Heart structure dynamic reconstruction method and device based on electrophysiological signals

Abstract

The invention discloses a heart structure dynamic reconstruction method and device based on electrophysiological signals, relates to the technical field of image processing, and mainly aims to solve the problem that existing heart three-dimensional image construction is not clear and inaccurate. Comprising the following steps: acquiring electrophysiological signals and medical image data acquired from a heart organ in different directions; in the direction, constructing a dipole model based on the electrophysiological signals; constructing a second-class heart electromagnetic model based on the electrophysiological signals and the medical image data; and performing parameter adjustment on the first-class dipole model and the second-class heart electromagnetic model on the basis of a reinforcement learning algorithm in combination with the real-time data of the heart organ, and reconstructing a heart structure model of the heart organ by using the first-class dipole model and the second-class heart electromagnetic model after parameter adjustment.

Description

Technical Field

[0001] This application relates to the field of image processing technology, and in particular to a method and apparatus for dynamic reconstruction of cardiac structure based on electrophysiological signals. Background Technology

[0002] The cardiac magnetic field is generated by the periodic contraction and relaxation of the atria and ventricles of the heart, accompanied by complex alternating bioelectric currents. This cardiac magnetic field is an electrophysiological signal that, after signal processing, can be used to monitor the internal structure of the heart. Specifically, a three-dimensional model of cardiac electrophysiology and the body surface magnetic field can be constructed using cardiac magnetic signals to describe the cardiac electrophysiological activity triggered by the pacemaker.

[0003] Currently, in the process of detecting cardiac magnetic fields, because the cardiac magnetic field is active, multiple sets of data need to be collected. Therefore, multiple measurements are required using the detection equipment, significantly increasing the detection time and the contact time between the equipment and the human body, thus increasing the burden on the patient. Furthermore, as an electrophysiological signal, cardiac magnetic field signals are typically modeled using Doppler 3D images during 3D modeling. However, because modeling requires combining data from multiple acquisitions, artifacts may occur in the depiction of the heart's contours and structures, resulting in unclear or inaccurate cardiac models. Summary of the Invention

[0004] In view of this, this application provides a method and device for dynamic reconstruction of cardiac structure based on electrophysiological signals, the main purpose of which is to solve the problems of unclear and inaccurate construction of existing three-dimensional cardiac images.

[0005] According to one aspect of this application, a method for dynamic reconstruction of cardiac structure based on electrophysiological signals is provided, comprising: Acquire electrophysiological signals and medical imaging data of the heart organ from different directions, wherein the electrophysiological signals are used to characterize the electrical signals generated by the physiological activities of the heart in different spatial directions; In the stated direction, a class of dipole models is constructed based on the electrophysiological signals. The class of dipole models includes a single-current dipole model, a multi-current dipole model, and a magnetic dipole model. Based on the electrophysiological signals and the medical imaging data, two types of cardiac electromagnetic models are constructed, including a three-dimensional geometric model of the heart trunk, a cardiac electrophysiological diffusion model, and an external cardiac magnetic field model. Based on reinforcement learning algorithms and real-time data of the heart organ, the parameters of the first type of dipole model and the second type of cardiac electromagnetic model are adjusted, and the adjusted first type of dipole model and the second type of cardiac electromagnetic model are used to reconstruct the heart structure model of the heart organ.

[0006] Furthermore, the acquisition of electrophysiological signals and medical imaging data of the heart organ from different directions includes: Electrophysiological signals during cardiac organ activity are collected by multiple data detectors in different directions mounted on the support of the cardiac magnetic detection device. The electrophysiological signals include cardiac magnetic field data and cardiac current source data corresponding to different spatial coordinates. The data detectors are evenly distributed at equal angles on the support.

[0007] Furthermore, constructing a dipole-type model based on the electrophysiological signal in the stated direction includes: Determine the plane current intensity corresponding to the cardiac current source data, and construct a single current dipole model based on the plane current intensity and the distance to the plane magnetic field source; Determine the multi-source current intensity corresponding to the cardiac current source data, and construct a multi-current dipole model based on the multi-source current intensity and the distance between the multi-source magnetic field sources; The loop current of the cardiac current source data is determined, and a magnetic dipole model is constructed based on the loop current and the loop area.

[0008] Furthermore, the medical data includes computed tomography (CT) images, magnetic resonance imaging (MRI) images, and transmembrane potential imaging data, and the method further includes: Computed tomography (CT) image data of the heart organ activity were acquired using a computed tomography (CT) scanner. Magnetic resonance imaging data of the heart organ activity were acquired using a magnetic resonance device; Transmembrane potential image data of the heart organ activity were acquired using a transmembrane potential measurement device; The method further includes: The electrocardiogram (ECG) signals of the heart organ are acquired using ECG monitoring equipment.

[0009] Furthermore, the construction of the second type of cardiac electromagnetic model based on the electrophysiological signals and the medical imaging data includes: Image segmentation of the computed tomography (CT) image data and / or magnetic resonance (MRI) image data is performed using an adaptive threshold and a generative adversarial network to obtain segmented organ region images. The segmented organ region image is reconstructed based on a moving average filter to obtain the three-dimensional structure of the organ. The three-dimensional structure of the organ is then modeled based on the three-dimensional structure and the spherical region of the heart to obtain a three-dimensional geometric model of the heart trunk. Based on the current, voltage, and diffusion tensor corresponding to the cardiac current source data, an electrical signal propagation characterization is generated, and the diffusion coefficient is used to describe the electrical signal propagation characterization to generate a cardiac electrophysiological diffusion model. Data on the surface magnetic field strength were obtained, and numerical simulations of the surface magnetic field strength were performed based on the finite element method to construct a model of the external cardiac magnetic field.

[0010] Furthermore, the method also includes: The cardiac magnetic field data is processed by time series filtering using the Kalman filter algorithm, and feature extraction is performed on the filtered cardiac magnetic field data to obtain magnetic field features. The abnormal magnetic field characteristics are obtained by using the abnormal magnetic field recognition model that has completed model training to identify the abnormal magnetic field features and electrocardiogram signals. The abnormal magnetic field features are located using the heart image to obtain the region of abnormal heart activity.

[0011] Furthermore, after extracting features from the filtered cardiac magnetic field data to obtain magnetic field features, the method further includes: The magnetic field characteristics are predicted based on the time-series prediction model that has completed model training, and the magnetic field characteristic prediction results are obtained. After the magnetic field prediction results are matched with preset risk characteristics, early warning information is generated.

[0012] According to another aspect of this application, a device for dynamic reconstruction of cardiac structure based on electrophysiological signals is provided, comprising: The acquisition module is used to acquire electrophysiological signals and medical imaging data collected from different directions of the heart organ. The electrophysiological signals are used to characterize the electrical signals generated by the physiological activities of the heart in different spatial directions. The first construction module is used to construct a class of dipole models based on the electrophysiological signals in the direction mentioned above. The class of dipole models includes a single-current dipole model, a multi-current dipole model, and a magnetic dipole model. The second construction module is used to construct two types of cardiac electromagnetic models based on the electrophysiological signals and the medical imaging data. The two types of cardiac electromagnetic models include a three-dimensional geometric model of the heart trunk, a cardiac electrophysiological diffusion model, and an external cardiac magnetic field model. The generation module is used to adjust the parameters of the first type of dipole model and the second type of cardiac electromagnetic model based on reinforcement learning algorithm combined with real-time data of the heart organ, and to reconstruct the heart structure model of the heart organ using the parameter-adjusted first type of dipole model and the second type of cardiac electromagnetic model.

[0013] Furthermore, The acquisition module is further configured to acquire electrophysiological signals of the heart organ activity through multiple data detectors set in different directions on the support of the magnetic heart detection device. The electrophysiological signals include cardiac magnetic field data and cardiac current source data corresponding to different spatial coordinates. The data detectors are evenly distributed at equal angles on the support.

[0014] Furthermore, The first construction module is specifically used to determine the plane current intensity corresponding to the cardiac current source data, and construct a single current dipole model based on the plane current intensity and the distance between the plane magnetic field sources; determine the multi-source current intensity corresponding to the cardiac current source data, and construct a multi-current dipole model based on the multi-source current intensity and the distance between the multi-source magnetic field sources; determine the loop current of the cardiac current source data, and construct a magnetic dipole model based on the loop current and the loop area.

[0015] Furthermore, the medical data includes computed tomography (CT) images, magnetic resonance imaging (MRI) data, and transmembrane potential imaging data, and the device further includes: The acquisition module is used to acquire computed tomography (CT) image data of the heart organ activity using a computed tomography (CT) scanner; acquire magnetic resonance (MRI) image data of the heart organ activity using a magnetic resonance (MRI) scanner; and acquire transmembrane potential image data of the heart organ activity using a transmembrane potential measurement device. The acquisition module is also used to acquire electrocardiogram (ECG) signals from the heart organ based on the ECG monitoring device.

[0016] Furthermore, The second construction module is specifically used to perform image segmentation on the computed tomography (CT) image data and / or magnetic resonance imaging (MRI) image data using adaptive thresholding and generative adversarial networks to obtain segmented organ region images; to reconstruct the segmented organ region images based on a moving average filter to obtain the three-dimensional structure of the organ, and to model the three-dimensional structure of the organ based on the three-dimensional structure and the spherical region of the heart to obtain a three-dimensional geometric model of the heart trunk; to generate an electrical signal propagation characterization based on the current, voltage, and diffusion tensor corresponding to the cardiac current source data, and to describe the electrical signal propagation characterization using a diffusion coefficient to generate a cardiac electrophysiological diffusion model; and to acquire surface magnetic field strength data, and to perform numerical simulation of the surface magnetic field strength data based on finite element method to construct an external cardiac magnetic field model.

[0017] Furthermore, the device also includes: The filtering module is used to perform time-series filtering on the cardiac magnetic field data using the Kalman filtering algorithm, and to extract features from the filtered cardiac magnetic field data to obtain magnetic field features. The identification module is used to identify abnormal magnetic field features and electrocardiogram signals based on the abnormal magnetic field identification model that has been trained, and to obtain abnormal magnetic field features. The positioning module is used to locate the abnormal magnetic field features using the heart image to obtain the abnormal heart activity area.

[0018] Furthermore, the device also includes: The prediction module is used to predict the magnetic field features based on the time-series prediction model that has been trained, obtain the magnetic field feature prediction results, and generate early warning information after the magnetic field prediction results match the preset risk features.

[0019] According to another aspect of this application, a storage medium is provided, wherein at least one executable instruction is stored therein, the executable instruction causing a processor to perform operations corresponding to the above-described method for dynamic reconstruction of cardiac structures based on electrophysiological signals.

[0020] According to another aspect of this application, a terminal is provided, comprising: a processor, a memory, a communication interface, and a communication bus, wherein the processor, the memory, and the communication interface communicate with each other through the communication bus; The memory is used to store at least one executable instruction, which causes the processor to perform the operation corresponding to the above-described method for dynamic reconstruction of cardiac structure based on electrophysiological signals.

[0021] By employing the above technical solutions, the technical solutions provided in the embodiments of this application have at least the following advantages: This application provides a method and apparatus for dynamic reconstruction of cardiac structure based on electrophysiological signals. Compared with the prior art, the embodiments of this application acquire electrophysiological signals and medical imaging data collected from different directions of the heart organ. The electrophysiological signals are used to characterize the electrical signals generated by cardiac physiological activities in different spatial directions. In each direction, a dipole-type model is constructed based on the electrophysiological signals. This dipole-type model includes a single-current dipole model, a multi-current dipole model, and a magnetic dipole model. Based on the electrophysiological signals and the medical imaging data, two types of cardiac electromagnetic models are constructed. The electromagnetic model includes a three-dimensional geometric model of the heart trunk, a cardiac electrophysiological diffusion model, and an external cardiac magnetic field model. Based on reinforcement learning algorithms and combined with real-time data of the heart organ, the parameters of the first-class dipole model and the second-class cardiac electromagnetic model are adjusted. The adjusted first-class dipole model and the second-class cardiac electromagnetic model are then used to reconstruct the cardiac structure model of the heart organ. This reduces the magnetic field burden on the human body during multiple detection processes, avoids the frequent occurrence of magnetic field artifacts, and improves the accuracy of magnetic field-based modeling of the heart structure, thereby achieving the goal of high-precision display of the heart structure and contour.

[0022] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application. Attached Figure Description

[0023] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the scope of this application. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings: Figure 1 A flowchart of a method for dynamic reconstruction of cardiac structure based on electrophysiological signals provided in an embodiment of this application is shown. Figure 2 This illustration shows a three-dimensional structural diagram of the front of a cardiac magnetic field detection device provided in an embodiment of this application; Figure 3 A three-dimensional structural schematic diagram of the back of a cardiac magnetic field detection device provided in an embodiment of this application is shown; Figure 4 A three-dimensional structural schematic diagram of a grounding bracket provided in an embodiment of this application is shown; Figure 5 A schematic diagram of the structure of a sliding drive component provided in an embodiment of this application is shown; Figure 6 The diagram shows a structural schematic of three cross-sections of a fixed platform provided in an embodiment of this application; Figure 7 This paper illustrates a flowchart of another method for dynamic reconstruction of cardiac structure based on electrophysiological signals provided in an embodiment of this application. Figure 8 This paper illustrates a flowchart of another method for dynamic reconstruction of cardiac structure based on electrophysiological signals provided in an embodiment of this application. Figure 9 This illustration shows a block diagram of a dynamic reconstruction device for cardiac structure based on electrophysiological signals, provided in an embodiment of this application. Figure 10 This illustration shows a structural diagram of a terminal provided in an embodiment of this application; in, Figures 2 to 6 The correspondence between the reference numerals and component names in the attached drawings is as follows: 1. Magnetic heart rate testing equipment with support frame; 2. Grounding support; 201. Counterweight T-shaped column; 202. Fixed platform; 2021. Short slide rail; 2022. Long slide rail; 3. Sliding drive assembly; 301. Limiting slide rail; 302. Driven sliding foot; 303. Driven sliding foot; 304. Spring component; 305. Drive cylinder; 4. Movable testing bed; 5. Stabilizing auxiliary support; 51. Limiting channel. Detailed Implementation

[0024] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

[0025] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0026] The embodiments of this invention can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence (AI) refers to the theories, methods, technologies, and application systems that use digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0027] Foundational technologies for artificial intelligence generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing, operating / interactive systems, and mechatronics. AI software technologies mainly encompass computer vision, robotics, biometrics, speech processing, natural language processing, and machine learning / deep learning.

[0028] Based on this, in one embodiment, the present invention provides a method for dynamic reconstruction of cardiac structure based on electrophysiological signals. Taking the application of this method to computer devices such as servers as an example, the server can be an independent server or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDN), and big data and artificial intelligence platforms, such as intelligent medical systems.

[0029] This application provides a method for dynamic reconstruction of cardiac structure based on electrophysiological signals, such as... Figure 1 As shown, the method includes: 101. Acquire electrophysiological signals and medical imaging data of the heart organ from different directions.

[0030] In this embodiment, the current execution end, as the main body for generating cardiac images, is a processing end that can interact with the cardiac magnetic field detection device, including but not limited to intelligent medical systems. The cardiac magnetic field detection device is used to scan the magnetic field of the heart organ to obtain electrophysiological signals. These electrophysiological signals characterize the electrical signals generated by cardiac physiological activities in different spatial directions. The electrophysiological signals include cardiac magnetic field data and cardiac current source data corresponding to different spatial coordinates. The cardiac magnetic field data represents the intensity of the cardiac magnetic field at different locations, and the cardiac current source data includes the location, intensity, and changes of the cardiac current source. This embodiment does not impose specific limitations. Medical imaging data includes computed tomography (CT) images, magnetic resonance imaging (MRI) images, and transmembrane potential (TMP) images, which are data collected by the CT scanner, MRI scanner, and TMP measuring device, respectively, serving as the image basis for model construction.

[0031] In another embodiment of this application, for further definition and explanation, the step of acquiring electrophysiological signals and medical imaging data of the heart organ from different directions includes: Electrophysiological signals during cardiac organ activity are collected by multiple data detectors positioned in different directions on the support of the magnetic heart detection device.

[0032] In order to accurately capture the electrophysiological signals generated during cardiac organ activity and improve the accuracy of cardiac image generation based on three-dimensional models, the current execution end collects the electrophysiological signals during cardiac organ activity through multiple data detectors set in different directions on the support of the cardiac magnetic detection device. At this time, the data detectors are evenly distributed at equal angles on the support.

[0033] In a specific embodiment, such as Figure 2-6As shown, the magnetic heart detection device mounting bracket 1 has several data detectors fixedly installed inside, and the data detectors are evenly distributed at equal angles. A grounding bracket 2 is provided on one side of the magnetic heart detection device mounting bracket 1. A sliding drive assembly 3 is provided on the grounding bracket 2, and a movable detection bed 4 is slidably connected to it through the sliding drive assembly 3. The movable detection bed 4 is located directly above the grounding bracket 2 and is parallel to the grounding bracket 2. A stabilizing auxiliary bracket 5 is provided on the other side of the magnetic heart detection device mounting bracket 1. The grounding support 2 includes a counterweight T-shaped column 201, which is flush with the ground. A fixed platform 202 is fixedly installed at the top of the counterweight T-shaped column 201. A short slide groove 2021 is formed in the middle of the upper end of the fixed platform 202. Long slide grooves 2022 are formed on both sides of the upper end of the fixed platform 202, and the two long slide grooves 2022 are symmetrically distributed on both sides of the short slide groove 2021. The short slide groove 2021 and the long slide groove 2022 are arranged in parallel. The sliding drive assembly 3 includes two limiting slide rails 301, which are fixed inside the corresponding long slide grooves 2022. Driven sliding feet 302 are slidably connected to the limiting slide rails 301. The two driven sliding feet 302 are fixed to one side of the bottom of the movable testing bed 4. A drive-type sliding foot 303 is fixedly installed at the center of one side of the bottom of the movable testing bed 4. The drive-type sliding foot 303 is slidably connected to the inside of the short slide groove 2021, and a spring member 304 is fixedly installed between the side of the drive-type sliding foot 303 and the inner end of the short slide groove 2021. A drive cylinder 305 is fixedly installed inside the fixed platform 202, and the output end of the drive cylinder 305 is fixed to the side of the drive-type sliding foot 303 through a connecting shaft. A limit channel 51 is fixedly installed at the upper end of the stabilizing auxiliary support 5, and the width of the limit channel 51 is the same as the width of the movable testing bed 4.

[0034] It should be noted that the data detectors can sequentially collect data on the changes in the magnetic field strength and direction of the heart at each instant, forming time-series data to analyze the temporal changes in the heart's electrical activity and capture its instantaneous magnetic field changes. Furthermore, because the data detectors are evenly distributed at equal angles, such as one detector every 60 degrees, the collected data will form spatial data covering different regions of the heart, providing precise spatial coordinates to reflect the location of different current sources.

[0035] 102. In the stated direction, a dipole-type model is constructed based on the electrophysiological signal.

[0036] In this application embodiment, a type of dipole model includes a single-current dipole model, a multi-current dipole model, and a magnetic dipole model. The single-current dipole model measures the magnetic field value at any point when the human body is an infinitely uniform conductor. The multi-current dipole model is used to reflect changes in heart activity. The magnetic dipole model calculates the magnetic dipole moment and deduces the magnetic dipole source model.

[0037] 103. Construct a second type of cardiac electromagnetic model based on the electrophysiological signals and the medical imaging data.

[0038] In this embodiment, the two types of cardiac electromagnetic models include a three-dimensional geometric model of the cardiac trunk, a cardiac electrophysiological diffusion model, and an external cardiac magnetic field model. The three-dimensional geometric model of the cardiac trunk is used for image segmentation and three-dimensional reconstruction of medical images, i.e., to establish a personalized three-dimensional geometric model of the cardiac trunk. The cardiac electrophysiological diffusion model is used to simulate the response of cardiac electrophysiological activities during changes in cardiac depolarization, repolarization, and TMP. The external cardiac magnetic field model is used to calculate the frequency of the human cardiac electromagnetic field during the propagation of the magnetic field generated by cardiac electrical excitation within the body.

[0039] 104. Based on reinforcement learning algorithm and combined with real-time data of the heart organ, the parameters of the first type of dipole class model and the second type of cardiac electromagnetic model are adjusted, and the first type of dipole class model and the second type of cardiac electromagnetic model with adjusted parameters are used to reconstruct the heart structure model of the heart organ.

[0040] In this embodiment, after the current execution end constructs a type I dipole model and a type II cardiac electromagnetic model, to better match the user's physiological movement and ensure the accuracy and reliability of the cardiac magnetic field model, a reinforcement learning algorithm is used to dynamically adjust the parameters of the type I dipole model and the type II cardiac electromagnetic model based on real-time cardiac organ data to achieve optimal model support. The reinforcement learning algorithm may include, but is not limited to, Q-Learning, A3C (Asynchronous Advantage Actor-Critic), and DQN (Deep Q-Network), etc. Real-time cardiac organ data includes real-time heart rate, real-time exercise intensity, etc., and this embodiment does not impose specific limitations. Finally, the type I dipole model and the type II cardiac electromagnetic model are bound together as the reconstructed cardiac structure model. For example, a one-to-one correspondence is established between the three-dimensional magnetic field coordinates of the type I dipole model and the spatial coordinates of the type II cardiac electromagnetic model, so that when a corresponding position is selected, the model content can be switched, or an image of the magnetic field or cardiac structure corresponding to the selected position can be output. This embodiment does not impose specific limitations.

[0041] It should be noted that when training model parameters based on reinforcement learning algorithms, historical measured data can be used as samples, such as electrocardiogram data and magnetocardiogram data as training inputs, and model parameters as training outputs. The constructed Q-Learning model is used to learn from the training samples until the preset conditions are met. This application does not impose specific limitations on the embodiments.

[0042] In one specific embodiment, adaptive optimization of parameters in cardiac magnetic field modeling is implemented to generate an accurate, individualized cardiac magnetic field model for each patient by combining personalized modeling strategies. When the current execution end trains the model based on a reinforcement learning algorithm, the optimization objective formula is expressed as: ; in, Here, L represents the model parameters, and L is the error loss function. Furthermore, historical cardiac data can be used as input to the model, automatically correcting the results to improve the accuracy of the personalized model.

[0043] In another embodiment of this application, for further definition and explanation, the step of constructing a dipole-type model based on the electrophysiological signal in the said direction includes: Determine the plane current intensity corresponding to the cardiac current source data, and construct a single current dipole model based on the plane current intensity and the distance to the plane magnetic field source; Determine the multi-source current intensity corresponding to the cardiac current source data, and construct a multi-current dipole model based on the multi-source current intensity and the distance between the multi-source magnetic field sources; The loop current of the cardiac current source data is determined, and a magnetic dipole model is constructed based on the loop current and the loop area.

[0044] To construct an effective model of the heart in a magnetic field, the collected medical data includes computed tomography (CT) images, magnetic resonance imaging (MRI) images, and transmembrane potential (TMP) images. When constructing a dipole-type model, the current execution device first determines the plane current intensity corresponding to the cardiac current source data. Then, based on the multi-source current intensity and the distance between the plane magnetic field sources, a single-current dipole model is constructed, represented as: ,in, The magnetic field strength is characterized by a single-current dipole model. For multi-source current intensity, The distance between the multi-source magnetic field sources is the distance between the magnetic field emitted by the magnetic field source and the object to be measured, as set by the magnetic field detection device. This application embodiment does not make specific limitations.

[0045] Meanwhile, for combinations of multiple current sources, when constructing a multi-current dipole model to reflect changes in cardiac activity, after determining the intensity of the multiple current sources corresponding to the cardiac current source data, a multi-current dipole model is constructed based on the intensity of the multiple current sources and the distance between the multiple magnetic field sources, expressed as: ,in, The magnetic field strength is characterized by a single-current dipole model. Let be the current intensity of the i-th current source. Let be the distance from the i-th current source to the detection point.

[0046] Furthermore, after determining the loop current of the cardiac current source data, a magnetic dipole model is constructed based on the loop current and loop area to calculate the magnetic dipole moment, which can be expressed as: ,in, It is the magnetic dipole moment. Let A be the loop current intensity and A be the loop area.

[0047] It should be noted that in this embodiment, the human body is set as an infinitely uniform conductor. When constructing the above model to detect the magnetic field strength at any point, the electrical activity of the heart can be equivalent to a current source, and the position of the current source can be regarded as the position of a dipole. The electromagnetic field generated by the current source propagates in space. Since the human body is a uniform conductor, the distribution of the magnetic field generated by the current source in space can be calculated using the above model. Especially in the theoretical model, the magnetic field generated by the current source is related to its distance, direction, and other factors. In some embodiments, in the single-current dipole model, the general steps for calculating the magnetic field value are to obtain the strength and direction of the magnetic field at a specific point by integration according to Biot-Savart's law or Ampere's law. Specifically, the magnetic field expression of the single-current dipole model is usually derived by calculating the contribution of the current source to a point in space. Assuming the current source is a uniformly distributed point current, the calculation formula will be adjusted according to the distribution and position of the current source and the spatial coordinates of the target point. Additionally, the distance from the current source to the detection point... This refers to the spatial location of each current source, usually represented by three-dimensional coordinates (x, y, z). Each current source represents the electrical activity of a region of the heart and reflects the spatial distribution of cardiac electrical activity. As the heart depolarizes and repolarizes, the location of the current source changes over time. For example, cardiac depolarization begins at the sinoatrial node and proceeds downwards to the atria and ventricles, thus changing the location and intensity of the current source. The current intensity of the current source... This refers to the intensity of each current source, which is usually related to the "amplitude" of cardiac electrical activity. That is, changes in current intensity are closely related to the depolarization or repolarization state of the heart. During depolarization, myocardial cells generate current with relatively high intensity; during repolarization, the current intensity gradually decreases. Therefore, changes in current intensity can reflect the intensity of the heart's electrophysiological activity.

[0048] Furthermore, the change in current source intensity over time can simulate the periodic changes in cardiac electrical activity. The temporal characteristics of cardiac electrical activity, such as depolarization, repolarization, and rhythmic variations, reflect the intensity and trend of cardiac electrical activity at different points in time, particularly the changes in current intensity within each heartbeat cycle. The spatial distribution of current sources refers to the changing positions of multiple current sources over time, forming a propagation pattern of cardiac electrical activity. During the conduction of cardiac electrical activity, depolarization and repolarization are processes that propagate gradually downwards from the apex of the heart (such as the sinoatrial node) towards the ventricles. During this process, the spatial positions of multiple current sources change with the progress of cardiac electrophysiological activity, reflecting the propagation of cardiac electrical signals in different regions. At this point, the interaction between dipoles characterizes the relative positional relationships between different current sources, influencing their interaction. The distribution pattern of these current sources determines the overall shape of the cardiac magnetic field and the coordination and synchronicity of cardiac electrical activity. In normal cardiac electrical activity, the effects of each current source are orderly and synchronous. However, in cases of arrhythmia or cardiac lesions, the interaction between current sources may become disordered, leading to irregular propagation of cardiac electrical activity.

[0049] The depolarization and repolarization processes in different parts of the heart refer to the fact that, in a multi-current dipole model, the changes in each current source are not merely changes in magnitude, but also include the depolarization and repolarization processes in different regions of the heart. Specifically, the model reflects the electrophysiological activities of different parts of the heart, such as the atria, ventricles, and conduction system, based on the distribution of current sources in different cardiac regions. For example, the depolarization process in the heart begins at the sinoatrial node and gradually propagates to the atria, atrioventricular node, and then to the ventricles; these changes are all reflected through the activity of multiple current sources. The repolarization process is a process in which the intensity of the current sources gradually decreases and changes in the opposite direction. The phase difference and synchronicity of cardiac electrical activity refer to the fact that, since different current sources represent the electrical activity of different regions of the heart, their relative time changes and phase differences are also important parameters. This parameter reflects the coordination and synchronicity of the electrical activity in different parts of the heart. The phase difference during depolarization and repolarization can help identify abnormal cardiac states, such as arrhythmias.

[0050] In this embodiment, the calculation of the magnetic field value relies on the cardiac magnetic field data collected by the data detector in the aforementioned cardiac magnetodetection device. The data detector measures the magnetic field strength generated by the heart and records the magnetic field changes at different points. These data are calculated using the model formula constructed above to further deduce the magnetic field value at any point. In some embodiments, the data detector collects magnetic field strength and direction information at different locations (such as the body surface or a detection plane near the heart). This information is used in a single-current dipole model to calculate the magnetic field values ​​at other locations on that plane. In the single-current dipole model, an arbitrary point refers to any point on the detection plane where the magnetic field value is to be calculated. These points are typically points on the human body surface where data is collected, possibly including areas such as the chest and abdomen, used to analyze the distribution of the magnetic field generated by the heart on the body surface. Points on the detection plane at a certain distance from the heart are usually related to the magnetic field strength and direction generated by the heart; the magnetic field values ​​at other points are deduced from the magnetic field data measured at different detector positions. Points inside the heart are mainly measured on the body surface, but the magnetic field value at any point inside the heart can also be deduced through model calculation. For example, assuming a magnetic field generated by a current source inside the heart, the magnetic field strength at that point can be calculated.

[0051] In another embodiment of this application, for further definition and explanation, the steps also include: Computed tomography (CT) image data of the heart organ activity were acquired using a computed tomography (CT) scanner. Magnetic resonance imaging data of the heart organ activity were acquired using a magnetic resonance device; Transmembrane potential image data of the heart organ activity were acquired using a transmembrane potential measurement device; The method further includes: The electrocardiogram (ECG) signals of the heart organ are acquired using ECG monitoring equipment.

[0052] To achieve greater precision in constructing the cardiac structure, the construction of a type II cardiac electromagnetic model can be combined with medical imaging data. Computed tomography (CT) scans, magnetic resonance imaging (MRI), transmembrane potential measurement (TPT) devices, and electrocardiogram (ECG) monitoring devices can be used to acquire corresponding data, including CT images, MRI images, TPT images, and ECG signals. Specifically, CT and MRI provide high-resolution imaging data of the heart and surrounding tissues, including the heart's shape, size, location, and the structure of surrounding tissues. MRI can also provide information on the orientation of myocardial fibers (DTI – diffusion tensor imaging), a necessary parameter for studying the propagation of electrical signals in the heart.

[0053] In addition, ECG monitoring devices record the electrical activity signals of the heart, reflecting the electrophysiological activity at different time points during cardiac depolarization and repolarization, including the time series of potential changes. This allows for analysis of the electrical signal conduction paths and time delays in different regions of the heart. ECG data is then used to determine the propagation characteristics of depolarization and repolarization waves, including propagation speed, direction, and dynamic changes in electrophysiological activity. Combined with diffusion equation models, the diffusion process of electrical signals in cardiac tissue can be simulated. Transmembrane potential measurement devices (TMPs) record the transmembrane potential changes of cardiomyocytes during depolarization and repolarization, reflecting the potential fluctuations of individual cardiomyocytes during electrical activity. This provides information on the electrophysiological characteristics of cardiomyocytes during electrical activity, including the range and time series characteristics of transmembrane potential changes.

[0054] In another embodiment of this application, for further definition and explanation, such as Figure 7 As shown, the steps for constructing a type II cardiac electromagnetic model based on the electrophysiological signals and the medical imaging data include: 401. The computed tomography image data and / or magnetic resonance image data are segmented using an adaptive threshold and a generative adversarial network to obtain segmented organ region images; 402. Reconstruct the segmented organ region image based on the moving average filter to obtain the three-dimensional structure of the organ, and model the three-dimensional structure of the organ based on the three-dimensional structure and the spherical region of the heart to obtain a three-dimensional geometric model of the heart trunk. 403. Based on the current, voltage, and diffusion tensor corresponding to the cardiac current source data, generate an electrical signal propagation characterization, and use the diffusion coefficient to describe the electrical signal propagation characterization to generate a cardiac electrophysiological diffusion model. 404. Obtain the surface magnetic field strength data, and perform numerical simulation of the surface magnetic field strength data based on the finite element method to construct an external cardiac magnetic field model.

[0055] To construct a structural model of the heart and improve the accuracy of heart image generation, the current execution end constructs a three-dimensional geometric model of the heart trunk, a cardiac electrophysiological diffusion model, and an external cardiac magnetic field model when constructing two types of cardiac electromagnetic models.

[0056] In a specific embodiment, for a three-dimensional geometric model of the heart trunk, image segmentation is first performed on the computed tomography (CT) image data and / or magnetic resonance imaging (MRI) image data based on adaptive thresholding and a generative adversarial network (GAN) to obtain segmented organ region images. Specifically, when performing image segmentation using adaptive thresholding and a GAN, the adaptive thresholding is applied dynamically to different heart regions to address varying brightness and noise levels in the heart images. This region-adaptive approach adjusts the segmentation threshold for each region in the image, resulting in more refined segmentation of the background and target regions. Simultaneously, the GAN automatically generates high-quality segmented images through a game between a generator and a discriminator. The generator attempts to generate segmented images of the heart region, while the discriminator evaluates the segmentation results, optimizing the generator to produce more accurate heart segmentation images. By combining adaptive thresholding and the GAN, the segmentation of the heart region is ensured to be both fast and accurate, effectively handling complex heart morphology and noise issues in clinical images. In some embodiments, CT and / or MRI image data are first preprocessed, such as by bilateral filtering for denoising, to obtain a smooth image while preserving edge information. During adaptive thresholding, the denoised CT or MRI image is first divided into blocks, and each block can select a different threshold based on local brightness differences, thereby avoiding the limitations of a global threshold. The adaptive thresholding method can then calculate the threshold for each region, expressed by the following formula: ;in, For the local threshold at position (x, y), It is the average value of the pixels surrounding the region (x, y). The parameters are adjusted to control the range of the threshold. During training of the Generative Adversarial Network (GNA), the generator is based on the U-Net architecture. The input is computed tomography (CT) and / or magnetic resonance imaging (MRI) images of the original heart image, and the output is the segmented heart region. The discriminator determines whether the generated segmented image matches the actually labeled heart region. The quality of the generated image can be evaluated using a convolutional neural network. The GAN loss function can be defined as a combination of adversarial loss and L1 loss, expressed as: ; in, For generator output, It is the output of the discriminator. These are the weighting coefficients of the L1 loss. Additionally, morphological operations (such as dilation and erosion) can be used to refine the segmentation results, ensuring smooth and continuous edges in the heart region. Contour extraction can also be performed on the segmentation results to further refine the morphology of the heart region. Furthermore, the generative adversarial network can be optimized through cross-validation and error analysis to ensure segmentation accuracy. At this point, the generator and discriminator are periodically adjusted to achieve the best segmentation results.

[0057] It should be noted that after obtaining the segmented organ region image, a moving average filter can be used to reconstruct the segmented organ region image to obtain the organ's three-dimensional structure. Based on this three-dimensional structure and the spherical region of the heart, a three-dimensional geometric model of the heart trunk is then obtained. Specifically, when reconstructing the segmented organ region image using a moving average filter, firstly, three-dimensional reconstruction is performed using two-dimensional slices of the segmented organ region image. The segmentation results of each slice are integrated to form a complete three-dimensional heart geometric model. The slices are then stacked according to their actual spatial positions to generate three-dimensional volume data. The three-dimensional surface contour of the heart can be extracted using surface reconstruction algorithms (such as Marching Cubes) to form the three-dimensional volume data. Furthermore, to smooth the three-dimensional structure and remove noise, a moving average filter is used to average the data of neighboring points in the reconstructed three-dimensional model, reducing structural discontinuities caused by image noise and obtaining the organ's three-dimensional structure. Finally, based on the 3D structural space reconstructed from the original 3D model, a spherical region containing the heart is constructed to define the spatial extent of the heart model. The heart region can be combined with the spherical region to determine the overall geometric contour of the heart, generating the final 3D geometric model of the heart trunk. This provides a geometric framework for subsequent electrophysiological simulation and magnetic field modeling. The reconstructed 3D volume is represented as follows: ;in, For the reconstructed three-dimensional volume, The grayscale value of each pixel, where n is the total number of pixels.

[0058] In a specific embodiment, for the cardiac electrophysiological diffusion model, an electrical signal propagation characterization can be generated based on the current, voltage, and diffusion tensor corresponding to the cardiac current source data. This characterization is then described using a diffusion coefficient to generate the cardiac electrophysiological diffusion model. Specifically, the model employs the ion current diffusion process during cardiac depolarization and repolarization, combined with the anisotropic characteristics of the myocardium, to simulate electrophysiological signals and accurately reflect the electrophysiological changes of the heart. Specifically, the Action Potential (AP) model in electrophysiology can be used to describe the depolarization and repolarization processes of cardiac cells. During current diffusion, the electrical signal propagation characterization can be represented as: Where V is the voltage and D is the diffusion tensor. This is a current source. Furthermore, considering the directionality of heart muscle fibers, it can be represented using the anisotropic diffusion tensor as: ; in, The diffusion coefficient along the myocardial fibers, The diffusion coefficient is perpendicular to the direction of myocardial fibers.

[0059] In a specific embodiment, for the external cardiac magnetic field model, surface magnetic field strength data is acquired, and numerical simulation of the surface magnetic field strength data is performed based on the finite element method to construct the external cardiac magnetic field model. That is, after acquiring the surface magnetic field strength data, the external magnetic field generated by the heart is simulated numerically by combining the finite element method (FEM) with measured data, thereby achieving the goal of accurately describing the distribution of the surface magnetic field. Specifically, the external magnetic field calculation can use a magnetic field calculation formula to simulate the propagation of the magnetic field generated by electrical excitation within the body. The formula for the external cardiac magnetic field model is expressed as: ;in, The magnetic field strength on the body surface. , These are the components of the cardiac electrical source in the X and Y directions. Let be the vacuum permeability. A numerical simulation of the magnetic field generated by the electrophysiological activity of the heart is performed using the finite element method. A geometric model of the heart is established and coupled with the external electromagnetic field for calculation. Based on this, the numerical solution of the external cardiac magnetic field is obtained using finite element discretization technology. Furthermore, calibration using real-time electrocardiogram (ECG) data and magnetic field strategy data can further improve the accuracy of the simulation results.

[0060] In another embodiment of this application, for further definition and explanation, such as Figure 8 As shown, the steps also include: 501. The cardiac magnetic field data is subjected to time series filtering processing using the Kalman filtering algorithm, and the filtered cardiac magnetic field data is subjected to feature extraction to obtain magnetic field features; 502. Based on the abnormal magnetic field recognition model that has completed model training, perform abnormal identification on the magnetic field features and electrocardiogram signals to obtain abnormal magnetic field features; 503. Using the heart image, the abnormal magnetic field features are located to obtain the abnormal heart activity region.

[0061] To enhance the dynamic analysis capability of the cardiac magnetic field and improve the accuracy of capturing temporal changes in cardiac electrical activity, the current execution terminal performs time-series filtering processing on the cardiac magnetic field data using the Kalman filter algorithm. Specifically, when performing time-series analysis on the cardiac magnetic field data using the Kalman filter algorithm, dynamic noise is removed and features of cardiac electrical activity are extracted. The time series can be characterized as follows: ; ; in, It is in a magnetic field state. Let F be the observed values, F be the state transition matrix, and H be the observation matrix. and These are process noise and observation noise, respectively. Feature extraction is performed on the filtered cardiac magnetic field data to obtain magnetic field features. These features can include magnetic field strength over a predetermined time period. Then, based on an abnormal magnetic field recognition model that has completed model training, anomalies are identified in the magnetic field features and electrocardiogram (ECG) signals to obtain abnormal magnetic field features. This is achieved using a machine learning-based pattern recognition algorithm (such as Support Vector Machine (SVM) or deep learning). By training normal and abnormal cardiac magnetic field patterns, abnormal features in the magnetic field are automatically identified, and abnormal cardiac electrical activity regions are located. This embodiment does not impose specific limitations. Finally, the abnormal magnetic field features are located using cardiac images to obtain abnormal cardiac activity regions. At this point, the abnormal magnetic field features can be located from the corresponding cardiac images after model construction, based on the coordinates corresponding to the abnormal magnetic field features, to determine the abnormal cardiac electrical activity regions. This embodiment does not impose specific limitations.

[0062] It should be noted that when performing abnormal feature identification, data from different channels can be dynamically fused. For example, multi-channel data such as electrocardiogram (ECG) and magnetocardiogram (MCG) can be fused together to capture subtle changes in cardiac electrical activity through data correlation analysis. This application does not impose specific limitations on the embodiments.

[0063] In another embodiment of this application, for further definition and explanation, after extracting features from the filtered cardiac magnetic field data to obtain magnetic field features, the method further includes: The magnetic field characteristics are predicted based on the time-series prediction model that has completed model training, and the magnetic field characteristic prediction results are obtained. After the magnetic field prediction results are matched with preset risk characteristics, early warning information is generated.

[0064] To accurately predict electromagnetic data of cardiac activity and aid in its analysis, after obtaining magnetic field characteristics, the execution end can predict these characteristics based on a pre-trained time-series prediction model. This yields a magnetic field characteristic prediction result, and a warning is generated after the prediction result matches a preset risk feature. Specifically, the time-series prediction model can use a Long Short-Term Memory (LSTM) network for time-series prediction: the LSTM model is used to model the time-series data of the cardiac magnetic field to predict future cardiac electrical activity and calculate future trends in magnetic field changes. The LSTM prediction formula is expressed as follows: ; in, In hidden state, For the current input, This is the activation function.

[0065] In one specific embodiment, based on the predicted magnetic field characteristics, the risk of cardiac activity can be assessed, i.e., compared with preset risk characteristics. When a match is found, it indicates that there is an abnormality, and therefore, an early warning message is generated. At this time, the preset risk characteristics can be configured based on different cardiac activity abnormality detection needs, and this application embodiment does not make specific limitations.

[0066] In one specific embodiment, the generated cardiac image can be displayed using a high-resolution 3D visualization system, showing the spatial distribution and dynamic changes of the cardiac magnetic field, thus improving the intuitiveness and interpretability of the results. Specifically, volume rendering technology is used to display a 3D density map to show the spatial distribution of the cardiac magnetic field. The volume rendering formula is expressed as: ; in, To output color, To input the color of the heart image, For the color of the sampling point, For transparency. Additionally, real-time ray tracing technology can be used to achieve realistic rendering of the magnetic field distribution, improving visualization and facilitating observation of the detailed spatial distribution of the heart's magnetic field. Dynamic animations generated using time-series data can showcase the entire process of cardiac electrical activity, helping doctors intuitively analyze the activity state of the magnetic field.

[0067] This application provides a method for dynamic reconstruction of cardiac structure based on electrophysiological signals. Compared with existing technologies, this application acquires electrophysiological signals and medical imaging data collected from different directions of the heart. The electrophysiological signals are used to characterize the electrical signals generated by cardiac physiological activities in different spatial directions. In each direction, a dipole-type model is constructed based on the electrophysiological signals. This dipole-type model includes a single-current dipole model, a multi-current dipole model, and a magnetic dipole model. Based on the electrophysiological signals and the medical imaging data, two types of cardiac electromagnetic models are constructed. The electromagnetic model includes a three-dimensional geometric model of the heart trunk, a cardiac electrophysiological diffusion model, and an external cardiac magnetic field model. Based on reinforcement learning algorithms and combined with real-time data of the heart organ, the parameters of the first-class dipole model and the second-class cardiac electromagnetic model are adjusted. The adjusted first-class dipole model and the second-class cardiac electromagnetic model are then used to reconstruct the cardiac structure model of the heart organ. This reduces the magnetic field burden on the human body during multiple detection processes, avoids the frequent occurrence of magnetic field artifacts, and improves the accuracy of magnetic field-based modeling of the heart structure, thereby achieving the goal of high-precision display of the heart structure and contour.

[0068] Furthermore, as a response to the above Figure 1 The implementation of the method shown in this application provides a device for dynamic reconstruction of cardiac structure based on electrophysiological signals, such as... Figure 9 As shown, the device includes: The acquisition module 61 is used to acquire electrophysiological signals and medical imaging data collected from different directions of the heart organ. The electrophysiological signals are used to characterize the electrical signals generated by the physiological activities of the heart in different spatial directions. The first construction module 62 is used to construct a type of dipole model based on the electrophysiological signal in the direction, wherein the type of dipole model includes a single-current dipole model, a multi-current dipole model, and a magnetic dipole model. The second construction module 63 is used to construct two types of cardiac electromagnetic models based on the electrophysiological signals and the medical imaging data. The two types of cardiac electromagnetic models include a three-dimensional geometric model of the heart trunk, a cardiac electrophysiological diffusion model, and an external cardiac magnetic field model. The generation module 64 is used to adjust the parameters of the first type of dipole model and the second type of cardiac electromagnetic model based on reinforcement learning algorithm combined with real-time data of the cardiac organ, and to reconstruct the cardiac structure model of the cardiac organ using the parameter-adjusted first type of dipole model and the second type of cardiac electromagnetic model.

[0069] Furthermore, The acquisition module is further configured to acquire electrophysiological signals of the heart organ activity through multiple data detectors set in different directions on the support of the magnetic heart detection device. The electrophysiological signals include cardiac magnetic field data and cardiac current source data corresponding to different spatial coordinates. The data detectors are evenly distributed at equal angles on the support.

[0070] Furthermore, The first construction module is specifically used to determine the plane current intensity corresponding to the cardiac current source data, and construct a single current dipole model based on the plane current intensity and the distance between the plane magnetic field sources; determine the multi-source current intensity corresponding to the cardiac current source data, and construct a multi-current dipole model based on the multi-source current intensity and the distance between the multi-source magnetic field sources; determine the loop current of the cardiac current source data, and construct a magnetic dipole model based on the loop current and the loop area.

[0071] Furthermore, the medical data includes computed tomography (CT) images, magnetic resonance imaging (MRI) data, and transmembrane potential imaging data, and the device further includes: The acquisition module is used to acquire computed tomography (CT) image data of the heart organ activity using a computed tomography (CT) scanner; acquire magnetic resonance (MRI) image data of the heart organ activity using a magnetic resonance (MRI) scanner; and acquire transmembrane potential image data of the heart organ activity using a transmembrane potential measurement device. The acquisition module is also used to acquire electrocardiogram (ECG) signals from the heart organ based on the ECG monitoring device.

[0072] Furthermore, The second construction module is specifically used to perform image segmentation on the computed tomography (CT) image data and / or magnetic resonance imaging (MRI) image data using adaptive thresholding and generative adversarial networks to obtain segmented organ region images; to reconstruct the segmented organ region images based on a moving average filter to obtain the three-dimensional structure of the organ, and to model the three-dimensional structure of the organ based on the three-dimensional structure and the spherical region of the heart to obtain a three-dimensional geometric model of the heart trunk; to generate an electrical signal propagation characterization based on the current, voltage, and diffusion tensor corresponding to the cardiac current source data, and to describe the electrical signal propagation characterization using a diffusion coefficient to generate a cardiac electrophysiological diffusion model; and to acquire surface magnetic field strength data, and to perform numerical simulation of the surface magnetic field strength data based on finite element method to construct an external cardiac magnetic field model.

[0073] Furthermore, the device also includes: The filtering module is used to perform time-series filtering on the cardiac magnetic field data using the Kalman filtering algorithm, and to extract features from the filtered cardiac magnetic field data to obtain magnetic field features. The identification module is used to identify abnormal magnetic field features and electrocardiogram signals based on the abnormal magnetic field identification model that has been trained, and to obtain abnormal magnetic field features. The positioning module is used to locate the abnormal magnetic field features using the heart image to obtain the abnormal heart activity area.

[0074] Furthermore, the device also includes: The prediction module is used to predict the magnetic field features based on the time-series prediction model that has been trained, obtain the magnetic field feature prediction results, and generate early warning information after the magnetic field prediction results match the preset risk features.

[0075] This application provides a device for dynamic reconstruction of cardiac structure based on electrophysiological signals. Compared with the prior art, this application acquires electrophysiological signals and medical imaging data collected from different directions of the heart organ. The electrophysiological signals are used to characterize the electrical signals generated by cardiac physiological activities in different spatial directions. In each direction, a dipole-type model is constructed based on the electrophysiological signals. This dipole-type model includes a single-current dipole model, a multi-current dipole model, and a magnetic dipole model. Two types of cardiac electromagnetic models are constructed based on the electrophysiological signals and the medical imaging data. The electromagnetic model includes a three-dimensional geometric model of the heart trunk, a cardiac electrophysiological diffusion model, and an external cardiac magnetic field model. Based on reinforcement learning algorithms and combined with real-time data of the heart organ, the parameters of the first-class dipole model and the second-class cardiac electromagnetic model are adjusted. The adjusted first-class dipole model and the second-class cardiac electromagnetic model are then used to reconstruct the cardiac structure model of the heart organ. This reduces the magnetic field burden on the human body during multiple detection processes, avoids the frequent occurrence of magnetic field artifacts, and improves the accuracy of magnetic field-based modeling of the heart structure, thereby achieving the goal of high-precision display of the heart structure and contour.

[0076] According to one embodiment of this application, a storage medium is provided, the storage medium storing at least one executable instruction, the computer-executable instruction being able to execute the method for dynamic reconstruction of cardiac structure based on electrophysiological signals in any of the above method embodiments.

[0077] Figure 10 The diagram shows a structural schematic of a terminal according to one embodiment of the present application. The specific embodiments of the present application do not limit the specific implementation of the terminal.

[0078] like Figure 10 As shown, the terminal may include: a processor 702, a communications interface 704, a memory 706, and a communications bus 708.

[0079] The processor 702, communication interface 704, and memory 706 communicate with each other via communication bus 708.

[0080] The communication interface 704 is used to communicate with other network elements such as clients or other servers.

[0081] The processor 702 is used to execute program 710, specifically to execute the relevant steps in the above-described embodiment of the dynamic reconstruction method for cardiac structure based on electrophysiological signals.

[0082] Specifically, program 710 may include program code that includes computer operation instructions.

[0083] The processor 702 may be a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of this application. The terminal includes one or more processors, which may be processors of the same type, such as one or more CPUs; or they may be processors of different types, such as one or more CPUs and one or more ASICs.

[0084] Memory 706 is used to store program 710. Memory 706 may include high-speed RAM memory, and may also include non-volatile memory, such as at least one disk storage device.

[0085] Specifically, program 710 can be used to cause processor 702 to perform the following operations: Acquire electrophysiological signals and medical imaging data of the heart organ from different directions, wherein the electrophysiological signals are used to characterize the electrical signals generated by the physiological activities of the heart in different spatial directions; In the stated direction, a class of dipole models is constructed based on the electrophysiological signals. The class of dipole models includes a single-current dipole model, a multi-current dipole model, and a magnetic dipole model. Based on the electrophysiological signals and the medical imaging data, two types of cardiac electromagnetic models are constructed, including a three-dimensional geometric model of the heart trunk, a cardiac electrophysiological diffusion model, and an external cardiac magnetic field model. Based on reinforcement learning algorithms and real-time data of the heart organ, the parameters of the first type of dipole model and the second type of cardiac electromagnetic model are adjusted, and the adjusted first type of dipole model and the second type of cardiac electromagnetic model are used to reconstruct the heart structure model of the heart organ.

[0086] Obviously, those skilled in the art should understand that the modules or steps of this application described above can be implemented using general-purpose computing devices. They can be centralized on a single computing device or distributed across a network of multiple computing devices. Optionally, they can be implemented using computer-executable program code, thereby storing them in a storage device for execution by a computing device. In some cases, the steps shown or described can be performed in a different order than those presented here, or they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. Thus, this application is not limited to any particular combination of hardware and software.

[0087] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for dynamic reconstruction of cardiac structure based on electrophysiological signals, characterized in that, include: Acquire electrophysiological signals and medical imaging data of the heart organ from different directions, wherein the electrophysiological signals are used to characterize the electrical signals generated by the physiological activities of the heart in different spatial directions; In the stated direction, a class of dipole models is constructed based on the electrophysiological signals. The class of dipole models includes a single-current dipole model, a multi-current dipole model, and a magnetic dipole model. Based on the electrophysiological signals and the medical imaging data, two types of cardiac electromagnetic models are constructed, including a three-dimensional geometric model of the heart trunk, a cardiac electrophysiological diffusion model, and an external cardiac magnetic field model. Based on reinforcement learning algorithms and real-time data of the heart organ, the parameters of the first type of dipole model and the second type of cardiac electromagnetic model are adjusted, and the adjusted first type of dipole model and the second type of cardiac electromagnetic model are used to reconstruct the heart structure model of the heart organ.

2. The method according to claim 1, characterized in that, The acquisition of electrophysiological signals and medical imaging data of the heart organ from different directions includes: Electrophysiological signals during cardiac organ activity are collected by multiple data detectors in different directions mounted on the support of the cardiac magnetic detection device. The electrophysiological signals include cardiac magnetic field data and cardiac current source data corresponding to different spatial coordinates. The data detectors are evenly distributed at equal angles on the support.

3. The method according to claim 2, characterized in that, The construction of a dipole-type model based on the electrophysiological signal in the stated direction includes: Determine the plane current intensity corresponding to the cardiac current source data, and construct a single current dipole model based on the plane current intensity and the distance to the plane magnetic field source; Determine the multi-source current intensity corresponding to the cardiac current source data, and construct a multi-current dipole model based on the multi-source current intensity and the distance between the multi-source magnetic field sources; The loop current of the cardiac current source data is determined, and a magnetic dipole model is constructed based on the loop current and the loop area.

4. The method according to claim 3, characterized in that, The medical data includes computed tomography (CT) images, magnetic resonance imaging (MRI) images, and transmembrane potential imaging data; the method further includes: Computed tomography (CT) image data of the heart organ activity were acquired using a computed tomography (CT) scanner. Magnetic resonance imaging data of the heart organ activity were acquired using a magnetic resonance device; Transmembrane potential image data of the heart organ activity were acquired using a transmembrane potential measurement device; The method further includes: The electrocardiogram (ECG) signals of the heart organ are acquired using ECG monitoring equipment.

5. The method according to claim 4, characterized in that, The construction of the second type of cardiac electromagnetic model based on the electrophysiological signals and the medical imaging data includes: Image segmentation of the computed tomography (CT) image data and / or magnetic resonance (MRI) image data is performed using an adaptive threshold and a generative adversarial network to obtain segmented organ region images. The segmented organ region image is reconstructed based on a moving average filter to obtain the three-dimensional structure of the organ. The three-dimensional structure of the organ is then modeled based on the three-dimensional structure and the spherical region of the heart to obtain a three-dimensional geometric model of the heart trunk. Based on the current, voltage, and diffusion tensor corresponding to the cardiac current source data, an electrical signal propagation characterization is generated, and the diffusion coefficient is used to describe the electrical signal propagation characterization to generate a cardiac electrophysiological diffusion model. Data on the surface magnetic field strength were obtained, and numerical simulations of the surface magnetic field strength were performed based on the finite element method to construct a model of the external cardiac magnetic field.

6. The method according to claim 5, characterized in that, The method further includes: The cardiac magnetic field data is processed by time series filtering using the Kalman filter algorithm, and feature extraction is performed on the filtered cardiac magnetic field data to obtain magnetic field features. The abnormal magnetic field characteristics are obtained by using the abnormal magnetic field recognition model that has completed model training to identify the abnormal magnetic field features and electrocardiogram signals. The abnormal magnetic field features are located using the heart image to obtain the region of abnormal heart activity.

7. The method according to claim 6, characterized in that, After extracting features from the filtered cardiac magnetic field data to obtain magnetic field features, the method further includes: The magnetic field characteristics are predicted based on the time-series prediction model that has completed model training, and the magnetic field characteristic prediction results are obtained. After the magnetic field prediction results are matched with preset risk characteristics, early warning information is generated.

8. A device for dynamic reconstruction of cardiac structure based on electrophysiological signals, characterized in that, include: The acquisition module is used to acquire electrophysiological signals and medical imaging data collected from different directions of the heart organ. The electrophysiological signals are used to characterize the electrical signals generated by the physiological activities of the heart in different spatial directions. The first construction module is used to construct a class of dipole models based on the electrophysiological signals in the direction mentioned above. The class of dipole models includes a single-current dipole model, a multi-current dipole model, and a magnetic dipole model. The second construction module is used to construct two types of cardiac electromagnetic models based on the electrophysiological signals and the medical imaging data. The two types of cardiac electromagnetic models include a three-dimensional geometric model of the heart trunk, a cardiac electrophysiological diffusion model, and an external cardiac magnetic field model. The generation module is used to adjust the parameters of the first type of dipole model and the second type of cardiac electromagnetic model based on reinforcement learning algorithm combined with real-time data of the heart organ, and to reconstruct the heart structure model of the heart organ using the parameter-adjusted first type of dipole model and the second type of cardiac electromagnetic model.

9. A computer-readable storage medium having a computer program / instructions stored thereon, characterized in that, When the computer program / instructions are executed by the processor, they implement the steps of the method of claim 1.

10. A computer device, comprising a memory, a processor, and a computer program stored in the memory, characterized in that, The processor executes the computer program to implement the steps of the method of claim 1.

Images