Matching method and device of magnetic resonance image and electric mark image
By reconstructing cardiac magnetic resonance and electrophysiological mapping images using elastic deformation and ICP registration algorithms, the problems of low accuracy and high subjectivity in cardiac image matching in existing technologies are solved, and automated, accurate image fusion and quantitative analysis are achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- PEKING UNION MEDICAL COLLEGE HOSPITAL
- Filing Date
- 2026-02-14
- Publication Date
- 2026-05-19
AI Technical Summary
In existing technologies, the matching of cardiac magnetic resonance images and electrophysiological mapping images mainly relies on visual comparison, resulting in low comparison accuracy, high subjectivity, lack of quantitative analysis tools, and inability to achieve accurate matching and fusion of myocardial fibrosis areas and low voltage areas.
A method based on elastic deformation and iterative nearest point ICP registration algorithm is used to reconstruct and match cardiac magnetic resonance images and electrophysiological mapping images. Automatic and accurate fusion is achieved through coordinate transformation to generate a fusion model containing fibrosis information and voltage information.
It enables automated and precise matching and fusion of cardiac magnetic resonance images and electrophysiological mapping images, provides quantitative indicators of fibrosis-electrophysiological consistency, and promotes the development of clinical research.
Smart Images
Figure CN122066907A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of medical image processing technology, and more particularly to a method and apparatus for matching magnetic resonance images and electrical reference images. Background Technology
[0002] In the diagnosis and treatment of patients with atrial fibrillation, clinical assessment mainly relies on two complementary imaging techniques:
[0003] 1. Cardiac magnetic resonance imaging: It can non-invasively and with high resolution visualize the spatial distribution of myocardial fibrosis or scar tissue. These areas are the pathological matrix that produces abnormal electrical activity.
[0004] 2. Intracardiac electrophysiological mapping: By measuring multiple points within the cardiac chamber through interventional catheters, a "voltage map" representing the electrical activity of the myocardium can be accurately drawn, and "low voltage areas" with slow conduction or inactivation can be identified. These areas are key targets for ablation therapy.
[0005] Currently, these two technologies exist as information silos in clinical practice. Doctors typically rely on visual comparison or experience-based judgment, observing the two images separately on independent MRI workstations and electrophysiological 3D mapping systems, making it impossible to accurately establish a correlation. The specific drawbacks are as follows:
[0006] 1. Low accuracy and high subjectivity in comparison: The position and shape of the heart in the chest cavity and its state at different examination times (such as heart rate and respiratory phase) vary. Spatial matching of two independent three-dimensional models based solely on vision and experience results in large errors and cannot achieve precise point-to-point comparison of anatomical positions.
[0007] 2. Lack of quantitative analysis tools: It is impossible to accurately match fibrotic areas with low voltage areas, making it difficult to scientifically assess the consistency of the two types of information, which limits the development of large-scale clinical studies. Summary of the Invention
[0008] To overcome the shortcomings of the prior art, the technical problem to be solved by the present invention is to provide a method for matching magnetic resonance images and electrophysiological mapping images, which can be used for multimodal image fusion of the left atrium in patients with atrial fibrillation, so as to realize automatic and accurate matching and fusion of cardiac magnetic resonance images and intracardiac electrophysiological mapping images.
[0009] The technical solution of this invention is: a method for matching magnetic resonance images with electrical reference images, comprising the following steps: (1) Based on cardiac magnetic resonance images, a first three-dimensional cardiac model containing the anatomical structure of the left atrium and the intensity of fibrosis was reconstructed; (2) The data points with three-dimensional spatial coordinates (X, Y, Z) and corresponding voltage values (V) collected by the electrophysiological mapping system are used to reconstruct a second three-dimensional heart model that reflects the geometry and voltage distribution of the left atrium. (3) The second three-dimensional heart model is elastically deformed to correct the overall displacement and rotation, and to compensate for the morphological differences caused by different examination conditions, so that it achieves the best match with the first three-dimensional heart model in terms of overall shape and local details. (4) The voltage information of each data point on the second three-dimensional heart model is mapped to the corresponding anatomical position of the first three-dimensional heart model through coordinate transformation, thereby generating a fusion model that simultaneously carries fibrosis information from magnetic resonance and voltage information from electrophysiology.
[0010] This invention enables multimodal image fusion of the left atrium in patients with atrial fibrillation, achieving automatic and precise matching and fusion of cardiac magnetic resonance imaging (MRI) images and intracardiac electrophysiological mapping images. It replaces subjective and coarse visual comparison, realizing automated spatial matching of two heterogeneous images and elevating comparative analysis from a qualitative to a precise quantitative level. It provides a series of objective and reproducible quantitative indicators for fibrosis-electrophysiological consistency in clinical research, promoting the development of evidence-based medicine in this field.
[0011] A matching device for magnetic resonance images and electrical standard images is also provided, which includes: The first model reconstruction module is configured to reconstruct a first three-dimensional heart model based on cardiac magnetic resonance images, including the left atrial anatomy and the intensity of fibrosis. The second model reconstruction module is configured to reconstruct a second three-dimensional heart model reflecting the geometry and voltage distribution of the left atrium from the data points with three-dimensional spatial coordinates (X,Y,Z) and corresponding voltage values (V) collected by the electrophysiological mapping system. The registration module is configured to elastically deform the second three-dimensional heart model, correct the overall displacement and rotation, compensate for the morphological differences caused by different examination conditions, and make it optimally matched with the first three-dimensional heart model in terms of overall shape and local details. The fusion module is configured to map the voltage information of each data point on the second three-dimensional heart model to the corresponding anatomical position on the first three-dimensional heart model through coordinate transformation, thereby generating a fusion model that simultaneously carries fibrosis information from magnetic resonance imaging and voltage information from electrophysiology. Attached Figure Description
[0012] Figure 1 This is a flowchart of a method for matching magnetic resonance images and electrical reference images according to the present invention. Detailed Implementation
[0013] like Figure 1As shown, this method for matching magnetic resonance images with electrical target images includes the following steps: (1) Based on cardiac magnetic resonance images, a first three-dimensional cardiac model containing the anatomical structure of the left atrium and the intensity of fibrosis was reconstructed; (2) The data points with three-dimensional spatial coordinates (X, Y, Z) and corresponding voltage values (V) collected by the electrophysiological mapping system are used to reconstruct a second three-dimensional heart model that reflects the geometry and voltage distribution of the left atrium. (3) The second three-dimensional heart model is elastically deformed to correct the overall displacement and rotation, and to compensate for the morphological differences caused by different examination conditions, so that it achieves the best match with the first three-dimensional heart model in terms of overall shape and local details. (4) The voltage information of each data point on the second three-dimensional heart model is mapped to the corresponding anatomical position of the first three-dimensional heart model through coordinate transformation, thereby generating a fusion model that simultaneously carries fibrosis information from magnetic resonance and voltage information from electrophysiology.
[0014] This invention enables multimodal image fusion of the left atrium in patients with atrial fibrillation, achieving automatic and precise matching and fusion of cardiac magnetic resonance imaging (MRI) images and intracardiac electrophysiological mapping images. It replaces subjective and coarse visual comparison, realizing automated spatial matching of two heterogeneous images and elevating comparative analysis from a qualitative to a precise quantitative level. It provides a series of objective and reproducible quantitative indicators for fibrosis-electrophysiological consistency in clinical research, promoting the development of evidence-based medicine in this field.
[0015] Furthermore, step (1) includes the following sub-steps: (1.1) Cardiac magnetic resonance imaging (MRI) was used as the raw data; (1.2) In CEMRG software, the first three-dimensional heart model of the left atrium was reconstructed using magnetic resonance images. The data included three-dimensional spatial coordinates and corresponding fibrosis data. (1.3) Import the first three-dimensional heart model into MATLAB software for subsequent analysis.
[0016] Furthermore, step (2) includes the following sub-steps: (2.1) Data acquired by the CARTO electrophysiological mapping system were used as raw data; (2.2) Mark the anatomical structures in CARTO, including the ablation zone of the left pulmonary vein, the ablation zone of the right pulmonary vein, the mitral valve annulus, and the left atrial appendage, and then export the data; (2.3) Reconstruct the second three-dimensional heart model in MATLAB and color-code the voltage distribution.
[0017] Furthermore, step (3) includes the following sub-steps: (3.1) Data preparation: In MATLAB, the second three-dimensional heart model and the first three-dimensional heart model are stored as data structures containing three-dimensional spatial coordinates and their corresponding physical properties. Each set of coordinates in the second three-dimensional heart model is associated with a voltage measurement value, while the coordinates of the first three-dimensional heart model correspond to its fibrosis degree quantification value. (3.2) Size standardization: The three-dimensional coordinate data of the first three-dimensional heart model and the second three-dimensional heart model are simultaneously imported into CloudCompare software. Using its registration and matching scale function, the spatial scale of the first three-dimensional heart model is used as the reference to perform elastic deformation of the second three-dimensional heart model with scaling as the core operation. (3.3) Deformation process record: Scale matching and deformation operation generate a complete transformation matrix, which mathematically describes the geometric transformation process from the second three-dimensional heart model to the first three-dimensional heart model, and serves as key data for subsequent reproduction and verification; (3.4) Precise spatial registration: Load the two model data after size standardization and the corresponding transformation matrix in MATLAB to reproduce the deformation process. Then, use the iterative nearest point ICP registration algorithm to perform high-precision spatial matching of the two sets of data.
[0018] Furthermore, step (4) includes the following sub-steps: (4.1) Establish spatial correspondence and data fusion: Perform elastic deformation and spatial registration to achieve consistency between the second three-dimensional heart model and the first three-dimensional heart model in terms of geometric shape and spatial scale; (4.2) Point-by-point mapping to generate a new model: In the MATLAB environment, for each spatial data point e in the second three-dimensional heart model, based on the nearest neighbor search algorithm, the nearest point m in the three-dimensional space of the first three-dimensional heart model is found, and the fiber data of point m is assigned to point e, thereby generating a new data model, which simultaneously contains the original spatial coordinates, voltage values and corresponding fiber values. (4.3) Output and application: The voltage distribution characteristics of the second three-dimensional heart model are spatially synchronized with the fibrosis distribution of the first three-dimensional heart model to form a composite model with multi-dimensional physical property information, providing a unified data foundation for cross-modal data comparison, joint modeling and analysis.
[0019] Furthermore, the method also includes step (5), which quantitatively analyzes the spatial overlap and consistency between the fibrous region and the low voltage region on the fused image.
[0020] A matching device for magnetic resonance images and electrical standard images is also provided, which includes: The first model reconstruction module is configured to reconstruct a first three-dimensional heart model based on cardiac magnetic resonance images, including the left atrial anatomy and the intensity of fibrosis. The second model reconstruction module is configured to reconstruct a second three-dimensional heart model reflecting the geometry and voltage distribution of the left atrium from the data points with three-dimensional spatial coordinates (X,Y,Z) and corresponding voltage values (V) collected by the electrophysiological mapping system. The registration module is configured to elastically deform the second three-dimensional heart model, correct the overall displacement and rotation, compensate for the morphological differences caused by different examination conditions, and make it optimally matched with the first three-dimensional heart model in terms of overall shape and local details. The fusion module is configured to map the voltage information of each data point on the second three-dimensional heart model to the corresponding anatomical position on the first three-dimensional heart model through coordinate transformation, thereby generating a fusion model that simultaneously carries fibrosis information from magnetic resonance imaging and voltage information from electrophysiology.
[0021] Furthermore, the first model reconstruction module executes: (1.1) Cardiac magnetic resonance imaging (MRI) was used as the raw data; (1.2) In CEMRG software, the first three-dimensional heart model of the left atrium was reconstructed using magnetic resonance images. The data included three-dimensional spatial coordinates and corresponding fibrosis data. (1.3) Import the first three-dimensional heart model into MATLAB software for subsequent analysis; The second model reconstruction module executes: (2.1) Data acquired by the CARTO electrophysiological mapping system were used as raw data; (2.2) Mark the anatomical structures in CARTO, including the ablation zone of the left pulmonary vein, the ablation zone of the right pulmonary vein, the mitral valve annulus, and the left atrial appendage, and then export the data; (2.3) Reconstruct the second three-dimensional heart model in MATLAB and color-code the voltage distribution.
[0022] Furthermore, the registration module performs: (3.1) Data preparation: In MATLAB, the second three-dimensional heart model and the first three-dimensional heart model are stored as data structures containing three-dimensional spatial coordinates and their corresponding physical properties. Each set of coordinates in the second three-dimensional heart model is associated with a voltage measurement value, while the coordinates of the first three-dimensional heart model correspond to its fibrosis degree quantification value. (3.2) Size standardization: The three-dimensional coordinate data of the first three-dimensional heart model and the second three-dimensional heart model are simultaneously imported into CloudCompare software. Using its registration and matching scale function, the spatial scale of the first three-dimensional heart model is used as the reference to perform elastic deformation of the second three-dimensional heart model with scaling as the core operation. (3.3) Deformation process record: Scale matching and deformation operation generate a complete transformation matrix, which mathematically describes the geometric transformation process from the second three-dimensional heart model to the first three-dimensional heart model, and serves as key data for subsequent reproduction and verification; (3.4) Precise spatial registration: Load the two model data after size standardization and the corresponding transformation matrix in MATLAB to reproduce the deformation process. Then, use the iterative nearest point ICP registration algorithm to perform high-precision spatial matching of the two sets of data.
[0023] Furthermore, the fusion module performs: (4.1) Establish spatial correspondence and data fusion: Perform elastic deformation and spatial registration to achieve consistency between the second three-dimensional heart model and the first three-dimensional heart model in terms of geometric shape and spatial scale; (4.2) Point-by-point mapping to generate a new model: In the MATLAB environment, for each spatial data point e in the second three-dimensional heart model, based on the nearest neighbor search algorithm, the nearest point m in the three-dimensional space of the first three-dimensional heart model is found, and the fiber data of point m is assigned to point e, thereby generating a new data model, which simultaneously contains the original spatial coordinates, voltage values and corresponding fiber values. (4.3) Output and application: The voltage distribution characteristics of the second three-dimensional heart model are spatially synchronized with the fibrosis distribution of the first three-dimensional heart model to form a composite model with multi-dimensional physical property information, providing a unified data foundation for cross-modal data comparison, joint modeling and analysis.
[0024] The present invention will now be described in more detail.
[0025] This invention provides a method and system for left atrial multimodal image fusion in patients with atrial fibrillation, aiming to achieve automatic and accurate matching and fusion of cardiac magnetic resonance images and intracardiac electrophysiological mapping images. Its core process is as follows:
[0026] 1. Based on cardiac magnetic resonance images: Using MRI images, a refined three-dimensional cardiac model (Model M) is reconstructed, including the anatomical structure of the left atrium and the intensity of fibrosis.
[0027] (1.1) The original data is a cardiac magnetic resonance tomography image; (1.2) In CEMRG software, the model M of the left atrium was reconstructed using nuclear magnetic resonance images (data including three-dimensional spatial coordinates and corresponding fibrosis data). (1.3) Import the model M into the MATLAB software for subsequent analysis.
[0028] 2. Based on electrophysiological mapping point cloud: The data points with three-dimensional spatial coordinates (X, Y, Z) and corresponding voltage values (V) collected by the electrophysiological mapping system are used to reconstruct another three-dimensional heart model (model E) that can reflect the geometry and voltage distribution of the left atrium.
[0029] (2.1) The raw data are those collected by the CARTO electrophysiological mapping system; (2.2) Mark the anatomical structures (including the ablation rings of the left pulmonary vein, right pulmonary vein, mitral valve annulus, and left atrial appendage) in CARTO → export the data; (2.3) Reconstruct a three-dimensional heart model in MATLAB and color-code the voltage distribution.
[0030] 3. Apply elastic deformation (scaling, displacement) to model E to achieve optimal matching with model M in terms of overall shape and local details. This process corrects for overall displacement and rotation, and compensates for shape differences caused by different inspection states.
[0031] (3.1) Data preparation: In MATLAB, model E and model M are stored as data structures containing three-dimensional spatial coordinates and their corresponding physical properties. Each set of coordinates in model E is associated with a voltage measurement value, while the coordinates in model M correspond to its fiber degree quantification value.
[0032] (3.2) Size Standardization: The three-dimensional coordinate data of the two models are simultaneously imported into CloudCompare software. Using its "Registration-Scale Matching" function, model E is subjected to elastic deformation with scaling as the core operation, with the spatial scale of model M as the reference. This step ensures that the two have consistent geometric dimensions, laying the foundation for subsequent spatial comparison.
[0033] (3.3) Deformation process record: The above scale matching and deformation operations will generate a complete transformation matrix. This matrix mathematically describes the geometric transformation process from model E to model M and is the key data for subsequent reproduction and verification.
[0034] (3.4) Precise Spatial Registration: Load the two size-normalized model data and their corresponding transformation matrices into MATLAB to reproduce the deformation process. Then, use the Iterative Closest Point (ICP) registration algorithm to perform high-precision spatial matching between the two sets of data. This step aims to further optimize the spatial alignment between the models, achieve optimal correspondence at the data point level, and ensure the accuracy and reliability of subsequent comparative analysis.
[0035] 4. After registration, the voltage information of each data point on model E is precisely mapped to the corresponding anatomical location on model M through coordinate transformation, thereby generating a fusion model. This model simultaneously carries "fibrosis information" from magnetic resonance imaging and "voltage information" from electrophysiology.
[0036] (4.1) Establishing spatial correspondence and data fusion: After completing elastic deformation and spatial registration, model E and model M achieve consistency in geometric shape and spatial scale, and have the basis for data association and fusion.
[0037] For example, suppose the coordinates of a point in model E are (15.2, 8.3, 6.5). After scaling and registration adjustments, its corresponding spatial location can be found in model M, let's say (15.1, 8.4, 6.5). By establishing this spatial mapping relationship, the physical properties (fiberization values) of model M can be accurately assigned to the corresponding coordinates in model E.
[0038] (4.2) Point-by-point mapping to generate a new model: In the MATLAB environment, for each spatial data point e in model E, based on the nearest neighbor search algorithm, its nearest point m in three-dimensional space is found in model M, and the fiberized data of point m is assigned to point e. After this process is completed, a new data model will be generated, which simultaneously contains the original spatial coordinates, voltage values, and corresponding fiberized values.
[0039] For example, suppose a data point in model E has coordinates (12.0, 9.5, 4.2) and a voltage value of 2.1 mV. Its nearest neighbor in model M has coordinates (11.9, 9.4, 4.3) and a fibrillation value of 0.76. After data fusion, this point in the new model will integrate the following complete information: coordinates (12.0, 9.5, 4.2), voltage 2.1 mV, and fibrillation value 0.76.
[0040] (4.3) Output and application: This method can spatially synchronize the voltage distribution characteristics of model E with the fiber distribution of model M to form a composite model with multi-dimensional physical property information, providing a unified data foundation for cross-modal data comparison, joint modeling and analysis.
[0041] 5. The spatial overlap and consistency between the left atrial fibrosis region and the low voltage region can be directly and quantitatively analyzed on the fused image.
[0042] The advantages of this invention compared to the prior art are as follows:
[0043] 1. It achieves the integration of automation and high precision: It replaces subjective and rough visual comparison, realizes automated spatial matching of two different source images, and improves the comparative analysis from "qualitative" to "precise quantitative".
[0044] 2. It provides powerful quantitative analysis tools: It provides a series of objective and reproducible quantitative indicators for the consistency of fibrosis and electrophysiology in clinical research, promoting the development of evidence-based medicine in this field.
[0045] The reasons for these advantages are as follows:
[0046] 1. Innovative Technical Approach: The core algorithm employed non-rigid registration, rather than simple rigid alignment. This fully considers and compensates for the morphological changes of the heart as a flexible organ under different examinations, which is the fundamental technical reason for achieving high-precision matching.
[0047] 2. Deep Fusion at the Data Level: This invention does not simply overlay images, but rather fuses information at the level of the reconstructed 3D model. This allows for more complex geometric processing and attribute mapping, laying the foundation for subsequent precise spatial analysis.
[0048] 3. Clinically problem-oriented design: The entire design closely revolves around the core clinical scientific question of "the relationship between fibrosis and low voltage zone". The developed quantitative tools directly serve to answer this question, thus having extremely high clinical practical value and guiding significance.
[0049] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Any simple modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of the present invention shall still fall within the protection scope of the present invention.
Claims
1. A method for matching magnetic resonance images with electrical target images, characterized in that: It includes the following steps: (1) Based on cardiac magnetic resonance images, a first three-dimensional cardiac model containing the anatomical structure of the left atrium and the intensity of fibrosis was reconstructed; (2) The data points with three-dimensional spatial coordinates (X, Y, Z) and corresponding voltage values (V) collected by the electrophysiological mapping system are used to reconstruct a second three-dimensional heart model that reflects the geometry and voltage distribution of the left atrium. (3) The second three-dimensional heart model is elastically deformed to correct the overall displacement and rotation, and to compensate for the morphological differences caused by different examination conditions, so that it achieves the best match with the first three-dimensional heart model in terms of overall shape and local details. (4) The voltage information of each data point on the second three-dimensional heart model is mapped to the corresponding anatomical position of the first three-dimensional heart model through coordinate transformation, thereby generating a fusion model that simultaneously carries fibrosis information from magnetic resonance and voltage information from electrophysiology.
2. The method for matching magnetic resonance images and electrical target images according to claim 1, characterized in that: Step (1) includes the following sub-steps: (1.1) Cardiac magnetic resonance imaging (MRI) was used as the raw data; (1.2) In CEMRG software, the first three-dimensional heart model of the left atrium was reconstructed using magnetic resonance images. The data included three-dimensional spatial coordinates and corresponding fibrosis data. (1.3) Import the first three-dimensional heart model into MATLAB software for subsequent analysis.
3. The method for matching magnetic resonance images and electrical target images according to claim 2, characterized in that: Step (2) includes the following sub-steps: (2.1) Data acquired by the CARTO electrophysiological mapping system were used as raw data; (2.2) Mark the anatomical structures in CARTO, including the ablation zone of the left pulmonary vein, the ablation zone of the right pulmonary vein, the mitral valve annulus, and the left atrial appendage, and then export the data; (2.3) Reconstruct the second three-dimensional heart model in MATLAB and color-code the voltage distribution.
4. The method for matching magnetic resonance images and electrical target images according to claim 3, characterized in that: Step (3) includes the following sub-steps: (3.1) Data preparation: In MATLAB, the second three-dimensional heart model and the first three-dimensional heart model are stored as data structures containing three-dimensional spatial coordinates and their corresponding physical properties. Each set of coordinates in the second three-dimensional heart model is associated with a voltage measurement value, while the coordinates of the first three-dimensional heart model correspond to its fibrosis degree quantification value. (3.2) Size standardization: The three-dimensional coordinate data of the first three-dimensional heart model and the second three-dimensional heart model are simultaneously imported into CloudCompare software. Using its registration and matching scale function, the spatial scale of the first three-dimensional heart model is used as the reference to perform elastic deformation of the second three-dimensional heart model with scaling as the core operation. (3.3) Deformation process record: Scale matching and deformation operation generate a complete transformation matrix, which mathematically describes the geometric transformation process from the second three-dimensional heart model to the first three-dimensional heart model, and serves as key data for subsequent reproduction and verification; (3.4) Precise spatial registration: Load the two model data after size standardization and the corresponding transformation matrix in MATLAB to reproduce the deformation process. Then, use the iterative nearest point ICP registration algorithm to perform high-precision spatial matching of the two sets of data.
5. The method for matching magnetic resonance images and electrical target images according to claim 4, characterized in that: Step (4) includes the following sub-steps: (4.1) Establish spatial correspondence and data fusion: Perform elastic deformation and spatial registration to achieve consistency between the second three-dimensional heart model and the first three-dimensional heart model in terms of geometric shape and spatial scale; (4.2) Point-by-point mapping to generate a new model: In the MATLAB environment, for each spatial data point e in the second three-dimensional heart model, based on the nearest neighbor search algorithm, the nearest point m in the three-dimensional space of the first three-dimensional heart model is found, and the fiber data of point m is assigned to point e, thereby generating a new data model, which simultaneously contains the original spatial coordinates, voltage values and corresponding fiber values. (4.3) Output and application: The voltage distribution characteristics of the second three-dimensional heart model are spatially synchronized with the fibrosis distribution of the first three-dimensional heart model to form a composite model with multi-dimensional physical property information, providing a unified data foundation for cross-modal data comparison, joint modeling and analysis.
6. The method for matching magnetic resonance images and electrical target images according to claim 5, characterized in that: The method also includes step (5), which involves quantitatively analyzing the spatial overlap and consistency between the fibrous region and the low-voltage region on the fused image.
7. A matching device for magnetic resonance images and electrical standard images, characterized in that: It includes: The first model reconstruction module is configured to reconstruct a first three-dimensional heart model based on cardiac magnetic resonance images, including the left atrial anatomy and the intensity of fibrosis. The second model reconstruction module is configured to reconstruct a second three-dimensional heart model reflecting the geometry and voltage distribution of the left atrium from the data points with three-dimensional spatial coordinates (X, Y, Z) and corresponding voltage values (V) collected by the electrophysiological mapping system. The registration module is configured to elastically deform the second three-dimensional heart model, correct the overall displacement and rotation, compensate for the morphological differences caused by different examination conditions, and make it optimally matched with the first three-dimensional heart model in terms of overall shape and local details. The fusion module is configured to map the voltage information of each data point on the second three-dimensional heart model to the corresponding anatomical position on the first three-dimensional heart model through coordinate transformation, thereby generating a fusion model that simultaneously carries fibrosis information from magnetic resonance imaging and voltage information from electrophysiology.
8. The matching device for magnetic resonance images and electrical standard images according to claim 7, characterized in that: The first model reconstruction module executes: (1.1) Cardiac magnetic resonance imaging (MRI) was used as the raw data; (1.2) In CEMRG software, the first three-dimensional heart model of the left atrium was reconstructed using magnetic resonance images. The data included three-dimensional spatial coordinates and corresponding fibrosis data. (1.3) Import the first three-dimensional heart model into MATLAB software for subsequent analysis; The second model reconstruction module executes: (2.1) Data acquired by the CARTO electrophysiological mapping system were used as raw data; (2.2) Mark the anatomical structures in CARTO, including the ablation zone of the left pulmonary vein, the ablation zone of the right pulmonary vein, the mitral valve annulus, and the left atrial appendage, and then export the data; (2.3) Reconstruct the second three-dimensional heart model in MATLAB and color-code the voltage distribution.
9. The matching device for magnetic resonance images and electrical target images according to claim 8, characterized in that: The registration module performs the following: (3.1) Data preparation: In MATLAB, the second three-dimensional heart model and the first three-dimensional heart model are stored as data structures containing three-dimensional spatial coordinates and their corresponding physical properties. Each set of coordinates in the second three-dimensional heart model is associated with a voltage measurement value, while the coordinates of the first three-dimensional heart model correspond to its fibrosis degree quantification value. (3.2) Size standardization: The three-dimensional coordinate data of the first three-dimensional heart model and the second three-dimensional heart model are simultaneously imported into CloudCompare software. Using its registration and matching scale function, the spatial scale of the first three-dimensional heart model is used as the reference to perform elastic deformation of the second three-dimensional heart model with scaling as the core operation. (3.3) Deformation process record: Scale matching and deformation operation generate a complete transformation matrix, which mathematically describes the geometric transformation process from the second three-dimensional heart model to the first three-dimensional heart model, and serves as key data for subsequent reproduction and verification; (3.4) Precise spatial registration: Load the two model data after size standardization and the corresponding transformation matrix in MATLAB to reproduce the deformation process. Then, use the iterative nearest point ICP registration algorithm to perform high-precision spatial matching of the two sets of data.
10. The matching device for magnetic resonance images and electrical target images according to claim 9, characterized in that: The fusion module performs the following: (4.1) Establish spatial correspondence and data fusion: Perform elastic deformation and spatial registration to achieve consistency between the second three-dimensional heart model and the first three-dimensional heart model in terms of geometric shape and spatial scale; (4.2) Point-by-point mapping to generate a new model: In the MATLAB environment, for each spatial data point e in the second three-dimensional heart model, based on the nearest neighbor search algorithm, the nearest point m in the three-dimensional space of the first three-dimensional heart model is found, and the fiber data of point m is assigned to point e, thereby generating a new data model, which simultaneously contains the original spatial coordinates, voltage values and corresponding fiber values. (4.3) Output and application: The voltage distribution characteristics of the second three-dimensional heart model are spatially synchronized with the fibrosis distribution of the first three-dimensional heart model to form a composite model with multi-dimensional physical property information, providing a unified data foundation for cross-modal data comparison, joint modeling and analysis.