Magnetocardiogram feature analysis system, method and equipment and storage medium

Abstract

The invention provides a magnetocardiogram feature analysis system, method and device and a storage medium, and the system comprises a data collection module which is used for collecting multichannel magnetocardiogram signals; the data preprocessing module is used for carrying out noise reduction and filtering on the collected data, drawing a one-dimensional magnetocardiogram of each channel and carrying out waveband segmentation on one-dimensional magnetocardiogram signals; the feature extraction and analysis module is used for carrying out key feature point calibration and feature value extraction on each channel and carrying out difference analysis on calibrated features; the model training and verification module is used for inputting the effective features of the training set into a machine learning model for training, inputting the effective features of the magnetocardiogram signals of the verification set into the model, and adjusting model parameters; and the model test and prediction module is used for acquiring the test set magnetocardiogram signal for analysis, and the model outputs a label of an analysis result. According to the invention, the accuracy and efficiency of magnetocardiogram signal analysis can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of magnetocardiogram signal processing, and more specifically, to a magnetocardiogram feature analysis system, method, device and storage medium. Background Art

[0002] Magnetocardiography is a non-invasive cardiac monitoring technology that analyzes physiological conditions by measuring changes in the magnetic field generated by the heart's bioelectrical activity. Compared with traditional electrocardiography, MCG can provide more three-dimensional information about cardiac activity, which helps to more accurately locate abnormal areas of the heart. However, magnetic cardiography signals are usually very weak and easily interfered by environmental noise and other biomagnetic fields, which complicates signal acquisition and analysis.

[0003] Existing magnetocardiogram analysis systems often require complex preprocessing steps to improve the signal-to-noise ratio when processing these weak signals, including noise reduction and filtering. In addition, feature extraction and analysis of magnetocardiogram signals require precise algorithms to identify key feature points and extract useful physiological information from them. During the training and validation process of machine learning models, choosing the right algorithm is crucial to improving the model's predictive accuracy. However, existing systems may suffer from inefficiency, lack of accuracy, or complex operation in these aspects.

[0004] In the process of implementing the embodiments of the present invention, the inventors found that there are at least the following problems or defects in the prior art: the existing magnetocardiogram feature analysis system may lack integrated and automated processing procedures in data acquisition, preprocessing, feature extraction and model training, resulting in low analysis efficiency, and there may be problems of improper data management and inconsistent analysis results when processing large amounts of data. In addition, the existing system may not fully utilize advanced machine learning algorithms to improve the predictive performance of the model, or the user interface design is not intuitive enough, making it difficult for operators to manage and view the analysis results. Summary of the invention

[0005] The present invention provides a magnetocardiogram feature analysis system, method, device and storage medium.

[0006] In a first aspect of the present invention, a magnetocardiogram feature analysis system is provided, comprising:

[0007] Data acquisition module, used for collecting multi-channel magnetic cardiology signals;

[0008] The data preprocessing module is used to reduce noise and filter the collected data, draw the one-dimensional magnetocardiographic signal diagram of each channel, and perform band segmentation of the one-dimensional magnetocardiographic signal;

[0009] The feature extraction and analysis module calibrates the key feature points and extracts the feature values ​​of each channel, and performs difference analysis on the calibrated features;

[0010] The model training and verification module is used to input the effective features of the training set into the machine learning model for training, and input the effective features of the magnetic cardiotoc signal of the verification set into the model to adjust the model parameters;

[0011] The model testing and prediction module is used to obtain the magnetic cardiotonic signals of the test set for analysis, and the model output is the label of the analysis result.

[0012] Furthermore, the data management module is used to manage the data division in the entire system, ensure that the data set is correctly divided into training set, validation set and test set according to the specified ratio, and implement unified data normalization processing.

[0013] Furthermore, the feature extraction module includes extracting maximum amplitude, minimum amplitude, average amplitude, amplitude change rate, amplitude ratio, waveform duration, peak interval and waveform morphological features.

[0014] Furthermore, the model training and verification module can be selected from one or a combination of support vector machines, decision trees, random forests, gradient boosting trees, K nearest neighbors, naive Bayes, logistic regression, LASSO regression or other supervised learning algorithms suitable for processing one-dimensional magnetocardiogram features.

[0015] Furthermore, it also includes a user interface module for providing a user interface for operators to manage data and view analysis results.

[0016] Furthermore, the data preprocessing module includes:

[0017] A band segmentation unit, used to perform band segmentation on the one-dimensional magnetocardiographic signal and mark it according to time;

[0018] A preprocessing model unit, used for denoising and filtering the segmented signal;

[0019] The feature visualization unit is used to visualize the features of the processed signal.

[0020] Furthermore, the noise reduction method is EMD noise reduction, comprising the steps of:

[0021] S01. Find all the maximum and minimum points in the acquired magnetic cardiotoc signal and record them as local maximum and local minimum;

[0022] S02. Obtaining an upper envelope and a lower envelope by spline curve fitting;

[0023] S03. Subtract the upper envelope from the lower envelope to obtain a new component;

[0024] S04. Check whether the new component obtained in step S03 is an intrinsic mode function. If not, loop through steps S01-S03 again. If yes, proceed to the next step.

[0025] S05. Determine whether the number of components reaches a preset value, if yes, save all components, if not, return to step S01;

[0026] S06. Performing noise reduction processing on the magnetocardiographic signal according to the number of components and the high frequency band of the noise signal, wherein: the more the number of components, the more the high frequency band of the signal after noise reduction, and vice versa.

[0027] In a second aspect of the present invention, a method for analyzing magnetocardiogram characteristics is provided, comprising:

[0028] Collect multi-channel magnetic cardiology signals;

[0029] The collected data is denoised and filtered, and the one-dimensional magnetocardiographic signal diagram of each channel is drawn to perform band segmentation of the one-dimensional magnetocardiographic signal;

[0030] Calibrate key feature points and extract feature values ​​for each channel, and perform difference analysis on the calibrated features;

[0031] Input the effective features of the training set into the machine learning model for training, and input the effective features of the magnetic cardiotoc signal of the validation set into the model to adjust the model parameters;

[0032] The test set magnetocardiographic signals are obtained for analysis, and the model output is the label of the analysis result.

[0033] In a third aspect of the present invention, an electronic device is provided, comprising: at least one processor, a memory and an input-output unit; wherein the memory is used to store a computer program, and the processor is used to call the computer program stored in the memory to execute any one of the methods described in the first aspect.

[0034] In a fourth aspect of the present invention, a computer-readable storage medium is provided, which includes instructions, and when the instructions are executed on a computer, the computer is caused to execute the method according to any one of the first aspects.

[0035] According to the above-mentioned embodiments of the present invention, at least the following beneficial effects are achieved: the magnetocardiogram feature analysis system can realize efficient acquisition, accurate preprocessing and in-depth analysis of magnetocardiogram signals through an integrated data acquisition module, a data preprocessing module, a feature extraction and analysis module, a model training and verification module, and a model testing and prediction module. The system can automatically perform noise reduction and filtering, draw a one-dimensional magnetocardiogram signal graph for each channel, and perform band segmentation, thereby improving the accuracy and efficiency of signal processing. The feature extraction and analysis module can calibrate key feature points and extract feature values, and perform difference analysis, which can enhance the accuracy of feature recognition. The model training and verification module uses a variety of machine learning algorithms to improve the generalization ability and prediction accuracy of the model.

[0036] In addition, the data management module of the system can ensure that the data set is correctly divided into training set, validation set and test set according to the specified ratio, and implement unified data normalization processing to optimize the model training process. The user interface module can provide an intuitive operation interface, allowing operators to easily manage data and view analysis results, which can improve the user experience. The system's EMD denoising method can effectively extract intrinsic mode functions from magnetic heart signals, reduce noise interference, and further improve signal quality. Overall, the system can provide a more accurate and efficient technical support for cardiac physiological research. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] The above and other objects, features and advantages of the exemplary embodiments of the present invention will become readily understood by reading the detailed description below with reference to the accompanying drawings. In the accompanying drawings, several embodiments of the present invention are shown in an exemplary and non-limiting manner, in which:

[0038] Figure 1 A schematic diagram of the structure of a magnetocardiogram feature analysis system provided by an embodiment of the present invention;

[0039] Figure 2 A schematic diagram of a flow chart of a magnetocardiogram feature analysis method provided in one embodiment of the present invention;

[0040] Figure 3 The schematic diagram schematically shows the structure of an electronic device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0041] The principles and spirit of the present invention will be described below with reference to several exemplary embodiments. It should be understood that these embodiments are provided only to enable those skilled in the art to better understand and implement the present invention, and are not intended to limit the scope of the present invention in any way. On the contrary, these embodiments are provided to make the present invention more thorough and complete, and to fully convey the scope of the present invention to those skilled in the art.

[0042] Those skilled in the art know that the embodiments of the present invention can be implemented as a system, device, apparatus, method or computer program product. Therefore, the present invention can be specifically implemented in the following forms, namely: complete hardware, complete software (including firmware, resident software, microcode, etc.), or a combination of hardware and software.

[0043] It should be noted that any number of elements in the drawings is for illustration rather than limitation, and any naming is only for distinction and does not have any limiting meaning.

[0044] Reference below Figure 1 , Figure 1 FIG. 1 is a schematic diagram of the structure of a magnetocardiogram feature analysis system provided by an embodiment of the present invention. Figure 1 As shown, a magnetocardiogram feature analysis system 100 includes:

[0045] The data acquisition module 101 is used to acquire multi-channel magnetocardiographic signals;

[0046] The data preprocessing module 102 is used to reduce noise and filter the collected data, draw a one-dimensional magnetocardiographic signal diagram of each channel, and perform band segmentation of the one-dimensional magnetocardiographic signal;

[0047] The feature extraction and analysis module 103 calibrates key feature points and extracts feature values ​​for each channel, and performs difference analysis on the calibrated features;

[0048] The model training and verification module 104 is used to input the effective features of the training set into the machine learning model for training, and input the effective features of the magnetic cardiotonic signals of the verification set into the model to adjust the model parameters;

[0049] The model testing and prediction module 105 is used to obtain the magnetic cardiotonic signal of the test set for analysis, and the model output is the label of the analysis result.

[0050] It should be noted that the system includes a data acquisition module, whose function is to collect magnetic heart signals from multiple channels. The magnetic heart signal refers to the magnetic field changes generated by the heart's bioelectric activity, and multi-channel means that the system can capture the heart's magnetic field information from different positions or angles at the same time. This multi-angle data acquisition can provide more comprehensive information on heart activity.

[0051] Specifically, the data acquisition module uses a specific sensor array to capture cardiac magnetic signals. These sensors need to be precisely placed at specific locations around the patient's body to ensure that the magnetic field signals of the heart can be captured from different angles. During the acquisition process, the system may set a specific sampling rate and bandwidth to ensure the accuracy and integrity of the signal. For example, the sampling rate may be set to 1000Hz to capture the details of cardiac activity, while the bandwidth may be set to 0.1-100Hz to filter out irrelevant high-frequency noise.

[0052] The data acquisition module can use sensors including but not limited to superconducting quantum interference devices, highly sensitive optically pumped magnetometers, etc., which can detect very weak magnetic field changes. In addition, in order to further improve the signal quality, the system may integrate a preamplifier to amplify the signal before it is transmitted to the data processing unit. Such a design can reduce the loss of the signal during transmission and ensure that the collected magnetic cardio signals are as close to the original signals as possible, thereby providing more accurate data for subsequent analysis.

[0053] In some embodiments, the data management module is used to manage data partitioning in the entire system, ensure that the data set is correctly divided into training sets, validation sets, and test sets according to specified proportions, and implement unified data normalization processing.

[0054] It should be noted that the system includes a data management module, whose core function is to divide and manage the data in the entire system, ensuring that the data set is correctly divided into training sets, validation sets, and test sets according to the specified ratio. The division of the data set is an important step in machine learning, which determines the effect of model training and evaluation. The training set is used to train the model, the validation set is used to adjust the model parameters, and the test set is used to evaluate the final performance of the model.

[0055] Specifically, the data management module involves data normalization, which is a preprocessing technique used to scale the data to a uniform range or distribution, such as between 0 and 1 or -1 and 1. Normalization helps to improve the convergence speed of the algorithm and prevents certain features from dominating the training process due to their large numerical range. In actual operation, the data management module may use the minimum-maximum normalization method to convert the raw data into

[0056]

[0057] In the form of, X is the original data, X min and X max are the minimum and maximum values ​​in the data set, respectively.

[0058] Preferably, the data management module may also include a data augmentation function, which is a technique for increasing the size of a data set by creating variations of the data, which helps to improve the generalization ability of the model. For example, new magnetic cardiotoc signal samples may be generated by adding noise, adjusting the amplitude or time delay of the signal, etc.

[0059] Furthermore, the data management module can also realize automatic data partitioning. By setting the partitioning ratio parameters, such as 70% for the training set, 15% for the validation set, and 15% for the test set, the system can automatically divide the data set according to these ratios to ensure that the data in each set is random and representative. Such automated processing not only improves efficiency, but also reduces the possibility of human error.

[0060] In some embodiments, the feature extraction module includes extracting maximum amplitude, minimum amplitude, average amplitude, amplitude change rate, amplitude ratio, waveform duration, peak interval and waveform morphology features.

[0061] It should be noted that the feature extraction module of the system includes extracting the maximum amplitude, minimum amplitude, average amplitude, amplitude change rate, amplitude ratio, waveform duration, peak interval and waveform morphological features. These features are key parameters in the analysis of magnetic cardiosignals, and they can reflect the physiological characteristics and physiological changes of cardiac activity. Waveform morphological features refer to the shape and structural characteristics of the waveform, which are particularly important for identifying abnormal cardiac rhythms.

[0062] Specifically, the feature extraction module analyzes the magnetic cardio signals of each channel to identify and quantify the above parameters. For example, the maximum amplitude and minimum amplitude can represent the peak and valley values ​​of the signal, the average amplitude can reflect the overall strength of the signal, and the amplitude change rate can indicate the speed of change of the signal strength. The waveform duration refers to the duration of a single waveform cycle, and the peak interval refers to the time interval between consecutive peaks. The extraction of these parameters can be achieved through digital signal processing techniques, such as Fourier transform, wavelet transform or peak detection algorithm.

[0063] Preferably, the feature extraction module can use automated algorithms to identify key points of the waveform, such as peaks and troughs, and calculate the area and width of the waveform, etc. These algorithms can be based on machine learning techniques, such as support vector machines or neural networks, to improve the accuracy and robustness of feature extraction.

[0064] Furthermore, the feature extraction module can also include a threshold setting function, allowing the operator to adjust the sensitivity of feature extraction according to actual needs to adapt to different signal qualities and analysis requirements. For example, a threshold for the amplitude change rate can be set to distinguish between normal and abnormal changes in the magnetic cardiosignal. Through these refined operating steps and alternatives, the system can more flexibly adapt to different analysis scenarios.

[0065] In some embodiments, the model training and validation module can be selected from one or a combination of support vector machines, decision trees, random forests, gradient boosting trees, K nearest neighbors, naive Bayes, logistic regression, LASSO regression or other supervised learning algorithms suitable for processing one-dimensional magnetocardiogram features.

[0066] It should be noted that the system model training and verification module can choose to use one or a combination of support vector machines, decision trees, random forests, gradient boosting trees, K nearest neighbors, naive Bayes, logistic regression, LASSO regression or other supervised learning algorithms suitable for processing one-dimensional magnetocardiogram features. Supervised learning algorithm is a method in machine learning that learns the relationship between input data and output labels through training data sets in order to predict new data. Here, one-dimensional magnetocardiogram features refer to data points extracted from magnetocardiogram signals for training models.

[0067] Specifically, the model training and validation module selects the appropriate algorithm based on the characteristics of the magnetic heart signal and the prediction target. For example, if the goal is to classify the type of cardiac image, support vector machine or random forest may be selected; if the goal is to predict the severity of the lesion, logistic regression or LASSO regression may be selected.

[0068] More specifically, the parameter settings of these algorithms, such as the depth of the tree, the number of neighbors, the regularization strength, etc., need to be adjusted according to the characteristics of the data and the experimental results. Parameter adjustment can be achieved through cross-validation, grid search and other techniques to find the best model performance.

[0069] Preferably, the model training and validation module can integrate an automated model selection and parameter tuning process. This process can include a predefined algorithm library, and the system can automatically select the best algorithm and parameter combination based on performance indicators (such as accuracy, recall, F1 score).

[0070] Furthermore, the system can also achieve model integration, that is, combining the analysis results of multiple models to improve the overall prediction performance. For example, methods such as voting, stacking, or weighted averaging can be used to fuse the outputs of different models. Through these detailed operation steps and alternatives, the system can train and verify the model more efficiently to obtain the best prediction effect.

[0071] In some embodiments, a user interface module is also included to provide a user interface for operators to manage data and view analysis results.

[0072] It should be noted that the system includes a user interface module, whose main function is to provide an intuitive operating interface for operators to manage data and view analysis results. The user interface is a bridge between human and computer interaction, which allows users to interact with the system, input instructions and view output results. In this context, data management involves the organization, editing and processing of magnetic cardiotoc signal data, while analysis result viewing refers to the display of results provided by the system after analysis.

[0073] Specifically, the user interface module can include multiple functional areas, such as data browsing area, parameter setting area, result display area and help document area. The data browsing area allows users to view and select different data sets, and the parameter setting area allows users to adjust the parameters of model training and testing, such as learning rate, number of iterations, etc. The result display area will display the analysis results of the model in the form of charts or lists, including classification labels, prediction probabilities, etc. The help document area provides system operation guides and algorithm descriptions so that users can better understand the working principle and usage of the system.

[0074] Preferably, the user interface module can be designed using a graphical user interface (GUI) to provide a more user-friendly experience. The GUI can include controls such as buttons, sliders, and drop-down menus, so that users can complete complex data management and model configuration through simple operations such as clicking and dragging.

[0075] Furthermore, the system can also provide real-time feedback mechanisms, such as progress bars and status messages, to let users know the progress and results of the current operation. In order to meet the needs of different users, the user interface module can also provide multi-language support and custom theme functions, allowing users to adjust the interface according to personal preferences and language habits. Through these detailed operation steps and alternatives, the system can provide a more flexible and personalized user interaction experience.

[0076] In some embodiments, the data preprocessing module includes:

[0077] A band segmentation unit, used to perform band segmentation on the one-dimensional magnetocardiographic signal and mark it according to time;

[0078] A preprocessing model unit, used for denoising and filtering the segmented signal;

[0079] The feature visualization unit is used to visualize the features of the processed signal.

[0080] It should be noted that the data preprocessing module of the system includes a band segmentation unit, a preprocessing model unit and a feature visualization unit. The band segmentation unit refers to the process of segmenting the continuous magnetic cardio signals into multiple bands according to specific time intervals or feature points. This is a common step in signal analysis and helps to identify and analyze specific events or cycles in the signal. The preprocessing model unit refers to the process of denoising and filtering these segmented signals to improve the quality of the signal. The feature visualization unit refers to the graphical display of the processed signal so that the operator can intuitively observe the characteristics of the signal.

[0081] Specifically, the band segmentation unit can determine the timing and length of segmentation based on the physiological characteristics of the magnetic cardio signal, such as the heartbeat cycle. The preprocessing model unit may use a bandpass filter to remove high-frequency noise and low-frequency drift in the signal, and may apply noise reduction algorithms, such as wavelet transform or singular value decomposition, to reduce the impact of environmental noise. The feature visualization unit may generate time domain waveforms, frequency domain spectra, or other statistical charts to display the key features of the signal.

[0082] Preferably, the band segmentation unit of the data preprocessing module can be configured to automatically identify the R wave peak in the signal, which is a key feature point in the electrocardiogram and magnetocardiogram, and usually marks the beginning of the heartbeat cycle. The preprocessing model unit can be further refined into a processing flow including multiple stages, for example, first performing bandpass filtering, then noise suppression, and finally performing baseline drift correction. The feature visualization unit can provide interactive tools that allow users to zoom in, zoom out, or pan the graph to carefully observe the details of the signal.

[0083] Furthermore, it can also provide statistical feature analysis of the signal, such as mean, standard deviation, peak frequency, etc., as well as phase and amplitude information of the signal, so as to facilitate a more comprehensive understanding of the signal characteristics. Through these refined operation steps and alternatives, the system can process and display the magnetic cardiogram signal data more accurately.

[0084] In some embodiments, the EMD noise reduction method comprises the steps of:

[0085] S01. Find all the maximum and minimum points in the acquired magnetic cardiotoc signal and record them as local maximum and local minimum;

[0086] S02. Obtaining an upper envelope and a lower envelope by spline curve fitting;

[0087] S03. Subtract the upper envelope from the lower envelope to obtain a new component;

[0088] S04. Check whether the new component obtained in step S03 is an intrinsic mode function. If not, loop through steps S01-S03 again. If yes, proceed to the next step.

[0089] S05. Determine whether the number of components reaches a preset value, if yes, save all components, if not, return to step S01;

[0090] S06. Performing noise reduction processing on the magnetocardiographic signal according to the number of components and the high frequency band of the noise signal, wherein: the more the number of components, the more the high frequency band of the signal after noise reduction, and vice versa.

[0091] It should be noted that the EMD denoising method of this system is an algorithm for processing magnetic cardio signals, and its purpose is to identify and remove noise by decomposing the signal. EMD (Empirical Mode Decomposition) is an adaptive time-frequency analysis method that can decompose complex signals into a series of intrinsic mode functions (IMFs), which are inherent components of the signal and can be trend terms, periodic components or instantaneous components. In this way, the noise component can be separated from the original signal, thereby achieving noise reduction.

[0092] Specifically, the steps of the EMD denoising method include: first, in the acquired magnetic heart signal, find all the maximum and minimum points, which are called local maximum and local minimum points, which are the key feature points in the signal decomposition process. Then, fit these extreme points through spline curves to obtain the upper envelope and lower envelope of the signal. Then, subtract the upper envelope from the lower envelope to obtain a new component. This process can be repeated until the new component is an intrinsic mode function. Finally, the magnetic heart signal is denoised according to the number of components and the high frequency band of the noise signal. The more components there are, the more high frequency bands there are in the denoised signal, and vice versa.

[0093] Preferably, the EMD denoising method can be further refined to include automated parameter settings, for example, a threshold can be set to determine whether a new component is an intrinsic mode function, or an upper limit on the number of iterations can be set to control the depth of the decomposition process. In addition, a verification step can be introduced to check whether the decomposed IMFs meet the preset physical or statistical properties to ensure the quality of the denoised signal.

[0094] Furthermore, as an alternative, other noise reduction techniques, such as wavelet transform or filter design, can be considered to improve the noise reduction effect. Through these refined operation steps and alternatives, the system can more effectively remove noise from the magnetic cardio signal and improve the usability of the signal.

[0095] The above-mentioned embodiments of the present invention have the following beneficial effects: The magnetocardiogram feature analysis system described in the present invention can provide a comprehensive solution for the acquisition, preprocessing, feature extraction, model training and verification, and testing and prediction of magnetocardiogram signals. The system can effectively improve the degree of automation of data processing and the accuracy of analysis through integrated modular design. The data acquisition module can collect multi-channel magnetocardiogram signals, while the data preprocessing module is responsible for noise reduction and filtering to ensure signal quality. At the same time, through the band segmentation unit, the preprocessing model unit and the feature visualization unit, the signal can be deeply analyzed and visualized, enhancing the depth and breadth of signal processing.

[0096] In addition, the feature extraction and analysis module can accurately calibrate key feature points and extract feature values ​​for differential analysis, which is crucial for identifying and understanding the characteristics of cardiac activity. The model training and verification module allows the use of a variety of supervised learning algorithms, such as support vector machines, decision trees, etc., to adapt to different data processing requirements, which can improve the flexibility and adaptability of the model. Through the collaborative work of these modules, the system can analyze the magnetic heart signal more accurately.

[0097] like Figure 2 As shown, a magnetocardiogram feature analysis method 200 of some embodiments includes:

[0098] Step 201, collecting multi-channel magnetocardiographic signals;

[0099] Step 202, denoising and filtering the collected data, and drawing a one-dimensional magnetocardiographic signal graph of each channel, and performing band segmentation of the one-dimensional magnetocardiographic signal;

[0100] Step 203, calibrating key feature points and extracting feature values ​​for each channel, and performing difference analysis on the calibrated features;

[0101] Step 204, input the effective features of the training set into the machine learning model for training, input the effective features of the magnetic cardiotonic signals of the validation set into the model, and adjust the model parameters;

[0102] Step 205: obtain the test set magnetocardiographic signal for analysis, and the model output is a label of the analysis result.

[0103] It is understandable that the steps described in the magnetocardiogram feature analysis method 200 are similar to those described in the reference Figure 1 Therefore, the modules, features and beneficial effects described above for the magnetocardiogram feature analysis system are also applicable to the magnetocardiogram feature analysis method 200 and the operations contained therein, and will not be described in detail here.

[0104] Reference below Figure 3, which shows a schematic diagram of a structure 300 of an electronic device suitable for implementing some embodiments of the present invention. The electronic devices in some embodiments of the present invention may include, but are not limited to, mobile terminals such as mobile phones, notebook computers, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 3 The terminal device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0105] like Figure 3 As shown, the electronic device 300 may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 301, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 302 or a program loaded from a storage device 308 into a random access memory (RAM) 303. In the RAM 303, various programs and data required for the operation of the electronic device 300 are also stored. The processing device 301, the ROM 302, and the RAM 303 are connected to each other via a bus 304. An input / output (I / O) interface 305 is also connected to the bus 304.

[0106] Typically, the following devices may be connected to the I / O interface 305: input devices 306 including, for example, a touch screen, a touch pad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; output devices 307 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; storage devices 308 including, for example, a magnetic tape, a hard disk, etc.; and communication devices 309. The communication devices 309 may allow the electronic device 300 to communicate with other devices wirelessly or by wire to exchange data. Although Figure 3 The electronic device 300 is shown with various devices, but it should be understood that it is not required to implement or possess all the devices shown. More or fewer devices may be implemented or possessed instead. Figure 3 Each block shown in the figure may represent one device, or may represent multiple devices as required.

[0107] Furthermore, the storage medium of the embodiment of the present application stores program instructions that can implement all the above methods, wherein the program instructions can be stored in the above storage medium in the form of a software product, including several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) or a processor to execute all or part of the steps of the methods 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, or terminal devices such as a computer, a server, a mobile phone, and a tablet.

[0108] The above descriptions are only some preferred embodiments of the present invention and an explanation of the technical principles used. Those skilled in the art should understand that the scope of the invention involved in the embodiments of the present invention is not limited to the technical solutions formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above inventive concept. For example, the above features are replaced with (but not limited to) technical features with similar functions disclosed in the embodiments of the present invention to form a technical solution.

Claims

1. A magnetocardiogram feature analysis system, characterized in that: include: Data acquisition module, used for collecting multi-channel magnetic cardiology signals; The data preprocessing module is used to reduce noise and filter the collected data, draw the one-dimensional magnetocardiographic signal diagram of each channel, and perform band segmentation of the one-dimensional magnetocardiographic signal; The feature extraction and analysis module calibrates the key feature points and extracts the feature values ​​of each channel, and performs difference analysis on the calibrated features; The model training and verification module is used to input the effective features of the training set into the machine learning model for training, and input the effective features of the magnetic cardiotoc signal of the verification set into the model to adjust the model parameters; The model testing and prediction module is used to obtain the magnetic cardiotonic signals of the test set for analysis, and the model output is the label of the analysis result.

2. The magnetocardiogram feature analysis system according to claim 1, characterized in that: The data management module is used to manage the data division in the entire system, ensure that the data set is correctly divided into training set, validation set and test set according to the specified ratio, and implement unified data normalization processing.

3. The magnetocardiogram feature analysis system according to claim 1, characterized in that: The feature extraction module includes extracting maximum amplitude, minimum amplitude, average amplitude, amplitude change rate, amplitude ratio, waveform duration, peak interval and waveform morphological features.

4. The magnetocardiogram feature analysis system according to claim 1, characterized in that: The model training and validation module can be selected from one or a combination of support vector machines, decision trees, random forests, gradient boosting trees, K nearest neighbors, naive Bayes, logistic regression, LASSO regression or other supervised learning algorithms suitable for processing one-dimensional magnetocardiogram features.

5. The magnetocardiogram feature analysis system according to claim 1, characterized in that: It also includes a user interface module for providing a user interface for operators to manage data and view analysis results.

6. The magnetocardiogram feature analysis system according to claim 1, characterized in that: The data preprocessing module comprises: A band segmentation unit, used to perform band segmentation on the one-dimensional magnetocardiographic signal and mark it according to time; A preprocessing model unit, used for denoising and filtering the segmented signal; The feature visualization unit is used to visualize the features of the processed signal.

7. The magnetocardiogram feature analysis system according to claim 1, characterized in that: The noise reduction method is an EMD noise reduction method, comprising the steps of: S01. Find all the maximum and minimum points in the acquired magnetic cardiotoc signal and record them as local maximum and local minimum; S02. Obtaining an upper envelope and a lower envelope by spline curve fitting; S03. Subtract the upper envelope from the lower envelope to obtain a new component; S04. Check whether the new component obtained in step S03 is an intrinsic mode function. If not, loop through steps S01-S03 again. If yes, proceed to the next step. S05. Determine whether the number of components reaches a preset value, if yes, save all components, if not, return to step S01; S06. Performing noise reduction processing on the magnetocardiographic signal according to the number of components and the high frequency band of the noise signal, wherein: the more the number of components, the more the high frequency band of the signal after noise reduction, and vice versa.

8. A method for analyzing magnetocardiogram characteristics, characterized in that: include: Collect multi-channel magnetic cardiogram signals; The collected data is denoised and filtered, and the one-dimensional magnetocardiographic signal diagram of each channel is drawn to perform band segmentation of the one-dimensional magnetocardiographic signal; Calibrate key feature points and extract feature values ​​for each channel, and perform difference analysis on the calibrated features; Input the effective features of the training set into the machine learning model for training, and input the effective features of the magnetic cardiotoc signal of the validation set into the model to adjust the model parameters; The test set magnetocardiographic signals are obtained for analysis, and the model output is the label of the analysis result.

9. An electronic device, characterized in that: The method comprises a memory and a processor, wherein the memory stores a computer program, and when the computer program is run on the processor, the method described in claim 8 is executed.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method of claim 8 is implemented.

Images