Three-dimensional watertight blood vessel geometric reconstruction method based on scanning ball method

Abstract

The invention discloses a three-dimensional watertight blood vessel geometric reconstruction method based on a scanning ball method, and relates to the technical field of graphics three-dimensional geometric reconstruction, and the method comprises the following steps: collecting three-dimensional watertight blood vessel geometric reconstruction data including medical image data, scanning ball image data and physiological data, and carrying out the preprocessing of the collected data; extracting blood vessel center line point set data from the preprocessed medical image data; in combination with the preprocessed physiological data and a multiple linear regression algorithm, a radius correction coefficient is obtained, and then blood vessel center line point set data is updated; according to the method, the multiple linear regression technology, the implicit function form fitting technology and the scanning ball modeling technology in the method are closely combined with the modern information technology, and the three-dimensional watertight blood vessel geometric model is obtained. The intelligent degree in the three-dimensional watertight blood vessel geometric reconstruction process based on the scanning ball method is obviously enhanced.

Description

Technical Field

[0001] The present invention relates to the technical field of 3D geometric reconstruction in graphics, and particularly to a 3D watertight vascular geometric reconstruction method based on the scanning sphere method. Background Art

[0002] In the medical field, vascular geometric reconstruction is of great significance for disease diagnosis, surgical planning, and biomedical research. Traditional vascular geometric reconstruction methods, such as Delaunay triangulation and Alpha shape method, have many limitations when dealing with complex vascular networks. Blood vessels have the characteristics of multiple bifurcations and curvatures, and these complex structures make it difficult for traditional methods to accurately describe. In areas where the blood vessel diameter changes significantly, the models generated by traditional methods are prone to deviations and cannot truly reflect the actual morphology of blood vessels. At the same time, traditional methods often have problems of geometric discontinuity during the reconstruction process, resulting in holes or uneven connections in the models. This not only reduces the authenticity of the vascular model and makes it greatly discounted in subsequent applications. In addition, when dealing with the reconstruction of large-scale 3D vascular networks, the existing methods have a high computational complexity, require a large amount of time and computing resources, and are difficult to meet the real-time and high-efficiency clinical application requirements.

[0003] Although there have been great advancements in the direction of 3D watertight vascular geometric reconstruction in the prior art, there are still some problems to be optimized. Blood vessels have the characteristics of multiple bifurcations and curvatures. Traditional vascular geometric reconstruction techniques are difficult to accurately describe the characteristics of blood vessels and are prone to geometric discontinuity, resulting in problems of uneven connections in the constructed geometric models. Therefore, how to combine the scanning sphere method with implicit functions to perform geometric reconstruction on 3D watertight blood vessels is the problem to be solved by the present invention. For this reason, a 3D watertight vascular geometric reconstruction method based on the scanning sphere method is proposed. Summary of the Invention

[0004] To achieve the above objectives, the present invention is realized through the following technical solutions: A 3D watertight vascular geometric reconstruction method based on the scanning sphere method, including the following steps:

[0005] Step 1: Collect 3D watertight vascular geometric reconstruction data including medical image data, scanning sphere image data, and physiological data, and preprocess the collected data to provide a data basis for the implementation of subsequent steps;

[0006] Step 2: Extract the vascular centerline point set data from the preprocessed medical image data;

[0007] Step 3: Combine the preprocessed physiological data with the multiple linear regression algorithm to obtain the radius correction coefficient, and then update the vascular centerline point set data;

[0008] Step 4: Based on the updated centerline point set data and the scanning sphere image data, generate the vascular center spline curve, and calculate the vascular shape fitting implicit function and the scanning sphere shape fitting implicit function of each vascular center spline curve;

[0009] Step 5: Through the generated vascular center spline curve, combined with the scanning sphere method, construct the local geometric model of the blood vessel;

[0010] Step 6: Through the vascular shape fitting implicit function, fuse each local geometric model of the blood vessel to obtain a three-dimensional watertight vascular geometric model.

[0011] A further improvement of the technical solution of the present invention is that in the step 1, the acquisition process of the three-dimensional watertight vascular geometric reconstruction data includes:

[0012] Deploy different types of acquisition devices to acquire medical image data, scanning sphere image data and physiological data. The acquisition devices include CT angiography devices, three-dimensional laser scanners, 3D laser scanners, electronic sphygmomanometers, electrocardiographs, blood analyzers and rheometers;

[0013] The medical image data is the two-dimensional slice image of the blood vessel and the two-dimensional slice layer spacing; the scanning sphere image data is the three-dimensional image of the scanning sphere; the physiological data includes the blood pressure, heart rate, hematocrit and blood viscosity of the blood vessel detector;

[0014] Perform image denoising and image enhancement processing on the acquired two-dimensional slice image of the blood vessel and the three-dimensional image of the scanning sphere, perform data cleaning and data standardization processing on the acquired two-dimensional slice layer spacing and physiological data, integrate the physiological data, generate a three-dimensional watertight vascular geometric reconstruction data set, and divide the three-dimensional watertight vascular geometric reconstruction data set into a training set and a test set, and the ratio of the training set to the test set is 7:3.

[0015] A further improvement of the technical solution of the present invention is that in the step 2, the extraction process of the vascular centerline point set data includes:

[0016] The vascular centerline point set data includes the three-dimensional coordinates and radii of each centerline point in the blood vessel, the blood vessel width in the three-dimensional coordinate direction, and the three-dimensional coordinates of the blood vessel bifurcation points;

[0017] Perform grayscale processing on the acquired two-dimensional slice image of the blood vessel, and use the threshold segmentation algorithm. According to the gray value difference between the blood vessel and its surrounding tissues, set the segmentation threshold to segment the blood vessel area from the two-dimensional slice image of the blood vessel;

[0018] Adopt the Zhang-Sune thinning algorithm to thin the blood vessel area into a single-pixel-width line, and the pixel points on the single-pixel-width line are the vascular centerline points;

[0019] Based on the distance transformation method, calculate the distance from each pixel point on the single-pixel-width line to the blood vessel boundary, take the maximum distance from each pixel point on the single-pixel-width line to the blood vessel boundary as the radius of each centerline point in the blood vessel, and use the single-pixel-width line corresponding to this maximum distance as the blood vessel centerline;

[0020] Perform image registration on the two-dimensional slice image of the blood vessel, extract the centerline points of the blood vessel in each slice, establish a three-dimensional spatial model of the blood vessel, determine the X-axis, Y-axis, and Z-axis of the three-dimensional spatial model of the blood vessel, use the two-dimensional slice image of the blood vessel as the section parallel to the XY plane, and use the Z-axis direction as the stacking direction of the two-dimensional slices. The slice interval is the unit length in the Z-axis direction, and then obtain the three-dimensional coordinates of each centerline point in the blood vessel and the three-dimensional coordinates of the blood vessel bifurcation points;

[0021] Use Mimics software to measure the blood vessel widths in the X-axis, Y-axis, and Z-axis directions, and then obtain the blood vessel widths in the three-dimensional coordinate directions.

[0022] A further improvement of the technical solution of the present invention lies in: in the third step, the process of obtaining the radius correction coefficient includes:

[0023] Extract the physiological data from the three-dimensional watertight blood vessel geometric reconstruction dataset. Using the training set data and the multiple linear regression algorithm, take the physiological data as the input and the radius correction coefficient as the output, learn the linear relationship between the physiological data and the radius correction coefficient, and train the blood vessel centerline point set data correction model;

[0024] Input the test set data into the blood vessel centerline point set data correction model, adjust the intercept term and regression coefficients of the blood vessel centerline point set data correction model, optimize the performance of the blood vessel centerline point set data correction model, obtain the final blood vessel centerline point set data correction model, and combine the physiological data to output the corresponding radius correction coefficient;

[0025] The expression of this blood vessel centerline point set data correction model is as follows:

[0026] U = α0 + α1u1 + α2u2 + α3u3 + α4u4 + θ

[0027] Where U is the radius correction coefficient, α1, α2, α3, and α4 are the regression coefficients of the blood pressure, heart rate, hematocrit, and blood viscosity of the blood vessel detector respectively, u1, u2, u3, and u4 are the blood pressure, heart rate, hematocrit, and blood viscosity of the blood vessel detector respectively, and α0 and θ are the intercept term and error term of the blood vessel centerline point set data correction model respectively.

[0028] A further improvement of the technical solution of the present invention lies in: in the third step, the process of updating the blood vessel centerline point set data includes:

[0029] Set a first radius correction threshold and a second radius correction threshold, and correct the radius of each centerline point in the blood vessel according to the output result of the correction model for the blood vessel centerline point set data;

[0030] Specifically, when the radius correction coefficient is lower than the first radius correction threshold, the radius of each centerline point in the blood vessel is not corrected;

[0031] When the radius correction coefficient is between the first radius correction threshold and the second radius correction threshold, calculate the absolute value of the difference between the radius correction coefficient and 1, and use the product of the radius of each centerline point in the blood vessel and the calculated absolute value to replace the radius of the corresponding centerline point in the blood vessel;

[0032] When the radius correction coefficient is higher than the second radius correction threshold, calculate the product of the radius of each centerline point in the blood vessel and the radius correction coefficient, and use the calculated result to replace the radius of the corresponding centerline point in the blood vessel;

[0033] Use the radius of each centerline point in the corrected blood vessel to update the radius of each centerline point in the blood vessel centerline point set data, so as to realize the update of the blood vessel centerline point set data.

[0034] A further improvement of the technical solution of the present invention is that in the fourth step, the generation process of the blood vessel center spline curve includes:

[0035] Select the cubic spline interpolation method, and for the three-dimensional coordinates of each centerline point in the blood vessel, interpolate the X-axis, Y-axis, and Z-axis coordinates respectively, and fit the smooth curve functions about the X-axis, Y-axis, and Z-axis respectively, so as to generate the blood vessel center spline curve;

[0036] According to the three-dimensional coordinates of the blood vessel bifurcation point, retrieve the corresponding position in the generated blood vessel center spline curve, and calculate the distance from the blood vessel bifurcation point to each centerline point in the blood vessel. The calculation process includes:

[0037]

[0038] Among them, d0 is the distance from the blood vessel bifurcation point to each centerline point in the blood vessel, (x i , y i , z i ) is the three-dimensional coordinates of the blood vessel bifurcation point, (x j , y j , z j ) are the three-dimensional coordinates of each centerline point in the blood vessel;

[0039] Determine the position of the vascular bifurcation point on the vascular center spline curve based on the distances from the vascular bifurcation point to each centerline point in the blood vessel. According to the vascular bifurcation point, divide the vascular center spline curve into multiple independent line segments, and each line segment represents a branch in the vascular center spline curve.

[0040] A further improvement of the technical solution of the present invention lies in: in the step four, the calculation process of the vascular shape fitting implicit function of each vascular center spline curve includes:

[0041] Obtain the three-dimensional coordinates of each vascular center spline curve point through the vascular three-dimensional space model, use the three-dimensional coordinates of the vascular center spline curve point as the center point of the Gaussian distribution, and then obtain the three-dimensional coordinates of the Gaussian distribution points;

[0042] Combined with the vascular width distributed in the three-dimensional coordinate directions, calculate the vascular shape fitting implicit function of each vascular center spline curve, and its calculation process is as follows:

[0043]

[0044] Among them, F1(x, y, z) is the vascular shape fitting implicit function, (x0, y0, z0) is the three-dimensional coordinates of the Gaussian distribution point, and (σ x , σ y , σ z ) is the vascular width distributed in the three-dimensional coordinate directions.

[0045] A further improvement of the technical solution of the present invention lies in: in the step four, the calculation process of the scanning sphere shape fitting implicit function includes:

[0046] Use MeshLab software to extract the center coordinates and radius of the scanning sphere from the three-dimensional image of the scanning sphere;

[0047] According to the geometric definition of the three-dimensional space sphere, calculate the scanning sphere shape fitting implicit function, and its calculation process is as follows:

[0048] F2(x, y, z) = (x - x t ) 2 + (y - y t ) 2 + (z - z t ) 2 - r 2

[0049] Among them, F2(x, y, z) is the scanning sphere shape fitting implicit function, (x t , y t , z t ) is the center coordinates of the scanning sphere, and r is the radius of the scanning sphere.

[0050] A further improvement of the technical solution of the present invention lies in: in the step five, the construction process of the local vascular geometric model includes:

[0051] Compare the three-dimensional coordinates and radii of each centerline point in the blood vessel with the center coordinates and radius of the scanning sphere respectively, and select the matching scanning sphere;

[0052] Set the morphological fitting implicit function of the scanning sphere to zero, and solve the equation to obtain the coordinates of the points on the surface of the scanning sphere. According to the matching between the scanning sphere and the blood vessel, the coordinates of the points on the surface of the scanning sphere are used as the coordinates of the points on the blood vessel wall;

[0053] Based on the coordinates of the points on the blood vessel wall and the center coordinates of the scanning sphere, calculate the distance from the points on the blood vessel wall to the center of the scanning sphere, and establish a linear transformation function. Specifically, the calculation process of the distance from the points on the blood vessel wall to the center of the scanning sphere and the establishment process of the linear transformation function include:

[0054]

[0055] r new = r + k(D - r)

[0056] where D is the distance from the point on the blood vessel wall to the center of the scanning sphere; r new is the radius of the scanning sphere after dynamic adjustment; (x p , y p , z p ) are the coordinates of the points on the blood vessel wall; (x t , y t , z t ) are the center coordinates of the scanning sphere; r is the radius of the scanning sphere; k is the proportionality coefficient;

[0057] Traverse several points on the blood vessel wall, calculate the radius of the scanning sphere after adjustment according to the linear transformation function, and dynamically adjust the radius of the scanning sphere as the scanning sphere moves along the central spline curve of the blood vessel, thereby constructing a local vascular geometric model to ensure the fitting degree between the scanning sphere and the blood vessel wall.

[0058] A further improvement of the technical solution of the present invention lies in: in the step six, the process of obtaining the three-dimensional watertight vascular geometric model includes:

[0059] Use the morphological fitting implicit function of each central spline curve of the blood vessel to fuse the local vascular geometric model to generate a global vascular geometric model;

[0060] Combine the mesh generation algorithm to convert the global vascular geometric model into discrete grid cells, fit the boundaries of each discrete grid cell, and refine and coarsen each discrete grid cell according to the requirement of the analysis accuracy of the global vascular geometric model for subsequent medical analysis, thereby obtaining a three-dimensional watertight vascular geometric model.

[0061] The beneficial effects of the present invention are as follows: In the three-dimensional watertight vascular geometry reconstruction method based on the scanning sphere method of the present invention, compared with the traditional three-dimensional watertight vascular geometry reconstruction method, the multiple linear regression technology, implicit function morphology fitting technology, scanning sphere modeling technology in the method of the present invention are closely combined with modern information technology to accurately capture medical image data, scanning sphere image data and physiological data, and then obtain the radius correction coefficient. Using the radius correction coefficient, the vascular centerline point set data is updated, and the vascular local geometry model is constructed through the implicit function of the scanning sphere morphology fitting. Combining the implicit function of the vascular morphology fitting of each vascular center spline curve with the mesh generation algorithm, a three-dimensional watertight vascular geometry model is constructed, which solves the problems that the traditional method is difficult to accurately describe the multiple bifurcations and bending characteristics of blood vessels and is prone to geometric discontinuities, ensuring that the method in the present invention can refine the dynamic monitoring standard of the three-dimensional watertight vascular geometry reconstruction method based on the scanning sphere method within a more accurate range, making the monitored data a more accurate index under the same conditions. The research and application of this method significantly enhance the degree of intelligence in the three-dimensional watertight vascular geometry reconstruction process based on the scanning sphere method. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments recorded in the present invention, and those of ordinary skill in the art can also obtain other drawings based on these drawings.

[0063] Figure 1 It is a flowchart of the three-dimensional watertight vascular geometry reconstruction method based on the scanning sphere method of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0064] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0065] As Figure 1 shown, the present invention provides a three-dimensional watertight vascular geometry reconstruction method based on the scanning sphere method, which consists of the following steps:

[0066] Step 1: Collect three-dimensional watertight vascular geometry reconstruction data including medical image data, scanning sphere image data, and physiological data, and preprocess the collected data to provide a data basis for the implementation of subsequent steps;

[0067] Step 2: Extract the vascular centerline point set data from the preprocessed medical image data;

[0068] Step 3: Combine the preprocessed physiological data with the multiple linear regression algorithm to obtain the radius correction coefficient, and then update the vascular centerline point set data;

[0069] Step 4: Based on the updated centerline point set data and the scanning sphere image data, generate the vascular center spline curve, and calculate the vascular shape fitting implicit function and the scanning sphere shape fitting implicit function of each vascular center spline curve;

[0070] Step 5: Through the generated vascular center spline curve, combine with the scanning sphere method to construct the local vascular geometric model;

[0071] Step 6: Through the vascular shape fitting implicit function, fuse each local vascular geometric model to obtain the three-dimensional watertight vascular geometric model.

[0072] In Step 1, the acquisition process of the three-dimensional watertight vascular geometric reconstruction data includes:

[0073] Deploy different types of acquisition devices to acquire medical image data, scanning sphere image data and physiological data. The acquisition devices include CT angiography devices, three-dimensional laser scanners, 3D laser scanners, electronic sphygmomanometers, electrocardiographs, blood analyzers and rheometers;

[0074] Among them, the medical image data is the two-dimensional slice image of the blood vessel and the two-dimensional slice layer spacing; the scanning sphere image data is the three-dimensional image of the scanning sphere; the physiological data includes the blood pressure, heart rate, hematocrit, and blood viscosity of the blood vessel detector;

[0075] Perform image denoising and image enhancement processing on the acquired two-dimensional slice image of the blood vessel and the three-dimensional image of the scanning sphere, perform data cleaning and data standardization processing on the acquired two-dimensional slice layer spacing and physiological data, integrate the physiological data, generate the three-dimensional watertight vascular geometric reconstruction data set, and divide the three-dimensional watertight vascular geometric reconstruction data set into a training set and a test set, and the ratio of the training set to the test set is 7:3.

[0076] In Step 2, the extraction process of the vascular centerline point set data includes:

[0077] Among them, the vascular centerline point set data includes the three-dimensional coordinates and radii of each centerline point in the blood vessel, the blood vessel width in the three-dimensional coordinate direction, and the three-dimensional coordinates of the blood vessel bifurcation points;

[0078] The gray-scale processing is performed on the collected two-dimensional vascular slice images. Using the threshold segmentation algorithm, according to the gray-scale value difference between the blood vessels and their surrounding tissues, the segmentation threshold is set to segment the blood vessel region from the two-dimensional vascular slice images;

[0079] The Zhang-Suen thinning algorithm is adopted to thin the blood vessel region into single-pixel-width lines, and the pixel points on the single-pixel-width lines are the blood vessel centerline points;

[0080] Based on the distance transformation method, the distance from each pixel point on the single-pixel-width line to the blood vessel boundary is calculated, and the maximum distance from each pixel point on the single-pixel-width line to the blood vessel boundary is taken as the radius of each centerline point in the blood vessel, and the single-pixel-width line corresponding to the maximum distance is taken as the blood vessel centerline;

[0081] Image registration is performed on the two-dimensional vascular slice images, the blood vessel centerline points in each slice are extracted, a three-dimensional vascular spatial model is established, the X-axis, Y-axis, and Z-axis of the three-dimensional vascular spatial model are determined, the two-dimensional vascular slice images are used as the cross-sections parallel to the XY plane, the Z-axis direction is used as the stacking direction of the two-dimensional slices, and the slice spacing is the unit length in the Z-axis direction, so as to obtain the three-dimensional coordinates of each centerline point in the blood vessel and the three-dimensional coordinates of the blood vessel bifurcation points;

[0082] Using Mimics software, the blood vessel widths in the X-axis, Y-axis, and Z-axis directions are measured, so as to obtain the blood vessel widths in the three-dimensional coordinate directions.

[0083] In step three, the process of obtaining the radius correction coefficient includes:

[0084] Extract the physiological data from the three-dimensional watertight vascular geometry reconstruction dataset. Using the training set data and the multiple linear regression algorithm, taking the physiological data as the input and the radius correction coefficient as the output, learn the linear relationship between the physiological data and the radius correction coefficient, and train the blood vessel centerline point set data correction model;

[0085] Input the test set data into the blood vessel centerline point set data correction model, adjust the intercept term and regression coefficient of the blood vessel centerline point set data correction model, optimize the performance of the blood vessel centerline point set data correction model, obtain the final blood vessel centerline point set data correction model, and combine the physiological data to output the corresponding radius correction coefficient;

[0086] The expression of the blood vessel centerline point set data correction model is as follows:

[0087]

[0088] Among them, U is the radius correction coefficient, α1, α2, α3, and α4 are the regression coefficients of the blood pressure, heart rate, hematocrit, and blood viscosity of the vascular examiner respectively, u1, u2, u3, and u4 are the blood pressure, heart rate, hematocrit, and blood viscosity of the vascular examiner respectively, α0 and are the intercept term and the error term of the correction model of the vascular centerline point set data respectively.

[0089] In step three, the update process of the vascular centerline point set data includes:

[0090] Set the first radius correction threshold and the second radius correction threshold, and correct the radius of each centerline point in the blood vessel according to the output result of the correction model of the vascular centerline point set data;

[0091] Specifically, when the radius correction coefficient is lower than the first radius correction threshold, the radius of each centerline point in the blood vessel is not corrected;

[0092] When the radius correction coefficient is between the first radius correction threshold and the second radius correction threshold, calculate the absolute value of the difference between the radius correction coefficient and 1, and use the product of the radius of each centerline point in the blood vessel and the calculated absolute value to replace the radius of the corresponding centerline point in the blood vessel;

[0093] When the radius correction coefficient is higher than the second radius correction threshold, calculate the product of the radius of each centerline point in the blood vessel and the radius correction coefficient, and use the calculated result to replace the radius of the corresponding centerline point in the blood vessel;

[0094] Use the radius of each centerline point in the corrected blood vessel to update the radius of each centerline point in the vascular centerline point set data, so as to realize the update of the vascular centerline point set data.

[0095] In step four, the generation process of the vascular center spline curve includes:

[0096] Select the cubic spline interpolation method, and for the three-dimensional coordinates of each centerline point in the blood vessel, interpolate the X-axis, Y-axis, and Z-axis coordinates respectively, and fit the smooth curve functions about the X-axis, Y-axis, and Z-axis respectively, so as to generate the vascular center spline curve;

[0097] According to the three-dimensional coordinates of the vascular bifurcation point, retrieve the corresponding position in the generated vascular center spline curve, and calculate the distance from the vascular bifurcation point to each centerline point in the blood vessel. The calculation process includes:

[0098]

[0099] Among them, d0 is the distance from the vascular bifurcation point to each centerline point in the blood vessel, (x i , y i , z i) are the three-dimensional coordinates of the blood vessel bifurcation point, (x j , y j , z j ) are the three-dimensional coordinates of each centerline point in the blood vessel;

[0100] Determine the position of the blood vessel bifurcation point on the blood vessel center spline curve by the distance from the blood vessel bifurcation point to each centerline point in the blood vessel. Based on the blood vessel bifurcation point, the blood vessel center spline curve is segmented into multiple independent line segments, and each line segment represents a branch in the blood vessel center spline curve.

[0101] In step four, the calculation process of the blood vessel shape fitting implicit function of each blood vessel center spline curve includes:

[0102] Obtain the three-dimensional coordinates of each blood vessel center spline curve point through the blood vessel three-dimensional space model, and use the three-dimensional coordinates of the blood vessel center spline curve point as the center point of the Gaussian distribution, and then obtain the three-dimensional coordinates of the Gaussian distribution points;

[0103] Combine the blood vessel widths distributed in the three-dimensional coordinate directions to calculate the blood vessel shape fitting implicit function of each blood vessel center spline curve. The calculation process is as follows:

[0104]

[0105] Among them, F1(x, y, z) is the blood vessel shape fitting implicit function, (x0, y0, z0) are the three-dimensional coordinates of the Gaussian distribution points, (σ x , σ y , σ z ) are the blood vessel widths distributed in the three-dimensional coordinate directions.

[0106] In step four, the calculation process of the scanning sphere shape fitting implicit function includes:

[0107] Use MeshLab software to extract the center coordinates and radius of the scanning sphere from the three-dimensional image of the scanning sphere;

[0108] According to the geometric definition of the three-dimensional space sphere, calculate the scanning sphere shape fitting implicit function. The calculation process is as follows:

[0109] F2(x, y, z) = (x - x t ) 2 + (y - y t ) 2 + (z - z t ) 2 - r 2

[0110] Among them, F2(x, y, z) is the scanning sphere shape fitting implicit function, (x t , y t , zt ) is the central coordinate of the scanning sphere, and r is the radius of the scanning sphere.

[0111] In step five, the construction process of the local vascular geometric model includes:

[0112] Compare the three-dimensional coordinates and radii of each centerline point in the blood vessel with the central coordinate and radius of the scanning sphere respectively, and select the matching scanning sphere;

[0113] Let the morphological fitting implicit function of the scanning sphere be equal to zero. By solving the equation, obtain the coordinates of the points on the surface of the scanning sphere. According to the matching between the scanning sphere and the blood vessel, take the coordinates of the points on the surface of the scanning sphere as the coordinates of the points on the blood vessel wall;

[0114] Based on the coordinates of the points on the blood vessel wall and the central coordinate of the scanning sphere, calculate the distance from the points on the blood vessel wall to the center of the scanning sphere, and establish a linear transformation function. Specifically, the calculation process of the distance from the points on the blood vessel wall to the center of the scanning sphere and the establishment process of the linear transformation function include:

[0115]

[0116] r new = r + k(D - r)

[0117] where D is the distance from the point on the blood vessel wall to the center of the scanning sphere; r new is the radius of the scanning sphere after dynamic adjustment; (x p , y p , z p ) are the coordinates of the points on the blood vessel wall; (x t , y t , z t ) are the central coordinates of the scanning sphere; r is the radius of the scanning sphere; k is the proportionality coefficient;

[0118] Traverse several points on the blood vessel wall, calculate the radius of the adjusted scanning sphere according to the linear transformation function, and dynamically adjust the radius of the scanning sphere as the scanning sphere moves along the central spline curve of the blood vessel, thereby constructing a local vascular geometric model to ensure the fitting degree between the scanning sphere and the blood vessel wall.

[0119] In step six, the process of obtaining the three-dimensional watertight vascular geometric model includes:

[0120] Use the morphological fitting implicit function of each central spline curve of the blood vessel to fuse the local vascular geometric model to generate a global vascular geometric model;

[0121] Combined with the grid generation algorithm, the global vascular geometry model is transformed into discrete grid cells, the boundaries of each discrete grid cell are fitted, and according to the requirements of the global vascular geometry model for analysis accuracy, each discrete grid cell is refined and coarsened for subsequent medical analysis, thereby obtaining a three-dimensional watertight vascular geometry model.

[0122] As described above, only the specific implementation manners of the present application are provided, but the protection scope of the present application is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present application can easily think of changes or substitutions, which should all be covered within the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.

Claims

1. A three-dimensional watertight vascular geometry reconstruction method based on the scanning sphere method, characterized in that: It includes the following steps: Step 1: Collect three-dimensional watertight vascular geometric reconstruction data including medical image data, scanning sphere image data, and physiological data, and preprocess the collected data; Step 2: Extract vascular centerline point set data from the preprocessed medical image data; Step 3: Combine the preprocessed physiological data with the multiple linear regression algorithm to obtain the radius correction coefficient, and then update the vascular centerline point set data; Step 4: Based on the updated centerline point set data and scanning sphere image data, generate vascular center spline curves, and calculate the vascular shape fitting implicit function and scanning sphere shape fitting implicit function of each vascular center spline curve; Step 5: Through the generated vascular center spline curves, combine with the scanning sphere method to construct a local vascular geometric model; Step 6: Through the vascular shape fitting implicit function, fuse each local vascular geometric model to obtain a three-dimensional watertight vascular geometric model.

2. The three-dimensional watertight vascular geometry reconstruction method based on the scanning sphere method according to claim 1, wherein: In the above Step 1, the collection process of the three-dimensional watertight vascular geometric reconstruction data includes: Deploy different types of collection devices to collect medical image data, scanning sphere image data, and physiological data. The collection devices include CT angiography devices, three-dimensional laser scanners, 3D laser scanners, electronic sphygmomanometers, electrocardiographs, blood analyzers, and blood rheometers; The medical image data is the two-dimensional slice images of blood vessels and the two-dimensional slice layer spacing; the scanning sphere image data is the three-dimensional image of the scanning sphere; the physiological data includes the blood pressure, heart rate, hematocrit, and blood viscosity of the blood vessel detector; Perform image denoising and image enhancement processing on the collected two-dimensional slice images of blood vessels and three-dimensional scanning sphere images, perform data cleaning and data standardization processing on the collected two-dimensional slice layer spacing and physiological data, integrate the physiological data, generate a three-dimensional watertight vascular geometric reconstruction data set, and divide the three-dimensional watertight vascular geometric reconstruction data set into a training set and a test set.

3. The three-dimensional watertight vascular geometry reconstruction method based on the scanning sphere method according to claim 2, characterized in that: In the above Step 2, the extraction process of the vascular centerline point set data includes: The vascular centerline point set data includes the three-dimensional coordinates and radii of each centerline point in the blood vessel, the blood vessel width in the three-dimensional coordinate direction, and the three-dimensional coordinates of the blood vessel bifurcation points; Perform grayscale processing on the collected two-dimensional slice images of blood vessels, and use the threshold segmentation algorithm. According to the gray value difference between the blood vessel and its surrounding tissues, set the segmentation threshold to segment the blood vessel area from the two-dimensional slice images of blood vessels; Adopt the Zhang-Sune thinning algorithm to thin the blood vessel area into a single-pixel-width line, and the pixel points on this single-pixel-width line are the vascular centerline points; Based on the distance transformation method, calculate the distance from each pixel point on the single-pixel-width line to the blood vessel boundary, and take the maximum distance from each pixel point on the single-pixel-width line to the blood vessel boundary as the radius of each centerline point in the blood vessel, and use the single-pixel-width line corresponding to this maximum distance as the vascular centerline; Perform image registration on two-dimensional vascular slice images, extract the centerline points of blood vessels in each slice, establish a three-dimensional vascular spatial model, determine the X-axis, Y-axis, and Z-axis of the three-dimensional vascular spatial model, use the two-dimensional vascular slice image as a cross-section parallel to the XY plane, and use the Z-axis direction as the stacking direction of the two-dimensional slices. The slice spacing is the unit length in the Z-axis direction, and then obtain the three-dimensional coordinates of each centerline point in the blood vessel and the three-dimensional coordinates of the blood vessel bifurcation points; Use Mimics software to measure the blood vessel widths in the X-axis, Y-axis, and Z-axis directions, and then obtain the blood vessel widths in the three-dimensional coordinate directions.

4. The three-dimensional watertight vascular geometry reconstruction method based on the scanning sphere method according to claim 3, characterized in that: In step three, the process of obtaining the radius correction coefficient includes: Extract the physiological data from the three-dimensional watertight vascular geometric reconstruction dataset. Using the training set data and the multiple linear regression algorithm, take the physiological data as the input and the radius correction coefficient as the output, learn the linear relationship between the physiological data and the radius correction coefficient, and train the blood vessel centerline point set data correction model; Input the test set data into the blood vessel centerline point set data correction model, adjust the intercept term and regression coefficient of the blood vessel centerline point set data correction model, optimize the performance of the blood vessel centerline point set data correction model, obtain the final blood vessel centerline point set data correction model, and combine the physiological data to output the corresponding radius correction coefficient.

5. The three-dimensional watertight blood vessel geometric reconstruction method based on the scanning sphere method according to claim 4, characterized in that: In step three, the process of updating the blood vessel centerline point set data includes: Set the first radius correction threshold and the second radius correction threshold, and correct the radius of each centerline point in the blood vessel according to the output result of the blood vessel centerline point set data correction model; Use the radius of each centerline point in the corrected blood vessel to update the radius of each centerline point in the blood vessel centerline point set data, and realize the update of the blood vessel centerline point set data.

6. The three-dimensional watertight vascular geometry reconstruction method based on the scanning sphere method according to claim 5, characterized in that: In step four, the process of generating the blood vessel center spline curve includes: Select the cubic spline interpolation method. For the three-dimensional coordinates of each centerline point in the blood vessel, interpolate the X-axis, Y-axis, and Z-axis coordinates respectively, and fit the smooth curve functions about the X-axis, Y-axis, and Z-axis respectively, and then generate the blood vessel center spline curve; According to the three-dimensional coordinates of the blood vessel bifurcation points, retrieve the corresponding positions in the generated blood vessel center spline curve, and calculate the distances from the blood vessel bifurcation points to each centerline point in the blood vessel; Determine the positions of the blood vessel bifurcation points in the blood vessel center spline curve through the distances from the blood vessel bifurcation points to each centerline point in the blood vessel. Based on the blood vessel bifurcation points, divide the blood vessel center spline curve into multiple independent line segments, and each line segment represents a branch in the blood vessel center spline curve.

7. The three-dimensional watertight vascular geometry reconstruction method based on the scanning sphere method according to claim 6, wherein: In step four, the calculation process of the blood vessel shape fitting implicit function of each blood vessel center spline curve includes: Through the three-dimensional vascular spatial model, obtain the three-dimensional coordinates of each blood vessel center spline curve point, and use the three-dimensional coordinates of the blood vessel center spline curve point as the center point of the Gaussian distribution, and then obtain the three-dimensional coordinates of the Gaussian distribution points; Combine the blood vessel widths in the three-dimensional coordinate directions to calculate the blood vessel shape fitting implicit function of each blood vessel center spline curve. The calculation process is as follows: Among them, F1(x, y, z) is the implicit function for fitting the blood vessel morphology, (x0, y0, z0) is the three-dimensional coordinate of the Gaussian distribution point, and (σ x , σ y , σ z ) is the blood vessel width distributed in the three-dimensional coordinate directions.

8. The three-dimensional watertight vascular geometry reconstruction method based on the scanning sphere method according to claim 7, wherein: In step four, the calculation process of the scanning sphere shape fitting implicit function includes: Using MeshLab software, extract the center coordinates and radius of the scanning sphere from the three-dimensional image of the scanning sphere; According to the geometric definition of a sphere in three-dimensional space, calculate the implicit function for fitting the shape of the scanning sphere. The calculation process is as follows: F2(x, y, z) = (x - x t ) 2 + (y - y t ) 2 + (z - z t ) 2 - r 2 Among them, F2(x, y, z) is the implicit function for fitting the shape of the scanning sphere, (x t , y t , z t ) is the center coordinate of the scanning sphere, and r is the radius of the scanning sphere.

9. The three-dimensional watertight vascular geometry reconstruction method based on the scanning sphere method according to claim 8, characterized in that: In the fifth step described above, the process of constructing the local geometric model of the blood vessel includes: Compare the three-dimensional coordinates and radii of each centerline point in the blood vessel with the center coordinates and radius of the scanning sphere respectively, and select the matching scanning sphere; Set the implicit function for fitting the shape of the scanning sphere to zero, and by solving the equation, obtain the coordinates of the points on the surface of the scanning sphere. According to the matching between the scanning sphere and the blood vessel, take the coordinates of the points on the surface of the scanning sphere as the coordinates of the points on the blood vessel wall; Based on the coordinates of the points on the blood vessel wall and the center coordinates of the scanning sphere, calculate the distance from the points on the blood vessel wall to the center of the scanning sphere, and establish a linear transformation function; Traverse several points on the blood vessel wall, and according to the linear transformation function, calculate the radius of the adjusted scanning sphere. As the scanning sphere moves along the central spline curve of the blood vessel, dynamically adjust the radius of the scanning sphere, and then construct the local geometric model of the blood vessel.

10. The three-dimensional watertight vascular geometry reconstruction method based on the scanning sphere method according to claim 9, wherein: In the sixth step described above, the process of obtaining the three-dimensional watertight blood vessel geometric model includes: Use the implicit function for fitting the blood vessel shape of each central spline curve of the blood vessel to fuse the local geometric model of the blood vessel and generate the global blood vessel geometric model; Combined with the mesh generation algorithm, convert the global blood vessel geometric model into discrete grid cells, fit the boundaries of each discrete grid cell, and perform refinement and coarsening processing on each discrete grid cell according to the requirement of the analysis accuracy of the global blood vessel geometric model, so as to obtain the three-dimensional watertight blood vessel geometric model.

Images