A method and system for cardiac three-dimensional mapping registration

Abstract

The application discloses a kind of heart three-dimensional mapping registration method and system, belong to data processing technical field, method includes: acquisition multiple heart two-dimensional slice, reconstructs heart three-dimensional model;Select heart three-dimensional model coordinate system, obtain heart three-dimensional model data point set;With heart three-dimensional model coordinate system as reference coordinate system, with mapping catheter coordinate system as floating coordinate system, in combination with heart three-dimensional model data point set and heart three-dimensional mapping data point set, construct including mobile variable and matrix variable affine transformation function;In succession from heart three-dimensional model data point set find the shortest corresponding point of distance heart three-dimensional mapping data point set, establish corresponding relationship;Based on corresponding relationship, calculate average registration error;Set pre-set iteration number, iteration is carried out to mobile variable and matrix variable, optimize average registration error;When iteration number reaches pre-set iteration number or when average registration error is less than pre-set error, output after registration heart three-dimensional mapping data point set.

Description

Technical Field

[0001] This invention belongs to the field of data processing technology, specifically relating to a method and system for cardiac three-dimensional mapping and registration. Background Technology

[0002] In interventional cardiology, catheters are commonly used for cardiac mapping and treatment. To accurately measure cardiac structural and functional parameters, a fixed coordinate system needs to be established, namely the mapping catheter fixed coordinate system. The mapping catheter fixed coordinate system is a three-dimensional coordinate system established with the catheter as the reference, where the X-axis, Y-axis, and Z-axis represent the direction of the catheter axis, the lateral direction of the catheter, and the direction perpendicular to the catheter, respectively.

[0003] Generally, the catheter axis is defined as the positive Z-axis, the catheter lateral direction as the positive X-axis, and the direction perpendicular to the catheter as the positive Y-axis. In the fixed coordinate system of the mapping catheter, various cardiac structural and functional parameters can be measured, such as heart wall thickness, volume, systolic and end-diastolic volume, and ejection fraction. Furthermore, catheters can be used to perform cardiac electrophysiological examinations and radiofrequency ablation procedures inside the heart.

[0004] Currently, medical imaging technology plays an increasingly important role in the diagnosis and treatment of heart diseases. However, mapping and registration are usually performed using manual matching methods. This method requires manual selection of mapping points, which suffers from subjectivity and high error rates. The registration process is also slow and cannot meet the real-time requirements of three-dimensional cardiac mapping. Summary of the Invention

[0005] To address the problems of existing technologies that typically employ manual matching methods for mapping and registration, which require manual selection of mapping points, suffer from subjectivity and high error rates, and have a slow registration process, thus failing to meet the real-time requirements of cardiac three-dimensional mapping, this invention provides a method and system for cardiac three-dimensional mapping and registration.

[0006] First aspect

[0007] This invention provides a method for cardiac three-dimensional mapping and registration, comprising:

[0008] S101: Collect multiple two-dimensional slices of the patient's heart and reconstruct a three-dimensional model of the heart;

[0009] S102: Select the coordinate system of the three-dimensional heart model and obtain the data point set of the three-dimensional heart model;

[0010] S103: Obtain a set of three-dimensional cardiac mapping data points, wherein the set of three-dimensional cardiac mapping data points is obtained based on the mapping catheter coordinate system;

[0011] S104: Using the coordinate system of the three-dimensional cardiac model as the reference coordinate system and the coordinate system of the mapping catheter as the floating coordinate system, and combining the data point set of the three-dimensional cardiac model and the data point set of the three-dimensional cardiac mapping, construct an affine transformation function that includes moving variables and matrix variables.

[0012] S105: Sequentially find the corresponding point with the shortest distance from the three-dimensional model data point set of the heart to the three-dimensional mapping data point set of the heart, and establish a one-to-one correspondence.

[0013] S106: Calculate the average registration error based on the correspondence;

[0014] S107: Set a preset number of iterations, use the gradient descent algorithm to iterate over the moving variables and matrix variables, and optimize the average registration error;

[0015] S108: Output the registered cardiac 3D mapping data point set when the number of iterations reaches the preset number of iterations or when the average registration error is less than the preset error.

[0016] Second aspect

[0017] This invention provides a cardiac three-dimensional mapping and registration system, comprising:

[0018] The acquisition module is used to acquire multiple two-dimensional slices of the patient's heart and reconstruct a three-dimensional model of the heart.

[0019] The first acquisition module is used to select the coordinate system of the three-dimensional heart model and acquire the data point set of the three-dimensional heart model;

[0020] The second acquisition module is used to acquire a set of three-dimensional cardiac mapping data points, which is acquired based on the mapping catheter coordinate system.

[0021] The module is used to construct an affine transformation function that combines the coordinate system of the 3D cardiac model as the reference coordinate system and the coordinate system of the mapping catheter as the floating coordinate system, along with the data point set of the 3D cardiac model and the data point set of the 3D cardiac mapping. This function includes both translational and matrix variables.

[0022] The lookup module is used to sequentially search the heart 3D model data point set for the shortest corresponding point to the heart 3D mapping data point set, and establish a one-to-one correspondence.

[0023] The calculation module is used to calculate the average registration error based on the correspondence relationship;

[0024] The setting module is used to set the preset number of iterations and use the gradient descent algorithm to iterate over the moving variables and matrix variables to optimize the average registration error;

[0025] The output module is used to output the registered cardiac 3D mapping data point set when the number of iterations reaches the preset number of iterations or when the average registration error is less than the preset error.

[0026] Compared with the prior art, the present invention has at least the following beneficial effects:

[0027] In this invention, multiple two-dimensional slices of the patient's heart are collected to reconstruct a three-dimensional model of the heart. Combined with a set of three-dimensional cardiac mapping data points, an affine transformation function is constructed to register the mapping catheter coordinate system (i.e., the floating coordinate system) to the reference coordinate system of the three-dimensional cardiac model. A one-to-one correspondence is established based on the shortest distance corresponding points, and gradient descent algorithm is used for iterative optimization to minimize the average registration error, thereby improving the accuracy and reliability of registration. After registration, the registered three-dimensional cardiac mapping data point set is output to allow surgical personnel to more accurately determine the location of the mapping catheter. This provides doctors with more accurate and reliable information on cardiac anatomy, thereby improving diagnostic efficiency and accuracy. During the mapping process, mapping data is acquired and processed in real time to improve registration speed and avoid excessive subjective factors affecting the accuracy of the mapping results, thus reducing surgical risks and increasing the success rate of surgery. Attached Figure Description

[0028] The preferred embodiments will now be described in a clear and easy-to-understand manner, in conjunction with the accompanying drawings, to further explain the above-mentioned characteristics, technical features, advantages, and implementation methods of the present invention.

[0029] Figure 1 This is a flowchart illustrating a cardiac three-dimensional mapping and registration method provided by the present invention;

[0030] Figure 2 This is a schematic diagram of the structure of a cardiac three-dimensional mapping and registration system provided by the present invention. Detailed Implementation

[0031] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the specific implementation methods of the present invention will be described below with reference to the accompanying drawings. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings and other implementation methods can be obtained based on these drawings without any creative effort.

[0032] In one embodiment, refer to the appendix to the specification. Figure 1 The present invention provides a flowchart of a three-dimensional cardiac mapping and registration method.

[0033] This invention provides a method for cardiac three-dimensional mapping and registration, comprising:

[0034] S101: Collect multiple two-dimensional slices of the patient's heart and reconstruct a three-dimensional model of the heart.

[0035] In one possible implementation, S101 specifically includes:

[0036] S1011: Multiple two-dimensional slices of the patient's heart are acquired through multi-slice spiral CT cardiac enhancement scan;

[0037] S1012: The moving cube iso-face drawing algorithm is used to reconstruct multiple two-dimensional slices of the heart to obtain a three-dimensional model of the heart.

[0038] The Moving Cube Isosurface rendering algorithm is a computer graphics algorithm used to convert 3D data into a visualized surface model. This algorithm divides the 3D data into a set of small cubic cells and calculates a surface within each cell using isosurfaces. Then, all surfaces are merged to form a continuous 3D model. In the reconstruction of a 3D heart model, the Moving Cube Isosurface rendering algorithm can reconstruct a continuous 3D heart model from multiple 2D slices. Compared to other reconstruction algorithms, such as triangulation and cube meshing, the Moving Cube Isosurface rendering algorithm can handle irregular data and produces 3D models with higher accuracy and better performance. Simultaneously, this algorithm also has high computational efficiency and scalability, making it suitable for processing large-scale data. Therefore, using the Moving Cube Isosurface rendering algorithm for 3D heart model reconstruction can provide a more accurate and aesthetically pleasing 3D heart model and can meet the needs of large-scale data processing.

[0039] It should be noted that for cardiac 3D mapping surgery, patients typically undergo a contrast-enhanced cardiac scan of at least 18 slices on a spiral CT or MRI at least 1.6T the day before surgery. Under ECG-gated technology, patients with permanent atrial fibrillation during diastole obtain raw computer data of the cardiac chambers and the great vessels entering and exiting the heart during systole. Through image processing, 3D reconstruction of the cardiac chambers and great vessels is performed, obtaining 3D graphics of various cardiac anatomical structures, including the superior vena cava, right atrium, right ventricle, pulmonary artery, pulmonary vein, left atrium, left ventricle, and aorta. During the surgery, the 3D cardiac graphic needs to be registered with the cardiac anatomy graphic constructed from the mapping points. In practical use, current technology can already obtain relatively accurate 3D models.

[0040] S102: Select the coordinate system of the three-dimensional heart model and obtain the data point set of the three-dimensional heart model.

[0041] Understandably, acquiring a set of data points from a 3D cardiac model is crucial. This step involves determining the coordinate system of the 3D cardiac model and obtaining data points on it for subsequent registration. The choice of the coordinate system and the acquisition of the data point set are critical. Selecting a suitable coordinate system and acquiring a sufficient number of data points with adequate accuracy improves the accuracy and efficiency of subsequent registration, thereby enhancing the diagnostic and therapeutic capabilities of medical imaging. Furthermore, defining the coordinate system of the 3D cardiac model allows for a unified description of its information and facilitates comparison, processing, and analysis with other models. This is extremely important for both physicians and researchers.

[0042] S103: Obtain a set of three-dimensional cardiac mapping data points, which are obtained based on the mapping catheter coordinate system.

[0043] It's important to note that the cardiac 3D mapping data set is acquired based on the mapping catheter coordinate system. This data set represents specific measurements of the heart, reflecting key information such as its shape, size, and location. During cardiac registration, the 3D mapping data set needs to be mapped to the cardiac 3D model for matching and adjustment. Acquiring this data set not only improves the accuracy and efficiency of cardiac registration but also provides physicians with more accurate and comprehensive information about the patient's condition, helping them make better treatment decisions.

[0044] In one possible implementation, S103 specifically includes:

[0045] S1031: Using the three-dimensional data of the heart collected by the mapping catheter in the fixed coordinate system of the mapping catheter during the movement of the cardiac endothelium, a map of the internal structure of the heart is established.

[0046] S1032: Feature points are extracted from the internal structure map of the heart using the Harris corner detection algorithm to obtain a set of three-dimensional mapping data points of the heart.

[0047] The Harris corner detection algorithm is a classic computer vision algorithm used to detect corner features in images. Its basic idea is to use the gray-level changes in the neighborhood of a pixel to determine whether the pixel is a corner. If there are large gray-level changes in both the horizontal and vertical directions in the neighborhood of a pixel, then this pixel is likely a corner. The algorithm determines which points in the image are corners by calculating the gray-level gradient in the pixel's neighborhood and the corner response function.

[0048] The advantage of this algorithm lies in its ability to quickly detect corner points in images and its invariance to image rotation and scaling transformations. Therefore, it is widely used in computer vision tasks such as feature point extraction and object tracking. This algorithm's invariance to image rotation and scaling allows it to handle feature point extraction tasks well in various scenarios. The Harris corner detection algorithm has relatively low computational cost, enabling it to quickly extract corner points in images, making it suitable for real-time applications such as cardiac 3D mapping.

[0049] S104: Using the coordinate system of the three-dimensional cardiac model as the reference coordinate system and the coordinate system of the mapping catheter as the floating coordinate system, and combining the data point set of the three-dimensional cardiac model and the data point set of the three-dimensional cardiac mapping, construct an affine transformation function that includes moving variables and matrix variables.

[0050] It should be noted that the reason for using the cardiac 3D model coordinate system as the reference coordinate system and the mapping catheter coordinate system as the floating coordinate system is that the number of feature points in the cardiac 3D model is far greater than the number of mapping points. Obviously, the cardiac 3D model is more reliable. Therefore, in the registration process, we use the cardiac 3D coordinate system where the cardiac 3D model is located as the reference coordinate system and the coordinate system where the mapping catheter is located as the floating coordinate system. We then map the mapping points in the floating coordinate system to the cardiac 3D model in the reference coordinate system to avoid the problem of coordinate system inconsistency and large data acquisition errors.

[0051] In one possible implementation, S104 specifically includes:

[0052] S1041: Constructing the affine transformation function:

[0053] Q = WQ' + b

[0054]

[0055] Where Q represents the 3D reference coordinates in the cardiac 3D data point set, Q' represents the 3D floating coordinates in the cardiac 3D mapping data point set, b represents the translation variable between the mapping catheter coordinate system and the cardiac 3D model coordinate system, W represents the rotation and stretching matrix variable between the mapping catheter coordinate system and the cardiac 3D model coordinate system, and let R... 3 Let Q and Q' ∈ R be a set of points in three-dimensional space. 3 .

[0056] It should be noted that transforming the cardiac 3D mapping data point set from the catheter coordinate system to the cardiac 3D model coordinate system, and then matching and registering it with the cardiac 3D model data point set, establishes the correspondence between them. This allows for accurate calculation and optimization of the registration error. By establishing an affine transformation function, the function of transforming the cardiac 3D mapping data point set from the catheter coordinate system to the cardiac 3D model coordinate system can be achieved. The registration effect can also be optimized by adjusting the shift variables and matrix variables in the function.

[0057] S105: Sequentially find the corresponding point with the shortest distance from the three-dimensional model data point set of the heart to the three-dimensional mapping data point set of the heart, and establish a one-to-one correspondence.

[0058] In one possible implementation, S105 specifically includes:

[0059] S1051: Sequentially find the points in the cardiac 3D model data set that are closest to the cardiac 3D mapping data set.

[0060]

[0061] Where i represents the number of mapping points in the cardiac 3D mapping data point set, j represents the number of 3D data points in the cardiac 3D model data point set, and ||·||2 represents the Euclidean norm;

[0062] S1052: Label the shortest distance corresponding point with the three-dimensional coordinates of the cardiac three-dimensional mapping data point set to obtain the correspondence between all three-dimensional data points in the cardiac three-dimensional mapping data point set and the three-dimensional data points in the cardiac three-dimensional mapping data point set.

[0063] In medical image processing, various factors such as noise and motion artifacts can cause discrepancies between the acquired 3D cardiac mapping data set and the reconstructed 3D cardiac model data set, necessitating the establishment of a correspondence. Based on the distance information between the two data sets, a one-to-one correspondence is established, providing a foundation for subsequent affine transformations and achieving registration between the 3D cardiac model data set and the 3D cardiac mapping data set. Registration is a crucial step in medical image processing, improving the accuracy and precision of data processing and providing more reliable data support for doctors to make correct diagnoses.

[0064] S106: Calculate the average registration error based on the correspondence.

[0065] In one possible implementation, S106 specifically includes:

[0066] S1061: Calculate the average registration error D using the one-to-one correspondence between all three-dimensional data points in the obtained cardiac three-dimensional mapping data set and the three-dimensional data points in the cardiac three-dimensional data set.

[0067]

[0068] It's important to note that the average registration error reflects the average distance between the acquired mapping points and feature points in the 3D cardiac model, i.e., the registration error. As the average registration error approaches zero, the mapping points and model points tend to coincide. By evaluating the registration accuracy between the 3D cardiac model and the mapping data and calculating the average registration error, we can obtain the current registration error situation, thus guiding subsequent optimization algorithms for iterative optimization. If the average registration error is large, further optimization is needed until satisfactory registration accuracy is achieved.

[0069] S107: Set a preset number of iterations, and use the gradient descent algorithm to iterate over the moving variables and matrix variables to optimize the average registration error.

[0070] Gradient descent is an optimization method for finding the extrema of a function. When solving the problem of finding the minimum value of a function, this algorithm iteratively and automatically updates the parameters, causing the function value to continuously approach the minimum value until the convergence condition is met.

[0071] In one possible implementation, S107 specifically includes:

[0072] S1071: Used to set the preset number of iterations, establishing the iteration formula as follows:

[0073]

[0074] Where μ1 and μ2 represent the iteration step size;

[0075] S1072: Based on the iteration results, calculate the optimized average registration error:

[0076]

[0077] It should be noted that iterating over the shifting and matrix variables using the gradient descent algorithm can quickly find the minimum value of a multivariate function near its extreme points. In this approach, gradient descent automatically adjusts the shifting and matrix variables to continuously reduce the average registration error, thereby achieving optimal registration results. It effectively improves registration accuracy and efficiency, automatically adjusts registration parameters without human intervention, and has a fast convergence speed, finding the optimal registration result in a short time. Therefore, in fields such as medical imaging where high-precision registration is required, automated registration using the gradient descent algorithm is highly valuable.

[0078] S108: Output the registered cardiac 3D mapping data point set when the number of iterations reaches the preset number of iterations or when the average registration error is less than the preset error.

[0079] In one possible implementation, S108 specifically includes:

[0080] S1081: The registered cardiac 3D mapping data point set is as follows:

[0081] Q i =W'Q i '+b', i = 1, 2, ... m.

[0082] Understandably, accurate matching between the 3D cardiac model and mapping data is achieved through iterative optimization of the shift and matrix variables. Iterating on the shift and matrix variables continuously optimizes the registration error, thus achieving precise registration. Through iterative optimization, the average registration error can be continuously reduced, resulting in higher registration accuracy. Even with less-than-ideal initial registration parameters, better registration results can be obtained through iterative optimization, enhancing the robustness of this method. By setting preset iteration counts and preset errors, the system can meet registration requirements within a certain timeframe or accuracy range, thereby improving system reliability. In practical use, operators can manually stop the iteration based on observed results, saving registration time, improving registration efficiency, and enhancing the controllability of this method.

[0083] Compared with the prior art, the present invention has at least the following beneficial effects:

[0084] In this invention, multiple two-dimensional slices of the patient's heart are collected to reconstruct a three-dimensional model of the heart. Combined with a set of three-dimensional cardiac mapping data points, an affine transformation function is constructed to register the mapping catheter coordinate system (i.e., the floating coordinate system) to the reference coordinate system of the three-dimensional cardiac model. A one-to-one correspondence is established based on the shortest distance corresponding points, and gradient descent algorithm is used for iterative optimization to minimize the average registration error, thereby improving the accuracy and reliability of registration. After registration, the registered three-dimensional cardiac mapping data point set is output to allow surgical personnel to more accurately determine the location of the mapping catheter. This provides doctors with more accurate and reliable information on cardiac anatomy, thereby improving diagnostic efficiency and accuracy. During the mapping process, mapping data is acquired and processed in real time to improve registration speed and avoid excessive subjective factors affecting the accuracy of the mapping results, thus reducing surgical risks and increasing the success rate of surgery.

[0085] Example 2

[0086] In one embodiment, refer to the appendix to the specification. Figure 2 The present invention provides a schematic diagram of the structure of a cardiac three-dimensional mapping and registration system.

[0087] The present invention provides a cardiac three-dimensional mapping and registration system 20, comprising:

[0088] The acquisition module 201 is used to acquire multiple two-dimensional slices of the patient's heart and reconstruct a three-dimensional model of the heart.

[0089] The first acquisition module 202 is used to select the coordinate system of the three-dimensional heart model and acquire the data point set of the three-dimensional heart model;

[0090] The second acquisition module 203 is used to acquire a set of three-dimensional cardiac mapping data points, wherein the set of three-dimensional cardiac mapping data points is acquired based on the mapping catheter coordinate system;

[0091] Module 204 is used to construct an affine transformation function that includes moving variables and matrix variables by using the coordinate system of the three-dimensional cardiac model as the reference coordinate system, the coordinate system of the mapping catheter as the floating coordinate system, and combining the three-dimensional cardiac model data point set and the three-dimensional cardiac mapping data point set.

[0092] The lookup module 205 is used to sequentially search the heart 3D model data point set for the shortest corresponding point to the heart 3D mapping data point set, and establish a one-to-one correspondence.

[0093] Calculation module 206 is used to calculate the average registration error based on the correspondence relationship;

[0094] The setting module 207 is used to set the preset number of iterations and to use the gradient descent algorithm to iterate the moving variables and matrix variables to optimize the average registration error.

[0095] The output module 208 is used to output the registered cardiac three-dimensional mapping data point set when the number of iterations reaches the preset number of iterations or when the average registration error is less than the preset error.

[0096] In one possible implementation, the acquisition module specifically includes:

[0097] The acquisition submodule is used to acquire multiple two-dimensional slices of the patient's heart through multi-slice spiral CT cardiac enhancement scans.

[0098] The reconstruction submodule is used to reconstruct multiple two-dimensional slices of the heart using a moving cube iso-face drawing algorithm to obtain a three-dimensional model of the heart.

[0099] In one possible implementation, the first acquisition module specifically includes:

[0100] A submodule is established to create a map of the internal structure of the heart using three-dimensional heart data collected in a fixed coordinate system of the mapping catheter during the movement of the mapping catheter through the heart endothelium.

[0101] The extraction submodule is used to extract feature points from the internal structure map of the heart using the Harris corner detection algorithm to obtain a set of three-dimensional mapping data points of the heart.

[0102] In one possible implementation, the building module specifically includes:

[0103] Construct a submodule for building affine transformation functions:

[0104] Q = WQ' + b

[0105]

[0106] Where Q represents the 3D reference coordinates in the cardiac 3D data point set, Q' represents the 3D floating coordinates in the cardiac 3D mapping data point set, b represents the translation variable between the mapping catheter coordinate system and the cardiac 3D model coordinate system, W represents the rotation and stretching matrix variable between the mapping catheter coordinate system and the cardiac 3D model coordinate system, and let R... 3 Let Q and Q' ∈ R be a set of points in three-dimensional space. 3 .

[0107] In one possible implementation, the lookup module specifically includes:

[0108] The search submodule is used to sequentially find the corresponding point in the set of 3D heart model data points that is closest to the set of 3D heart mapping data points.

[0109]

[0110] Where i represents the number of mapping points in the cardiac 3D mapping data point set, j represents the number of 3D data points in the cardiac 3D model data point set, and ||·||2 represents the Euclidean norm;

[0111] The annotation submodule is used to annotate the shortest distance corresponding point with the three-dimensional coordinates of the cardiac three-dimensional mapping data point set, so as to obtain the correspondence between all three-dimensional data points in the cardiac three-dimensional mapping data point set and the three-dimensional data points in the cardiac three-dimensional data point set.

[0112] In one possible implementation, the computing module specifically includes:

[0113] The first calculation submodule uses the one-to-one correspondence between all three-dimensional data points in the acquired cardiac three-dimensional mapping data point set and the three-dimensional data points in the cardiac three-dimensional data point set to calculate the average registration error D:

[0114]

[0115] In one possible implementation, the setting module specifically includes:

[0116] The settings submodule is used to set the preset number of iterations, and the iteration formula is established as follows:

[0117]

[0118] Where μ1 and μ2 represent the iteration step size;

[0119] The second calculation submodule is used to calculate the optimized average registration error based on the iteration results:

[0120]

[0121] In one possible implementation, the output module specifically includes:

[0122] The output submodule is used to output the registered cardiac 3D mapping data point set:

[0123] Q i =W'Q i '+b', i = 1, 2, ... m.

[0124] The contract image recognition system 20 provided by the present invention can implement the various processes implemented in the above method embodiments, and will not be described again here to avoid repetition.

[0125] The virtual system provided by this invention can be a system, or a component, integrated circuit, or chip in a terminal.

[0126] Compared with the prior art, the present invention has at least the following beneficial effects:

[0127] In this invention, multiple two-dimensional slices of the patient's heart are collected to reconstruct a three-dimensional model of the heart. Combined with a set of three-dimensional cardiac mapping data points, an affine transformation function is constructed to register the mapping catheter coordinate system (i.e., the floating coordinate system) to the reference coordinate system of the three-dimensional cardiac model. A one-to-one correspondence is established based on the shortest distance corresponding points, and gradient descent algorithm is used for iterative optimization to minimize the average registration error, thereby improving the accuracy and reliability of registration. After registration, the registered three-dimensional cardiac mapping data point set is output to allow surgical personnel to more accurately determine the location of the mapping catheter. This provides doctors with more accurate and reliable information on cardiac anatomy, thereby improving diagnostic efficiency and accuracy. During the mapping process, mapping data is acquired and processed in real time to improve registration speed and avoid excessive subjective factors affecting the accuracy of the mapping results, thus reducing surgical risks and increasing the success rate of surgery.

[0128] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0129] The above embodiments merely illustrate several implementation methods of the present invention, and their descriptions are relatively specific and detailed, but they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these all fall within the protection scope of the present invention. Therefore, the protection scope of this invention patent should be determined by the appended claims.

Claims

1. A method for three-dimensional mapping and registration of the heart, characterized in that, include: S101: Collect multiple two-dimensional slices of the patient's heart and reconstruct a three-dimensional model of the heart; S102: Select the coordinate system of the three-dimensional heart model and obtain the data point set of the three-dimensional heart model; S103: Obtain a set of three-dimensional cardiac mapping data points, wherein the set of three-dimensional cardiac mapping data points is obtained based on the mapping catheter coordinate system; S104: Using the coordinate system of the three-dimensional cardiac model as the reference coordinate system and the coordinate system of the mapping catheter as the floating coordinate system, and combining the data point set of the three-dimensional cardiac model and the data point set of the three-dimensional cardiac mapping, construct an affine transformation function including moving variables and matrix variables. S105: Sequentially search the set of three-dimensional heart model data points for the corresponding point with the shortest distance to the set of three-dimensional heart mapping data points, and establish a one-to-one correspondence. S106: Calculate the average registration error based on the aforementioned correspondence; S107: Set a preset number of iterations, and use the gradient descent algorithm to iterate over the moving variable and the matrix variable to optimize the average registration error; S108: When the number of iterations reaches the preset number of iterations or when the average registration error is less than the preset error, output the registered cardiac three-dimensional mapping data point set; Specifically, S104 is: S1041: Construct the affine transformation function: ； ， ， ， ； Where Q represents the three-dimensional reference coordinates in the set of three-dimensional cardiac data points. Let R represent the three-dimensional floating coordinates in the cardiac three-dimensional mapping data point set, b represent the translation variable between the mapping catheter coordinate system and the cardiac three-dimensional model coordinate system, W represent the rotation and stretching matrix variable between the mapping catheter coordinate system and the cardiac three-dimensional model coordinate system, and let R 3 To represent a set of points in three-dimensional space, then ; Specifically, S105 includes: S1051: Sequentially find the point with the shortest distance from the three-dimensional model data point set of the heart to the three-dimensional mapping data point set of the heart. ； Where i represents the number of mapping points in the cardiac 3D mapping data point set, and j represents the number of 3D data points in the cardiac 3D model data point set. Denotes the Euclidean norm; S1052: Mark the corresponding point with the shortest distance with the three-dimensional coordinates of the three-dimensional mapping data point set of the heart, and obtain the correspondence between all three-dimensional data points in the three-dimensional mapping data point set of the heart and the three-dimensional data points in the three-dimensional mapping data point set of the heart.

2. The cardiac three-dimensional mapping and registration method according to claim 1, characterized in that, S101 specifically includes: S1011: Multiple two-dimensional slices of the patient's heart are acquired through multi-slice spiral CT cardiac enhancement scan; S1012: The heart is reconstructed from multiple two-dimensional slices using a moving cube iso-face drawing algorithm to obtain a three-dimensional model of the heart.

3. The cardiac three-dimensional mapping and registration method according to claim 1, characterized in that, S103 specifically includes: S1031: Using the three-dimensional data of the heart collected by the mapping catheter in the fixed coordinate system of the mapping catheter during the movement of the heart endothelium, a map of the internal structure of the heart is established. S1032: The feature points of the internal structure map of the heart are extracted by the Harris corner detection algorithm to obtain the three-dimensional mapping data point set of the heart.

4. The cardiac three-dimensional mapping and registration method according to claim 1, characterized in that, Specifically, S106 is: S1061: Calculate the average registration error D using the one-to-one correspondence between all three-dimensional data points in the obtained cardiac three-dimensional mapping data point set and the three-dimensional data points in the cardiac three-dimensional data point set. 。 5. The cardiac three-dimensional mapping and registration method according to claim 1, characterized in that, Specifically, S107 includes: S1071: Used to set the preset number of iterations, establishing the iteration formula as follows: ； in, and Indicates the iteration step size; S1072: Based on the iteration results, calculate the optimized average registration error: 。 6. The cardiac three-dimensional mapping and registration method according to claim 1, characterized in that, Specifically, S108 is: S1081: The registered cardiac three-dimensional mapping data point set is as follows: 。 7. A cardiac three-dimensional mapping and registration system, characterized in that, include: The acquisition module is used to acquire multiple two-dimensional slices of the patient's heart and reconstruct a three-dimensional model of the heart. The first acquisition module is used to select the coordinate system of the three-dimensional heart model and acquire the data point set of the three-dimensional heart model; The second acquisition module is used to acquire a set of three-dimensional cardiac mapping data points, wherein the set of three-dimensional cardiac mapping data points is acquired based on the mapping catheter coordinate system; The construction module is used to take the coordinate system of the three-dimensional cardiac model as the reference coordinate system, the coordinate system of the mapping catheter as the floating coordinate system, and combine the data point set of the three-dimensional cardiac model and the data point set of the three-dimensional cardiac mapping to construct an affine transformation function including moving variables and matrix variables. The lookup module is used to sequentially search the heart 3D model data point set for the corresponding point with the shortest distance to the heart 3D mapping data point set, and establish a one-to-one correspondence. The calculation module is used to calculate the average registration error based on the correspondence. The setting module is used to set a preset number of iterations and to iterate the moving variable and the matrix variable using the gradient descent algorithm to optimize the average registration error; The output module is used to output the registered cardiac three-dimensional mapping data point set when the number of iterations reaches the preset number of iterations or when the average registration error is less than the preset error. The building blocks specifically include: Construct a submodule for building affine transformation functions: ； ， ， ， ； Where Q represents the three-dimensional reference coordinates in the three-dimensional data point set of the heart. Let R represent the three-dimensional floating coordinates in the cardiac three-dimensional mapping data point set, b represent the translation variable between the mapping catheter coordinate system and the cardiac three-dimensional model coordinate system, W represent the rotation and stretching matrix variable between the mapping catheter coordinate system and the cardiac three-dimensional model coordinate system, and let R... 3 To represent a set of points in three-dimensional space, then ; The search module specifically includes: The search submodule is used to sequentially find the corresponding point in the set of 3D heart model data points that is closest to the set of 3D heart mapping data points. ； Where i represents the number of mapping points in the cardiac 3D mapping data point set, and j represents the number of 3D data points in the cardiac 3D model data point set. This represents the Euclidean norm.

8. The cardiac three-dimensional mapping and registration system according to claim 7, characterized in that, The acquisition module specifically includes: The acquisition submodule is used to acquire multiple two-dimensional slices of the patient's heart through multi-slice spiral CT cardiac enhancement scans. The reconstruction submodule is used to reconstruct multiple two-dimensional slices of the heart using a moving cube iso-face drawing algorithm to obtain a three-dimensional model of the heart.

Images