodiments provided by the present invention, the parent artery is located at the origin of the internal carotid artery.

[0073] In some possible embodiments provided by the present invention, reference is made to Figure 4 Combined with Figure 2 , S02: Extract the region of interest in the initial mask image, including:

[0074] S022: Extract the starting point and the ending point of the region of interest.

[0075] The starting point of the ROI is the first masked pixel in each of the three-dimensional directions of the initial mask image, i.e., the point in each of the X, Y, and Z directions. The ending point of the ROI is the first non-masked pixel in each of the three-dimensional directions of the initial mask image. By extracting the starting and ending points of the ROI, the scope of the ROI can be accurately determined, reducing subsequent computational effort and improving efficiency.

[0076] In some possible embodiments provided by the present invention, reference is made to Figure 4 Combined with Figure 2 , S022: Before extracting the starting point and the ending point of the region of interest, the method further includes:

[0077] S021: Perform maximum connected component processing on the initial mask image.

[0078] Step S021 eliminates the noise caused by the initial mask image processing algorithm, solving the problem of inaccurate extracted regions of interest.

[0079] In some possible embodiments provided by the present invention, reference is made to Figure 5 Combined with Figure 1 , S2: Refine the target mask image and preliminarily extract the center line of the cerebral artery to obtain a refined image, including:

[0080] S21: Use the distance transform algorithm to refine the target mask image to obtain a distance transform grayscale image.

[0081] The distance transform algorithm is used to refine the target mask image, which has a fast operation speed but causes the center lines of the extracted cerebral arteries to be discontinuous.

[0082] S22: performing a point deletion operation based on the distance transformed grayscale image through 26-neighborhood analysis, including: determining whether each pixel point is 26-neighborhood connected: if yes, retaining the pixel point; if not, deleting the pixel point.

[0083] The multiple neighborhoods of a three-dimensional image include 6 neighborhoods, 18 neighborhoods, 26 neighborhoods, etc., and the present invention does not limit this. The 6 neighborhoods of a three-dimensional image indicate that the 6 faces can be connected. The 18 neighborhoods indicate that all directions except the four corners of the cube can be connected. The 26 neighborhoods, that is, the 26 voxel points in the three-dimensional space adjacent to the center point are considered to be its neighborhood points, including the voxels directly connected to it and diagonally connected to it. For example, if the coordinates of the central voxel point are (i, j, k), then there are 27 points in total (i±1, j±1, k±1). Excluding the original center point (i, j, k), there are a total of 26 points around it. Among the 26 points, there are 6 points with a distance of 1 from the center point, that is, there are 6 points directly connected to the center point (which just corresponds to the 6 neighborhoods, indicating that the 6 faces of the three-dimensional image can be connected); there are 6 points with a distance of 1 from the center point, that is, there are 6 points directly connected to the center point (which just corresponds to the 6 neighborhoods, indicating that the 6 faces of the three-dimensional image can be connected); There are 12 points that are diagonally opposite to the center point (the four corners of the three faces passing through the center point); the distance from the center point is That is, there are 8 points on the diagonal of a small cube with the center point (two layers, upper and lower, with 4 cubes on each layer).

[0084] Compared with the 6-neighborhood and 18-neighborhood, the 26-neighborhood ensures that pixels can be connected in any direction and can achieve 360-degree directionality. Based on the arbitrary directionality of the centerline of the cerebral artery, analysis is performed through the 26-neighborhood to ensure that the 26-neighborhood property of deleting pixels meets the definition of extracting the centerline of the cerebral artery.

[0085] In step S22, a point deletion operation is performed based on the distance transformation grayscale image through 26-neighborhood analysis, which solves the problem of discontinuity of the center line of the cerebral artery extracted in step S21, making the extracted center line of the cerebral artery a continuous center line.

[0086] In some possible embodiments provided by the present invention, reference is made to Figure 6 Combined with Figure 1 , S3: Determine the starting point of the internal carotid artery according to the refined image, and construct a connectivity graph according to the refined image, including:

[0087] S31: Determine all end points based on the refined image through 26-neighborhood analysis.

[0088] After thinning the target mask image in step S2, the refined image obtained maintains connectivity. Since the end points of the refined image usually have only one neighboring node within the 26-neighborhood, all the end points of the refined image can be found through the 26-neighborhood analysis.

[0089] S32: Determine the starting point of the internal carotid artery based on all the end points. The starting point of the internal carotid artery is the end point farthest from the center of all the end points.

[0090] The starting point of the internal carotid artery is defined as the end point farthest from the center of all end points. The starting point of the internal carotid artery can be automatically determined through step S32, without the need to manually determine the starting point of the internal carotid artery based on all end points, thereby reducing the increase in complexity caused by the end point calculation.

[0091] In some possible embodiments provided by the present invention, in step S32, the starting point of the internal carotid artery is determined based on all the end points. The formula for determining the starting point is:

[0092]

[0093] Among them, Max{} means taking the maximum value, || means Euclidean distance, Ei means the i-th end point, i is a positive integer, The formula for determining the starting point shows that the starting point S of the internal carotid artery is the distance from the center of all the end points to the starting point S. The farthest end point.

[0094] In some possible embodiments provided by the present invention, reference is made to Figure 6 Combined with Figure 1 , S3: Determine the starting point of the internal carotid artery according to the refined image, and construct a connectivity graph according to the refined image, including:

[0095] S33: Based on the refined image, a distance transformation algorithm is used to construct a weighted connected graph.

[0096] A connectivity graph is constructed to facilitate subsequent extraction of cerebral artery centerlines. A distance transform algorithm is used to calculate weights to construct a weighted connectivity graph, aligning the extracted cerebral artery centerlines closer to the center. Weight calculations are based on a distance transform graph. The principle of the distance transform graph is that the farther away from the edge of the refined image, the greater the weight of the connectivity graph, and the smaller the value; the closer to the edge of the refined image, the smaller the weight of the connectivity graph, and the greater the value.

[0097] In some possible embodiments provided by the present invention, in step S33, the weight calculation formula of the connectivity graph is as follows:

[0098] W i =-W i

[0099] Among them, W represents the weight, the actual value is the distance transformation grayscale image value DF(i), i is the position of all mask pixel values ​​in the refined image, W i Represents the weight value of the i-th pixel.

[0100] In some possible embodiments provided by the present invention, reference is made to Figure 7 Combined with Figure 1 , S4: Calculate all minimum cost paths as the target centerline based on the starting point and connectivity graph of the internal carotid artery, including:

[0101] S41: Use Dijkstra's algorithm to calculate the minimum cost path between the starting point and each end point of the internal carotid artery as the target centerline.

[0102] The starting point and each end point of the internal carotid artery are clearly defined and accurately determined to accurately extract the minimum cost path as the target centerline, making the extracted target centerline more accurate and closer to the center.

[0103] S42: until the calculation of the minimum cost path between the starting point and all end points of the internal carotid artery is completed, all target center lines are obtained.

[0104] The minimum cost path between the starting point and all end points of the internal carotid artery is iteratively extracted to obtain all accurate target centerlines.

[0105] In some other possible embodiments provided by the present invention, other algorithms such as depth-first search are used to extract the path between the starting point and each end point of the internal carotid artery as the target centerline.

[0106] In some possible embodiments provided by the present invention, reference is made to Figure 7 S4: After calculating all minimum cost paths as target center lines according to the starting point and connectivity graph of the internal carotid artery, the method further includes:

[0107] S43: Filtering is performed on all target centerlines (not shown in the figure) to smooth all extracted target centerlines. Further, the 6D pose of subsequent navigation planning can be guided based on all filtered target centerlines, thereby improving the robustness of the algorithm.

[0108] For example, the filtering algorithm is a mean filter; for another example, the filtering algorithm is an operation of further filtering based on the edge of the mask image. The present invention does not limit the filtering algorithm used to smooth the center line.

[0109] refer to Figures 1 to 9 , the following introduces an exemplary implementation of the centerline extraction method of the cerebral artery.

[0110] refer to Figure 3 , S011: Input a three-dimensional image of cerebral arteries.

[0111] S012: Obtain an initial mask image of the cerebral aneurysm blood vessels and surrounding blood vessels using a segmentation algorithm based on the three-dimensional image of the cerebral artery blood vessels.

[0112] In this embodiment, the initial mask image starts at the starting point of the internal carotid artery proximal to the parent artery, and ends at a predetermined multiple of the parent artery diameter distal to the parent artery at the location of the tumor. The diameter of the cerebral artery branches in the initial mask image is greater than a first predetermined value, and the length of the cerebral artery branches is no less than a second predetermined value. The initial mask image obtained in step S012 removes redundant bifurcated vessels from the three-dimensional image of the cerebral arteries, thereby reducing subsequent computational complexity and improving computational efficiency.

[0113] refer to Figure 2 , executing step S01: acquiring an initial mask image of the 3D image of the cerebral arteries, in preparation for extracting the region of interest.

[0114] Executing step S02: extracting the region of interest in the initial mask image, including:

[0115] refer to Figure 4 , executing step S021: performing maximum connected component processing on the initial mask image to eliminate the noise caused by the initial mask image processing algorithm, so that the extracted region of interest is accurate.

[0116] S022: Extract the starting and ending points of the region of interest. The starting point of the region of interest is the first point in the initial mask image in the X, Y, and Z directions that contains a mask pixel value. The ending point of the region of interest is the first point in the initial mask image in the X, Y, and Z directions that contains a non-mask pixel value. By extracting the starting and ending points of the region of interest, the scope of the region of interest is accurately determined, reducing subsequent computational effort and improving efficiency.

[0117] Continue to refer Figure 2 , execute step S03: fill the holes in the cerebral blood vessel image in the region of interest to form a target mask image of the cerebral artery. Figure 9 , holes like Figure 9 As shown in B. As mentioned above, the hole filling is performed in step S03 to prepare for the subsequent extraction of the accurate center line of the cerebral artery.

[0118] refer to Figure 1 , executing step S1: acquiring a target mask image of the cerebral artery, extracting the centerline of the cerebral artery from the target mask image, reducing the amount of computation required for centerline extraction and saving time for centerline extraction.

[0119] Execute step S2: refine the target mask image and preliminarily extract the center line of the cerebral artery to obtain a refined image, including: Figure 5 Steps S21 and S22 are shown.

[0120] Specifically, refer to Figure 5 S21: Use the distance transform algorithm to refine the target mask image to obtain a distance transform grayscale image. Although the distance transform algorithm is fast, the extracted center line is discontinuous, so step S22 needs to be continued to overcome the problem of discontinuous center line extraction.

[0121] S22: Performing a point deletion operation based on the distance-to-grayscale image through 26-neighborhood analysis includes: determining whether each pixel point is 26-neighborhood connected; if so, retaining the pixel point; if not, deleting the pixel point, thereby extracting the continuous centerline of the cerebral artery.

[0122] By adopting the above technical solution, the thinning process ensures that the center line of the initially extracted cerebral artery is a continuous center line and ensures the connectivity of the thinned image.

[0123] refer to Figure 1 , execute step S3: determine the starting point of the internal carotid artery according to the refined image, and construct a connectivity map according to the refined image, including Figure 6 Steps S31, S32 and S33 are shown.

[0124] Specifically, refer to Figure 6 , executing step S31: determining all end points through 26-neighborhood analysis based on the refined image.

[0125] The refined image obtained by step S2 maintains connectivity. Since an end point of the refined image usually has only one neighboring node within the 26-neighborhood, all end points of the refined image can be found through the 26-neighborhood analysis.

[0126] Execute step S32: Determine the starting point of the internal carotid artery based on all the end points. The starting point of the internal carotid artery is the end point farthest from the center of all the end points. Determining the starting point of the internal carotid artery prepares for the subsequent extraction of the target centerline through the minimum cost path.

[0127] In this embodiment, the formula Determine the starting point S. Where Max{} represents the maximum value, || represents the Euclidean distance, Ei represents the i-th end point, i is a positive integer, The formula for determining the starting point shows that the starting point S of the internal carotid artery is the distance from the center of all the end points to the starting point S. The farthest end point. Figure 8 The origin of the internal carotid artery is Figure 8 Point A shown in FIG.

[0128] refer to Figure 6 Then, step S33 is continued: a distance transformation algorithm is used to construct a weighted connectivity graph based on the refined image, so that the center line of the extracted cerebral artery is closer to the center.

[0129] In this embodiment, the weight calculation formula of the connectivity graph is: i =-W i Where W represents the weight, the actual value is the distance transformation grayscale image value DF(i), i is the position of all mask pixel values ​​in the refined image, W i Represents the weight value of the i-th pixel.

[0130] Continue to refer Figure 1 , execute step S4: calculate all minimum cost paths as the target center line according to the starting point and connectivity graph of the internal carotid artery. Step S4 includes the following: Figure 7 Step S41 and step S42 are shown, and step S43 is not shown in the figure.

[0131] S41: Use the Dijkstra algorithm to calculate the minimum cost path between the starting point and each end point of the internal carotid artery as the target centerline. Specifically, using the starting point of the internal carotid artery as the starting point and each end point as the end point, the Dijkstra algorithm is used to calculate the minimum cost path between the starting point and each end point as the target centerline. This ensures that the extracted target centerline is more accurate and closer to the center.

[0132] S42: The minimum cost path between the starting point of the internal carotid artery and all end points is calculated until all target center lines are obtained. That is, after each minimum cost path between an end point (i.e., end point) and the starting point of the internal carotid artery (i.e., starting point) is calculated, it is determined whether there are any end points (i.e., end points) for which the minimum cost path has not yet been calculated:

[0133] If so, step S41 is executed: using Dijkstra algorithm to calculate the minimum cost path between the starting point and each end point of the internal carotid artery as the target centerline;

[0134] If not, step S42 is executed to obtain all target center lines.

[0135] Next, step S43 is executed: all target centerlines are averaged (not shown in the figure) to smooth all extracted target centerlines. Furthermore, the 6D pose of subsequent navigation planning can be guided based on all filtered target centerlines, thereby improving the robustness of the algorithm.

[0136] refer to Figure 1 , execute step S5: output all target center lines.

[0137] The above technical solution not only incorporates pre-processing steps into the cerebral artery centerline extraction process, such as noise removal in step S01 to obtain an initial mask image and hole filling in step S03, addressing data inaccuracies during data source acquisition while also improving computational speed and algorithm robustness. Post-processing steps, such as filtering all target centerlines in step S43, also enhance the robustness of the algorithm.

[0138] The cerebral artery centerline extraction method realizes the automatic and rapid extraction of accurate, close-to-center centerlines, while increasing the robustness of the centerline extraction algorithm and optimizing the extraction of vascular centerlines in cerebral aneurysm planning software. It can accurately determine the specific location of the aneurysm, calculate the degree of aneurysm lesions, estimate the risk of rupture, and plan interventional surgical paths.

[0139] Second, reference Figure 10 The present invention provides a cerebral artery centerline extraction system 1, comprising: a data receiving module 11, a data processing module 12 and a data output module 13, wherein:

[0140] The data receiving module 11 is used to receive a three-dimensional image of a cerebral artery input by a user;

[0141] The data processing module 12 is connected to the data receiving module 11 and is used to

[0142] Acquire a target mask image of cerebral arteries;

[0143] Performing thinning processing on the target mask image and preliminarily extracting the center line of the cerebral artery to obtain a thinned image;

[0144] Determine the starting point of the internal carotid artery based on the refined image, and construct a connectivity map based on the refined image;

[0145] Calculate all minimum cost paths as the target centerline based on the starting point and connectivity graph of the internal carotid artery;

[0146] The data output module 13 is connected to the data processing module 12 and is used to output all target center lines.

[0147] By adopting the above technical solution, under the coordinated action of the data receiving module 11, the data processing module 12 and the data output module 13, the center line of the cerebral artery is extracted fully automatically through the center line extraction system 1 of the cerebral artery, which reduces the amount of calculation, improves the calculation speed, and the extracted center line of the cerebral artery is closer to the center.

[0148] Thirdly, reference Figure 11 The present invention provides an electronic device 2, comprising 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, the method for extracting the centerline of a cerebral artery according to any of the aforementioned embodiments is implemented. The memory 21 may include, for example, a system memory, a fixed non-volatile storage medium, or the like. The system memory may store, for example, an operating system, application programs, a boot loader, and other programs.

[0149] In this embodiment, the electronic device 2 automatically and quickly extracts the centerline of the cerebral artery that is accurate and close to the center, thereby improving efficiency and saving time.

[0150] In a fourth aspect, an embodiment of the present invention further provides 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 cerebral artery.

[0151] It will be understood by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable non-transitory storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0152] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts 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, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0153] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0154] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 The steps for the function specified in one or more boxes.

[0155] Although the present invention has been illustrated and described with reference to certain preferred embodiments thereof, it should be understood by those skilled in the art that the above description is provided as a further detailed description of the present invention in conjunction with specific embodiments thereof, and that the specific implementation of the present invention is not limited to these descriptions. Those skilled in the art may make various changes in form and details, including simple deductions or substitutions, without departing from the spirit and scope of the present invention.

Claims

1. A method for extracting the centerline of a cerebral artery, characterized in that: The method comprises: Acquire a target mask image of cerebral arteries; performing a thinning process on the target mask image and preliminarily extracting the centerline of the cerebral artery to obtain a thinned image; determining a starting point of the internal carotid artery according to the refined image, and constructing a connectivity map according to the refined image; Calculating all minimum cost paths as target center lines according to the starting point of the internal carotid artery and the connectivity graph; Outputting all target center lines; Determining the starting point of the internal carotid artery according to the refined image and constructing a connectivity map according to the refined image includes: Determine all end points by 26-neighborhood analysis based on the thinned image; Determine the starting point of the internal carotid artery based on all the end points, wherein the starting point of the internal carotid artery is the end point farthest from the center of all the end points; Determining the starting point of the internal carotid artery according to the refined image and constructing a connectivity map according to the refined image includes: According to the thinned image, a distance transformation algorithm is used to construct a weighted connected graph; The farther away from the edge of the refined image, the greater the weight of the connectivity graph and the smaller the weight value; the closer to the edge of the refined image, the smaller the weight of the connectivity graph and the greater the weight value.

2. The method for extracting the centerline of a cerebral artery according to claim 1, wherein: The starting point of the internal carotid artery is determined based on all the end points, and the formula is: Among them, Max{} means taking the maximum value, || means Euclidean distance, Ei means the i-th end point, i is a positive integer, Represents the mean of all endpoints.

3. The method for extracting the centerline of a cerebral artery according to claim 1, wherein: The weight calculation formula of the connected graph is as follows: IN i =-W i Where W represents the weight, the actual value is the distance transformation grayscale image value DF(i), i is the position of all mask pixel values ​​in the refined image, W i Represents the weight value of the i-th pixel.

4. The method for extracting the centerline of a cerebral artery according to claim 3, wherein: The step of calculating all minimum cost paths as target centerlines based on the starting point of the internal carotid artery and the connectivity graph includes: Calculate the minimum cost path between the starting point of the internal carotid artery and each of the end points as the target centerline using the Dijkstra algorithm; The calculation of the minimum cost path between the starting point of the internal carotid artery and all the end points is completed, and all target center lines are obtained.

5. The method for extracting the centerline of a cerebral artery according to claim 1, wherein: Before acquiring the target mask image of the cerebral artery, the method further includes: obtaining an initial mask image of a three-dimensional image of cerebral arteries; extracting a region of interest from the initial mask image; Holes are filled in the cerebral blood vessel image in the region of interest to form a target mask image of the cerebral artery.

6. The method for extracting the centerline of a cerebral artery according to claim 5, wherein: The step of obtaining an initial mask image for a three-dimensional image of a cerebral artery comprises: Input a three-dimensional image of cerebral arteries; Acquiring an initial mask image of the cerebral aneurysm blood vessel and surrounding blood vessels using a segmentation algorithm based on the three-dimensional image of the cerebral artery; Among them, the starting position of the initial mask image is the starting point of the internal carotid artery proximal to the tumor-bearing artery, and the ending position is the distal end of the tumor-bearing artery that is not shorter than a preset multiple of the diameter of the tumor-bearing artery at the tumor position. The diameter of the cerebral artery branch in the initial mask image is greater than a first preset value, and the length of the cerebral artery branch is not less than a second preset value.

7. The method for extracting the centerline of a cerebral artery according to claim 6, wherein: The tumor-bearing artery is located at the starting point of the internal carotid artery.

8. The method for extracting the centerline of a cerebral artery according to claim 5, wherein: The extracting the region of interest from the initial mask image includes: Extracting the starting point and the ending point of the region of interest; The starting point of the region of interest is the first point in the three-dimensional direction of the initial mask image that is a mask pixel value, and the ending point is the first point in the three-dimensional direction that is not a mask pixel value.

9. The method for extracting the centerline of a cerebral artery according to claim 1, wherein: The thinning process of the target mask image to preliminarily extract the centerline of the cerebral artery to obtain a thinned image includes: Using a distance transformation algorithm to refine the target mask image to obtain a distance transformation grayscale image; Performing a point deletion operation based on the distance transformed grayscale image through 26-neighborhood analysis includes: determining whether each pixel point is 26-neighborhood connected: if yes, retaining the pixel point; if not, deleting the pixel point.

10. A system for extracting the centerline of a cerebral artery, characterized in that: include: A data receiving module, configured to receive a three-dimensional image of a cerebral artery input by a user; A data processing module is connected to the data receiving module and is used to Acquire a target mask image of cerebral arteries; performing a thinning process on the target mask image and preliminarily extracting the centerline of the cerebral artery to obtain a thinned image; determining a starting point of the internal carotid artery according to the refined image, and constructing a connectivity map according to the refined image; Calculating all minimum cost paths as target center lines according to the starting point of the internal carotid artery and the connectivity graph; Determining the starting point of the internal carotid artery according to the refined image and constructing a connectivity map according to the refined image includes: Determine all end points by 26-neighborhood analysis based on the thinned image; Determine the starting point of the internal carotid artery based on all the end points, wherein the starting point of the internal carotid artery is the end point farthest from the center of all the end points; Determining the starting point of the internal carotid artery according to the refined image and constructing a connectivity map according to the refined image includes: According to the thinned image, a distance transformation algorithm is used to construct a weighted connected graph; The farther away from the edge of the refined image, the greater the weight of the connectivity graph and the smaller the weight value; the closer to the edge of the refined image, the smaller the weight of the connectivity graph and the greater the weight value; A data output module is connected to the data processing module and is used to output all the target center lines.

11. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the method for extracting the cerebral artery centerline according to any one of claims 1 to 9 is implemented.

12. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method for extracting the centerline of a cerebral artery according to any one of claims 1 to 9 is implemented.

Images