Multistage vascular network reconstruction and segmentation method based on drainage basin analysis

Abstract

The invention relates to the technical field of medical image processing, in particular to a multilevel vascular network reconstruction and segmentation method based on drainage basin analysis, which comprises the following steps: acquiring a 3D CT (three-dimensional computed tomography angiography) scanning result; based on the 3DCTA scanning result, reconstructing a 3D blood vessel network; performing image segmentation and enhancement processing on the 3DCTA scanning result to obtain an original blood vessel center line; constructing a pressure field based on the original blood vessel center line; determining a watershed position in the blood vessel network based on the pressure field; segmenting the blood vessel network according to the watershed position; and outputting a reconstructed 3D blood vessel network and a blood vessel segmentation result, and by ingeniously combining a drainage basin analysis theory and a traditional image processing technology, considering the geometric morphology and hemodynamic characteristics of the blood vessel in the blood vessel reconstruction and segmentation process. The accuracy of reconstruction and segmentation is improved, and a new view angle is provided for vascular function evaluation.

Description

Technical Field

[0001] The present invention relates to the technical field of medical image processing, and more specifically to a multi-level vascular network reconstruction and segmentation method based on watershed analysis. Background Art

[0002] With the continuous advancement of medical imaging technology, angiographic imaging plays an increasingly important role in clinical diagnosis and treatment. Accurate vascular network reconstruction and segmentation can not only provide doctors with valuable diagnostic information, but also provide important support for surgical planning and navigation. However, due to the complexity of vascular structure and various interference factors in the imaging process, accurate vascular network reconstruction and segmentation remains a challenging problem.

[0003] Traditional vascular reconstruction methods mainly rely on basic image processing techniques such as threshold segmentation and morphological processing. These methods perform well when dealing with simple vascular structures, but are often unable to cope with complex multi-level vascular networks. Especially when dealing with small blood vessels, bifurcations, and lesion areas, these methods often cause problems such as breakage, over-segmentation, or under-segmentation, resulting in inaccurate and incomplete reconstruction results.

[0004] In recent years, model-based methods and machine learning techniques have been widely used in the field of vascular reconstruction. These methods have improved the accuracy of vascular reconstruction to a certain extent by introducing prior knowledge or using a large amount of labeled data for training. However, these methods also have some inherent limitations. Model-based methods often require carefully designed mathematical models, which have limited generalization capabilities and are difficult to adapt to changes in vascular morphology under different cases and imaging conditions. Machine learning methods are highly dependent on the quality and quantity of training data and may perform poorly when faced with rare cases or new imaging modalities.

[0005] In addition, most existing vascular reconstruction methods focus on the geometric reconstruction of blood vessels, while ignoring the hemodynamic characteristics inside the blood vessels. This reconstruction method based solely on morphology is difficult to reflect the functional characteristics of the vascular network, limiting its value in certain clinical applications. For example, when assessing the risk of aneurysm rupture or planning vascular interventional surgery, relying solely on the geometric information of the blood vessels is far from enough.

[0006] Another issue that deserves attention is the accuracy and consistency of vascular segmentation. Existing methods often have difficulty accurately identifying the boundaries between vessels of different levels when dealing with complex multi-level vascular networks, resulting in inaccurate segmentation results. This not only affects subsequent quantitative analysis, but may also mislead clinical decision-making.

[0007] In summary, the existing vascular network reconstruction and segmentation methods still have many deficiencies when dealing with complex multi-level vascular networks, and it is difficult to meet the growing clinical needs. Therefore, there is an urgent need for vascular network reconstruction and segmentation methods that can comprehensively consider vascular geometry and hemodynamic characteristics, and have high accuracy, high reliability and good adaptability. Summary of the invention

[0008] The present invention is proposed to solve the above technical problems. The present invention proposes a multi-level vascular network reconstruction and segmentation method based on watershed analysis, aiming to overcome the limitations of the prior art and achieve more accurate, reliable and comprehensive vascular network reconstruction and segmentation.

[0009] In order to solve the above technical problems, the present invention adopts the following technical solutions: Multi-level vascular network reconstruction and segmentation method based on watershed analysis, including: The acquisition steps include: Obtain three-dimensional computed tomography angiography (3DCTA) scan results; Processing steps include: Reconstructing a 3D vascular network based on the 3D CTA scanning results; Performing image segmentation and enhancement processing on the 3DCTA scanning result to obtain the original blood vessel centerline; constructing a pressure field based on the original blood vessel centerline; determining a watershed location in the vascular network based on the pressure field; Segmenting the vascular network according to the watershed position; Output steps include: Output the reconstructed 3D vascular network and vascular segmentation results.

[0010] Preferably, the reconstructing the 3D vascular network specifically includes: A series of vascular regions are obtained by using a vascular extraction algorithm; Based on a vascular centerline extraction algorithm, a 3D vascular network is reconstructed according to the skeleton of the vascular region.

[0011] Preferably, the blood vessel area is generated by calculating the accumulation of water volume in the watershed to form different watershed areas.

[0012] Preferably, obtaining the original blood vessel centerline specifically includes: A vascular enhancement algorithm is used to perform nonlinear enhancement on the 3DCTA scan result to highlight the vascular structure and details; The maximum radius filtering technique is used to identify the centerline of the blood vessel and filter out non-vascular regions of interest and small blood vessels; Obtain a rough original blood vessel centerline based on a three-dimensional reconstruction algorithm; The refined vessel centerlines are obtained using a neighborhood-based skeleton thinning technique.

[0013] Preferably, the constructing of the pressure field specifically includes: constructing a blood pressure field by combining the original blood vessel centerline and blood vessel imaging parameters; generating a gradient field of the blood pressure field and optimizing the gradient field; The pressure gradient of the blood pressure field is calculated.

[0014] Preferably, the determining of the watershed position in the vascular network specifically includes: Projecting the original blood vessel centerline into two-dimensional space by a pressure gradient calculation method to obtain a projected blood vessel centerline; The projected blood vessel centerline is projected onto the blood vessel image to mark the position of the watershed.

[0015] As a preference, it also includes: Performing connectivity analysis on the blood vessel segmentation results; Based on the connectivity analysis result and the blood vessel segmentation result, a fully reconstructed multi-level blood vessel segmentation result is obtained.

[0016] Preferably, the connectivity analysis specifically includes: Performing a voxelization operation on the blood vessel segmentation result, marking each voxel as an independent region; Scan all connected areas and test the connection between each two connected areas; Based on the connection test results and the input image blood vessels, the connected areas are merged.

[0017] As a preference, it also includes: Based on the reconstructed 3D vascular network, the vessel length and volume are calculated.

[0018] As a preference, it also includes: When intraoperative detection is required, segmentation results are obtained based on the vascular centerline of each independent image; According to the segmentation result, adjusting the segmentation position of each blood vessel; When the vessel segmentation result and the vessel centerline change, the watershed is reconstructed to obtain an updated segmentation result.

[0019] The method of the present invention has the following significant technical effects: The method of the present invention cleverly combines the theory of watershed analysis with traditional image processing technology to simultaneously consider the geometric morphology and hemodynamic characteristics of blood vessels during vascular reconstruction and segmentation. This innovative method not only improves the accuracy of reconstruction and segmentation, but also provides a new perspective for vascular function assessment. In particular, the method of the present invention shows significant advantages when dealing with complex multi-level vascular networks.

[0020] First, by introducing the theory of watershed analysis, this method can better capture the topological structure of the vascular network, effectively solving the difficulties of traditional methods in dealing with bifurcations and small blood vessels. Secondly, this method takes into account the hemodynamic characteristics during the reconstruction process, and provides a more reliable basis for vascular segmentation by constructing the pressure field and calculating the pressure gradient. This not only improves the accuracy of the segmentation, but also lays the foundation for subsequent hemodynamic analysis.

[0021] In addition, the method of the present invention has good adaptability and robustness. By adopting a multi-step, multi-scale processing strategy, the method can effectively cope with different imaging conditions and changes in vascular morphology. In particular, when dealing with noise and artifacts, the method shows excellent anti-interference ability. This feature makes the method widely applicable in clinical practice and can provide reliable support for the diagnosis and treatment of various vascular-related diseases.

[0022] Another significant advantage of the method of the present invention is its comprehensiveness and systematicness. From the initial image segmentation to the final blood vessel segmentation, the method provides a complete solution. This end-to-end processing flow not only improves the overall efficiency, but also ensures the consistency and coherence between the various processing steps. In particular, in the key steps such as blood vessel centerline extraction, pressure field construction and watershed position determination, the method adopts innovative algorithms and optimization strategies to achieve the organic combination and mutual promotion between these steps.

[0023] It is worth mentioning that the method of the present invention has also made considerable progress in computational efficiency. By optimizing the algorithm design and utilizing parallel computing technology, the method can complete the reconstruction and segmentation tasks of complex vascular networks in a reasonable time. This feature is of great significance for supporting real-time or quasi-real-time clinical applications, such as intraoperative navigation and interventional treatment.

[0024] In summary, the multi-level vascular network reconstruction and segmentation method based on watershed analysis proposed in the present invention achieves high accuracy, high reliability and comprehensiveness of vascular network reconstruction and segmentation by innovatively combining geometric morphology analysis and hemodynamic simulation. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 The figure is an overall flow chart of the method of the present invention.

[0026] Figure 2 FIG. 1 is a flow chart of the present invention for reconstructing a 3D vascular network.

[0027] Figure 3 The flowchart of the present invention for obtaining the original blood vessel centerline.

[0028] Figure 4 This is a flow chart of constructing a pressure field according to the present invention.

[0029] Figure 5 The flowchart of the present invention is to determine the watershed position and the vascular network segmentation. DETAILED DESCRIPTION

[0030] like Figure 1-5 As shown, the present invention discloses a multi-level vascular network reconstruction and segmentation method based on watershed analysis. The method first obtains the three-dimensional computed tomography angiography (3DCTA) scan results, and then reconstructs the 3D vascular network based on the scan results. Next, the 3DCTA scan results are subjected to image segmentation and enhancement processing to obtain the original vascular centerline. Subsequently, a pressure field is constructed based on the obtained original vascular centerline, and the watershed position in the vascular network is determined using the pressure field. Finally, the vascular network is segmented according to the determined watershed position, and the reconstructed 3D vascular network and vascular segmentation results are output.

[0031] Preferably, in one embodiment of the present invention, the process of reconstructing a 3D vascular network specifically includes obtaining a series of vascular regions through a vascular extraction algorithm, and then reconstructing the 3D vascular network based on the skeleton of the obtained vascular regions based on a vascular centerline extraction algorithm. This method can effectively capture the geometric features of blood vessels and improve the accuracy of reconstruction. For example, when processing a cerebral vascular network, a multi-scale vesselness filter can be used as a vascular extraction algorithm, and the scale range of the filter can be set to 0.5mm to 2mm with a step size of 0.1mm. Such parameter settings can effectively process blood vessels of different calibers, thereby improving the comprehensiveness of the reconstruction results.

[0032] Furthermore, the method of the present invention also includes forming different watershed areas by calculating the accumulation of water volume in the watershed, thereby generating the above-mentioned vascular area. This method based on watershed analysis can better reflect the branching structure of blood vessels, especially when dealing with complex vascular networks. For example, when calculating the water volume in the watershed, the following formula can be used: , Where W(x) is the accumulated water volume at position x, U(x) is the upstream area of ​​x, and f(y) is the local water volume at position y. In practical applications, the grayscale value of the blood vessel or the vesselness response value can be used as a measure of the local water volume. By setting an appropriate threshold (for example, the accumulated water volume is greater than 1.5 times the average grayscale value of the image), the blood vessel area and the background can be effectively distinguished.

[0033] In the process of obtaining the original vascular centerline, the present invention adopts a multi-step processing method. First, the 3DCTA scan result is nonlinearly enhanced using a vascular enhancement algorithm to highlight the vascular structure and details. Here, the Frangi filter can be used as the vascular enhancement algorithm, and its mathematical expression is: , Where V(s) is the vesselness metric, , , is the eigenvalue of the Hessian matrix, , and Respectively represent the cross-sectional ratio, sphericity and structural degree of blood vessels. , and Controls the sensitivity of the filter, which can usually be set to =0.5, =0.5, c=half maximum intensity width / 2.

[0034] Next, the present invention uses the maximum radius filtering technique to identify the centerline of the blood vessel and filter out the non-vascular region of interest and small blood vessels. This step can effectively reduce the complexity of subsequent processing and improve the efficiency of the algorithm. In practice, the maximum radius threshold can be set to 1% of the length of the image diagonal line, so that the main blood vessels can be retained while effectively removing noise and small branches.

[0035] Subsequently, a rough original blood vessel centerline is obtained based on the 3D reconstruction algorithm. Here, the minimum path method can be used to obtain the centerline by solving the Eikonal equation: , Among them, T(x) is the shortest path length to the starting point, and P(x) is the cost function, which can be defined as , V(x) is the vesselness response, and k is a tuning parameter that can be set to a value between 2 and 5.

[0036] Finally, the neighborhood-based skeleton thinning technique is used to obtain the thinned blood vessel centerline. This step can further improve the accuracy of the centerline. A parallel thinning algorithm can be used to iteratively delete boundary points until no more can be deleted. In each iteration, a 3x3x3 neighborhood can be used to determine whether to delete the current point. The judgment conditions include connectivity preservation, endpoint preservation, etc.

[0037] Through the above steps, the method of the present invention can effectively reconstruct a complex vascular network and accurately extract the vascular centerline. This is of great significance for subsequent vascular analysis, disease diagnosis and surgical planning. For example, in the diagnosis of cerebral aneurysms, accurate vascular reconstruction and centerline extraction can help doctors better observe the location, size and shape of aneurysms, thereby formulating a more reasonable treatment plan.

[0038] The method of the present invention is not only applicable to the analysis of cerebral blood vessels, but can also be extended to the analysis of blood vessels of other organs, such as coronary arteries, pulmonary blood vessels, etc. By adjusting relevant parameters, it can adapt to the characteristics of blood vessels of different organs and achieve a wide range of applications.

[0039] In summary, the multi-level vascular network reconstruction and segmentation method based on watershed analysis proposed in this invention achieves high-precision and high-efficiency vascular network reconstruction and segmentation by combining a variety of advanced image processing technologies and watershed analysis theory. This provides a powerful tool support for medical image analysis and clinical diagnosis.

[0040] In a preferred embodiment of the present invention, the process of reconstructing the 3D vascular network specifically includes obtaining a series of vascular regions through a vascular extraction algorithm, and then reconstructing the 3D vascular network based on the skeletons of the obtained vascular regions based on a vascular centerline extraction algorithm. This method can effectively capture the geometric features of the blood vessels and improve the accuracy of reconstruction.

[0041] Specifically, the blood vessel extraction algorithm can use a multi-scale Hessian filter. The basic principle of this algorithm is to use the tubular structure characteristics of the blood vessels and enhance the blood vessel structure by analyzing the second-order derivative information of the image. The main steps of the algorithm are as follows: 1. For the input image Perform Gaussian filtering to obtain smooth images of different scales: ,in, is a 3D Gaussian kernel, is the standard deviation of the Gaussian kernel.

[0042] 2. Calculate the Hessian matrix at each scale: ,in, , etc. represent the second-order partial derivatives of the image.

[0043] 3. Perform eigenvalue decomposition on the Hessian matrix to obtain the eigenvalue , , (Assumption . , Half maximum intensity width / 2.

[0044] 4. Calculate the vesselness response at multiple scales and take the maximum value as the final result , by setting an appropriate threshold (e.g., vesselness response greater than 0.3), a preliminary vascular region can be obtained.

[0045] Next, based on the blood vessel centerline extraction algorithm, the position and shape of the blood vessel can be further refined. The commonly used method is the centerline extraction algorithm based on distance transformation. Its main steps are as follows: 1. Perform distance transformation on the blood vessel area to obtain the distance map .

[0046] 2. Calculate the gradient field of the distance map: ; 3. Calculate the divergence of the gradient field: ; 4. Find the local maximum points in the divergence field, which constitute the initial center line.

[0047] 5. Use the minimum spanning tree algorithm to connect these points to form a continuous center line.

[0048] In another embodiment of the present invention, the accumulation of water volume in the watershed is calculated to form different watershed regions to generate blood vessel regions. This method based on watershed analysis can better reflect the branching structure of blood vessels. Specifically, the cumulative water volume can be calculated using the following formula: , in, For location The accumulated water volume, for The upstream area, For location In practical applications, the vesselness response value of the blood vessel can be used as a measure of the local water volume.

[0049] Preferably, in the process of obtaining the original vascular centerline, the present invention adopts a multi-step processing method. First, a vascular enhancement algorithm is used to perform nonlinear enhancement on the 3DCTA scan result to highlight the vascular structure and details. Here, a Frangi filter can be selected as a vascular enhancement algorithm, and its mathematical expression has been described in detail above.

[0050] Next, the present invention uses the maximum radius filtering technique to identify the centerline of the blood vessel and filter out the non-vascular region of interest and small blood vessels. The basic idea of ​​the maximum radius filtering is to calculate the maximum sphere radius that can be completely contained inside the blood vessel at each voxel position. The specific algorithm is as follows: 1. Perform distance transformation on the binary vascular image to obtain the distance from each voxel to the nearest background point .

[0051] 2. At each voxel position , calculate the maximum inscribed sphere radius centered at this point

[0052] , in, for 's neighborhood.

[0053] 3. Set the radius threshold (e.g. 1% of the image diagonal length), retaining points with a radius greater than the threshold: .

[0054] 4. Perform morphological refinement on the results to obtain preliminary vascular centerlines.

[0055] Subsequently, a rough original blood vessel centerline is obtained based on the 3D reconstruction algorithm. Here, the minimum path method can be used to obtain the centerline by solving the Eikonal equation: , in, is the shortest path length to the starting point, As the cost function, it can be defined as , For vesselness response, To adjust the parameter, it can be set to a value between 2 and 5. The Fast Marching Method can be used to solve the Eikonal equation.

[0056] Finally, the neighborhood-based skeleton refinement technique is used to obtain the refined blood vessel centerline. The commonly used parallel refinement algorithms are as follows: 1. Initialize the boundary point set B, which contains all foreground points adjacent to the background.

[0057] 2. For each point p in B, mark it as deletable if the following conditions are met: the deletion of p does not change the local connectivity; p is not an endpoint; the deletion of p does not cause excessive erosion; 3. Delete all marked points.

[0058] 4. Update the boundary point set B.

[0059] 5. Repeat steps 2-4 until there are no points left to delete.

[0060] When determining local connectivity, the Euler number can be used to ensure that the topology remains unchanged. For a 3x3x3 neighborhood, the calculation formula for the Euler number E is: , Where V is the number of vertices, E is the number of edges, F is the number of faces, and C is the number of volumes. If the Euler number remains unchanged before and after deleting the point, the local connectivity is considered unchanged.

[0061] Through the above detailed algorithm description, the method of the present invention can effectively reconstruct complex vascular networks and accurately extract vascular centerlines. This is of great significance for subsequent vascular analysis, disease diagnosis and surgical planning. For example, in the diagnosis of cerebral aneurysms, accurate vascular reconstruction and centerline extraction can help doctors better observe the location, size and shape of aneurysms, thereby formulating more reasonable treatment plans.

[0062] In addition, the method of the present invention can be further optimized and expanded. For example, machine learning technology can be introduced to automatically adjust algorithm parameters to improve the adaptability and robustness of the method. At the same time, considering the characteristics of blood vessels in different organs, specific pre-processing and post-processing steps can be designed to adapt to different application scenarios.

[0063] In summary, the multi-level vascular network reconstruction and segmentation method based on watershed analysis proposed in this invention achieves high-precision and high-efficiency vascular network reconstruction and segmentation by combining a variety of advanced image processing technologies and watershed analysis theory. This provides a powerful tool support for medical image analysis and clinical diagnosis, and is expected to play an important role in the fields of cerebrovascular diseases, cardiovascular diseases, etc.

[0064] In another preferred embodiment of the present invention, the process of constructing the pressure field includes multiple steps, and the ingenious combination of these steps can effectively simulate the pressure distribution in the blood vessel, providing an important basis for the subsequent determination of the watershed position.

[0065] First, the method of the present invention combines the original vascular centerline and vascular imaging parameters to construct a blood pressure field. The core idea of ​​this step is to use the geometric characteristics of the blood vessels and the principles of hemodynamics to estimate the pressure distribution in the blood vessels. Specifically, the following model can be used: , in, Indicates location Blood pressure at is the blood pressure at the entrance (usually set to 120 mmHg, the average value of normal human arterial blood pressure), From the entrance to the location The distance along the centerline of the vessel, For location The radius of the blood vessel at is the attenuation coefficient (which can be set to a value between 0.1 and 0.5 based on experience).

[0066] It is worth noting that the vessel radius It can be estimated by analyzing the grayscale distribution around the center line of the original blood vessel. For example, a Gaussian fit can be performed on a plane perpendicular to the center line, and the standard deviation of the fitting curve can be used as an estimate of the blood vessel radius.

[0067] Next, the method of the present invention generates a gradient field of the blood pressure field and optimizes it. The calculation of the gradient field can adopt the central difference method: , in, , , , Here, h is the step size, which can usually be set to the voxel size.

[0068] In order to optimize the gradient field, anisotropic diffusion filtering can be used. This method can smooth the noise while maintaining edge information. Its mathematical expression is: , Where I is the image intensity and c(x,y,z,t) is the diffusion coefficient, which can be defined as: , K is a control parameter that can be set according to image characteristics and is usually 10% to 20% of the image gradient amplitude.

[0069] Finally, the method of the present invention calculates the pressure gradient of the blood pressure field. The pressure gradient reflects the rate of change of blood pressure in space and is crucial for understanding the hemodynamic characteristics. The pressure gradient can be calculated by the following formula: , in, is the blood density (about 1060kg / m³), is the blood flow velocity vector. The blood flow velocity can be obtained through phase contrast magnetic resonance imaging (PC-MRI) and other techniques, or estimated based on the geometric characteristics of blood vessels and the principles of fluid mechanics.

[0070] The pressure field constructed through the above steps provides an important basis for the subsequent determination of the watershed position. This pressure field construction method based on the principle of fluid mechanics can better reflect the physiological characteristics of the vascular network and help improve the accuracy and physiological relevance of vascular segmentation.

[0071] In another embodiment of the present invention, the process of determining the watershed position in a vascular network includes the following steps: first, the original vascular centerline is projected into a two-dimensional space by a pressure gradient calculation method to obtain a projected vascular centerline; then, the projected vascular centerline is projected onto a vascular image to mark the position of the watershed.

[0072] Specifically, the pressure gradient calculation method can adopt the following formula: , in, is the pressure gradient, is the blood density, is the acceleration due to gravity, is the height difference. In practical applications, the points on the centerline of the blood vessel can be projected onto a two-dimensional plane along the direction of the pressure gradient. The projection process can be numerically integrated using the Runge-Kutta method: , in, , , , , is the pressure gradient field, is the step length.

[0073] When projecting the projected vascular centerline onto the vascular image, the nearest neighbor interpolation or bilinear interpolation method can be used. The watershed position usually corresponds to the saddle point or local minimum point in the pressure gradient field. These feature points can be identified by analyzing the topological structure of the pressure gradient field.

[0074] In another embodiment of the present invention, a connectivity analysis is performed on the blood vessel segmentation result, and based on the connectivity analysis result and the blood vessel segmentation result, a fully reconstructed multi-level blood vessel segmentation result is obtained. The specific steps of the connectivity analysis are as follows: 1. Perform voxelization on the vascular segmentation results and mark each voxel as an independent region.

[0075] 2. Scan all connected areas and test the connection between each two connected areas. The connection test can adopt the following criteria: , in, and For two regions, Indicates area The set of boundary voxels of represents the cardinality of a set. If Greater than the preset threshold .

[0076] 3. Based on the connection test results and the input image blood vessels, the connected areas are merged. The merging process can be efficiently implemented using the Union-Find data structure. Through this connectivity analysis, the problem of blood vessel breakage caused by noise or segmentation errors can be effectively handled, and the integrity and continuity of blood vessel segmentation can be improved.

[0077] Finally, the method of the present invention also includes calculating the length and volume of blood vessels based on the reconstructed 3D blood vessel network. The length of the blood vessel can be calculated by accumulating the Euclidean distance between adjacent points on the center line of the blood vessel: , in, The first The coordinates of the points. The blood vessel volume can be calculated by integration: , in, is the blood vessel length, is the radius function along the centerline of the blood vessel. In the discrete case, it can be approximated as: , in, For the The average radius of the segments, For the The length of a segment.

[0078] By calculating these geometric parameters, doctors can have a more comprehensive understanding of the morphological characteristics of blood vessels, providing important reference for disease diagnosis and treatment plan formulation.

[0079] In summary, the multi-level vascular network reconstruction and segmentation method based on watershed analysis proposed in this invention achieves high-precision and high-reliability vascular network reconstruction and segmentation through a series of innovative steps such as constructing an accurate pressure field, determining the watershed position, and performing connectivity analysis. This method not only takes into account the geometric characteristics of the blood vessels, but also incorporates the principles of hemodynamics, so it can better reflect the physiological characteristics of the vascular network.

[0080] In another preferred embodiment of the present invention, the length and volume of blood vessels are further calculated based on the reconstructed 3D blood vessel network. The accurate calculation of these geometric parameters is of great significance for comprehensively evaluating the morphological characteristics of blood vessels, assisting disease diagnosis and formulating treatment plans.

[0081] Specifically, the calculation of the blood vessel length adopts the accumulation method, that is, the Euclidean distance between adjacent points along the center line of the blood vessel is accumulated. Its mathematical expression is as follows: , in, is the total length of the blood vessel, Indicates the first The three-dimensional coordinates of the points, is the total number of points on the center line. In order to improve the calculation accuracy, the B-spline interpolation method can be used to smooth the center line to reduce the error caused by discrete sampling.

[0082] The calculation of blood vessel volume uses the integration method, which regards the blood vessel as a series of continuous cylinders. Its mathematical expression is: , in, is the total vascular volume, is the blood vessel length, is the radius function along the centerline of the blood vessel. In practical applications, since the blood vessel radius is discretely sampled, the numerical integration method can be used for approximate calculation: , in, For the The average radius of the segments, For the In order to improve the calculation accuracy, a high-order numerical integration method can be used, such as the Simpson method or the Gauss-Legendre integration method.

[0083] It is worth noting that the estimation of the vessel radius is crucial to the accuracy of volume calculation. The present invention uses multi-planar reconstruction technology to improve the accuracy of radius estimation. Specifically, at each point on the centerline of the vessel, a plane perpendicular to the centerline is constructed, and the vessel contour is fitted on the plane. The fitting can use a circular or elliptical model, where the elliptical model can better adapt to blood vessels with non-circular cross-sections. The mathematical model of elliptical fitting is as follows: , in, are the major and minor axes of the ellipse, is the tilt angle of the ellipse. The fitting process can use the least squares method or RANSAC algorithm to improve robustness.

[0084] In addition, the present invention also considers the special treatment of the bifurcation point of the blood vessel. Near the bifurcation point, a simple cylindrical model may cause deviation in volume calculation. To this end, the present invention introduces a bifurcation point modeling method based on the Voronoi diagram. The specific steps are as follows: 1. Identify bifurcation points in the vascular network.

[0085] 2. Construct a three-dimensional Voronoi diagram around the bifurcation point.

[0086] 3. Use the boundaries of the Voronoi diagram to define the geometry of the bifurcation region.

[0087] 4. Calculate the volume of the Voronoi polyhedron as the volume of the bifurcation region.

[0088] This approach can more accurately describe complex bifurcation structures, thereby improving the accuracy of overall vessel volume calculations.

[0089] In another embodiment of the present invention, when intraoperative detection is required, the method further includes the following steps: first, a segmentation result is obtained based on the vascular centerline of each independent image; then, the position of each vascular segmentation is adjusted according to the obtained segmentation result; finally, when the vascular segmentation result and the vascular centerline change, the watershed is reconstructed to obtain an updated segmentation result.

[0090] The introduction of this dynamic update mechanism greatly improves the flexibility and adaptability of this method in actual clinical applications. Specifically, the intraoperative detection process can be described as follows: 1. Obtain segmentation results based on the vascular centerline of each independent image. This step uses a fast centerline tracking algorithm, the core idea of ​​which is to use the known centerline information to perform local search and matching on the new image. Mathematically, it can be expressed as minimizing the following energy function: , in, represents the parameterized centerline curve, , and Represent image term, smoothing term and prior term respectively, , and is the weight coefficient.

[0091] 2. Adjust the position of each blood vessel segment. This step uses deformation registration technology to deform the previous segmentation results into the new image space. The deformation field can be obtained by solving the following partial differential equation: , in, is the deformation field, As driving force, , and is the control parameter.

[0092] 3. Reconstruct the watershed. When the blood vessel segmentation results and the blood vessel centerline change significantly, the watershed needs to be reconstructed to ensure the accuracy of the segmentation. The fast watershed algorithm is used here. The core idea is to use the previous watershed results as initialization and only perform local updates in the changed areas. The main steps of the algorithm are as follows: a. Construct the minimum spanning tree (MST) of the image gradient. The local watershed algorithm can be expressed as the following optimization problem: b. Extract the watershed line from the MST. c. Compare the old and new watershed lines to identify the areas that need to be updated. d. Apply the local watershed algorithm to the areas that need to be updated. , in, Indicates segmentation, For areas that need to be updated, For the watershed basin, Pixel The gray value of for The average gray value.

[0093] Through this dynamic update mechanism, the method of the present invention can adapt to changes in vascular morphology during surgery in real time, providing doctors with continuously updated vascular segmentation information, thereby supporting more accurate surgical navigation and decision making.

[0094] In addition, in order to further improve the robustness and adaptability of the method, the present invention also introduces an adaptive parameter adjustment mechanism. Specifically, the system continuously monitors the quality of vessel segmentation and centerline extraction, and automatically adjusts the algorithm parameters based on the quality assessment results. The quality assessment can be based on the following indicators: 1. Segmentation consistency: Use the Dice coefficient or Jaccard index to measure the consistency of segmentation results between consecutive frames.

[0095] 2. Centerline stability: Calculate the average displacement of the centerline between consecutive frames.

[0096] 3. Topological consistency: Check whether the topological structure of the vascular network has changed significantly.

[0097] Based on these indicators, the system uses a reinforcement learning algorithm to optimize parameter settings. The state space of the reinforcement learning model includes the current image features and algorithm parameters, the action space is the direction and amplitude of parameter adjustment, and the reward function is defined based on the above quality assessment indicators. In this way, the system can automatically select the optimal parameter settings under different imaging conditions and vascular morphology, thereby ensuring the stability and reliability of the method.

[0098] In summary, the multi-level vascular network reconstruction and segmentation method based on watershed analysis proposed in the present invention can not only accurately calculate the geometric parameters of the blood vessels, but also has the ability of dynamic update and adaptive adjustment. These characteristics enable this method to better adapt to complex clinical scenarios.

[0099] The above description is only a preferred specific implementation manner of the present invention; however, the protection scope of the present invention is not limited thereto; any person familiar with the art who, within the scope disclosed by the present invention, makes equivalent replacements or changes based on the scheme and improved concepts of the present invention shall be covered within the protection scope of the present invention.

Claims

1. A multi-level vascular network reconstruction and segmentation method based on watershed analysis, characterized in that: include: The acquisition steps include: Obtain three-dimensional computed tomography angiography (3DCTA) scan results; Processing steps include: Reconstructing a 3D vascular network based on the 3D CTA scanning results; Performing image segmentation and enhancement processing on the 3DCTA scanning result to obtain the original blood vessel centerline; constructing a pressure field based on the original blood vessel centerline; determining a watershed location in the vascular network based on the pressure field; Segmenting the vascular network according to the watershed position; Output steps include: Output the reconstructed 3D vascular network and vascular segmentation results.

2. The method according to claim 1, characterized in that The reconstruction of the 3D vascular network specifically includes: A series of vascular regions are obtained by using a vascular extraction algorithm; Based on a vascular centerline extraction algorithm, a 3D vascular network is reconstructed according to the skeleton of the vascular region.

3. The method according to claim 2, characterized in that The vascular area is generated by calculating the accumulation of water volume in the watershed and forming different watershed areas.

4. The method according to claim 1, characterized in that: The obtaining of the original blood vessel centerline specifically includes: A vascular enhancement algorithm is used to perform nonlinear enhancement on the 3DCTA scan result to highlight the vascular structure and details; The maximum radius filtering technique is used to identify the centerline of the blood vessel and filter out non-vascular regions of interest and small blood vessels; Obtain a rough original blood vessel centerline based on a three-dimensional reconstruction algorithm; The refined vessel centerlines are obtained using a neighborhood-based skeleton thinning technique.

5. The method according to claim 1, characterized in that The construction of the pressure field specifically includes: constructing a blood pressure field by combining the original blood vessel centerline and blood vessel imaging parameters; generating a gradient field of the blood pressure field and optimizing the gradient field; The pressure gradient of the blood pressure field is calculated.

6. The method according to claim 1, characterized in that Determining the watershed position in the vascular network specifically includes: Projecting the original blood vessel centerline into two-dimensional space by a pressure gradient calculation method to obtain a projected blood vessel centerline; The projected blood vessel centerline is projected onto the blood vessel image to mark the position of the watershed.

7. The method according to claim 1, characterized in that Also includes: Performing connectivity analysis on the blood vessel segmentation results; Based on the connectivity analysis result and the blood vessel segmentation result, a fully reconstructed multi-level blood vessel segmentation result is obtained.

8. The method according to claim 7, characterized in that The connectivity analysis specifically includes: Performing a voxelization operation on the blood vessel segmentation result, marking each voxel as an independent region; Scan all connected areas and test the connection between each two connected areas; Based on the connection test results and the input image blood vessels, the connected areas are merged.

9. The method according to claim 1, characterized in that: Also includes: Based on the reconstructed 3D vascular network, the vessel length and volume are calculated.

10. The method according to claim 1, characterized in that Also includes: When intraoperative detection is required, segmentation results are obtained based on the vascular centerline of each independent image; According to the segmentation result, adjusting the segmentation position of each blood vessel; When the vessel segmentation result and the vessel centerline change, the watershed is reconstructed to obtain an updated segmentation result.

Images