Image mark point adaptive registration method based on triangular mesh model

Abstract

The invention discloses an image mark point self-adaptive registration method based on a triangular mesh model, which adopts a computed tomography image as input data and performs surgical navigation image registration by using a clear display effect of the computed tomography image. Thirdly, preliminarily simplifying the CT image by using a computer image processing technology, and extracting a triangular mesh model from a binary image obtained by processing to realize efficient acquisition of mark points; adaptive grouping is carried out on the extracted mark points according to the area sizes of the mark points, then geometric center point extraction is carried out on the grouped mark points, and the complexity of subsequent data processing is reduced; and finally, performing matrix conversion and error value calculation according to the obtained center point and the input actual coordinate point to obtain an optimal matching result. The automatic registration method can ensure shorter registration time, a more stable registration process and a more accurate registration effect, and is adaptive to a surgical navigation system which needs to improve image registration efficiency.

Description

Technical Field

[0001] The present invention relates to the technical field of medical image registration, and in particular to an image landmark point adaptive registration method based on a triangular mesh model. Background Art

[0002] With the increasing popularity of precision medicine and the development of minimally invasive surgical techniques, medical image-based navigation systems are playing an increasingly important role in surgical procedures. Medical image registration is a core technology for surgical navigation. Its task is to accurately map sensor positioning data to medical images by calculating the spatial transformation matrix from the sensor coordinate system to the image navigation coordinate system. Traditional medical image registration methods often rely on the manual selection of special markers for registration, which has drawbacks such as time-consuming operations, reliance on physician experience, and poor registration stability. In areas such as orthopedics and neurosurgery, where spatial positioning accuracy is extremely high, existing methods struggle to meet the dual requirements of registration efficiency and accuracy for real-time intraoperative navigation. Although automatic registration algorithms based on landmark recognition have emerged in recent years, they still face challenges in processing complex image structures, such as low landmark recognition efficiency and long corresponding point matching time. Furthermore, existing methods lack effective utilization of landmark features, making it difficult to achieve highly robust registration parameter estimation.

[0003] Therefore, it is necessary to develop an image landmark adaptive registration method based on a triangular mesh model to solve the above problems. Summary of the Invention

[0004] The purpose of the present invention is to design an image landmark point adaptive registration method based on a triangular mesh model in order to solve the above problems.

[0005] The present invention achieves the above-mentioned purpose through the following technical solutions:

[0006] An image landmark adaptive registration method based on a triangular mesh model, comprising:

[0007] S1, input image data and sensor coordinates;

[0008] S2, image preprocessing using thresholding, erosion and dilation operations;

[0009] S3, using the isosurface extraction algorithm to extract the triangulated network model of the landmarks;

[0010] S4, the triangular mesh is divided into a number of independent marker points through the connected domain, and it is determined whether the number of marker points is greater than the actual number of points. If so, the process proceeds to step S5, otherwise, the process proceeds to step S6;

[0011] S5, using the adaptive grouping algorithm to group the landmarks with similar area sizes into one group, and then proceed to the next step;

[0012] S6. Find the geometric center point of the marker point;

[0013] S7, matching and combining the center point set and the sensor coordinate set;

[0014] S8, calculating the transformation matrix and error value of each combination;

[0015] S9. Output the optimal matrix and error value to complete the registration.

[0016] Furthermore, step S1 includes: selecting a 3D CT image file containing multiple sensor positioning patches as input. The 3D CT image is volume data formed by stacking multiple layers of 2D CT slices in the same direction, which can be expressed as:

[0017] in, are the number of voxels in the transverse, sagittal, and coronal planes of the three-dimensional volume data, respectively; Represents voxel position The CT value at .

[0018] Furthermore, step S2 includes: thresholding the image to pre-process it into binary segmentation, the formula is:

[0019]

[0020] In the above formula, is the upper threshold value; after completing the thresholding, a 3×3×3 cubic structural element is used Perform erosion operation on a binary image:

[0021]

[0022] In the above formula, the structural element , yes The offset relative to the center point The displacement, Indicates the minimum value; its mathematical morphology abbreviation is:

[0023]

[0024] In the above formula, represents the corrosion operation;

[0025] Subsequent use of 5×5×5 cubic structural elements Perform dilation on the erosion result:

[0026]

[0027] In the above formula, the structural element , yes The offset relative to the center point The displacement, Indicates the maximum value; its mathematical morphology abbreviation is:

[0028]

[0029] in, Represents the dilation operation, using Represents the final result of the preprocessed image; the complete process is shown in the following formula:

[0030] .

[0031] Furthermore, step S3 includes: using an isosurface extraction algorithm to extract the binary image Extracting triangular mesh models ; Among them, the vertex set Generated by voxel edge interpolation:

[0032]

[0033] In the above formula, Represents the vertex position of the triangle mesh, only when Calculate when and and Represents voxel edges The coordinates of the two endpoints, represents the isosurface threshold, Indicates taking all The union of; the set of patches in a triangle mesh Then we can get it by looking up the voxel state index table:

[0034]

[0035] in, is the Heaviside function, represents the isosurface threshold, Represents voxels in a 3D image The corner position of , is a predefined triangulated lookup table, It means taking the union of all results and finally generating a set of triangle faces.

[0036] Furthermore, in step S5: using an adaptive grouping algorithm to group landmarks with similar area sizes into one group, including:

[0037] definition is the input data, where is the region identification number, is the area size, Indicates the total number of mark points; according to Arrange in descending order to get an ordered sequence :

[0038]

[0039] right Calculate statistics:

[0040]

[0041] in, , define the grouping threshold , get the group critical point index set:

[0042]

[0043] set up is the group number, and , then the grouping rule for:

[0044] .

[0045] Furthermore, step S6 includes: after the marker points are grouped, their specific coordinate values ​​in the image need to be obtained, and the geometric center point is calculated using the vertices of the marker point area:

[0046]

[0047] in, Indicates area The set of all vertices of Indicates area The number of vertices, Indicates area The geometric center coordinates of all landmark points are called the center point set.

[0048] Furthermore, step S7 includes: matching the center point set with the actual coordinate point set input by the sensor according to the grouping, selecting center points from each group with the same number as the actual coordinate points to form multiple unique combinations.

[0049] Furthermore, steps S8 and S9 include: calculating the conversion matrix and error value for all unique combinations of all groups, and taking the conversion matrix and center point group with the minimum error value among all combinations as the final result; the calculation formula of the conversion matrix is:

[0050]

[0051] in, The number of representative points, is the actual coordinate point, As the center point, is a 4×4 transformation matrix, is the optimal transformation matrix.

[0052] The beneficial effects of the present invention are:

[0053] In this application, the final center point group obtained after image preprocessing is accurately positioned, and its transformation matrix can achieve a lower error value than traditional methods, with the error value reaching below 1.0. Using adaptive grouping and isosurface extraction algorithms, this method has a shorter processing time and can achieve a several-fold reduction in processing time when there are many redundant points. Validated on different types of data, this method as a whole can achieve highly robust registration parameter estimation, meeting the practical needs of surgical navigation. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 This is a flowchart of the application. DETAILED DESCRIPTION

[0055] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more apparent, the technical solutions of the embodiments of the present invention will be described clearly and completely below in conjunction with the accompanying drawings of the embodiments of the present invention. It should be understood that the described embodiments are only a portion of the embodiments of the present invention, not all of them. Generally, the components of the embodiments of the present invention described and illustrated in the drawings herein may be arranged and designed in a variety of different configurations.

[0056] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention as claimed, but rather merely represents selected embodiments of the present invention. All other embodiments derived by persons of ordinary skill in the art based on the embodiments of the present invention without creative effort are intended to fall within the scope of protection of the present invention.

[0057] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, it does not need to be further defined or explained in subsequent drawings.

[0058] In the description of the present invention, it should be understood that the terms "upper", "lower", "inside", "outside", "left", "right", etc. indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, or are the orientations or positional relationships in which the inventive product is conventionally placed when in use, or are the orientations or positional relationships conventionally understood by those skilled in the art. These are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or component referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as a limitation on the present invention.

[0059] Furthermore, the terms “first”, “second”, etc. are merely used for distinguishing descriptions and should not be understood as indicating or implying relative importance.

[0060] In the description of the present invention, it should also be noted that, unless otherwise expressly specified or limited, terms such as "disposed" and "connected" should be understood in a broad sense. For example, "connected" can mean a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium; it can also mean internal communication between two components. Those skilled in the art will be able to understand the specific meanings of the above terms in the present invention based on specific circumstances.

[0061] The specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0062] like Figure 1 As shown, an image landmark adaptive registration method based on a triangular mesh model includes:

[0063] S1, input image data and sensor coordinates;

[0064] A 3D CT image file containing four sensor positioning patches is selected as input. The 3D CT image is a volume data formed by stacking multiple 2D CT slices in the same direction, which can be expressed as:

[0065]

[0066] in, are the number of voxels in the transverse, sagittal, and coronal planes of the three-dimensional volume data, respectively; Represents voxel position The CT value at .

[0067] S2, image preprocessing using thresholding, erosion and dilation operations;

[0068] 3D CT images describe anatomical structures through voxel positions and CT values, and are 3D grayscale data. To improve registration accuracy, image processing techniques can be used to screen, filter, and refine the landmarks contained in CT images. The initial operation of image preprocessing is image thresholding, also known as binary segmentation, which is expressed as:

[0069]

[0070] In the above formula, is the upper threshold value; after completing the thresholding, a 3×3×3 cubic structural element is used Perform erosion operation on a binary image:

[0071]

[0072] In the above formula, the structural element , yes The offset relative to the center point The displacement, Indicates the minimum value; its mathematical morphology abbreviation is:

[0073]

[0074] In the above formula, represents the corrosion operation;

[0075] Subsequent use of 5×5×5 cubic structural elements Perform dilation on the erosion result:

[0076]

[0077] In the above formula, the structural element , yes The offset relative to the center point The displacement, Indicates the maximum value; its mathematical morphology abbreviation is:

[0078]

[0079] in, Represents the dilation operation, using Represents the final result of the preprocessed image; the complete process is shown in the following formula:

[0080] .

[0081] S3, using the isosurface extraction algorithm to extract the triangulated network model of the landmarks;

[0082] The result of image preprocessing only contains all possible landmarks. In order to improve the efficiency of landmark recognition, an isosurface extraction algorithm is used to extract the landmarks from the binary image. Extracting triangular mesh models ; Among them, the vertex set Generated by voxel edge interpolation:

[0083]

[0084] In the above formula, Represents the vertex position of the triangle mesh, only when Calculate when and and Represents voxel edges The coordinates of the two endpoints, represents the isosurface threshold, Indicates taking all The union of; the set of patches in a triangle mesh Then we can get it by looking up the voxel state index table:

[0085]

[0086] in, is the Heaviside function, represents the isosurface threshold, Represents voxels in a 3D image The corner position of , is a predefined triangulated lookup table, It means taking the union of all results and finally generating a set of triangle faces.

[0087] S4, the triangular mesh is divided into a number of independent marker points through the connected domain, and it is determined whether the number of marker points is greater than the actual number of points. If so, the process proceeds to step S5, otherwise, the process proceeds to step S6;

[0088] S5. Use the adaptive grouping algorithm to group the landmarks with similar area sizes into one group, thereby reducing the time consumption of corresponding point matching, and then proceed to the next step;

[0089] An adaptive grouping algorithm is used to group landmarks with similar area sizes into a group, including:

[0090] definition is the input data, where is the region identification number, is the area size, Indicates the total number of mark points; according to Arrange in descending order to get an ordered sequence :

[0091]

[0092] right Calculate statistics:

[0093]

[0094] in, , define the grouping threshold , get the group critical point index set:

[0095]

[0096] set up is the group number, and , then the grouping rule for:

[0097] .

[0098] S6. Find the geometric center point of the marker point;

[0099] After the marker points are grouped, their specific coordinate values ​​in the image need to be obtained. The geometric center point is calculated using the vertices of the marker area:

[0100]

[0101] in, Indicates area The set of all vertices of Indicates area The number of vertices, Indicates area The geometric center coordinates of all landmark points are called the center point set.

[0102] S7, matching and combining the center point set and the sensor coordinate set;

[0103] The center point set and the actual coordinate point set transmitted by the sensor are matched according to the grouping, and the same number of center points as the actual coordinate points are selected from each group to form multiple unique combinations.

[0104] S8, calculating the transformation matrix and error value of each combination;

[0105] S9. Output the optimal matrix and error value to complete the registration.

[0106] For all unique combinations of all groups, the conversion matrix and error value are calculated, and the conversion matrix and center point group with the minimum error value among all combinations are taken as the final result; the calculation formula of the conversion matrix is:

[0107]

[0108] in, The number of representative points, is the actual coordinate point, As the center point, is a 4×4 transformation matrix, is the optimal transformation matrix.

[0109] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as within the scope of protection of the present invention.

Claims

1. An adaptive registration method for image landmarks based on a triangular mesh model, characterized in that: include: S1, input image data and sensor coordinates; S2, image preprocessing using thresholding, erosion and dilation operations; S3, using the isosurface extraction algorithm to extract the triangulated network model of the landmarks; S4, the triangular mesh is divided into a number of independent marker points through the connected domain, and it is determined whether the number of marker points is greater than the actual number of points. If so, the process proceeds to step S5, otherwise, the process proceeds to step S6; S5, using the adaptive grouping algorithm to group the landmarks with similar area sizes into one group, and then proceed to the next step; S6. Find the geometric center point of the marker point; S7, matching and combining the center point set and the sensor coordinate set; S8, calculating the transformation matrix and error value of each combination; S9. Output the optimal matrix and error value to complete the registration.

2. The method for adaptive registration of image landmarks based on a triangular mesh model according to claim 1, characterized in that: Step S1 includes: selecting a 3D CT image file containing multiple sensor positioning patches as input. The 3D CT image is volume data formed by stacking multiple layers of 2D CT slices in the same direction, which can be expressed as: ； in, are the number of voxels in the transverse, sagittal, and coronal planes of the three-dimensional volume data, respectively; Represents voxel position The CT value at .

3. The method for adaptive registration of image landmarks based on a triangular mesh model according to claim 2, characterized in that: Step S2 includes: thresholding the image to pre-process it into binary segmentation, the formula is: ； In the above formula, is the upper threshold value; after completing the thresholding, a 3×3×3 cubic structural element is used Perform erosion operation on a binary image: ； In the above formula, the structural element , yes The offset relative to the center point The displacement, Indicates the minimum value; its mathematical morphology abbreviation is: ； In the above formula, represents the corrosion operation; Subsequent use of 5×5×5 cubic structural elements Perform dilation on the erosion result: ； In the above formula, the structural element , yes The offset relative to the center point The displacement, Indicates the maximum value; its mathematical morphology abbreviation is: ； in, Represents the dilation operation, using Represents the final result of the preprocessed image; the complete process is shown in the following formula: 。 4. The method for adaptive registration of image landmarks based on a triangular mesh model according to claim 3, characterized in that: Step S3 includes: using an isosurface extraction algorithm to extract the binary image Extracting triangular mesh models ; Among them, the vertex set Generated by voxel edge interpolation: ； In the above formula, Represents the vertex position of the triangle mesh, only when Calculate when and and Represents voxel edges The coordinates of the two endpoints, represents the isosurface threshold, Indicates taking all The union of; the set of patches in a triangle mesh Then we can get it by looking up the voxel state index table: ； in, is the Heaviside function, represents the isosurface threshold, Represents voxels in a 3D image The corner position of , is a predefined triangulated lookup table, It means taking the union of all results and finally generating a set of triangle faces.

5. The method for adaptive registration of image landmarks based on a triangular mesh model according to claim 4, characterized in that: In step S5: using an adaptive grouping algorithm to group landmarks with similar area sizes into a group, including: definition is the input data, where is the region identification number, is the area size, Indicates the total number of mark points; according to Arrange in descending order to get an ordered sequence : ； right Calculate statistics: ； in, , define the grouping threshold , get the group critical point index set: ； set up is the group number, and , then the grouping rule for: 。 6. The method for adaptive registration of image landmarks based on a triangular mesh model according to claim 5, characterized in that: Step S6 includes: after the marker points are grouped, their specific coordinate values ​​in the image need to be obtained, and the geometric center point is calculated using the vertices of the marker point area: ； in, Indicates area The set of all vertices of Indicates area The number of vertices, Indicates area The geometric center coordinates of all landmark points are called the center point set.

7. The method for adaptive registration of image landmarks based on a triangular mesh model according to claim 6, characterized in that: Step S7 includes: matching the center point set with the actual coordinate point set input by the sensor according to the grouping, selecting center points from each group with the same number as the actual coordinate points to form multiple unique combinations.

8. The method for adaptive registration of image landmarks based on a triangular mesh model according to claim 7, characterized in that: Steps S8 and S9 include: calculating the conversion matrix and error value for all unique combinations of all groups, and taking the conversion matrix and center point group with the minimum error value among all combinations as the final result; the calculation formula of the conversion matrix is: ； in, The number of representative points, is the actual coordinate point, As the center point, is a 4×4 transformation matrix, is the optimal transformation matrix.

Images