CN112450942B

CN112450942B - Electrocardiosignal monitoring method, system, device and medium

Info

Publication number: CN112450942B
Application number: CN202011350453.9A
Authority: CN
Inventors: 徐琳; 李爽; 姜文才; 秦长瑜; 陈梅香; 阮征
Original assignee: Southern Theater Command General Hospital of PLA
Current assignee: Southern Theater Command General Hospital of PLA
Priority date: 2020-11-26
Filing date: 2020-11-26
Publication date: 2023-01-24
Anticipated expiration: 2040-11-26
Also published as: CN112450942A

Abstract

The invention provides a method, a system, a device and a medium for monitoring electrocardiosignals, wherein the method comprises the steps of obtaining an electrocardiosignal sample, and segmenting according to the electrocardiosignal sample to obtain an electrocardiogram signal sample and heart beating cycle characteristics; performing wavelet transformation according to the electrocardiogram sample signal to obtain wavelet characteristics; inputting an electrocardiogram signal sample, heart beating cycle characteristics and wavelet characteristics into a first model to obtain arrhythmia types; inputting the electrocardiogram signal sample, the heart beat cycle characteristic and the wavelet characteristic into a second model to obtain a prediction data array; carrying out visual display by combining arrhythmia categories and prediction data arrays; compared with the existing ECG signal classification prediction method, the novel framework constructed by wavelet transformation and two learning models has higher performance, realizes the lightweight of the model, can realize off-line processing, saves a large amount of cost, and can be widely applied to the technical field of signal analysis processing.

Description

Electrocardiosignal monitoring method, electrocardiosignal monitoring system, electrocardiosignal monitoring device and electrocardiosignal monitoring medium

Technical Field

The invention belongs to the technical field of signal analysis and processing, and particularly relates to a method, a system, a device and a medium for monitoring electrocardiosignals.

Background

Electrocardiography (ECG) is a technique that uses an electrocardiograph to record from the body surface the pattern of electrical activity changes produced by each cardiac cycle of the heart. ECG signals have a stationary characteristic, and in the related art, there are many cases where other electrocardiographic signals such as arrhythmia are abnormal from ECG signals. However, the existing classification prediction method based on the ECG signal cannot process data in real time, and if the real-time processing of the signal needs to be realized, computing resources of a cloud or a server need to be used; the existing classification prediction method can not realize off-line processing, and uses a large model, so that the operation and maintenance work is complex and the cost is high.

Disclosure of Invention

In view of the above, to at least partially solve one of the above technical problems, embodiments of the present invention provide a method for monitoring an ecg signal, which is light and capable of performing offline processing; meanwhile, the invention also provides a system, a device and a computer readable storage medium which can correspondingly realize the method.

In a first aspect, an embodiment of the present invention provides a method for monitoring an electrocardiographic signal, including the following steps:

acquiring an electrocardiosignal sample, and segmenting according to the electrocardiosignal sample to obtain an electrocardiogram signal sample and heart beat cycle characteristics;

performing wavelet transformation according to the electrocardiogram sample signal to obtain wavelet characteristics;

inputting the electrocardiogram signal sample, the heart beating cycle characteristic and the wavelet characteristic into a first model to obtain arrhythmia types;

inputting the electrocardiogram signal sample, the heart beating cycle characteristic and the wavelet characteristic into a second model to obtain a prediction data array;

and carrying out visual display by combining the arrhythmia types and the prediction data array.

In some embodiments of the present invention, the step of performing wavelet transform to obtain wavelet features according to the electrocardiogram sample signals comprises:

discretizing the electrocardiogram sample signal to obtain a discrete signal;

performing empirical mode decomposition on the discrete signal to obtain first high-frequency information and first low-frequency information;

extracting a wavelet feature according to the first high-frequency information and the first low-frequency information, wherein the wavelet feature comprises: time domain features and frequency domain features.

In some embodiments of the present invention, the first model comprises a first branch model and a second branch model, and the step of inputting the electrocardiogram signal samples, the heart beat cycle characteristics and the wavelet characteristics into the first model to obtain the arrhythmia category comprises:

inputting the electrocardiogram signal sample and the heart beating cycle characteristics into the first branch model to obtain a first characteristic array;

inputting the heart beating cycle characteristics and the wavelet characteristics into the second branch model to obtain a second characteristic array;

and combining the first characteristic data and the second characteristic array to obtain the arrhythmia category through a fully-connected neural network.

In some embodiments of the present invention, the step of inputting the electrocardiogram signal samples, the heart beat cycle characteristics and the wavelet characteristics into a second model to obtain a prediction data array comprises:

generating a downsampled set from the electrocardiogram signal samples;

obtaining a third feature array through principal component analysis according to the downsampling set and the wavelet features;

and obtaining the prediction data array by the third feature array through a recurrent neural network in the second model.

In some embodiments of the present invention, the step of visually presenting in combination the arrhythmia category and the prediction data array comprises:

inputting the arrhythmia type and the prediction data array into a multilayer neural network to obtain an electrocardiosignal prediction result;

and carrying out visual display according to the electrocardiosignal prediction result.

In some embodiments of the present invention, the step of performing wavelet transform according to the electrocardiogram sample signal to obtain wavelet features further includes:

down-sampling the first high-frequency information to obtain second high-frequency information;

down-sampling the first low-frequency information to obtain second low-frequency information;

and extracting to obtain wavelet characteristics according to the second high-frequency information and the second low-frequency information.

In some embodiments of the invention, the first branch model comprises at least one recurrent neural network; the second branch model comprises at least one recurrent neural network.

In a second aspect, a technical solution of the present invention further provides a system for monitoring an electrocardiographic signal, including:

the signal acquisition module is used for acquiring an electrocardiosignal sample and segmenting according to the electrocardiosignal sample to obtain an electrocardiogram signal sample and heart beating cycle characteristics;

the characteristic extraction module is used for carrying out wavelet transformation according to the electrocardiogram sample signal to obtain wavelet characteristics;

the result prediction module is used for inputting the electrocardiogram signal samples, the heart beating cycle characteristics and the wavelet characteristics into a first model to obtain arrhythmia types; inputting the electrocardiogram signal sample, the heart beating cycle characteristic and the wavelet characteristic into a second model to obtain a prediction data array;

and the visualization module is used for carrying out visualization display by combining the arrhythmia types and the prediction data array.

In a third aspect, the present invention further provides a device for monitoring an electrocardiographic signal, including:

at least one processor;

at least one memory for storing at least one program;

when the at least one program is executed by the at least one processor, the at least one processor implements the method for monitoring an ecg signal according to the first aspect.

In a fourth aspect, the present invention also provides a storage medium in which a processor-executable program is stored, the processor-executable program being configured to implement the method as in the first aspect when executed by a processor.

Advantages and benefits of the present invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention:

according to the electrocardiosignal monitoring method provided by the invention, for each heartbeat signal, the input ECG sample, the heart beating period and the wavelet characteristics are provided for two independent machine learning models, so that the arrhythmia type and the data array of the electrocardiosignal are respectively obtained, and the real-time monitoring of the electrocardiosignal is realized; compared with the existing ECG signal classification prediction method, the novel framework constructed by wavelet transformation and two learning models provided by the scheme has higher performance, realizes the lightweight of the model, can realize off-line processing, and saves a large amount of cost.

Drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required to be used in the description of the embodiments are briefly introduced below, and it is obvious that the drawings in the description below are only some embodiments of the present application, and it is obvious for those skilled in the art to obtain other drawings without creative efforts.

FIG. 1 is a flow chart illustrating the steps of a method for monitoring electrocardiosignals according to the present invention;

FIG. 2 is a schematic diagram of a model of wavelet analysis in an embodiment of the present invention;

FIG. 3 is a schematic diagram of a down-sampling filter in an embodiment of the present invention;

FIG. 4 is a schematic diagram of a model of wavelet analysis in another embodiment of the present invention;

FIG. 5 is a diagram of an algorithm architecture in an embodiment of the invention.

Detailed Description

Reference will now be made in detail to the embodiments of the present invention, examples of which are illustrated in the accompanying drawings, wherein like reference numerals refer to the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are illustrative only for the purpose of explaining the present invention and are not to be construed as limiting the present invention. For the step numbers in the following embodiments, they are set for convenience of illustration only, the order between the steps is not limited at all, and the execution order of each step in the embodiments can be adapted according to the understanding of those skilled in the art.

First, the main terms referred to in this scheme are explained, and other non-main terms will first appear and will not be explained here:

electrocardiogram (ECG) is a technique for recording a pattern of changes in electrical activity generated in each cardiac cycle of the heart from the body surface using an electrocardiograph.

The RR period is the heart beat period.

In a first aspect, as shown in fig. 1, the technical solution of the present application provides a method for monitoring an electrocardiographic signal, which mainly includes steps S01-S05:

and S01, acquiring an electrocardiosignal sample, and segmenting according to the electrocardiosignal sample to obtain an electrocardiogram signal sample and heart beating cycle characteristics.

Specifically, according to the acquisition device of the electrocardiograph signal, for example, an electrocardiograph acquires an electrocardiograph signal sample, the acquired electrocardiograph signal sample is composed of a plurality of waves and wave bands, and in this embodiment, only the QRS complex in the electrocardiograph limit signal needs to be monitored. The QRS complex is formed by activation of the left and right ventricles via the his bundle, and simultaneous activation of the left and right brachial tracts. The QRS complex represents ventricular depolarization with an activation time interval of less than 0.11 seconds. When the conditions of conduction block, ventricular enlargement or hypertrophy and the like of the left and right fascicles of the heart occur, the QRS complex is widened, deformed and prolonged in time limit. More specifically, embodiments select a heart beat period, which is the R-wave duration of two QRS waves (RR period), as an input to a subsequent deep learning model, which can be used to assess specific heart rate variations.

And S02, performing wavelet transformation according to the electrocardiogram sample signal to obtain wavelet characteristics.

In particular, the ECG signal has non-stationary characteristics, so in order to obtain time and frequency domain information of the electrocardiogram sample signals, a discrete Wavelet transform is applied to the digitized sample signals in each heartbeat, and in particular, embodiments select the family of multi-beshies wavelets (Daubechies Wavelet), because the family of Daubechies wavelets is similar to the ECG signal, with lower order Daubechies wavelets having higher temporal resolution but lower frequency resolution, and higher order Daubechies wavelets having higher frequency resolution and lower temporal resolution.

In some alternative embodiments, step S02 can be further subdivided into steps S021-S023:

s021, discretizing the electrocardiogram sample signal to obtain a discrete signal;

s022, performing empirical mode decomposition on the discrete signal to obtain first high-frequency information and first low-frequency information;

s023, extracting the wavelet feature according to the first high-frequency information and the first low-frequency information.

The wavelet features comprise time domain features and frequency domain features, the first high-frequency information is high-frequency information obtained by dividing an input electrocardiogram sample signal, and the corresponding first low-frequency information is low-frequency information obtained by dividing. Specifically, in the processing of the electrocardiogram sample signal, a continuous Wavelet and its Wavelet transform need to be discretized, in the embodiment, binary Discrete processing is used, and the discretized Wavelet and its corresponding Wavelet transform are transformed into Discrete Wavelet Transform (DWT), and in the Discrete Wavelet transform process, the Discrete Wavelet transform is obtained by discretizing the scale of the continuous Wavelet transform and the displacement by the power of 2, and may also be referred to as binary Wavelet transform.

As shown in fig. 2, first define the signal and filter in the process; where x [ N ] is a discrete input signal, length N, g [ N ] is a low pass filter (low pass filter) that filters out the high frequency portion of the input signal and outputs the low frequency portion, and h [ N ] is a high pass filter (high pass filter) that filters out the low frequency portion and outputs the high frequency portion, as opposed to the low pass filter. As shown in fig. 3, in an embodiment, a down sampling filter (down sampling filter) is further included, and step S02 may further include step S024:

s024, down-sampling the first high-frequency information to obtain second high-frequency information, and extracting wavelet characteristics according to the second high-frequency information. With x [ n ] as an input, the output is y [ n ] = x [ Qn ], e.g., Q =2. The resulting wavelet transform is shown in fig. 4, where for the electrocardiogram sample signal, the low frequencies contain the features of the signal, while the high frequency components give details or differences of the signal. Because the approximate characteristics and detail characteristics of the electrocardiogram sample signals need to be obtained through wavelet analysis. Approximately represents the high-scale, i.e., low-frequency, information of the signal; the details represent the high scale, i.e. high frequency information, of the signal. Thus, the original signal passes through two mutual filters to produce two signals.

By continuously decomposing the approximate signal through a continuous decomposition process, the signal can be decomposed into a plurality of low-resolution components. In theory the decomposition can proceed without limitation, but in embodiments the decomposition can proceed until the details (i.e. the high frequency information) contain only a single sample. Therefore, the appropriate number of decomposition layers may be selected depending on the characteristics of the signal or an appropriate criterion.

S03, inputting an electrocardiogram signal sample, heart beating cycle characteristics and wavelet characteristics into the first model to obtain arrhythmia types;

s04, inputting the electrocardiogram signal sample, the heart beating cycle characteristic and the wavelet characteristic into a second model to obtain a prediction data array;

in particular, steps S03 and S04 are two independent steps performed in synchronism, for each heartbeat, an ECG sample to be input

Extracted RR interval features

And wavelet characteristics

Transmitting to two separate neural network deep learning models, e.g. RNN-based models, respectively called models

Namely the first model, is to be used,and a model

I.e. the second model. The overall view of the algorithm in the embodiment is shown in fig. 5, the two models respectively output the prediction of arrhythmia category and the prediction of specific data array, and then the prediction results obtained by the two models are fused to form the final prediction of each heartbeat.

In an embodiment, step S03 may be further subdivided into steps S031-S033:

s031, inputting the electrocardiogram signal sample and the heart beat cycle characteristics into the first branch model to obtain a first characteristic array;

s032, inputting the heart beating cycle characteristic and the wavelet characteristic into a second branch model to obtain a second characteristic array;

and S033, combining the first characteristic data and the second characteristic array, and obtaining the arrhythmia category through a full connection neural network.

Specifically, as shown in FIG. 5, the model

Consisting of two branches. Each branch comprising at least one RNN unit, each RNN further comprising a number of hidden units, a first branch model (i.e. left branch) input being provided by

Is shown and bound by

And

and (4) forming. RNN unit processing array in first branch model

And extracting n feature arrays, namely the first feature array. Likewise, the second branch model will

And

connected to an array

Then, this array is processed and m feature arrays, i.e., the second feature array, are extracted. The outputs of the two branches are connected and input to a fully connected neural network layer to generate the probabilities for all Ny output arrhythmia classes. Namely the model

The resulting maximum probability determines the arrhythmia class.

In an embodiment, step S04 may be further subdivided into steps S041-S043:

s041, generating a downsampling set according to the electrocardiogram signal sample;

s042, obtaining a third feature array through principal component analysis according to the downsampling set and the wavelet features;

and S043, obtaining a prediction data array by the third characteristic array through a recurrent neural network in the second model.

Specifically, as shown in FIG. 5, the model

Containing only one branch, but also at least one RNN unit, by

Down-sampled collection of

And

and

connected, and then the connected data sets are subjected to principal component analysis to form an array. Then by the model

The RNN in the data block is processed to obtain a feature array (namely, a third feature array), and the feature array is input to a full-connection neural network layer to obtain a prediction data array.

And S05, carrying out visual display by combining arrhythmia categories and prediction data arrays.

Specifically, a specific prediction result of the electrocardiosignal is generated according to the arrhythmia type prediction data array, the prediction result is displayed on an interaction interface of the mobile terminal device, an abnormal result exists in the prediction result, namely the prediction result comprises a specific arrhythmia type and a specific abnormal numerical value, and alarm information can be correspondingly generated for reminding. In some embodiments, step S05 may be further subdivided into:

s051, inputting the arrhythmia types and the prediction data array into a multilayer neural network to obtain an electrocardiosignal prediction result;

and S052, carrying out visual display according to the electrocardiosignal prediction result.

Specifically, the model is modeled by a Multi-Layer Perceptron (MLP)

And a model

The output of the fusion module is fused, and then the result obtained by the fusion is displayed visually.

In a second aspect, an embodiment of the present application further provides a system for monitoring an electrocardiograph signal, including:

the result prediction module is used for inputting the electrocardiogram signal sample, the heart beating cycle characteristics and the wavelet characteristics into the first model to obtain arrhythmia types; inputting the electrocardiogram signal sample, the heart beating cycle characteristic and the wavelet characteristic into a second model to obtain a prediction data array;

and the visualization module is used for carrying out visualization display by combining the arrhythmia types and the prediction data arrays.

In a third aspect, an embodiment of the present application further provides an apparatus for monitoring an ecg signal, which includes at least one processor; at least one memory for storing at least one program; when the at least one program is executed by the at least one processor, the at least one processor implements the method for monitoring cardiac electrical signals as in the first aspect.

Embodiments of the present invention further provide a storage medium, in which a program is stored, and the program is executed by a processor to perform the method in the first aspect.

From the above specific implementation process, it can be concluded that the technical solution provided by the present invention has the following advantages or advantages compared to the prior art:

1. compared with the existing ECG signal classification prediction method, the novel framework constructed by wavelet transformation and two learning models provided by the technical scheme of the application has higher performance.

2. According to the technical scheme, the lightweight of the model can be realized, the off-line processing can be realized, and a large amount of cost is saved.

In alternative embodiments, the functions/acts noted in the block diagrams may occur out of the order noted in the operational illustrations. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality/acts involved. Furthermore, the embodiments presented and described in the flow charts of the present invention are provided by way of example in order to provide a more comprehensive understanding of the technology. The disclosed methods are not limited to the operations and logic flows presented herein. Alternative embodiments are contemplated in which the order of various operations is changed and in which sub-operations described as part of larger operations are performed independently.

Furthermore, although the present invention is described in the context of functional modules, it should be understood that, unless otherwise stated to the contrary, one or more of the functions and/or features may be integrated in a single physical device and/or software module, or one or more of the functions and/or features may be implemented in a separate physical device or software module. It will also be appreciated that a detailed discussion of the actual implementation of each module is not necessary for an understanding of the present invention. Rather, the actual implementation of the various functional modules in the apparatus disclosed herein will be understood within the ordinary skill of an engineer, given the nature, function, and internal relationship of the modules. Accordingly, those skilled in the art can, using ordinary skill, practice the invention as set forth in the claims without undue experimentation. It is also to be understood that the specific concepts disclosed are merely illustrative of and not intended to limit the scope of the invention, which is to be determined from the appended claims along with their full scope of equivalents.

Wherein the functions, if implemented in the form of software functional units and sold or used as a stand-alone product, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.

The logic and/or steps represented in the flowcharts or otherwise described herein, e.g., an ordered listing of executable instructions that can be considered to implement logical functions, can be embodied in any computer-readable medium for use by or in connection with an instruction execution system, apparatus, or device, such as a computer-based system, processor-containing system, or other system that can fetch the instructions from the instruction execution system, apparatus, or device and execute the instructions. For the purposes of this description, a "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transport the program for use by or in connection with the instruction execution system, apparatus, or device.

More specific examples (a non-exhaustive list) of the computer-readable medium would include the following: an electrical connection (electronic device) having one or more wires, a portable computer diskette (magnetic device), a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM). Additionally, the computer-readable medium could even be paper or another suitable medium upon which the program is printed, as the program can be electronically captured, via for instance optical scanning of the paper or other medium, then compiled, interpreted or otherwise processed in a suitable manner if necessary, and then stored in a computer memory.

It should be understood that portions of the present invention may be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, various steps or methods may be implemented in software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, any one or combination of the following techniques, which are known in the art, may be used: a discrete logic circuit having a logic gate circuit for implementing a logic function on a data signal, an application specific integrated circuit having an appropriate combinational logic gate circuit, a Programmable Gate Array (PGA), a Field Programmable Gate Array (FPGA), or the like.

In the description herein, references to the description of the term "one embodiment," "some embodiments," "an example," "a specific example," or "some examples," etc., mean that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the invention. In this specification, the schematic representations of the terms used above do not necessarily refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.

While embodiments of the invention have been shown and described, it will be understood by those of ordinary skill in the art that: various changes, modifications, substitutions and alterations can be made to the embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

While the preferred embodiments of the present invention have been illustrated and described, it will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the invention as defined by the appended claims.

Claims

1. Electrocardiosignal's monitoring system, its characterized in that includes:

the characteristic extraction module is used for carrying out wavelet transformation according to the electrocardiogram signal sample to obtain wavelet characteristics;

the visualization module is used for carrying out visualization display by combining the arrhythmia types and the prediction data array;

the first model includes a first branch model and a second branch model, and the outcome prediction module includes:

the first submodule is used for inputting the electrocardiogram signal samples and the heart beating cycle characteristics into the first branch model to obtain a first characteristic array;

the second sub-module is used for inputting the heart beating cycle characteristics and the wavelet characteristics to the second branch model to obtain a second characteristic array;

the third submodule is used for combining the first characteristic array and the second characteristic array and obtaining the arrhythmia type through a fully-connected neural network;

the result prediction module further comprises:

a seventh sub-module for generating a down-sampling set from the electrocardiogram signal samples;

the eighth sub-module is used for obtaining a third feature array through principal component analysis according to the downsampling set, the heart beat interval feature and the wavelet feature;

and the ninth submodule is used for enabling the third feature array to obtain the prediction data array through a recurrent neural network in the second model.

2. The system for monitoring cardiac electrical signals according to claim 1, wherein the feature extraction module comprises:

the fourth sub-module is used for carrying out discretization processing on the electrocardiogram signal sample to obtain a discrete signal;

the fifth sub-module is used for carrying out empirical mode decomposition on the discrete signal to obtain first high-frequency information and first low-frequency information;

a sixth sub-module, configured to extract a wavelet feature according to the first high-frequency information and the first low-frequency information, where the wavelet feature includes: time domain features and frequency domain features.

3. The system for monitoring cardiac electrical signals according to claim 1, wherein the visualization module comprises:

a tenth submodule, configured to input the arrhythmia category and the prediction data array into a multilayer neural network, so as to obtain an electrocardiographic signal prediction result; and carrying out visual display according to the electrocardiosignal prediction result.

4. The system for monitoring cardiac electrical signals according to claim 2, wherein the feature extraction module further comprises:

the eleventh submodule is used for performing down-sampling on the first high-frequency information to obtain second high-frequency information;

the twelfth submodule is used for performing down-sampling on the first low-frequency information to obtain second low-frequency information;

and the thirteenth sub-module is used for extracting wavelet characteristics according to the second high-frequency information and the second low-frequency information.

5. The system for monitoring cardiac electrical signals according to claim 1, wherein the first branch model comprises at least one recurrent neural network; the second branch model comprises at least one recurrent neural network.

6. Electrocardiosignal's monitoring devices, its characterized in that includes:

at least one processor;

at least one memory for storing at least one program;

when executed by the at least one processor, cause the at least one processor to operate the system for monitoring cardiac electrical signals according to any one of claims 1-5.

7. A computer-readable storage medium in which a program executable by a processor is stored, characterized in that: the processor executable program when executed by the processor is for running a system for monitoring cardiac electrical signals as defined in any one of claims 1-5.