Heart current source positioning method

Abstract

The invention provides a heart current source positioning method, which comprises the following steps of: acquiring magnetocardiogram data of a target user, and performing contour reconstruction on a target contour region in a heart general model by using the magnetocardiogram data to acquire a personalized heart model; subdividing the personalized heart model to obtain a plurality of tetrahedrons, and selecting an optimal tetrahedron from the plurality of tetrahedrons according to the personalized heart model and the effective magnetocardiogram data; performing multiple rounds of first finite subdivision on the preferred tetrahedron, and when each round of first finite subdivision is finished, determining whether a preferred sub-tetrahedron obtained by the round of first finite subdivision meets a first subdivision stop condition; and if the first subdivision stop condition is met, determining the final current source position according to the central point position of the optimal sub-tetrahedron obtained by the first finite subdivision of the round. By adopting the positioning method of the heart current source, the problems of low positioning efficiency and low positioning accuracy in the positioning process of the existing heart current source are solved.

Description

Technical Field

[0001] The present application relates to the technical field of biomedical signal analysis, and in particular to a method for locating a cardiac current source. Background Art

[0002] Magnetocardiography is a non-invasive inspection technology with no radiation, no contact, and high spatial resolution. The inverse magnetocardiography problem is based on the principle of electromagnetics, and quickly restores the magnetocardiography data in the magnetocardiography into a current source inside the heart, which is of great significance for detecting arrhythmias and myocardial ischemia. At present, the restoration and positioning method of the inverse magnetocardiography problem is usually to establish a heart model based on human medical images (such as CT, MRI), and then restore and locate the inverse magnetocardiography problem based on the point set inside the heart. For example: the patent application with publication number CN117493812A discloses: the patient's heart model is obtained based on CT image segmentation and triangular meshing processing, and then the heart model is rasterized, divided into grids with a resolution of 5mm, and the inverse magnetocardiography problem is calculated one by one on the grid points to obtain the optimal current source position, and then a sphere with a radius of 2cm is taken based on the optimal current source position, and the grid is divided with a resolution of 1mm for further refined positioning.

[0003] However, when using the above-mentioned cardiac current source positioning method, on the one hand, it is necessary to establish a cardiac model through human medical images, which is not only time-consuming and costly, but also harmful to the human body. On the other hand, if the grid is divided at a resolution of 5mm, for an adult's heart, the cardiac magnetic inversion problem requires traversing tens of thousands of cubes to calculate the current source, which not only takes a lot of time, but also may lead to inaccurate current source restoration results because the cube cannot accurately describe the irregular heart. Summary of the invention

[0004] In view of this, the purpose of the present application is to provide a cardiac current source positioning method to solve the problems of low positioning efficiency and positioning accuracy in the existing cardiac current source positioning process.

[0005] In a first aspect, an embodiment of the present application provides a method for locating a cardiac current source, comprising:

[0006] Acquire the target user's magnetic heart data, and use the magnetic heart data to reconstruct the target contour area in the universal heart model to obtain a personalized heart model for the target user;

[0007] The personalized heart model is segmented to obtain a plurality of tetrahedrons, and a preferred tetrahedron is selected from the plurality of tetrahedrons according to the personalized heart model and the effective magnetic heart data;

[0008] Perform multiple rounds of first finite-time splitting on the preferred tetrahedron, and at the end of each round of first finite-time splitting, determine whether the preferred sub-tetrahedron obtained by the first finite-time splitting in this round meets the first splitting stop condition, wherein the first finite-time splitting is to perform finite-time splitting on the tetrahedron according to the center point of the preferred tetrahedron and the number of faces of the preferred tetrahedron;

[0009] If the first segmentation stop condition is met, the final current source position is determined according to the center point position of the preferred sub-tetrahedron obtained by the first finite segmentation in this round.

[0010] Optionally, the target contour area in the universal heart model is reconstructed using the magnetic heart data, including: filtering and denoising the magnetic heart data to obtain denoised magnetic heart data, and selecting the denoised magnetic heart data within an effective time as effective magnetic heart data, where the effective time is a portion of time selected from the time when the QRS complex occurs; based on the effective magnetic heart data, selecting a target voxel point with maximum current power at each moment within the effective time from a preset source space, and constructing an edge curve using the target voxel point at each moment; and reconstructing the target contour area of ​​the universal heart model using the edge curve to obtain a personalized heart model.

[0011] Optionally, an edge curve is constructed using the target voxel points at each moment, including: projecting multiple target voxel points into a preset plane, selecting edge projection points in a preset direction in the preset plane as contour points of the target contour area, the preset plane being a plane perpendicular to the direction of the line between the heart and the magnetic heart detector; and performing curve fitting on the contour points to obtain the edge curve.

[0012] Optionally, based on the effective magnetocardiographic data, a target voxel point with the maximum current power at each moment within the effective time is selected from the preset source space, including: for each moment, based on the effective magnetocardiographic data and the initial guidance field matrix corresponding to each voxel point in the preset source space, determining the minimum variance filter matrix; using the minimum variance filter matrix, determining the current power of each voxel point at that moment, and selecting the voxel point with the maximum current power as the target voxel point at that moment.

[0013] Optionally, the target contour area of ​​the universal heart model is reconstructed using edge curves to obtain a personalized heart model, including: projecting the universal heart model into a preset plane to obtain a two-dimensional universal heart contour; adjusting the edge shape of the two-dimensional universal heart contour to align the edge of the two-dimensional universal heart contour with the edge curve; and restoring the adjusted two-dimensional universal heart contour to a three-dimensional heart model to obtain a personalized heart model for the target user.

[0014] Optionally, the personalized heart model is segmented to obtain a plurality of tetrahedrons, including: reconstructing a triangular mesh on the surface of the personalized heart model using a mesh reconstruction algorithm; and performing Delaunay triangulation on the interior of the personalized heart model using the reconstructed triangular mesh to obtain a plurality of tetrahedrons.

[0015] Optionally, based on the personalized heart model and the effective magnetic heart data, a preferred tetrahedron is selected from multiple tetrahedrons, including: using a boundary element algorithm to discretize the magnetic field equation corresponding to the personalized heart model to obtain a guidance field matrix corresponding to the center point of each tetrahedron; based on the correlation between the guidance field matrix corresponding to the center point of each tetrahedron and the effective magnetic heart data, a preferred tetrahedron is selected from multiple tetrahedrons.

[0016] Optionally, multiple rounds of first finite-time splitting are performed on the preferred tetrahedron, including: for each round of first finite-time splitting, splitting the previous round of preferred sub-tetrahedrons based on each side face and the center point of the preferred sub-tetrahedron in the previous round to obtain a preset number of sub-tetrahedrons in this round; selecting the preferred sub-tetrahedron of this round from the preset number of sub-tetrahedrons of this round according to the correlation between the guiding field matrix corresponding to the center point of each sub-tetrahedron in this round and the effective magnetocardiographic data, so as to use the preferred sub-tetrahedron of this round to perform the next round of first finite-time splitting.

[0017] Optionally, the method also includes: for each tetrahedron, performing singular value decomposition on the guiding field matrix corresponding to the center point of the tetrahedron to obtain the left eigenvector of the guiding field matrix; performing singular value decomposition on the product of the transpose of the left eigenvector of the guiding field matrix and the normalized effective magnetocardio data to determine the correlation between the guiding field matrix of the center point of the tetrahedron and the effective magnetocardio data.

[0018] Optionally, the final current source position is determined based on the center point position of the preferred sub-tetrahedron obtained in this round of first finite splitting, including: taking the center point position of the optimal sub-tetrahedron as the first final current source position, and determining the first magnetic field generated by the first final current source; performing multiple rounds of second finite splitting on multiple tetrahedrons based on the residual of the effective magnetic heart data and the first magnetic field, and at the end of each round of second finite splitting, determining whether the second finite splitting of this round meets the second splitting stop condition; if the second splitting stop condition is met, determining the final current source position based on the center point position of the preferred sub-tetrahedron obtained in each round of second finite splitting.

[0019] The embodiments of the present application bring the following beneficial effects:

[0020] The embodiment of the present application provides a method for locating a cardiac current source, which can reconstruct a universal cardiac model using the target user's magnetic heart data to obtain a personalized cardiac model that reflects the characteristics of the target user, without the need to construct a cardiac model with the help of human medical images. While improving the efficiency of constructing the cardiac model, the efficiency of locating the current source is also improved. In addition, the first finite number of divisions is to divide the tetrahedron according to the number of faces of the tetrahedron, which can not only effectively reduce the number of divisions and quickly determine the final current source position, but also avoid the problem of inaccurate positioning caused by dividing the cube. Compared with the cardiac current source positioning method in the prior art, it solves the problem of low positioning efficiency and positioning accuracy.

[0021] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, preferred embodiments are specifically cited below and described in detail with reference to the attached drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0023] Figure 1 A flow chart showing a method for locating a cardiac current source provided in an embodiment of the present application is shown;

[0024] Figure 2 A flow chart showing a method for obtaining a personalized heart model provided in an embodiment of the present application is shown;

[0025] Figure 3 A schematic diagram showing the positions of contour points of a target contour area provided in an embodiment of the present application;

[0026] Figure 4 A schematic diagram of an edge curve provided in an embodiment of the present application is shown;

[0027] Figure 5 A schematic structural diagram of a cardiac current source positioning device provided in an embodiment of the present application is shown;

[0028] Figure 6 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application is shown. DETAILED DESCRIPTION

[0029] To make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments. The components of the embodiments of the present application usually described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the application claimed for protection, but merely represents the selected embodiments of the present application. Based on the embodiments of the present application, each other embodiment obtained by those skilled in the art without making creative work belongs to the scope of protection of the present application.

[0030] It is worth noting that before the present application was proposed, magnetocardiography was a non-invasive inspection technology with no radiation, no contact, and high spatial resolution. The inverse magnetocardiography problem is based on the principle of electromagnetics, and the magnetocardiography data in the magnetocardiography is quickly restored to the current source inside the heart, which is of great significance for the detection of arrhythmias and myocardial ischemia. At present, the restoration and positioning method of the inverse magnetocardiography problem is usually to establish a heart model based on human medical images (such as CT, MRI), and then restore and locate the inverse magnetocardiography problem based on the point set inside the heart. For example: the patent application with publication number CN117493812A discloses: the patient's heart model is obtained based on CT image segmentation and triangular meshing processing, and then the heart model is rasterized, and the grid is divided with a resolution of 5mm. The inverse magnetocardiography problem is calculated one by one on the grid points to obtain the optimal current source position, and then a sphere with a radius of 2cm is taken based on the optimal current source position, and the grid is divided with a resolution of 1mm for further refined positioning.

[0031] However, when using the above-mentioned cardiac current source positioning method, on the one hand, it is necessary to establish a cardiac model through human medical images, which is not only time-consuming and costly, but also causes radiation damage to the human body. On the other hand, for an adult, the long diameter of the heart is 12 cm, the horizontal diameter is 8 cm, and the anterior-posterior diameter is 6 cm. If the grid is divided at a resolution of 5 mm, at least 4,600 cubes may be divided. Then, a sphere with a radius of 2 cm is taken as the center of the optimal point, and the grid is divided at a resolution of 1 mm, which may divide at least 10,000 cubes. In other words, the cardiac magnetic inverse problem requires traversing tens of thousands of cubes in total to calculate the current source, which takes a lot of time. For some urgent patients, if the prediction result takes a long time, the best time for surgery may be missed. At the same time, the cube cannot accurately describe an irregular heart, which may lead to inaccurate current source restoration results.

[0032] Based on this, an embodiment of the present application provides a method for positioning a cardiac current source to improve the positioning efficiency and positioning accuracy of the current source.

[0033] See also Figure 1 , Figure 1 This is a flow chart of a method for locating a cardiac current source provided in an embodiment of the present application. Figure 1 As shown, the method for locating a cardiac current source provided in an embodiment of the present application includes:

[0034] Step S101, obtaining the target user's magnetocardiographic data, and reconstructing the target contour area in the universal heart model using the magnetocardiographic data to obtain a personalized heart model of the target user;

[0035] Step S102, dividing the personalized heart model into multiple tetrahedrons, and selecting a preferred tetrahedron from the multiple tetrahedrons according to the personalized heart model and the effective magnetic heart data;

[0036] Step S103, performing multiple rounds of first finite-time splitting on the preferred tetrahedron, and at the end of each round of first finite-time splitting, determining whether the preferred sub-tetrahedron obtained by the first finite-time splitting in this round satisfies the first splitting stop condition;

[0037] Step S104: if the first segmentation stop condition is met, the final current source position is determined according to the center point position of the preferred sub-tetrahedron obtained by the first finite segmentation in this round.

[0038] Through the above steps, the present application can reconstruct the universal heart model using the target user's magnetic heart data to obtain a personalized heart model that reflects the characteristics of the target user, without the need to construct the heart model with the help of human medical images. While improving the efficiency of building the heart model, it also improves the efficiency of positioning the current source. In addition, the first finite number of divisions is to divide the tetrahedron according to the number of faces of the tetrahedron, which can not only effectively reduce the number of divisions and quickly determine the final current source position, but also avoid the problem of inaccurate positioning caused by dividing the cube, solving the problem of low positioning efficiency and positioning accuracy.

[0039] Below Figure 1 Each step is described in detail.

[0040] In step S101, the magnetic heart data of the target user is obtained, and the magnetic heart data is used to reconstruct the contour of the target contour area in the universal heart model to obtain a personalized heart model of the target user.

[0041] Here, the target user may refer to the object of collecting the magnetic heart data. For example, the target user may be a patient suffering from heart disease.

[0042] The universal heart model may refer to a universal three-dimensional heart model. The universal heart model is not a heart model of the target user. The universal heart model may be obtained through historical CTA (traumatic vascular imaging technology) examination data of other hospitals or from other channels, for example, obtaining a heart model from the Internet.

[0043] The target contour area may refer to an edge contour area of ​​the left ventricle.

[0044] The personalized heart model may refer to a heart model that can reflect the heart contour characteristics of the target user.

[0045] Refer to the following Figure 2 Let's introduce the process of obtaining a personalized heart model.

[0046] Figure 2 A flowchart of a personalized heart model acquisition method provided in an embodiment of the present application is shown. Figure 2 As shown, the method for obtaining a personalized heart model includes:

[0047] Step S201 : filtering and denoising the magnetocardiographic data to obtain denoised magnetocardiographic data, and selecting the denoised magnetocardiographic data within a valid time as valid magnetocardiographic data.

[0048] In the embodiment of the present application, the target user is in a lying state, and the target user's magnetic heart data is collected by a magnetic heart detector, and the magnetic heart data is filtered and denoised to obtain the denoised magnetic heart data. The denoising process at this time is mainly used to reduce the environmental noise and physiological noise, for example, abnormal sounds in the room, and the target user's breathing and digestion noise can be reduced.

[0049] Subsequently, the filtered and denoised magnetic heart data is averaged, which is to average the magnetic heart data of multiple cycles of the target user to further suppress noise. The P wave, QRS complex and T wave are segmented from the magnetic heart data after the average processing to eliminate the magnetic heart data in a stable state when the heart is not beating, and the magnetic heart data of the P wave, QRS complex and T wave are retained, thereby obtaining the pre-processed denoised magnetic heart data.

[0050] For the preprocessed denoised magnetic heart data, according to the characteristic of the QRS complex that the current source moves upward in the ventricular septum-apex-lateral wall, the QRS time is a continuous time interval in the cycle. The effective time is selected in the QRS time, and the preprocessed denoised magnetic heart data within the effective time is selected as the effective magnetic heart data.

[0051] The effective time is a part of the time selected from the time when the QRS complex appears. The starting point of the effective time is the moment when the negative magnetic pole points to the positive magnetic pole by 135 degrees, at which time the current source of the QRS complex is at the apex position; the end point of the effective time is the moment when the positive magnetic pole points to the negative magnetic pole by less than 30 degrees, at which time the current source of the QRS complex is at the side wall position. Since the cardiac current in the effective time will flow upward through the apex, a point on the edge of the heart model corresponding to the magnetic heart data can be used to calibrate the apex position of the target user.

[0052] Step S202: based on the effective magnetocardiographic data, a target voxel point with the maximum current power at each moment within the effective time is selected from the preset source space, and an edge curve is constructed using the target voxel point at each moment.

[0053] Here, firstly, based on the relative position relationship between the heart of the target user and the magnetocardiographic detector, a three-dimensional source space is constructed. The source space is a rectangular space, and the smallest volume element in the source space is called a voxel. Specifically, the size range of the source space can be determined according to the detection range of the magnetocardiographic detector. For example, if the coordinates of the magnetocardiographic detector are (0, 0, 240 mm), the x-axis detection range of the magnetocardiographic detector is 0 mm to 200 mm, the y-axis detection range is 0 mm to 200 mm, and the z-axis detection range is 0 mm to 240 mm, then the x-axis direction of the source space can be set to 100 mm to 200 mm, the y-axis direction of the source space can be set to 0 mm to 200 mm, and the z-axis direction of the source space can be set to 100 mm to 150 mm. Among them, the target user lies flat under the magnetocardiographic detector, the positive direction of the x-axis is the direction from left to right when the human body lies flat, the positive direction of the y-axis is the direction from the feet to the head when the human body lies flat, and the positive direction of the z-axis is the direction from the heart to the magnetocardiographic detector. Since the heart is located to the right of the center of the human body, the range of the x-axis direction in the source space is 100 mm-200 mm.

[0054] After determining the size range of the source space, 21 voxel points can be uniformly selected in the x-axis direction of the source space, that is, one voxel point is selected every 5 mm; 21 voxel points can be uniformly selected in the y-axis direction of the source space, that is, one voxel point is selected every 10 mm, and 11 voxel points can be uniformly selected in the z-axis direction of the source space, that is, one voxel point is selected every 5 mm. In this way, there are more than 4,000 voxel points in the source space.

[0055] When selecting the target voxel point with the maximum current power at each moment in the effective time, the minimum variance filter matrix can be determined for each moment based on the effective magnetocardiographic data and the initial guidance field matrix corresponding to each voxel point in the preset source space; the minimum variance filter matrix is ​​used to determine the current power of each voxel point at that moment, and the voxel point with the maximum current power is selected as the target voxel point at that moment. Here, the effective magnetocardiographic data at moment a is denoted as: b, the minimum variance filter matrix is ​​denoted as: W, and the current power of each voxel point is denoted as: qq T , then the calculation formula of the minimum variance filter matrix and current power of a single voxel point at this moment is:

[0056]

[0057] QQ T =W T bb T W;

[0058] In the above formula, Lf1 represents the initial guidance field matrix of each voxel point, which can be obtained according to the horizontal layered conductor model corresponding to the source space. tr represents trace, that is, the calculation matrix trace. Time a is the time selected within the effective time.

[0059] In this way, the current power of each voxel point at that moment can be calculated, and then the voxel point with the maximum current power is selected as the target voxel point at that moment within the effective time.

[0060] When using the target voxel points to construct the edge curve, multiple target voxel points at different times can be projected into the preset plane. The preset plane is a plane perpendicular to the direction of the line between the heart and the magnetic heart detector, that is, the xoy plane of the source space, and the preset plane is parallel to the body plane of the target user when lying flat. Then, the edge projection points in the preset direction in the preset plane xoy are selected as the contour points of the target contour area. For example, in the preset plane xoy, the point with the largest x value among the target voxel points with the same y-axis value can be used as the contour point of the left ventricle. At this time, the preset direction is the x-axis direction.

[0061] Refer to the following Figure 3 Let's introduce contour points.

[0062] Figure 3 A schematic diagram showing the positions of contour points of the target contour area provided in the embodiment of the present application is shown in FIG. Figure 3As shown, a plurality of contour points are obtained in the xoy plane, and the plurality of contour points include contour point 311, contour point 312, contour point 313, contour point 314, contour point 315, and contour point 316, and the contour points are represented by white solid circles. It can be seen that the right edge of the subject's heart model 320 is substantially consistent with the shape of the curves corresponding to these contour points, and the shape of the right edge of the universal heart model can be adjusted based on these contour points to obtain a personalized heart model.

[0063] Refer to the following Figure 4 Let's introduce the process of obtaining edge curves.

[0064] Figure 4 FIG. 4 shows a schematic diagram of an edge curve provided in an embodiment of the present application, such as Figure 4 As shown, after acquiring a plurality of contour points, curve fitting is performed on these contour points to obtain an edge curve 410. For example, Bezier curves may be used to fit these contour points.

[0065] Step S203: reconstruct the target contour area of ​​the universal heart model using the edge curve to obtain a personalized heart model.

[0066] Specifically, the universal heart model can be projected into the preset plane xoy to obtain a two-dimensional universal heart contour, so as to align the right edge of the two-dimensional universal heart contour with the edge curve. Then, the edge shape of the two-dimensional universal heart contour is adjusted, for example: the right edge of the two-dimensional universal heart contour is scaled up, down, left, and right, the apex position of the two-dimensional universal heart contour is overlapped with the contour point at the bottom of the fitting curve, and the edge of the two-dimensional universal heart contour is aligned with the edge curve. The original z-axis data of the adjusted two-dimensional universal heart contour is restored, so that the adjusted two-dimensional universal heart contour is restored to a three-dimensional heart model to obtain a personalized heart model for the target user.

[0067] In step S102, the personalized heart model is segmented to obtain a plurality of tetrahedrons, and a preferred tetrahedron is selected from the plurality of tetrahedrons according to the personalized heart model and the effective magnetic heart data.

[0068] In an embodiment of the present application, when the personalized heart model is segmented, a mesh reconstruction algorithm can be used to reconstruct a triangular mesh of the surface of the personalized heart model. Among them, the mesh reconstruction algorithm includes but is not limited to: Poisson reconstruction algorithm, and can also be a ball rotation algorithm Ball-PivotingAlgorithm (BPA). Then, the reconstructed triangular mesh is used to perform Delaunay triangulation on the inside of the personalized heart model to obtain a plurality of tetrahedrons. Here, the triangular mesh segmentation of the personalized heart model not only avoids the problem that the irregular heart cannot be accurately described when the cube segmentation is used, improves the accuracy of the current source position restoration positioning, but also reduces the number of heart model segmentation and the number of tetrahedrons segmented, and improves the segmentation efficiency and positioning efficiency. Among them, the multiple tetrahedrons obtained by the segmentation of the present application can constitute a complete personalized heart model, and one face of the edge tetrahedron is a triangle, and the triangle is part of the heart surface.

[0069] After the personalized heart model is divided into multiple tetrahedrons, not every voxel point in the tetrahedron has a current source, and the specific location of the current source needs to be screened out from these tetrahedrons. To this end, it is necessary to first screen out the preferred tetrahedron, and then further divide the preferred tetrahedron to accurately locate the current source position.

[0070] When selecting a preferred tetrahedron from multiple tetrahedrons, the boundary element algorithm can be used to discretize the magnetic field equation corresponding to the personalized heart model to obtain the guidance field matrix corresponding to the center point of each tetrahedron. The magnetic field equation may refer to the quasistatic Maxwell equation, which is used to characterize the magnetic field and current measured by the magnetic heart detector at position r (x, y, z). The relationship between the magnetic field and the current represented by the magnetic field equation is as follows:

[0071]

[0072]

[0073]

[0074]

[0075] In the above formula, B(r) represents the magnetic field, B0(r) represents the magnetic field directly generated by the current, and V(r) represents the electric potential on different surfaces, including S′ k , k = 1, 2, 3, the latter part of the B(r) calculation formula except B0(r) represents the magnetic field generated by the body potential effect of the secondary current, S′1, S′2, S′3 represent the curved surfaces formed by the heart surface, the body surface, and the lungs surface, respectively, and They represent the internal and external conductivity of different parts, l = 1, 2, 3, V′ represents the personalized heart internal model, which is used to limit the integration inside the heart, μ0 represents the magnetic permeability, J P (r′) represents the intensity and direction of the current, r′ represents the location where the current is generated inside the heart, V ∞ (r) also represents the potential generated by the current.

[0076] Considering the equivalent dipole, the current J P (r′) is represented by J P (r′)=J(r″)δ(r″-r′), where r″ represents the center point of the tetrahedral mesh in the personalized heart model, J(r″) represents the current source on r″, and δ(r″-r′) represents the δ function calculation formula. Then, J P (r′)=J(r″)δ(r″-r′) is substituted into the magnetic field equation, and the boundary element method is used to replace the above J P The mapping relationship between the current and the magnetic field can be determined by discretizing the four magnetic field equations of (r′), and then the guiding field matrix at the center point of each tetrahedral grid can be determined based on the mapping relationship between the current and the magnetic field. That is, the mapping relationship between the current and the magnetic field is represented by the guiding field matrix. Among them, r′ represents a point inside the personalized heart model, but it is in the integral, so r″ can be used in the equivalent current model to represent the center point of the tetrahedron grid.

[0077] According to the guidance field matrix of each tetrahedron center point, the magnetic field B in the z-axis direction at a single moment in the effective magnetic heart data z and current, a volume conductor conduction model at the center point r″ of each tetrahedral grid is constructed. The volume conductor conduction model is a forward and inverse problem model and can be expressed as:

[0078] B z =(B z (r1), B z (r2),…,B z (r M )) T =Lf2(r″)J(r″);

[0079] In the above formula, B z is the vector composed of the magnetic field of M channels in the z direction at a single moment measured by the magnetocardiographic detector, and its dimension is M×1. The number of M is determined by the magnetocardiographic detector. Lf2(r″) is the guidance field matrix at the grid center point r″. The guidance field matrix at different grid centers (tetrahedral grid centers) is different and is determined by the relative position relationship between the grid center point and the magnetocardiographic detector. Among them, B zis a vector at a single moment in the effective time. If 20 moments are taken in the effective time, there are 20 B z .

[0080] In one example, if only the magnetic field B0(r) directly generated by the current is considered, a horizontal layered conductor model and a corresponding guide field matrix can be obtained. The horizontal layered conductor model is also a forward and inverse problem model.

[0081] After determining the guidance field matrix of each tetrahedron grid center point, for each moment in the effective time, the preferred tetrahedron can be selected from multiple tetrahedrons according to the correlation between the guidance field matrix corresponding to the center point of each tetrahedron and the effective magnetic heart data. Specifically, for each tetrahedron, the guidance field matrix Lf2 corresponding to the center point of the tetrahedron can be subjected to singular value decomposition to obtain the decomposed guidance field matrix. The decomposed guidance field matrix Lf2 = USV T Since Lf2 has only three columns, the first three columns of the left eigenvector U can be selected and recorded as U3. Then the effective magnetocardiographic data B at the time a after the transposition and normalization of the selected left eigenvector U3 of the guidance field matrix is ​​obtained. z The product of Perform singular value decomposition, and select the largest singular value from the decomposition result as the value of the correlation between the guidance field matrix at the center point of the tetrahedron and the effective magnetocardiometry data at the time a.

[0082] After determining the correlation between the guidance field matrix corresponding to the center point of each tetrahedron and the effective magnetic heart data, the numerical values ​​of the correlations corresponding to all tetrahedrons are compared, and the tetrahedron with the largest correlation value is selected as the preferred tetrahedron. In this way, assuming that 3000 tetrahedrons are obtained after the personalized heart model is segmented, a preferred tetrahedron is selected from these 3000 tetrahedrons, and the preferred tetrahedron is recorded as A0.

[0083] In step S103, multiple rounds of first finite subdivisions are performed on the preferred tetrahedron, and at the end of each round of first finite subdivisions, it is determined whether the preferred sub-tetrahedron obtained by the first finite subdivisions in this round meets the first subdivision stop condition.

[0084] Here, the position of the current source at the time a can only be roughly located by the preferred tetrahedron A0, and further precise positioning is required. To this end, the preferred tetrahedron A0 can be subjected to multiple rounds of first finite-time divisions to accurately locate the position of the current source according to the results of the first finite-time divisions, and the first finite-time divisions refer to the finite-time divisions of the tetrahedron according to the center point of the preferred tetrahedron and the number of faces of the preferred tetrahedron.

[0085] When performing multiple rounds of first finite-time partitioning on the preferred tetrahedron, for each round of first finite-time partitioning, the previous round of preferred sub-tetrahedron can be partitioned based on each side face and the center point of the previous round of preferred sub-tetrahedron to obtain a preset number of sub-tetrahedrons of this round; the guiding field matrix corresponding to the center point of each sub-tetrahedron of this round is determined; and according to the correlation between the guiding field matrix corresponding to the center point of each sub-tetrahedron of this round and the effective magnetic heart data, the preferred sub-tetrahedron of this round is selected from the preset number of sub-tetrahedrons of this round, so as to use the preferred sub-tetrahedron of this round to perform the first finite-time partitioning of the next round.

[0086] For example: when performing the first finite number of splits in each round, the preferred tetrahedron A0 has a total of 4 sides. For each side, the center point of the preferred tetrahedron A0 can be connected to the three vertices of the side to form a new tetrahedron. In this way, after the preferred tetrahedron A0 is split, a total of four new tetrahedrons can be split. These four new tetrahedrons are the sub-tetrahedrons obtained after the first finite number of splits in the first round. For each sub-tetrahedron obtained in the first round of splitting, the guidance field matrix corresponding to the center point of each sub-tetrahedron in this round is determined, and the correlation between the guidance field matrix corresponding to the center point of each sub-tetrahedron in the first round and the effective magnetic heart data is calculated, and the sub-tetrahedron with the largest correlation is selected as the preferred sub-tetrahedron in this round. The preferred sub-tetrahedron obtained in the first finite number of splits in the first round is recorded as: A1. The calculation method of the correlation is the same as the calculation method when selecting the preferred tetrahedron A0, which will not be repeated here.

[0087] After the first round of the first finite number of segmentations is completed, determine whether the correlation corresponding to the selected preferred tetrahedron A1 is greater than the correlation threshold. If it is greater than the correlation threshold, it is determined that the first segmentation stop condition is met; if it is less than or equal to the correlation threshold, it is determined that the first segmentation stop condition is not met, and the second round of the first finite number of segmentations is started.

[0088] When performing the first finite number of subdivisions in the second round, the previous round of preferred subtetrahedrons can be subdivided based on each side face of the previous round of preferred subtetrahedrons and the center point of the preferred subtetrahedrons to obtain a preset number of subtetrahedrons in this round. At this time, the preferred subtetrahedron A1 is the preferred subtetrahedron in the previous round. When performing the first finite number of subdivisions in the second round, the preferred subtetrahedron A1 is subdivided into four subtetrahedrons. For each subtetrahedron obtained by the second round of subdivision, the guidance field matrix corresponding to the center point of each subtetrahedron in this round is determined, and the correlation between the guidance field matrix corresponding to the center point of each subtetrahedron in the second round and the effective magnetic heart data is calculated, and the subtetrahedron with the largest correlation is selected as the second round of preferred subtetrahedrons. The preferred subtetrahedron obtained by the first finite number of subdivisions in the second round is recorded as: A2. The calculation method of the correlation is the same as the calculation method when selecting the preferred tetrahedron A1, which will not be repeated here.

[0089] Similarly, after the first finite number of segmentations in the second round is completed, it is determined whether the correlation corresponding to the selected preferred tetrahedron A2 is greater than the correlation threshold. If it is greater than the correlation threshold, it is determined that the first segmentation stop condition is met; if it is less than or equal to the correlation threshold, it is determined that the first segmentation stop condition is not met, and the third round of first finite number of segmentations is started.

[0090] And so on, until the first segmentation stop condition is met, it means that the center point position of the preferred tetrahedron obtained by segmentation is almost the same as the actual current source position, at which time the entire first finite segmentation process can be ended.

[0091] In step S104, if the first segmentation stop condition is met, the final current source position is determined according to the center point position of the preferred sub-tetrahedron obtained by the first finite segmentation in this round.

[0092] The final current source position may refer to the final current source position at a certain moment in the effective time, and the final current source position at a certain moment may be one or more.

[0093] In the embodiment of the present application, when the first segmentation stop condition is met, it means that the magnetic field generated by the center point position calculated at this moment a is the most matched with the magnetic field measured by the magnetocentrometry detector, that is, the correlation is the highest. Then, the optimal sub-tetrahedron A that meets the first segmentation stop condition can be N The center point position r N1 As the first final current source position at time a. At this time, since the first segmentation stop condition is met and the first final current source position is obtained, the first round of the second finite segmentation process ends. In other words, the second finite segmentation process is nested outside the first finite segmentation process. After each round of the first finite segmentation is performed and the first segmentation stop condition is met, a round of the second finite segmentation ends. At the same time, the current direction of the first final current source position is calculated. At this time, due to the magnetic field B measured at this moment z and the guidance field matrix Lf2(r″) at the first final current source position are known, then B z =(B z (r1), B z (r2),…,B z (r M )) T =Lf2(r″)J(r″) is calculated by least square method, and the obtained J(r″)=[Lf2(r″)Lf2(r″) T ] -1 Lf2(r″) T B z This is the current direction of the first final current source.

[0094] In one example, after the first final current source position is determined, there may be a residual between the magnetic field at the first final current source position and the magnetic field measured by the magnetic heart detector, and the two are not completely equal. This may be caused by the current generating magnetic field at other positions in the heart. To this end, it is necessary to determine whether there are current sources at other positions. In this case, the first final current source position r can be determined first. N1 The corresponding first magnetic field in, Then, the residual magnetic field is determined according to the first magnetic field, that is, If the residual magnetic field is less than or equal to the set residual threshold, it is determined that the second segmentation stop condition is met, and there is no need to determine the remaining final current source positions; if the residual magnetic field is greater than the set residual threshold, it is determined that the second segmentation stop condition is not met, indicating that there are current sources at other locations, and the current sources at other locations need to be further located.

[0095] When the second segmentation condition is not met, the second round of the second finite segmentation process is started. First, based on the residual of the effective magnetic field data at the time a and the first magnetic field, multiple rounds of the first finite segmentation can be performed on multiple tetrahedrons until the second segmentation stop condition is met to determine the remaining final current source positions.

[0096] For example: the residual magnetic field As the magnetic field B when calculating the correlation z , and for each of the 3000 tetrahedrons of the personalized heart model in the above example, calculate the correlation between the guidance field matrix corresponding to the center point of the tetrahedron and the residual magnetic field, select the tetrahedron with the largest correlation as the preferred tetrahedron, and record the preferred tetrahedron at this time as: C0.

[0097] The preferred tetrahedron C0 is subjected to multiple rounds of first finite times division, and four tetrahedrons are obtained in each round of first finite times division. At the end of each round of first finite times division, it is determined whether the first finite times division in this round satisfies the first division stop condition.

[0098] For example: After the first round of the first finite number of subdivisions, multiple sub-tetrahedrons are obtained, and the guide field matrix and residual magnetic field of the center point of each sub-tetrahedron are determined. The correlation between them is used to select the sub-tetrahedron with the largest correlation as the optimal sub-tetrahedron. At this time, the preferred sub-tetrahedron is recorded as: C1. Determine the guidance field matrix and residual magnetic field of the center point of the optimal sub-tetrahedron C1 Whether the correlation between them satisfies the first segmentation stop condition. If the first segmentation stop condition is not met, the first finite segmentation of the subsequent rounds will continue until the first segmentation stop condition is met. If the first segmentation stop condition is met, the center point position of the optimal sub-tetrahedron C1 is the second final current source position at time a, and at this time, the second round of the second finite segmentation process ends.

[0099] After the second round of the second finite subdivision process is completed, the second magnetic field at the center point of the priority tetrahedron C1 is determined Determine the latest residual magnetic field based on the second magnetic field Then, determine the latest residual magnetic field Whether the second segmentation stop condition is met, if the latest residual magnetic field If the residual magnetic field is greater than the set residual threshold, it is determined that the second segmentation stop condition is not met, and the third round of second finite segmentation begins; if the latest residual magnetic field If it is less than or equal to the set residual threshold, it is determined that the second segmentation stop condition is met.

[0100] When performing the third round of the second finite subdivision, the latest residual magnetic field As the magnetic field B when calculating the correlation z ,at this time For each of the 3000 tetrahedrons of the personalized heart model in the above example, the correlation between the guidance field matrix corresponding to the center point of the tetrahedron and the latest residual magnetic field is calculated, and the tetrahedron with the largest correlation is also selected as the preferred tetrahedron, and the preferred tetrahedron at this time is recorded as: D0.

[0101] The preferred tetrahedron D0 is subjected to multiple rounds of first finite times division, and four tetrahedrons are obtained in each round of first finite times division. At the end of each round of first finite times division, it is determined whether the first finite times division in this round satisfies the first division stop condition.

[0102] For example: After the first round of the first finite number of subdivisions, multiple sub-tetrahedrons are obtained, and the guidance field matrix of the center point of each sub-tetrahedron and the latest residual magnetic field are determined. The correlation between them is used to select the sub-tetrahedron with the largest correlation as the optimal sub-tetrahedron. At this time, the preferred sub-tetrahedron is recorded as: D1. Determine the guidance field matrix of the optimal sub-tetrahedron D1 and the latest residual magnetic field Whether the correlation between them satisfies the first segmentation stop condition. If the first segmentation stop condition is not met, the first finite segmentation of the subsequent rounds will continue until the first segmentation stop condition is met. If the first segmentation stop condition is met, the center point position of the optimal sub-tetrahedron D1 is the third final current source position at time a, and at this time, the third round of the second finite segmentation process ends.

[0103] Determine the third magnetic field at the center point of the preferred tetrahedron D1 According to the third magnetic field, determine the latest residual magnetic field Then, determine the latest residual magnetic field Whether the second segmentation stop condition is met, if the latest residual magnetic field If the residual magnetic field is greater than the set residual threshold, it is determined that the second segmentation stop condition is not met, and the fourth round of second finite segmentation is started, and so on, until the second segmentation stop condition is met; if the latest residual magnetic field If it is less than or equal to the set residual threshold, it is determined that the second segmentation stop condition is met.

[0104] When the second segmentation stop condition is met, the entire second finite segmentation process ends, and the center point positions of all optimal sub-tetrahedrons that meet the first segmentation stop condition are determined as the final current source positions at time a. Each time a final current source position is determined, the current direction here also needs to be determined. The method for determining the current direction at all final current source positions is the same as the method for determining the current direction at the first final current source position, which will not be repeated here.

[0105] Based on the same inventive concept, an embodiment of the present application also provides a cardiac current source positioning device corresponding to the cardiac current source positioning method. Since the principle of solving the problem by the device in the embodiment of the present application is similar to the above-mentioned cardiac current source positioning method in the embodiment of the present application, the implementation of the device can refer to the implementation of the method, and the repeated parts will not be repeated.

[0106] See also Figure 5 , Figure 5 This is a schematic diagram of the structure of a cardiac current source positioning device provided in an embodiment of the present application. Figure 5 As shown in FIG. , the cardiac current source positioning device 500 comprises:

[0107] The contour reconstruction module 501 is used to obtain the magnetic heart data of the target user, and use the magnetic heart data to reconstruct the contour of the target contour area in the universal heart model to obtain a personalized heart model of the target user;

[0108] A tetrahedron partitioning module 502 is used to partition the personalized heart model to obtain a plurality of tetrahedrons, and select a preferred tetrahedron from the plurality of tetrahedrons according to the personalized heart model and the effective magnetic heart data;

[0109] Iterative partitioning module 503, used for performing multiple rounds of first finite partitioning on the preferred tetrahedron, and determining whether the preferred sub-tetrahedron obtained by the first finite partitioning in this round meets the first partitioning stop condition at the end of each round of the first finite partitioning, wherein the first finite partitioning is to perform finite partitioning on the tetrahedron according to the center point of the preferred tetrahedron and the number of faces of the preferred tetrahedron;

[0110] The current source position determination module 504 is used to determine the final current source position according to the center point position of the preferred sub-tetrahedron obtained by the first finite number of subdivisions in this round if the first subdivision stop condition is met.

[0111] See also Figure 6 , Figure 6 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. Figure 6 As shown in , the electronic device 600 includes a processor 610 , a memory 620 and a bus 630 .

[0112] The memory 620 stores machine-readable instructions executable by the processor 610. When the electronic device 600 is running, the processor 610 communicates with the memory 620 via the bus 630. When the machine-readable instructions are executed by the processor 610, the above-mentioned Figure 1 The steps of the method for locating the cardiac current source in the method embodiment shown and the specific implementation thereof can be referred to the method embodiment, which will not be described in detail here.

[0113] The present application also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the computer program can execute the above-mentioned Figure 1 The steps of the method for locating the cardiac current source in the method embodiment shown and the specific implementation thereof can be referred to the method embodiment, which will not be described in detail here.

[0114] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0115] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. The device embodiments described above are merely schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some communication interfaces, and the indirect coupling or communication connection of devices or units can be electrical, mechanical or other forms.

[0116] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0117] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0118] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a non-volatile computer-readable storage medium that is executable by a processor. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0119] Finally, it should be noted that the above-described embodiments are only specific implementation methods of the present application, which are used to illustrate the technical solutions of the present application, rather than to limit them. The protection scope of the present application is not limited thereto. Although the present application is described in detail with reference to the above-mentioned embodiments, ordinary technicians in the field should understand that any technician familiar with the technical field can still modify the technical solutions recorded in the above-mentioned embodiments within the technical scope disclosed in the present application, or can easily think of changes, or make equivalent replacements for some of the technical features therein; and these modifications, changes or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be based on the protection scope of the claims.

Claims

1. A method for locating a cardiac current source, characterized in that: include: Acquire the target user's magnetocardiographic data, and use the magnetocardiographic data to reconstruct the target contour area in the universal heart model to obtain a personalized heart model of the target user; The personalized heart model is segmented to obtain a plurality of tetrahedrons, and a preferred tetrahedron is selected from the plurality of tetrahedrons according to the personalized heart model and the effective magnetic heart data; Performing multiple rounds of first finite-time splitting on the preferred tetrahedron, and at the end of each round of first finite-time splitting, determining whether the preferred sub-tetrahedron obtained by the first finite-time splitting in this round meets the first splitting stop condition, wherein the first finite-time splitting is to perform finite-time splitting on the tetrahedron according to the center point of the preferred tetrahedron and the number of faces of the preferred tetrahedron; If the first segmentation stop condition is met, the final current source position is determined according to the center point position of the preferred sub-tetrahedron obtained by the first finite segmentation in this round.

2. The method according to claim 1, characterized in that The method of reconstructing the contour of the target contour area in the universal heart model by using the magnetic heart data comprises: Performing filtering and noise reduction processing on the magnetic heart data to obtain the denoised magnetic heart data, and selecting the denoised magnetic heart data within a valid time as the valid magnetic heart data, wherein the valid time is a part of the time selected from the occurrence time of the QRS complex; Based on the effective magnetocardiographic data, a target voxel point with the maximum current power at each moment within the effective time is selected from a preset source space, and an edge curve is constructed using the target voxel point at each moment; The target contour region of the universal heart model is reconstructed using the edge curve to obtain a personalized heart model.

3. The method according to claim 2, characterized in that The edge curve is constructed by using the target voxel point at each moment, including: Projecting multiple target voxel points into a preset plane, selecting edge projection points in a preset direction in the preset plane as contour points of the target contour area, wherein the preset plane is a plane perpendicular to the direction of the line between the heart and the magnetic heart detector; Curve fitting is performed on the contour points to obtain an edge curve.

4. The method according to claim 2, characterized in that: The step of selecting a target voxel point having the maximum current power at each moment within the effective time from a preset source space based on the effective magnetocardiographic data comprises: At each moment, determining a minimum variance filter matrix based on the effective magnetocardiographic data and an initial guidance field matrix corresponding to each voxel point in the preset source space; The minimum variance filter matrix is ​​used to determine the current power of each voxel point at that moment, and the voxel point with the maximum current power is selected as the target voxel point at that moment.

5. The method according to claim 3, characterized in that: The step of reconstructing the target contour area of ​​the universal heart model by using the edge curve to obtain a personalized heart model includes: Projecting the universal heart model onto the preset plane to obtain a two-dimensional universal heart contour; Adjusting the edge shape of the two-dimensional universal heart contour so that the edge of the two-dimensional universal heart contour is aligned with the edge curve; The adjusted two-dimensional universal heart contour is restored to a three-dimensional heart model to obtain a personalized heart model of the target user.

6. The method according to claim 1, characterized in that The personalized heart model is segmented to obtain a plurality of tetrahedrons, including: Reconstructing a triangular mesh of the surface of the personalized heart model using a mesh reconstruction algorithm; The reconstructed triangular mesh is used to perform Delaunay triangulation on the interior of the personalized heart model to obtain a plurality of tetrahedrons.

7. The method according to claim 2, characterized in that: The step of selecting a preferred tetrahedron from the plurality of tetrahedrons according to the personalized heart model and the effective magnetic heart data comprises: Discretizing the magnetic field equation corresponding to the personalized heart model using a boundary element algorithm to obtain a guidance field matrix corresponding to the center point of each tetrahedron; A preferred tetrahedron is selected from a plurality of tetrahedrons according to the correlation between the guidance field matrix corresponding to the center point of each tetrahedron and the effective magnetocardiometry data.

8. The method according to claim 1, characterized in that The preferred tetrahedron is subjected to multiple rounds of first finite subdivisions, including: For each round of the first finite number of subdivisions, based on each side face of the previous round of preferred subtetrahedron and the center point of the preferred subtetrahedron, the previous round of preferred subtetrahedron is subdivided to obtain a preset number of subtetrahedrons of this round; According to the correlation between the guidance field matrix corresponding to the center point of each sub-tetrahedron of this round and the effective magnetic heart data, the preferred sub-tetrahedron of this round is selected from the preset number of sub-tetrahedrons of this round, so as to use the preferred sub-tetrahedron of this round to perform the first finite subdivision of the next round.

9. The method according to claim 1, characterized in that: The method further comprises: For each tetrahedron, singular value decomposition is performed on the guidance field matrix corresponding to the center point of the tetrahedron to obtain the left eigenvector of the guidance field matrix; Singular value decomposition is performed on the product of the transpose of the left eigenvector of the guidance field matrix and the normalized effective magnetocardio data to determine the correlation between the guidance field matrix of the center point of the tetrahedron and the effective magnetocardio data.

10. The method according to claim 1, characterized in that The method of determining the final current source position according to the center point position of the preferred sub-tetrahedron obtained by the first finite subdivision in this round includes: Taking the center point position of the optimal sub-tetrahedron as the first final current source position, and determining the first magnetic field generated by the first final current source; Based on the residual of the effective magnetocardiometry data and the first magnetic field, performing multiple rounds of second finite-time segmentation on the multiple tetrahedrons, and at the end of each round of second finite-time segmentation, determining whether the second finite-time segmentation in this round meets the second segmentation stop condition; If the second segmentation stop condition is met, the final current source position is determined according to the center point position of the preferred sub-tetrahedron obtained in each round of the second finite segmentation.

Images