A blood vessel geometry model repairing method based on point cloud data

Abstract

This invention provides a method for repairing vascular geometric models based on point cloud data, comprising: acquiring original three-dimensional data; smoothing the original three-dimensional data using a Gaussian smoothing filter to obtain smoothed data; constructing a vascular three-dimensional model based on the smoothed data; extracting the vascular skeleton from the vascular three-dimensional model using an exploration operator; repairing missing vascular models in the middle based on point cloud data; and repairing missing vascular models at the ends based on duct data. This invention achieves the repair of fractured vascular models, greatly improving the accuracy of vascular reconstruction and bringing significant benefits to computer-aided diagnosis of hepatic vascular diseases.

Description

Technical Field

[0001] This invention relates to the fields of computer-aided design and computer graphics, and in particular to a method for repairing a blood vessel geometric model based on point cloud data. Background Technology

[0002] Hepatic vascular diseases are seriously threatening human life and health, with a large number of new cases appearing every year, showing a continuous upward trend. Clinicians typically rely on computer-aided systems when diagnosing patients.

[0003] The application of computer-aided diagnosis in medical imaging has been demonstrated by relevant research, highlighting the crucial role of timely interventional treatment in the early stages of liver cancer in effectively reducing its incidence. Initially, interventional surgery was the preferred treatment method; however, after prolonged use, it has been found that this technique requires precise tumor location. Without adequate preoperative training and intraoperative planning, surgeons cannot perform such precise surgical tasks.

[0004] Currently, existing technologies disclose a scheme for repairing the micro-channels of vascular trees. This scheme involves user-interactive selection of interpolation methods to calculate discrete functions, and then using a nonlinear 3D reconstruction method based on fractals to ultimately repair the missing vascular model. However, during vascular segmentation, extremely small radii within the vessels can easily lead to partial or complete loss of the reconstructed vascular model, thus affecting the effectiveness of vascular reconstruction and subsequent processing such as blood flow analysis. Summary of the Invention

[0005] To overcome the shortcomings of existing technologies, the purpose of this invention is to provide a method for repairing vascular geometric models based on point cloud data.

[0006] To achieve the above objectives, the present invention provides the following solution:

[0007] A method for repairing blood vessel geometric models based on point cloud data includes:

[0008] Obtain the raw 3D data;

[0009] Based on the Gaussian smoothing filter method, the original three-dimensional data is smoothed to obtain smoothed data;

[0010] A three-dimensional model of blood vessels is constructed based on the smoothed data;

[0011] The vascular skeleton in the three-dimensional vascular model is extracted based on the exploration operator;

[0012] A point cloud-based model for repairing missing blood vessel structures;

[0013] Based on a model of missing blood vessels at the end of a duct repair.

[0014] Preferably, following the model of repairing missing blood vessels at the end of a duct, the method further includes:

[0015] A visual blood vessel model is generated based on the repaired model.

[0016] Preferably, the point cloud-based method for repairing missing blood vessel models includes:

[0017] Based on the surface rendering algorithm, the three-dimensional model volume data of both ends of the missing blood vessel is extracted to obtain the isosurface mesh;

[0018] The isosurface mesh is smoothed using the Laplace algorithm, and each point is replaced by the center of its neighborhood points to obtain a smooth mesh.

[0019] Based on the geometry of a smooth mesh, the vertices of the mesh are extracted to obtain a 3D point cloud;

[0020] Based on the three-dimensional point cloud, the center point of the three-dimensional point cloud is obtained by using the average value method to obtain the seed point;

[0021] Segmentation based on level set segmentation combined with seed points is used to repair missing blood vessel models.

[0022] Preferably, the model based on the missing blood vessel at the end of the duct repair includes:

[0023] Based on the existing 3D model of the blood vessel and the blood vessel skeleton, the end points of the channel are obtained;

[0024] Based on the existing 3D vascular model and the vascular skeleton, 2D slices are performed to calculate the vascular diameter.

[0025] Based on the vascular skeleton, calculate the local coordinate system of points on the skeleton line;

[0026] Based on the local coordinate system and the blood vessel diameter in the nb plane, the point cloud of the pipeline is obtained, and the missing blood vessel model is repaired.

[0027] According to specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0028] This invention provides a method for repairing vascular geometric models based on point cloud data, comprising: acquiring original three-dimensional data; smoothing the original three-dimensional data using a Gaussian smoothing filter to obtain smoothed data; constructing a vascular three-dimensional model based on the smoothed data; extracting the vascular skeleton from the vascular three-dimensional model using an exploration operator; repairing missing vascular models in the middle based on point cloud data; and repairing missing vascular models at the ends based on duct data. This invention achieves the repair of fractured vascular models, greatly improving the accuracy of vascular reconstruction and bringing significant benefits to computer-aided diagnosis of hepatic vascular diseases. Attached Figure Description

[0029] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0030] Figure 1 A flowchart of the method provided in an embodiment of the present invention;

[0031] Figure 2 This is a schematic diagram of the steps provided in the embodiments of the present invention;

[0032] Figure 3 This is a schematic diagram of the first final reconstruction result provided for an embodiment of the present invention;

[0033] Figure 4 This is a second schematic diagram of the final reconstruction result provided in an embodiment of the present invention;

[0034] Figure 5 The third schematic diagram of the final reconstruction result provided in this embodiment of the invention;

[0035] Figure 6 This is a fourth schematic diagram of the final reconstruction result provided in an embodiment of the present invention. Detailed Implementation

[0036] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0037] The purpose of this invention is to provide a method for repairing vascular geometric models based on point cloud data, which realizes the repair of ruptured vascular models, greatly improves the accuracy of vascular reconstruction, and brings significant benefits to computer-aided diagnosis of liver vascular diseases.

[0038] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0039] Figure 1 The method flowchart provided in the embodiments of the present invention is as follows: Figure 1 As shown, this invention provides a method for repairing a blood vessel geometric model based on point cloud data, including:

[0040] Step 100: Obtain the raw 3D data;

[0041] Step 200: Based on the Gaussian smoothing filter method, smooth the original three-dimensional data to obtain smoothed data;

[0042] Step 300: Construct a three-dimensional model of blood vessels based on the smoothed data;

[0043] Step 400: Extract the vascular skeleton from the 3D vascular model based on the exploration operator;

[0044] Step 500: Repair the missing blood vessel model in the middle based on point cloud;

[0045] Step 600: Based on the model of missing blood vessels at the end of the duct repair.

[0046] Figure 2 This is a schematic diagram of the steps provided in the embodiments of the present invention, as shown below. Figure 2 As shown, this embodiment uses the original data and the existing three-dimensional blood vessel model as the data input of the present invention, and repairs blood vessels based on three-dimensional point clouds and channels.

[0047] The specific steps are as follows:

[0048] S1. Obtain the raw 3D data;

[0049] S2. Smooth the original 3D data based on the Gaussian smoothing filter method;

[0050] S3. Read the existing 3D blood vessel model;

[0051] S4. Extracting the vascular skeleton based on the exploration operator;

[0052] S5. Point cloud-based model for repairing missing blood vessels in the middle;

[0053] S6. Based on a model of missing vessels at the end of a duct repair procedure;

[0054] S7. Generate blood vessel models based on visualization.

[0055] Optionally, in step S3, the existing 3D model is stored in an STL file format. This is an ASCII STL model, a text-based format for storing 3D models. It consists of a series of triangular facets, each facet having a normal vector and three vertex coordinates. The specific steps are as follows:

[0056] Open an ASCII STL file

[0057] Read the file line by line and parse it into the coordinates of the three vertices of a triangle and the normal vector. Combine the parsed vertices and normal vectors to form a triangle.

[0058] Add the parsed triangles to the triangle list.

[0059] Specifically, step S4 involves using existing exploration operators as control points, multiplying each control point by its weight and basis function, and then summing them to obtain a curve, which is the skeleton line.

[0060] The specific method for repairing the missing blood vessel model at both ends of the point cloud is as follows:

[0061] S51. Based on MarchingCubes, extract the three-dimensional model volume data of both ends of the missing blood vessel to obtain the isosurface mesh;

[0062] The steps for extracting the volumetric data of the three-dimensional model at both ends of the missing blood vessel include:

[0063] 1. Divide the three-dimensional voxel data into many small cubic units;

[0064] 2. Process each cubic unit to determine the location and shape of its isosurface.

[0065] 3. Use a triangular network to connect adjacent isosurfaces to generate a complete isosurface model.

[0066] In step 2, the isosurfaces of the cube are determined in the following way:

[0067] Compare the eight vertices of the cube with isosurfaces to determine which vertices the isosurfaces pass through. Mark each vertex as either "on the isosurface" or "not on the isosurface".

[0068] Based on the state of the eight vertices, determine the intersection of the isosurface with the cube, and determine the location and shape of the intersection point of the isosurface.

[0069] Based on the location and shape of the intersection points, a triangular network is generated to connect adjacent isosurfaces.

[0070] By repeating the above steps, a complete isosurface model can be extracted from the three-dimensional voxel data.

[0071] S52. Smooth the mesh based on Laplace's algorithm by replacing each point with the center of its neighborhood points to obtain a smoother mesh;

[0072] Laplacian smoothing achieves a smoothing effect by updating the position of each vertex in the mesh by taking a weighted average of the coordinates of its neighboring vertices. For each vertex in the network, the average coordinates of the vertices surrounding it are calculated as the new position, and the vertex is moved there.

[0073] S53. Based on the mesh geometry, extract the vertices of the mesh to obtain a 3D point cloud;

[0074] S54. Based on the 3D point cloud, use the average value method to find the center point of the 3D point cloud and obtain the seed point;

[0075] S55. Segmentation based on level set segmentation combined with seed points to repair missing blood vessel models.

[0076] Specifically, the original medical images are segmented into level sets, and the specific steps are as follows:

[0077] 1. Initialize the level set function, which can be a constant value or a region manually marked by the user.

[0078] 2. Calculate the gradient of the level set function and update the level set function based on the magnitude and direction of the gradient. This can be achieved using the PDE (Partial Differential Equation) algorithm.

[0079] 3. Iterate through the second step until the level set function converges or the maximum number of iterations is reached.

[0080] 4. Extract the segmentation results. Based on the value of the level set function, the image can be divided into two parts: the target blood vessel and the background.

[0081] 5. Reconstruct the segmented results, namely the 3D point cloud data and the 3D point cloud obtained by S53, to repair the missing blood vessel model.

[0082] The specific method for repairing the missing blood vessel model at the end of the duct is as follows:

[0083] S61. Obtain the pipe endpoints based on the existing 3D vascular model and vascular skeleton;

[0084] S62. Based on the existing three-dimensional blood vessel model and the blood vessel skeleton, perform two-dimensional slicing and calculate the blood vessel diameter;

[0085] S63. Based on the vascular skeleton, calculate the local coordinate system of points on the skeleton line;

[0086] Specifically, step S63 involves calculating the local coordinate system of each point on the skeleton line, which includes:

[0087] Based on the skeleton lines, calculate the tangent vector at each point on the skeleton lines using calculus.

[0088] The normal vector, which is the vector perpendicular to the tangent plane, is calculated using the cross product.

[0089] The binormal vector is calculated by taking the cross product of the tangent vector and the normal vector. The binormal vector is calculated using the cross product of vectors.

[0090] Optionally, step S64 involves normalizing the tangent vector, normal vector, and binormal vector to obtain a unit vector, specifically including:

[0091] Calculate the ellipse in the nb (normal vector - binormal vector) plane:

[0092] Calculate the nb plane: For each point, the nb plane is the plane formed by the n normal vector and the b subnormal vector, obtained by using the n normal vector and the b subnormal vector as basis vectors.

[0093] Determine the center point of the ellipse: a point on the skeleton line.

[0094] Determine the major and minor axes of the ellipse: the major axis is the binormal vector, and the minor axis is the normal vector.

[0095] Determine the size of the ellipse: major semi-axis length r * 1.1, minor semi-axis length r

[0096] Draw an ellipse: Draw an ellipse on the nb plane based on its center point, major axis, minor axis, and size.

[0097] 1. Save the points on the ellipse, i.e., the point cloud of the pipeline.

[0098] 2. Downsample the point cloud:

[0099] The original point cloud is projected onto a 2D grid, and one point is selected in each grid to represent the downsampled point cloud. Specific steps:

[0100] Determine the size of the voxels and divide the point cloud space into a regular voxel grid;

[0101] Construct a voxel grid and assign each point in the point cloud to its corresponding voxel;

[0102] Iterate through each voxel and select the point closest to the voxel center as the sampling point.

[0103] Delete the other points and combine all the sampled points into a new pipeline point cloud.

[0104] S64. Based on the local coordinate system, combine the vessel diameter in the nb plane to obtain the pipe point cloud and repair the missing vessel model.

[0105] Furthermore, the specific repair steps in this embodiment are as follows:

[0106] 1. Subdivide the obtained pipeline model into a mesh and optimize the pipeline model.

[0107] 2. Based on the geometry of the pipeline model, extract the vertices of the surface mesh of the pipeline model to obtain a 3D point cloud.

[0108] 3. Combine the obtained 3D point cloud with the 3D point cloud obtained by S53 to reconstruct the missing blood vessel model.

[0109] This embodiment uses the original data and an existing 3D blood vessel model as data input for the present invention. Based on 3D point clouds and vessel repair, the final reconstruction result is as follows: Figures 3 to 5 As shown.

[0110] The beneficial effects of this invention are as follows:

[0111] This invention utilizes a liver blood vessel reconstruction method based on three-dimensional point clouds to repair ruptured blood vessel models, significantly improving the accuracy of blood vessel reconstruction and bringing remarkable benefits to computer-aided diagnosis of liver vascular diseases.

[0112] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0113] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. Furthermore, those skilled in the art will recognize that, based on the ideas of the present invention, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A method for repairing a vascular geometric model based on point cloud data, characterized in that, include: Obtain the raw 3D data; Based on the Gaussian smoothing filter method, the original three-dimensional data is smoothed to obtain smoothed data; A three-dimensional model of blood vessels is constructed based on the smoothed data; The vascular skeleton in the three-dimensional vascular model is extracted based on the exploration operator; A point cloud-based model for repairing missing blood vessel structures; Based on a model of missing vessels at the end of a duct repair; The model based on the repair of missing blood vessels at the end of the duct includes: Based on the existing 3D model of the blood vessel and the blood vessel skeleton, the end points of the channel are obtained; Based on the existing 3D vascular model and the vascular skeleton, 2D slices are performed to calculate the vascular diameter. Based on the vascular skeleton, calculate the local coordinate system of points on the skeleton line; Based on the local coordinate system and combined with the blood vessel diameter in the nb plane, the point cloud of the pipeline is obtained, and the missing blood vessel model is repaired. Calculating the local coordinate system of points on the skeleton line includes: calculating the tangent vector of each point on the skeleton line according to the skeleton line of the vascular skeleton; calculating the normal vector perpendicular to the tangent plane; calculating the subnormal vector according to the cross product of the tangent vector and the normal vector; normalizing the tangent vector, the normal vector and the subnormal vector to obtain a unit vector; wherein, the nb plane is the plane formed by the normal vector and the subnormal vector.

2. The method for repairing a blood vessel geometric model based on point cloud data according to claim 1, characterized in that, Following the model based on the repair of missing blood vessels at the end of the duct, the following is also included: A visual blood vessel model is generated based on the repaired model.

3. The method for repairing a blood vessel geometric model based on point cloud data according to claim 1, characterized in that, The point cloud-based model for repairing missing blood vessels includes: Based on the surface rendering algorithm, the three-dimensional model volume data of both ends of the missing blood vessel is extracted to obtain the isosurface mesh; The isosurface mesh is smoothed using the Laplace algorithm, and each point is replaced by the center of its neighborhood points to obtain a smooth mesh. Based on the geometry of a smooth mesh, the vertices of the mesh are extracted to obtain a 3D point cloud; Based on the three-dimensional point cloud, the center point of the three-dimensional point cloud is obtained by using the average value method to obtain the seed point; Segmentation based on level set segmentation combined with seed points is used to repair missing blood vessel models.

Images