CN110522440B

CN110522440B - Electrocardiosignal recognition device based on grouping convolution neural network

Info

Publication number: CN110522440B
Application number: CN201910740289.3A
Authority: CN
Inventors: 王红梅
Original assignee: Guangzhou Shiyuan Electronics Thecnology Co Ltd
Current assignee: Guangzhou Shiyuan Electronics Thecnology Co Ltd
Priority date: 2019-08-12
Filing date: 2019-08-12
Publication date: 2021-04-13
Anticipated expiration: 2039-08-12
Also published as: CN110522440A; WO2021027224A1

Abstract

The application relates to an electrocardiosignal recognition device based on a grouping convolution neural network, a computer device and a storage medium. The apparatus is for: acquiring multi-lead electrocardiosignals, and splitting the multi-lead electrocardiosignals to obtain independent lead electrocardiosignals; inputting the independent lead electrocardiosignals into the convolution blocks respectively to obtain convolution characteristics; grouping the convolution characteristics to obtain grouping characteristics, and inputting the grouping characteristics into the grouping convolution block to obtain grouping convolution characteristics; combining the grouped convolution characteristics to obtain electrocardiosignal combination characteristics, and carrying out full-connection processing on the electrocardiosignal combination characteristics to obtain myocardial infarction abnormal probability; and judging the multi-lead electrocardiosignals to be myocardial infarction signals according to the myocardial infarction abnormal probability. The device can solve the problem that the existing electrocardiosignal identification method has inaccurate myocardial infarction abnormity identification.

Description

Electrocardiosignal recognition device based on grouping convolution neural network

Technical Field

The present application relates to the field of electrocardiographic signal identification, and in particular, to an electrocardiographic signal identification method and apparatus based on a packet convolutional neural network, a computer device, and a storage medium.

Background

Coronary Heart Disease (CHD) is the leading killer of human health in modern society. Myocardial infarction is the most serious consequence of coronary heart disease. Currently, the risk of myocardial infarction of a patient is generally predicted by means of recognizing electrocardiosignals.

The common electrocardiosignal identification method is mainly based on key point detection of electrocardiosignals. For example, features of ST-segment, T-wave, and R-wave of the electrocardiographic signal are extracted and detected.

However, the above method relies heavily on the detection of key points such as Q-waves, P-waves, J-points, S-points, T-waves, etc. When the quality of the electrocardiosignal is not good, the key point may not be accurately positioned, so that the myocardial infarction risk cannot be accurately identified from the electrocardiosignal.

Therefore, the current electrocardiosignal identification method has the problem of inaccurate myocardial infarction abnormity identification.

Disclosure of Invention

In view of the above, it is necessary to provide an electrocardiographic signal identification method, apparatus, computer device and storage medium based on a packet convolutional neural network.

In a first aspect, an apparatus for identifying an ecg signal based on a block convolutional neural network, the block convolutional neural network including a convolution block and a block convolution block, the apparatus comprising:

the signal acquisition module is used for acquiring multi-lead electrocardiosignals and splitting the multi-lead electrocardiosignals to obtain a plurality of groups of independent lead electrocardiosignals;

the convolution module is used for respectively inputting each group of the independent lead electrocardiosignals to the convolution block to obtain the convolution characteristics of the corresponding group of the independent lead electrocardiosignals;

the characteristic grouping module is used for grouping the convolution characteristics of each group of independent lead electrocardiosignals to obtain the grouping characteristics of the group of independent lead electrocardiosignals, and inputting the grouping characteristics into the grouping convolution block to obtain a plurality of grouping convolution characteristics of the group of independent lead electrocardiosignals;

the characteristic combination module is used for combining a plurality of grouped convolution characteristics of each group of independent lead electrocardiosignals to obtain the electrocardiosignal combination characteristics of the group of independent lead electrocardiosignals;

the full-connection processing module is used for performing full-connection processing on the electrocardiosignal combination characteristics of each group of independent lead electrocardiosignals to obtain the myocardial infarction abnormal probability;

and the judging module is used for judging the multi-lead electrocardiosignals to be myocardial infarction signals according to the myocardial infarction abnormal probability.

In another embodiment, the feature grouping module includes:

the group number obtaining submodule is used for obtaining the group convolution group number of the grouping convolution block;

and the grouping submodule is used for uniformly grouping the convolution characteristics of each group of independent lead electrocardiosignals according to the grouping convolution group number to obtain the grouping characteristics of the group of independent lead electrocardiosignals.

In another embodiment, the block convolution block includes a block convolution layer and a block max pooling layer, and the feature grouping module includes:

the grouping convolution characteristic output submodule is used for carrying out convolution, batch normalization and activation on the grouping characteristics through the grouping convolution layer to obtain the grouping convolution layer output characteristics;

and the pooling sub-module is used for performing maximal pooling on the output characteristics of the grouped convolutional layer through the grouped maximal pooling layer to obtain a plurality of grouped convolutional characteristics of the group of independent lead electrocardiosignals.

In another embodiment, the convolution block of the block convolutional neural network includes a first convolution block and a second convolution block, and the convolution module includes:

the first convolution sub-module is used for performing convolution and pooling on the independent lead electrocardiosignals through the first convolution block to obtain convolution block output characteristics;

and the second convolution submodule is used for performing convolution and pooling on the output bits of the convolution block through the second convolution block to obtain the convolution characteristics.

In another embodiment, the first convolution sub-module is specifically configured to:

performing convolution, batch normalization and activation on the independent lead electrocardiosignals through the first convolution layer to obtain a first convolution layer output characteristic; and performing maximum pooling on the output characteristic of the first convolution layer through the first maximum pooling layer to obtain the output characteristic of the convolution block.

In another embodiment, the second convolution block includes a second convolution layer and a second maximum pooling layer, and the second convolution sub-module is specifically configured to:

performing convolution, batch normalization and activation on the output characteristics of the convolution blocks through the second convolution layer to obtain second convolution layer output characteristics; and performing maximum pooling on the output characteristics of the second convolution layer through the second maximum pooling layer to obtain the convolution characteristics.

In another embodiment, the signal acquisition module includes:

the original signal receiving submodule is used for receiving an original signal;

the wavelet decomposition sub-module is used for performing wavelet decomposition on the original signal to obtain a wavelet decomposition signal; the wavelet decomposition signal has dimension X1;

the zero setting sub-module is used for setting the zero of the X2 dimensional signal in the wavelet decomposition signal to obtain a partial zero setting signal; wherein X2 is less than X1;

the inverse transformation submodule is used for carrying out wavelet inverse transformation on the partial zero-set signals to obtain de-noised signals; the de-noising signal is a signal after high-frequency noise and baseline drift are removed;

and the multi-lead acquisition sub-module is used for obtaining the multi-lead electrocardiosignals according to the de-noising signals.

In another embodiment, the multi-lead acquisition sub-module is specifically configured to:

determining the R wave position of the de-noising signal;

determining the first M1 positions of the R-wave position and the last M2 positions of the R-wave position;

and forming a structured signal matrix as the multi-lead electrocardiosignals by adopting the de-noised signals on the R wave position, the front M1 positions and the rear M2 positions.

In a second aspect, a method for training a packet convolutional neural network, the packet convolutional neural network comprising a convolutional block and a packet convolutional block, the method comprising:

acquiring electrocardiosignal training samples aiming at the grouped convolutional neural network;

performing machine training on the grouped convolutional neural network by adopting the electrocardiosignal training sample to obtain a trained grouped convolutional neural network; the trained packet convolution neural network comprises a pre-trained convolution block and a packet convolution block; the trained grouped convolutional neural network is used for respectively inputting each group of independent lead electrocardiosignals into the convolution block to obtain convolution characteristics of the corresponding group of independent lead electrocardiosignals; the independent lead electrocardiosignals are obtained by splitting multi-lead electrocardiosignals; the system is also used for grouping the convolution characteristics of each group of independent lead electrocardiosignals to obtain the grouping characteristics of the group of independent lead electrocardiosignals, and inputting the grouping characteristics into the grouping convolution block to obtain a plurality of grouping convolution characteristics of the group of independent lead electrocardiosignals; the system is also used for combining a plurality of grouped convolution characteristics of each group of independent lead electrocardiosignals to obtain the electrocardiosignal combination characteristics of the group of independent lead electrocardiosignals; the system is also used for carrying out full-connection processing on the electrocardiosignal combination characteristics of each group of independent lead electrocardiosignals to obtain the myocardial infarction abnormal probability; and the multi-lead electrocardiosignal is judged to be the myocardial infarction signal according to the myocardial infarction abnormal probability.

In a third aspect, a method for identifying an ecg signal of a block convolutional neural network, where the block convolutional neural network includes a convolutional block and a block convolutional block, the method includes:

acquiring multi-lead electrocardiosignals, and splitting the multi-lead electrocardiosignals to obtain a plurality of groups of independent lead electrocardiosignals;

inputting each group of the independent lead electrocardiosignals into the convolution block respectively to obtain convolution characteristics of the independent lead electrocardiosignals of the corresponding group;

grouping the convolution characteristics of each group of independent lead electrocardiosignals to obtain the grouping characteristics of the group of independent lead electrocardiosignals, and inputting the grouping characteristics into the grouping convolution block to obtain a plurality of grouping convolution characteristics of the group of independent lead electrocardiosignals;

combining a plurality of grouped convolution characteristics of each group of independent lead electrocardiosignals to obtain the electrocardiosignal combination characteristics of the group of independent lead electrocardiosignals;

fully connecting the electrocardiosignal combination characteristics of each group of independent lead electrocardiosignals to obtain the myocardial infarction abnormal probability;

and judging the multi-lead electrocardiosignals to be myocardial infarction signals according to the myocardial infarction abnormal probability.

In a fourth aspect, there is provided an apparatus for training a block convolutional neural network, the block convolutional neural network including a convolutional block and a block convolutional block, the apparatus comprising:

a training sample acquisition module, configured to acquire an electrocardiographic signal training sample for the packet convolutional neural network;

the machine training module is used for performing machine training on the packet convolutional neural network by adopting the electrocardiosignal training sample to obtain a trained packet convolutional neural network; the trained packet convolution neural network comprises a pre-trained convolution block and a packet convolution block; the trained grouped convolutional neural network is used for respectively inputting each group of independent lead electrocardiosignals into the convolution block to obtain convolution characteristics of the corresponding group of independent lead electrocardiosignals; the independent lead electrocardiosignals are obtained by splitting multi-lead electrocardiosignals; the system is also used for grouping the convolution characteristics of each group of independent lead electrocardiosignals to obtain the grouping characteristics of the group of independent lead electrocardiosignals, and inputting the grouping characteristics into the grouping convolution block to obtain a plurality of grouping convolution characteristics of the group of independent lead electrocardiosignals; the system is also used for combining a plurality of grouped convolution characteristics of each group of independent lead electrocardiosignals to obtain the electrocardiosignal combination characteristics of the group of independent lead electrocardiosignals; the system is also used for carrying out full-connection processing on the electrocardiosignal combination characteristics of each group of independent lead electrocardiosignals to obtain the myocardial infarction abnormal probability; and the multi-lead electrocardiosignal is judged to be the myocardial infarction signal according to the myocardial infarction abnormal probability.

In a fifth aspect, an electronic device is provided, which includes: a memory, one or more processors;

the memory for storing one or more programs;

when executed by the one or more processors, cause the one or more processors to perform operations comprising:

In a sixth aspect, there is provided a computer readable storage medium having stored thereon a computer program which when executed by a processor performs the steps of:

According to the electrocardiosignal identification method, the electrocardiosignal identification device, the computer equipment and the storage medium based on the packet convolutional neural network, the multiple-lead electrocardiosignals are obtained, and the multiple-lead electrocardiosignals are split to obtain the independent-lead electrocardiosignals; then, the independent lead electrocardiosignal is input into a convolution block to obtain convolution characteristics; then, grouping the convolution characteristics to obtain grouping characteristics, and inputting the grouping characteristics into a grouping convolution block to obtain grouping convolution characteristics; then, determining the myocardial infarction abnormal probability according to the grouping convolution characteristics; finally, according to the myocardial infarction abnormal probability, judging the multi-lead electrocardiosignals as myocardial infarction signals; therefore, the parameter quantity of convolution operation can be reduced, under the condition of not reducing the identification performance of the grouped convolution neural network, the overfitting of the grouped convolution neural network is reduced, and whether the multi-lead electrocardiosignals are myocardial infarction signals or not can be identified and judged more accurately; therefore, when the electrocardiosignals are identified, the accurate positioning of Q waves, P waves, J points, S points and T waves of key points of the electrocardiosignals is not needed, and even under the condition that the quality of the electrocardiosignals is poor and the key points of the electrocardiosignals cannot be accurately positioned, the grouped convolutional neural network is input, so that the myocardial infarction risk can be more accurately identified from the electrocardiosignals.

Drawings

Fig. 1 is a schematic structural diagram of an electrocardiographic signal recognition apparatus based on a packet convolutional neural network according to an embodiment of the present application;

fig. 2 is a schematic diagram of a neural network structure of an electrocardiographic signal identification method based on a packet convolutional network according to an embodiment of the present application;

fig. 3 is a schematic structural diagram of an electrocardiographic signal recognition apparatus based on a packet convolutional neural network according to a second embodiment of the present application;

FIG. 4A is a schematic representation of an original cardiac signal in one embodiment;

FIG. 4B is a diagram of a denoised signal in one embodiment;

FIG. 5 is a schematic diagram of a network structure of a method for identifying an electrocardiosignal based on a packet convolutional neural network in an embodiment;

FIG. 6 is a schematic diagram of feature dimension variation of electrocardiosignal identification based on a packet convolutional neural network in one embodiment;

FIG. 7 is a flow chart of cardiac signal identification based on neural networks;

FIG. 8 is a flowchart of a block convolutional neural network-based electrocardiosignal method according to a fourth embodiment of the present application;

fig. 9 is a schematic structural diagram of an electronic device according to a sixth embodiment of the present application;

FIG. 10 is a flowchart of a training method of a packet convolutional neural network according to a third embodiment of the present application;

fig. 11 is a schematic structural diagram of a training apparatus for a packet convolutional neural network according to a fifth embodiment of the present application.

Detailed Description

In order to make the objects, technical solutions and advantages of the present application more apparent, the present application is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.

Example one

Fig. 1 is a schematic structural diagram of an electrocardiograph signal recognition apparatus based on a packet convolutional neural network according to an embodiment of the present application. Specifically, referring to fig. 1, in an electrocardiographic signal recognition apparatus based on a packet convolutional neural network according to an embodiment of the present application, the packet convolutional neural network includes a pre-trained convolutional block and a packet convolutional block, and specifically includes:

the signal obtaining module 110 is configured to obtain multiple lead electrocardiographic signals, and split the multiple lead electrocardiographic signals to obtain multiple sets of independent lead electrocardiographic signals.

The multi-lead electrocardiographic signals can be signal data matrixes representing the multi-lead electrocardiographic signals. The multi-lead electrocardiosignals can be signals collected by a multi-lead electrocardio system.

The independent lead electrocardiographic signals may refer to electrocardiographic signals of each independent lead in the multi-lead electrocardiographic signals.

In specific implementation, the original signal can be collected, and the multi-lead electrocardiosignal can be obtained by preprocessing the original signal such as wavelet transformation and denoising. Then, the multi-lead electrocardiosignals are split to obtain a plurality of groups of independent lead electrocardiosignals.

In practical application, the multi-lead electrocardiosignals can be collected through the multi-lead electrocardio system. Currently, a common multi-lead electrocardiograph system is twelve leads. The twelve-lead multi-lead electrocardiograph signals comprise lead signals V1, V2, V3, V4, V5, V6, aVF, aVR, aVL, I, II and III. That is to say, the multi-lead electrocardiographic signals according to the embodiment of the present application may be two or more than two of the multi-lead electrocardiographic signals of the twelve leads. Then, the multi-lead electrocardiosignals are split according to respective lead types to obtain multiple groups of independent lead electrocardiosignals, and the number of the independent lead electrocardiosignals can be Z. For example, a twelve-lead multi-lead electrocardiographic signal is split to obtain 12 groups of independent lead electrocardiographic signals, namely, Z is 12.

And the convolution module 120 is configured to input each group of independent lead electrocardiographic signals to the convolution block, so as to obtain convolution characteristics of the corresponding group of independent lead electrocardiographic signals.

The volume Block (Basic relational Block) may be a set used to perform a series of operations such as one-dimensional convolution, batch normalization, activation, and pooling on the input features. The function of the series of operations in the neural network is named as a volume block.

The batch Normalization may be replaced by group Normalization, Instance Normalization, Layer Normalization, and other algorithms.

Among them, common activation functions include ReLU (an activation function), ELU (an activation function), SELU (an activation function), Sigmoid (an activation function), tanh (an activation function), and the like.

Pooling may refer to an operation method that reduces high-dimensional features into low-dimensional features and removes redundant features.

Wherein pooling may be at least one of a maximize pooling algorithm and an average pooling algorithm.

In the specific implementation, after the multi-lead electrocardiosignals are split to obtain a plurality of groups of independent lead electrocardiosignals, the independent lead electrocardiosignals are respectively input into corresponding convolution blocks, the independent lead electrocardiosignals are used as the input of the convolution blocks, the convolution blocks perform convolution and output, and the output data is used as the convolution characteristics of the independent lead electrocardiosignals of the corresponding groups. For example, the above-mentioned packet convolution network has Z signal input channels, and inputs the above-mentioned Z groups of independent lead electrocardiographic signals into corresponding signal input channels, so as to realize that the Z groups of independent lead electrocardiographic signals are input into corresponding convolution blocks, and obtain the convolution characteristics of each group. In practical application, the convolution block can be used for performing one-dimensional convolution on an input signal, performing batch normalization on the one-dimensional convolved features, and finally activating through an activation function, so that the representation of the features has nonlinearity and is no longer only 0 or 1 output, and the expression capacity of the model is improved. And pooling the output features, namely reducing the dimension of the high-dimensional features into low-dimensional features and removing redundant features to finally obtain the convolution features. In the above packet convolutional neural network, there may be one or more convolutional blocks, and those skilled in the art may design the number of convolutional blocks according to actual needs.

It should be noted that the processing process of each set of independent leads is similar to the above-mentioned embodiment, and is not described herein again.

The feature grouping module 130 is configured to group the convolution features of each group of independent lead electrocardiographic signals to obtain grouping features of the group of independent lead electrocardiographic signals, and input the grouping features to the grouping convolution block to obtain a plurality of grouping convolution features of the group of independent lead electrocardiographic signals.

The grouping feature may refer to a feature that performs feature grouping on the convolution feature.

The grouping convolution block may be a convolution block that performs a convolution operation on the grouping feature. In practical applications, the above-mentioned packet convolution block may be a set for performing a series of operations such as one-dimensional convolution and pooling on the input packet features.

In the specific implementation, the convolution characteristics of each group of independent lead electrocardiosignals output by the convolution block are equally grouped according to the set grouping number to obtain the grouping characteristics of the group of independent lead electrocardiosignals; meanwhile, the obtained grouping features are input into the grouping convolution block, the grouping features are used as input of the grouping convolution block, the grouping convolution block performs convolution and outputs, and the output data is used as a plurality of grouping convolution features of the independent lead electrocardiosignals.

In practical application, one-dimensional convolution can be performed on input grouping features through the grouping convolution block, pooling is performed on the features after one-dimensional convolution, the dimension of high-dimensional features is reduced to low-dimensional features, redundant features are removed, and finally the grouping convolution features are obtained.

For example, given that the preset grouping number is G, the convolution features have M feature planes, and the convolution features are grouped to obtain grouping features with the group number being G; wherein, each group of grouping features has M/G feature planes.

It should be noted that the convolution feature processing process corresponding to each individual lead electrocardiographic signal is similar to that in the above-mentioned embodiment, and is not described herein again.

The characteristic combination module 140 is configured to combine a plurality of grouped convolution characteristics of each group of independent lead electrocardiographic signals to obtain electrocardiographic signal combination characteristics of the group of independent lead electrocardiographic signals.

The electrocardiosignal combination feature may be a feature obtained by combining the above-mentioned respective grouped convolution features.

In the specific implementation, after obtaining a plurality of grouped convolution characteristics of each group of independent lead electrocardiosignals, the grouped convolution characteristics of each group of independent lead electrocardiosignals are combined, so as to obtain the electrocardiosignal combination characteristics of the group of independent lead electrocardiosignals.

For example, given that Z groups of independent lead electrocardiographic signals correspond to Z groups of packet convolution features, and feature dimensions of each group of packet convolution features are a × B, the above-mentioned group convolution features are subjected to feature combination to obtain electrocardiographic signal combination features with a group number of 1. Wherein, the characteristic dimension of the connected electrocardiosignal combination characteristic is (ZA) B.

And the full-connection processing module 150 is used for performing full-connection processing on the electrocardiosignal combination characteristics of each group of independent lead electrocardiosignals to obtain the myocardial infarction anomaly probability.

Wherein, the fully connected processing may refer to processing using a fully connected neural network classifier.

In the specific implementation, after a plurality of grouped convolution characteristics of each group of independent lead electrocardiosignals are combined to obtain electrocardiosignal combination characteristics, the electrocardiosignal combination characteristics are input into a fully-connected neural network classifier, and the fully-connected neural network classifier is used for performing fully-connected processing on the electrocardiosignal combination characteristics to obtain the myocardial infarction anomaly probability.

And the judging module 160 is used for judging the multi-lead electrocardiosignals to be myocardial infarction signals according to the myocardial infarction abnormal probability.

In a specific implementation, the number of input cells of the fully-connected neural network classifier is equal to the number of feature vectors of the electrocardiosignal combination features, and the number of output cells of the fully-connected neural network classifier is 2, so that two prediction results are represented. When the obtained myocardial infarction abnormal probability is higher than a preset abnormal probability threshold value, the predicted value output by the fully-connected neural network classifier is 1, and the myocardial infarction related abnormal performance of the myocardial infarction sample is represented; when the obtained myocardial infarction abnormal probability is lower than a preset abnormal probability threshold value, the predicted value output by the fully-connected neural network classifier is 0, and the heart beat sample is healthy.

Before the above-mentioned multi-lead electrocardiographic signals are subjected to feature recognition using the above-mentioned grouped convolutional neural network, it is necessary to train the above-mentioned grouped convolutional neural network using various multi-lead electrocardiographic signals having abnormal signals and known myocardial infarction types and normal multi-lead electrocardiographic signals as training samples, and to optimize the above-mentioned myocardial infarction recognition neural network.

In practice, training and testing may be performed via public databases such as PTB. More specifically, the myocardial infarction patient and non-myocardial infarction patient data sets can be randomly divided into a training set and a testing set according to a proportion, and the two data sets do not contain the data of the same person at the same time. The structured multi-lead electrocardiosignal is marked as X, and the abnormity of the characteristic change related to the myocardial infarction are marked as the output Y of the grouped convolutional neural network. The (X, Y) of the training set collectively comprise the training samples of the multi-lead multi-structure aggregation network. Inputting X into the packet convolutional neural network according to a certain batch size, obtaining a predicted value Pred _ Y of Y through forward propagation, calculating Y and Pred _ Y losses through a loss function, propagating the losses reversely, and training the network by using a gradient descent method to obtain an optimal packet convolutional neural network.

To facilitate a thorough understanding of the embodiments of the present application by those skilled in the art, the following description will be given with reference to a specific example.

Fig. 2 is a schematic diagram of a neural network structure of an electrocardiographic signal recognition apparatus based on a packet convolutional network according to an embodiment. As shown in fig. 2, firstly, obtaining a multi-lead electrocardiographic signal, and splitting the multi-lead electrocardiographic signal to obtain a plurality of groups of independent-lead electrocardiographic signals; then, inputting each group of independent lead electrocardiosignals into a convolution block respectively to obtain convolution characteristics of the independent lead electrocardiosignals of the corresponding group; then, grouping the convolution characteristics of each group of independent lead electrocardiosignals to obtain the grouping characteristics of the group of independent lead electrocardiosignals, inputting the grouping characteristics into a grouping convolution block as the input of the grouping convolution block to obtain a plurality of grouping convolution characteristics of the group of independent lead electrocardiosignals; then, grouping convolution characteristics output by a plurality of grouping convolution blocks of each group of independent lead electrocardiosignals are combined to obtain the electrocardiosignal combination characteristics of the group of independent lead electrocardiosignals, and full connection processing is carried out on the electrocardiosignal combination characteristics of each group of independent lead electrocardiosignals to obtain myocardial infarction abnormal probability; and finally, judging the multi-lead electrocardiosignals as myocardial infarction signals according to the myocardial infarction abnormal probability.

In the electrocardiosignal identification device based on the grouping convolution neural network, the independent lead electrocardiosignals are obtained by acquiring the multi-lead electrocardiosignals and splitting the multi-lead electrocardiosignals; then, the independent lead electrocardiosignal is input into a convolution block to obtain convolution characteristics; then, grouping the convolution characteristics to obtain grouping characteristics, and inputting the grouping characteristics into a grouping convolution block to obtain grouping convolution characteristics; then, determining the myocardial infarction abnormal probability according to the grouping convolution characteristics; finally, according to the myocardial infarction abnormal probability, judging the multi-lead electrocardiosignals as myocardial infarction signals; therefore, the parameter quantity of convolution operation can be reduced, under the condition of not reducing the identification performance of the grouped convolution neural network, the overfitting of the grouped convolution neural network is reduced, and whether the multi-lead electrocardiosignals are myocardial infarction signals or not can be identified and judged more accurately; therefore, when the electrocardiosignals are identified, the accurate positioning of Q waves, P waves, J points, S points and T waves of key points of the electrocardiosignals is not needed, and even under the condition that the quality of the electrocardiosignals is poor and the key points of the electrocardiosignals cannot be accurately positioned, the grouped convolutional neural network is input, so that the myocardial infarction risk can be more accurately identified from the electrocardiosignals.

Example two

Fig. 3 is a schematic structural diagram of an electrocardiographic signal recognition apparatus based on a packet convolutional neural network according to a second embodiment of the present application. Specifically, referring to fig. 3, the electrocardiosignal identification method based on the packet convolutional neural network according to the second embodiment of the present application specifically includes:

the signal obtaining module 310 is configured to obtain multiple lead electrocardiographic signals, and split the multiple lead electrocardiographic signals to obtain multiple sets of independent lead electrocardiographic signals.

Alternatively, the number of sets of individual lead cardiac signals may be Z. More specifically, Z ═ 12. The 12 groups of independent lead electrocardiosignals can be V1 lead electrocardiosignals, V2 lead electrocardiosignals, V3 lead electrocardiosignals, V4 lead electrocardiosignals, V5 lead electrocardiosignals, V6 lead electrocardiosignals, aVF lead electrocardiosignals, aVR lead electrocardiosignals, aVL lead electrocardiosignals, I lead electrocardiosignals, II lead electrocardiosignals and III lead electrocardiosignals.

Optionally, the signal obtaining module 310 includes:

the original signal receiving submodule is used for receiving an original signal; the wavelet decomposition sub-module is used for performing wavelet decomposition on the original signal to obtain a wavelet decomposition signal; the wavelet decomposition signal has dimension X1; the zero setting sub-module is used for setting the zero of the X2 dimensional signal in the wavelet decomposition signal to obtain a partial zero setting signal; wherein X2 is less than X1; the inverse transformation submodule is used for carrying out wavelet inverse transformation on the partial zero-set signals to obtain de-noised signals; the de-noising signal is a signal after high-frequency noise and baseline drift are removed; and the multi-lead acquisition sub-module is used for obtaining the multi-lead electrocardiosignals according to the de-noising signals.

The original signal can be an original signal acquired by a multi-lead electrocardiogram system.

The wavelet decomposition signal may be a signal obtained by performing wavelet decomposition on an original signal.

Wherein the partial nulling signal may be a signal in which a signal of a partial dimension is nulled. After wavelet decomposition, a wavelet decomposition signal of X1 dimension can be decomposed, and a signal of X2 dimension is set to zero to obtain a partially set zero signal.

In a specific implementation, the original signal may be resampled to a signal with a certain frequency, for example, to a signal of 1000 Hz.

Then, the resampled signal is subjected to an X1-dimensional wavelet decomposition using a wavelet basis function of a certain db (power gain unit) to obtain an X1-dimensional wavelet decomposed signal. For example, the wavelet decomposition may be performed with a wavelet basis function of preferably 6 db.

And zeroing the wavelet decomposition signal of the X2 dimension in the X1 dimension to obtain a partial zeroing signal. For example, when X1 is 10, X2 may be 3, and specifically, the zero setting may be performed on the 0 th, 9 th, and 10 th dimensional wavelet decomposition signals.

After the partial nulling signal is obtained, the partial nulling signal can be converted in a wavelet inverse transformation mode to obtain a signal which is used as a de-noising signal, the de-noising signal removes high-frequency noise and baseline drift, and finally the multi-lead electrocardiosignal can be obtained based on the de-noising signal.

FIG. 4A is a schematic representation of an original cardiac signal, under an embodiment. FIG. 4B is a diagram of a denoised signal according to one embodiment. As shown in the figure, the X axis and the Y axis respectively represent the acquisition time point (s, s) and the signal intensity (mV, millivolt) of the signal, and the comparison of the original electrocardiosignal and the de-noised signal is visible, so that the signal baseline of the de-noised signal tends to be flat, and the extraction and the detection of subsequent characteristics are facilitated.

According to the technical scheme of the embodiment of the application, the denoising signal for removing the high-frequency noise and the baseline wander is obtained by preprocessing means such as wavelet decomposition, signal nulling of partial dimensionality, inverse wavelet transformation and the like, the multi-lead electrocardiosignal is obtained based on the denoising signal, the interference of the high-frequency noise and the baseline wander can be avoided, the multi-lead electrocardiosignal with better signal quality is obtained, and the accuracy of electrocardiosignal identification is improved.

Optionally, the multi-lead obtaining sub-module is configured to, in particular, determine an R-wave position of the denoised signal; determining the first M1 positions of the R-wave position and the last M2 positions of the R-wave position; and forming a structured signal matrix as the multi-lead electrocardiosignals by adopting the de-noised signals on the R wave position, the front M1 positions and the rear M2 positions.

Wherein, the R-wave position may be a position where a maximum value of the R-wave appears in the signal.

The structured signal matrix may be a matrix formed by arranging values representing signals.

In a specific implementation, the R-wave position of each de-noised signal can be detected through a modified Pan-Tompkins (an algorithm for detecting QRS complexes) algorithm. The Pan-Tompkins algorithm may specifically include operations such as low-pass filtering, high-pass filtering, differentiation, squaring, integration, adaptive thresholding, and searching.

Then, with each R-wave position as a reference, determining the first M1 positions and the last M2 positions of the R-wave position, and using the denoising signals at the R-wave position, the first M1 positions and the last M2 positions to form signal data corresponding to one heart beat, which is composed of (M1+ M2+1) denoising signals, and for the same patient, acquiring the signal data of N heart beats and forming a structured signal matrix.

The matrix structure may be N × L (M1+ M2+1), where L represents the number of leads, and the specific values of M1 and M2 may be set according to actual needs.

And the convolution module 320 is configured to input each group of independent lead electrocardiographic signals to the convolution block, so as to obtain convolution characteristics of the corresponding group of independent lead electrocardiographic signals.

In the specific implementation, after the multi-lead electrocardiosignals are split to obtain a plurality of groups of independent lead electrocardiosignals, the independent lead electrocardiosignals are respectively input into corresponding convolution blocks, the independent lead electrocardiosignals are used as the input of the convolution blocks, the convolution blocks perform convolution and output, and the output data is used as the convolution characteristics of the independent lead electrocardiosignals of the corresponding groups.

Optionally, the convolution block of the block convolutional neural network includes a first convolution block and a second convolution block, and the convolution module 320 includes:

the first convolution block may be a set used for performing a series of operations such as one-dimensional convolution, batch normalization, activation, and pooling on the input features.

The first convolution submodule is used for performing convolution and pooling on the independent lead electrocardiosignals through the first convolution block to obtain the output characteristics of the convolution block; and the second convolution submodule is used for performing convolution and pooling on the output bits of the convolution block through the second convolution block to obtain convolution characteristics.

In the specific implementation, after the multi-lead electrocardiosignals are split to obtain independent lead electrocardiosignals, the independent lead electrocardiosignals are respectively input into corresponding first convolution blocks, the independent lead electrocardiosignals are used as the input of the first convolution blocks, the first convolution blocks perform one-dimensional convolution on the input independent lead electrocardiosignals, the features after the one-dimensional convolution are subjected to batch normalization, finally, the features are activated through an activation function, and the output features are pooled, for example, maximized pooling or average pooling is performed, so that the high-dimensional features are reduced into low-dimensional features and redundant features are removed, and finally, the output features of the convolution blocks are obtained.

Then, the output characteristics of the convolution blocks are input into corresponding second convolution blocks, the output characteristics of the convolution blocks are used as the input of the first convolution blocks, the second convolution blocks carry out one-dimensional convolution on the input output characteristics of the convolution blocks, the characteristics after one-dimensional convolution are subjected to batch normalization, finally activation is carried out through an activation function, pooling is carried out on the output characteristics, for example, maximization pooling or average pooling is carried out, the high-dimensional characteristics are reduced into the low-dimensional characteristics, redundant characteristics are removed, and finally convolution characteristics are obtained. Wherein, the activation function can be at least one of ReLU, ELU, SELU, Sigmoid and tanh.

It should be noted that, when the convolution block includes 2 or more convolution blocks, the processing procedure is similar to the above embodiment, and is not described herein again.

Optionally, the first convolution block includes a first convolution layer and a first maximum pooling layer, and the first convolution sub-module is specifically configured to: performing convolution, batch normalization and activation on the independent lead electrocardiosignals through the first convolution layer to obtain output characteristics of the first convolution layer; and performing maximum pooling on the output characteristic of the first convolution layer through the first maximum pooling layer to obtain the output characteristic of the convolution block.

Wherein Max pooling layer (Max Pool) may be an operation for pooling the maximum of the features of the input. According to the role of the operation in the neural network, the maximum pooling layer is named.

In the specific implementation, the independent lead electrocardiosignals are input to a first convolution layer to be used as input of the first convolution layer, then one-dimensional convolution is carried out on the input independent lead electrocardiosignals through the first convolution layer, batch normalization is carried out on the features after one-dimensional convolution, and finally activation is carried out through an activation function to obtain output features of the first convolution layer.

And then, inputting the first convolution layer output characteristic to a first maximum pooling layer as the input of the first maximum pooling layer, and then performing maximum pooling on the first convolution layer output characteristic through the first maximum pooling layer to reduce the high-dimensional characteristic into the low-dimensional characteristic and remove the redundant characteristic, thereby finally obtaining the convolution block output characteristic. Wherein, the activation function can be at least one of ReLU, ELU, SELU, Sigmoid and tanh.

In practical application, the characteristic dimension of the individual lead electrocardiograph signal may be 1 × 600, and the structural parameters of the first convolution layer may be: the convolution kernel size k is 61, the sliding step s is 1, the supplemental element p is 0, the number f of output feature planes is 12, and the feature dimension of the first convolution layer output feature of the first convolution layer output may be 12 × 540. And the sliding step s of the first maximum pooling layer is 3 the feature dimension of the convolution block output feature of the first maximum pooling layer output may be 12 x 180.

According to the technical scheme of the embodiment of the application, the convolution, batch normalization and activation are carried out on the independent lead electrocardiosignals by using the first convolution layer, so that the output characteristics of the first convolution layer are obtained; and the maximum pooling is carried out on the output characteristics of the first convolution layer through the first maximum pooling layer, so that the characteristic dimensionality of the independent lead electrocardiosignal is reduced to obtain the output characteristics of a convolution block, the parameter processing amount of subsequent myocardial infarction signal identification is reduced, and the identification efficiency of the myocardial infarction signal is improved.

Optionally, the second convolution block includes a second convolution layer and a second maximum pooling layer, and the second convolution sub-module is specifically configured to: performing convolution, batch normalization and activation on the output characteristics of the convolution blocks through a second convolution layer to obtain second convolution layer output characteristics; and performing maximum pooling on the output characteristics of the second convolution layer through the second maximum pooling layer to obtain convolution characteristics.

In the specific implementation, the output characteristics of the convolution block are input into the second convolution layer and used as the input of the second convolution layer, then the one-dimensional convolution is carried out on the input output characteristics of the convolution block through the second convolution layer, the characteristics after the one-dimensional convolution are subjected to batch normalization, and finally activation is carried out through an activation function, so that the output characteristics of the second convolution layer are obtained.

And then, inputting the second convolution layer output characteristics to a second maximum pooling layer as input of the second maximum pooling layer, and performing maximum pooling on the second convolution layer output characteristics through the second maximum pooling layer to reduce the high-dimensional characteristics into low-dimensional characteristics and remove redundant characteristics, thereby finally obtaining the convolution characteristics. Wherein, the activation function can be at least one of ReLU, ELU, SELU, Sigmoid and tanh.

In practical applications, the feature dimension of the convolution block output feature may be 12 × 180. The structural parameters of the second convolution layer may be: the convolution kernel size k is 31, the sliding step s is 1, the supplemental element p is 0, the number f of output feature planes is 16, and the feature dimension of the second convolution layer output feature of the second convolution layer output may be 16 × 150. And the second largest pooling layer has a sliding step s of 3. the feature dimension of the convolved features output by the second largest pooling layer may be 16 x 50.

According to the technical scheme of the embodiment of the application, after convolution operation of the first convolution layer is carried out, convolution, batch normalization and activation are carried out on the output characteristics of the convolution block through the second convolution layer, and the output characteristics of the second convolution layer are obtained; performing maximum pooling on the output characteristics of the second convolution layer through a second maximum pooling layer to obtain convolution characteristics; therefore, the characteristic dimension of the output characteristic of the convolution block is further reduced, the parameter processing amount of the subsequent myocardial infarction signal identification process is further reduced, and the myocardial infarction identification speed is improved; meanwhile, convolution operation is carried out on the independent lead electrocardiosignals by using the first convolution layer and the first convolution layer, so that feature loss caused by overlarge dimensionality reduction gradient in the process of reducing the feature dimensionality can be avoided, and the accuracy of myocardial infarction identification is improved.

The feature grouping module 330 is configured to group the convolution features of each group of independent lead electrocardiographic signals to obtain grouping features of the group of independent lead electrocardiographic signals, and input the grouping features to the grouping convolution block to obtain a plurality of grouping convolution features of the group of independent lead electrocardiographic signals.

Optionally, the feature grouping module 330 includes: the group number obtaining submodule is used for obtaining the group number of the packet convolution blocks; and the grouping submodule is used for uniformly grouping the convolution characteristics of each group of independent lead electrocardiosignals according to the grouping convolution group number to obtain the grouping characteristics of the group of independent lead electrocardiosignals.

In specific implementation, before grouping the convolution characteristics, firstly, the number of grouped convolution groups of a grouped convolution block needs to be acquired; then, the convolution characteristics are equally grouped according to the number of the group convolution groups to obtain grouping characteristics. For example, if the number of grouped convolution groups is known to be 4, and the convolution features have 16 feature planes, the convolution features are divided into four groups of grouped features; wherein the grouping feature has 4 feature planes.

In practical application, if the number of the grouped convolution groups of the grouped convolution blocks is g, the characteristic surface of the input characteristic is equally divided into g groups, and meanwhile, the characteristic surface of the output characteristic is equally divided into g groups, wherein the output ith group of characteristics are obtained by carrying out convolution calculation on the output ith group of characteristics; for example, if the number of grouped convolution groups is known to be 4, and the convolution features have 16 feature planes, the convolution features are divided into four groups of grouped features; wherein the grouping feature has 4 feature planes. Then, inputting the four groups of grouping features into a grouping convolution block, wherein the finally output grouping convolution features have 24 feature surfaces which are divided into four groups, each group has 6 feature surfaces, the feature surface of the 1 st group is obtained by convolution calculation of the first group of convolution groups in the grouping convolution block, and the rest 3 groups are analogized, and the description is omitted here.

Optionally, the block convolution block includes a block convolution layer and a block max-pooling layer, and the feature grouping module 330 includes: the grouping convolution characteristic output submodule is used for carrying out convolution, batch normalization and activation on the grouping characteristics through the grouping convolution layer to obtain the grouping convolution layer output characteristics; and the pooling sub-module is used for performing maximal pooling on the output characteristics of the grouping convolution layer through a grouping maximal pooling layer to obtain a plurality of grouping convolution characteristics of the group of independent lead electrocardiosignals.

In the specific implementation, the grouping features are input into the grouping convolution layer and used as the input of the grouping convolution layer, then the input grouping features are subjected to one-dimensional convolution through the grouping convolution layer, the features after the one-dimensional convolution are subjected to batch normalization, and finally activation is carried out through an activation function, so that the output features of the grouping convolution layer are obtained.

Then, the output characteristics of the grouped convolutional layer are input into a grouped maximum pooling layer as the input of the grouped maximum pooling layer, then the maximum pooling is carried out on the output characteristics of the grouped convolutional layer through the grouped maximum pooling layer, the high-dimensional characteristics are reduced into low-dimensional characteristics, redundant characteristics are removed, and finally the grouped convolutional characteristics are obtained. Wherein, the activation function can be at least one of ReLU, ELU, SELU, Sigmoid and tanh.

In practical applications, the feature dimension of the convolution block output feature of the first largest pooling layer output may be 16 × 50, and the feature dimension of the grouping feature may be 4 × 50. The structure parameters of the packet convolutional layer may be: the convolution kernel size k is 9, the sliding step length s is 1, the supplementary element p is 0, the number f of output characteristic surfaces is 24, and the number g of grouped convolution groups is 4; the feature dimension of the packet convolutional layer output feature of the packet convolutional layer output may be 24 × 42. And the sliding step s of the packet max pooling layer is 2, the characteristic dimension of the packet convolution feature of the packet max pooling layer output can be 24 × 14.

According to the technical scheme of the embodiment of the application, the convolution, batch normalization and activation are carried out on the grouping characteristics through the grouping convolution layer, and the output characteristics of the grouping convolution layer are obtained; performing maximum pooling on the output characteristics of the grouping convolution layer through a grouping maximum pooling layer to obtain a plurality of grouping convolution characteristics of the independent lead electrocardiosignals; therefore, the sub-group convolution block has less parameter processing amount in the convolution operation process, so that the processing speed of the myocardial infarction signal identification device is increased, and the myocardial infarction identification efficiency is improved.

To facilitate understanding of those skilled in the art, table 1 provides a network structure parameter table of an electrocardiographic signal identification method based on a packet convolutional neural network.

Table 1 network structure parameter table

Wherein k is the convolution kernel size, s is the sliding step length, p is the supplementary element, f is the number of output feature surfaces, and g is the number of grouped convolution groups.

Specifically, the characteristic dimension of the independent lead electrocardiographic signal, i.e., the input characteristic, may be 1 × 600, and the structural parameters of the first convolution layer may be: the convolution kernel size k is 61, the sliding step s is 1, the supplemental element p is 0, the number f of output feature planes is 12, and the feature dimension of the first convolution layer output feature of the first convolution layer output may be 12 × 540. The first convolution layer output feature is taken as the input of the first maximum pooling layer, the sliding step s of which is 3.

The feature dimensions of the convolution block output features are then taken as input to the second convolution layer. The structural parameters of the second convolution layer may be: the convolution kernel size k is 31, the sliding step s is 1, the supplemental element p is 0, the number f of output feature planes is 16, and the feature dimension of the second convolution layer output feature of the second convolution layer output may be 16 × 150. The second convolutional layer output characteristic is taken as an input of the second max-pooling layer. And the sliding step s of the second largest pooling layer is 3. the feature dimension of the convolved block output feature of the second largest pooling layer output may be 16 x 50.

And finally, taking the characteristic dimension of the output characteristic of the convolution block as the input of the grouping convolution layer. The feature dimension of the grouping feature may be 4 x 50. The structure parameters of the packet convolutional layer may be: the convolution kernel size k is 9, the sliding step length s is 1, the supplementary element p is 0, the number f of output characteristic surfaces is 24, and the number g of grouped convolution groups is 4; the feature dimension of the packet convolutional layer output feature of the packet convolutional layer output may be 24 × 42. The characteristic dimension of the convolution block output characteristic output by the grouping maximum pooling layer can be 24 × 14 by taking the grouping convolution layer output characteristic as the input of the grouping maximum pooling layer, wherein the sliding step s of the grouping maximum pooling layer is 2.

The characteristic combination module 340 is configured to combine a plurality of grouped convolution characteristics of each group of independent lead electrocardiographic signals to obtain electrocardiographic signal combination characteristics of the group of independent lead electrocardiographic signals.

And the full-connection processing module 350 is configured to perform full-connection processing on the electrocardiograph signal combination characteristics of each group of independent lead electrocardiograph signals to obtain the myocardial infarction abnormality probability.

And the myocardial infarction judging module 360 is used for judging the multi-lead electrocardiosignals as myocardial infarction signals when the myocardial infarction abnormal probability is higher than a preset abnormal probability threshold value.

In a specific implementation, the number of input cells of the fully-connected neural network classifier is equal to the number of feature vectors of the parietal region identification features at the same position, and the number of output cells of the fully-connected neural network classifier is 2, so that two prediction results are represented. When the obtained myocardial infarction abnormal probability is higher than a preset abnormal probability threshold value, the predicted value output by the fully-connected neural network classifier is 1, and the myocardial infarction related abnormal performance of the myocardial infarction sample is represented; when the obtained myocardial infarction abnormal probability is lower than a preset abnormal probability threshold value, the predicted value output by the fully-connected neural network classifier is 0, and the heart beat sample is healthy.

All or part of the modules in the electrocardiosignal identification device based on the packet convolutional neural network can be realized by software, hardware and a combination thereof. The modules can be embedded in a hardware form or independent from a processor in the computer device, and can also be stored in a memory in the computer device in a software form, so that the processor can call and execute operations corresponding to the modules.

FIG. 5 is a schematic diagram of a network structure of an electrocardiosignal recognition device based on a packet convolutional neural network; as shown in fig. 5, including a first convolution block, a second convolution block, and a grouping convolution block; wherein the first volume block comprises a first volume layer and a first max pooling layer; the second convolution block includes a second convolution layer and a second maximum pooling layer; the packet convolution block comprises a packet convolution layer and a packet maximum pooling layer; firstly, carrying out convolution, batch normalization and activation on the independent lead electrocardiosignals through a first convolution layer to obtain output characteristics of the first convolution layer; and performing maximum pooling on the output characteristics of the first convolution layer through the first maximum pooling layer to obtain the output characteristics of the convolution block.

Then, performing convolution, batch normalization and activation on the output characteristics of the convolution block through a second convolution layer to obtain second convolution layer output characteristics; and performing maximum pooling on the output characteristics of the second convolution layer by using a second maximum pooling layer to obtain convolution characteristics.

Then, carrying out convolution, batch normalization and activation on the grouping characteristics through the grouping convolution layer to obtain the output characteristics of the grouping convolution layer; performing maximum pooling on the output characteristics of the packet convolution layer through a packet maximum pooling layer to obtain packet convolution characteristics;

finally, obtaining the electrocardiosignal combination characteristics by the convolution characteristics of the combination groups, and carrying out full-connection processing on the electrocardiosignal combination characteristics to obtain the myocardial infarction abnormity probability; wherein, a fully connected neural network classifier is used for outputting an electrocardiosignal identification result; the number of output cells of the fully-connected neural network classifier is 2, and thus represents two prediction results. When the obtained myocardial infarction abnormal probability is higher than a preset abnormal probability threshold value, the predicted value output by the fully-connected neural network classifier is 1, and the myocardial infarction related abnormal performance of the myocardial infarction sample is represented; when the obtained myocardial infarction abnormal probability is lower than a preset abnormal probability threshold value, the predicted value output by the fully-connected neural network classifier is 0, and the heart beat sample is healthy.

EXAMPLE III

Fig. 11 is a structural diagram of a training apparatus for a packet convolutional neural network according to a third embodiment of the present application. Specifically, referring to fig. 11, a training method of a packet convolutional neural network according to a third embodiment of the present application specifically includes:

a training sample obtaining module 1110, configured to obtain an electrocardiographic signal training sample for the packet convolutional neural network.

In specific implementation, an electrocardiosignal training sample for a packet convolutional neural network is obtained through a public database such as a PTB (packet transport bus). The electrocardiosignal training sample comprises a multi-lead electrocardiosignal with an abnormal signal and a known myocardial infarction type and a normal multi-lead electrocardiosignal.

The machine training module 1120 is used for performing machine training on the packet convolutional neural network by adopting the electrocardiosignal training sample to obtain a trained packet convolutional neural network; the trained packet convolution neural network comprises a pre-trained convolution block and a packet convolution block; after training, grouping a convolutional neural network, and inputting each group of independent lead electrocardiosignals to a convolution block respectively to obtain convolution characteristics of the corresponding group of independent lead electrocardiosignals; the independent lead electrocardiosignals are obtained by splitting multi-lead electrocardiosignals; the system is also used for grouping the convolution characteristics of each group of independent lead electrocardiosignals to obtain the grouping characteristics of the group of independent lead electrocardiosignals, and inputting the grouping characteristics into a grouping convolution block to obtain a plurality of grouping convolution characteristics of the group of independent lead electrocardiosignals; the group convolution characteristics of each group of independent lead electrocardiosignals are combined to obtain the electrocardiosignal combination characteristics of the group of independent lead electrocardiosignals; the system is also used for carrying out full-connection processing on the electrocardiosignal combination characteristics of each group of independent lead electrocardiosignals to obtain the myocardial infarction abnormal probability; and the method is also used for judging the multi-lead electrocardiosignals to be myocardial infarction signals according to the myocardial infarction abnormal probability.

In the concrete implementation, after an electrocardiosignal training sample is obtained, machine training is carried out on the packet convolution neural network, more specifically, data sets of a myocardial infarction patient and a non-myocardial infarction patient can be randomly divided into a training set and a testing set according to a proportion, and the two data sets do not contain the same person data at the same time. The structured multi-lead electrocardiosignal is marked as X, and the abnormity of the characteristic change related to the myocardial infarction are marked as the output Y of the grouped convolutional neural network. The (X, Y) of the training set collectively comprise the training samples of the multi-lead multi-structure aggregation network. Inputting X into the packet convolutional neural network according to a certain batch size in batches, obtaining a predicted value Pred _ Y of Y through forward propagation, calculating Y and Pred _ Y losses through a loss function, propagating the losses in a reverse direction, training the network by using a gradient descent method, and obtaining an optimal packet convolutional neural network, namely the packet convolutional neural network after training.

The trained packet convolution neural network comprises a pre-trained convolution block and a packet convolution block; in the process of identifying the electrocardiosignals by using the grouping convolution neural network, firstly, the multi-lead electrocardiosignals can be obtained, and the multi-lead electrocardiosignals are split to obtain a plurality of groups of independent lead electrocardiosignals; then, inputting each group of independent lead electrocardiosignals into a convolution block respectively to obtain convolution characteristics of the independent lead electrocardiosignals of the corresponding group; then, grouping the convolution characteristics of each group of independent lead electrocardiosignals to obtain the grouping characteristics of the group of independent lead electrocardiosignals, inputting the grouping characteristics into a grouping convolution block as the input of the grouping convolution block to obtain a plurality of grouping convolution characteristics of the group of independent lead electrocardiosignals; then, grouping convolution characteristics output by a plurality of grouping convolution blocks of each group of independent lead electrocardiosignals are combined to obtain the electrocardiosignal combination characteristics of the group of independent lead electrocardiosignals, and full connection processing is carried out on the electrocardiosignal combination characteristics of each group of independent lead electrocardiosignals to obtain myocardial infarction abnormal probability; and finally, judging the multi-lead electrocardiosignals as myocardial infarction signals according to the myocardial infarction abnormal probability.

According to the technical scheme provided by the embodiment of the application, the grouped convolutional neural network is subjected to machine training by adopting an electrocardiosignal training sample aiming at the grouped convolutional neural network, so that the trained grouped convolutional neural network is obtained; the trained packet convolution neural network comprises a pre-trained convolution block and a packet convolution block; obtaining independent lead electrocardiosignals by obtaining the multi-lead electrocardiosignals and splitting the multi-lead electrocardiosignals; then, the independent lead electrocardiosignal is input into a convolution block to obtain convolution characteristics; then, grouping the convolution characteristics to obtain grouping characteristics, and inputting the grouping characteristics into a grouping convolution block to obtain grouping convolution characteristics; then, determining the myocardial infarction abnormal probability according to the grouping convolution characteristics; finally, according to the myocardial infarction abnormal probability, judging the multi-lead electrocardiosignals as myocardial infarction signals; therefore, the parameter quantity of convolution operation can be reduced, under the condition that the identification performance of the grouped convolution neural network is not reduced, the overfitting of the grouped convolution neural network is reduced, and whether the multi-lead electrocardiosignals are myocardial infarction signals or not can be identified and judged more accurately.

Furthermore, when the electrocardiosignals are identified, the accurate positioning of key points Q wave, P wave, J point, S point and T wave of the electrocardiosignals is not needed, and even under the condition that the quality of the electrocardiosignals is poor and the key points of the electrocardiosignals cannot be accurately positioned, the myocardial infarction risk can be more accurately identified by inputting the grouped convolutional neural network.

The modules in the training device of the packet convolutional neural network can be wholly or partially implemented by software, hardware and a combination thereof. The modules can be embedded in a hardware form or independent from a processor in the computer device, and can also be stored in a memory in the computer device in a software form, so that the processor can call and execute operations corresponding to the modules.

FIG. 6 is a schematic diagram of characteristic dimension variation of electrocardiosignal identification based on a packet convolutional neural network. As shown in fig. 6, firstly, a plurality of groups of independent lead electrocardiographic signals are input into a first convolution block through a plurality of independent channels as input features, wherein the feature dimension of the independent lead electrocardiographic signals may be 1 × 600, and then, convolution pooling is performed on the independent lead electrocardiographic signals through the first convolution block to obtain output features of the convolution block; wherein the feature dimension of the convolution block output feature may be 12 x 180; then, the output characteristic of the convolution block is used as the input of a second convolution block, and the convolution characteristic is obtained through the convolution pooling of the second convolution block; wherein the feature dimension of the convolution feature may be 16 x 50; then, if the number of the grouped convolution groups is known to be 4 and the convolution features have 16 feature surfaces, dividing the convolution features into four groups of grouped features; wherein the grouping feature has 4 feature planes. Then, inputting the four groups of grouping features into a grouping convolution block for grouping convolution pooling, and finally outputting grouping convolution features; wherein the feature dimension of the grouped convolution features may be 24 x 14; specifically, the grouped convolution features are divided into four groups of 6 feature planes each. Because 12 groups of independent lead electrocardiosignals correspond to 12 groups of convolution characteristics, and the feature dimension of each group of convolution characteristics is 24 × 14, the group number of the electrocardiosignal combination characteristics obtained by combining the characteristics of the group convolution characteristics is 1. Wherein, the characteristic dimension of the connected electrocardiosignal combination characteristic is 268 × 12. And finally, inputting the electrocardiosignal combination characteristics into a fully-connected neural network classifier, wherein the number of output cells of the fully-connected neural network classifier is 2, and then representing two prediction results. When the obtained myocardial infarction abnormal probability is higher than a preset abnormal probability threshold value, the predicted value output by the fully-connected neural network classifier is 1, and the myocardial infarction related abnormal performance of the myocardial infarction sample is represented; when the obtained myocardial infarction abnormal probability is lower than a preset abnormal probability threshold value, the predicted value output by the fully-connected neural network classifier is 0, and the heart beat sample is healthy.

FIG. 7 is a flow chart of cardiac signal identification based on neural networks. As shown in the figure, firstly, a multi-lead electrocardiosignal of a patient is collected through a multi-lead electrocardio system, the multi-lead electrocardiosignal is stored, then the multi-lead electrocardiosignal is preprocessed through wavelet decomposition, signal zeroing of partial dimensionality and the like, and the signal is subjected to structuring processing to obtain a structured signal matrix which is used as the input of a multi-lead multi-structure polymerization network. The multi-lead multi-structure aggregation network outputs a multi-network aggregation recognition result according to the input data, and generates a final report according to the multi-network aggregation recognition result to reflect whether the patient has myocardial infarction risks.

Example four

Fig. 8 is a flowchart of an electrocardiosignal identification method based on a packet convolutional neural network according to a fourth embodiment of the present application. Referring to fig. 8, the electrocardiosignal identification method based on the packet convolutional neural network provided in this embodiment specifically includes:

s810, acquiring multi-lead electrocardiosignals, and splitting the multi-lead electrocardiosignals to obtain multiple groups of independent lead electrocardiosignals;

s820, inputting each group of the independent lead electrocardiosignals into the convolution block respectively to obtain convolution characteristics of the independent lead electrocardiosignals of the corresponding group;

s830, grouping the convolution characteristics of each group of independent lead electrocardiosignals to obtain the grouping characteristics of the group of independent lead electrocardiosignals, and inputting the grouping characteristics into the grouping convolution block to obtain a plurality of grouping convolution characteristics of the group of independent lead electrocardiosignals;

s840, combining a plurality of grouping convolution characteristics of each group of independent lead electrocardiosignals to obtain the electrocardiosignal combination characteristics of the group of independent lead electrocardiosignals;

s850, carrying out full-connection processing on the electrocardiosignal combination characteristics of each group of independent lead electrocardiosignals to obtain the myocardial infarction abnormal probability;

s860, judging the multi-lead electrocardiosignal to be the myocardial infarction signal according to the myocardial infarction abnormal probability.

According to the technical scheme provided by the embodiment of the application, the multi-lead electrocardiosignals are obtained, and the multi-lead electrocardiosignals are split to obtain the independent-lead electrocardiosignals; then, the independent lead electrocardiosignal is input into a convolution block to obtain convolution characteristics; then, grouping the convolution characteristics to obtain grouping characteristics, and inputting the grouping characteristics into a grouping convolution block to obtain grouping convolution characteristics; then, determining the myocardial infarction abnormal probability according to the grouping convolution characteristics; finally, according to the myocardial infarction abnormal probability, judging the multi-lead electrocardiosignals as myocardial infarction signals; therefore, the parameter quantity of convolution operation can be reduced, under the condition that the identification performance of the grouped convolution neural network is not reduced, the overfitting of the grouped convolution neural network is reduced, and whether the multi-lead electrocardiosignals are myocardial infarction signals or not can be identified and judged more accurately.

In another embodiment, the grouping the convolution features of each group of the individual lead electrocardiographic signals to obtain the grouping features of the group of the individual lead electrocardiographic signals includes:

acquiring the number of packet convolution groups of the packet convolution block;

and uniformly grouping the convolution characteristics of each group of independent lead electrocardiosignals according to the grouping convolution group number to obtain the grouping characteristics of the group of independent lead electrocardiosignals.

In another embodiment, the grouping convolution block includes a grouping convolution layer and a grouping maximum pooling layer, and the inputting the grouping features into the grouping convolution block to obtain a plurality of grouping convolution features of the set of independent lead electrocardiographic signals includes:

performing convolution, batch normalization and activation on the grouping features through the grouping convolution layer to obtain output features of the grouping convolution layer;

and performing maximum pooling on the output characteristics of the grouping convolution layer through the grouping maximum pooling layer to obtain a plurality of grouping convolution characteristics of the independent lead electrocardiosignals.

In another embodiment, the convolution block of the grouped convolutional neural network includes a first convolution block and a second convolution block, and the separately inputting each group of the independent lead electrocardiographic signals into the convolution block to obtain the convolution characteristics of the corresponding group of independent lead electrocardiographic signals includes:

performing convolution and pooling on the independent lead electrocardiosignals through the first convolution block to obtain convolution block output characteristics;

and performing convolution and pooling on the output bits of the convolution block through the second convolution block to obtain the convolution characteristic.

In another embodiment, the first convolution block includes a first convolution layer and a first maximum pooling layer, and the convolving and pooling the independent lead electrocardiographic signals by the first convolution block to obtain convolution block output characteristics includes:

performing convolution, batch normalization and activation on the independent lead electrocardiosignals through the first convolution layer to obtain a first convolution layer output characteristic;

and performing maximum pooling on the output characteristic of the first convolution layer through the first maximum pooling layer to obtain the output characteristic of the convolution block.

In another embodiment, the convolving the second convolution block includes a second convolution layer and a second maximum pooling layer, and the convolving and pooling the convolution block output bits by the second convolution block to obtain the convolution characteristic includes:

performing convolution, batch normalization and activation on the output characteristics of the convolution blocks through the second convolution layer to obtain second convolution layer output characteristics;

and performing maximum pooling on the output characteristics of the second convolution layer through the second maximum pooling layer to obtain the convolution characteristics.

In another embodiment, the acquiring multi-lead cardiac electrical signals comprises:

receiving an original signal;

performing wavelet decomposition on the original signal to obtain a wavelet decomposition signal; the wavelet decomposition signal has dimension X1;

zeroing an X2 dimensional signal in the wavelet decomposition signal to obtain a partial zeroing signal; wherein X2 is less than X1;

performing wavelet inverse transformation on the partial zero-set signals to obtain de-noising signals; the de-noising signal is a signal after high-frequency noise and baseline drift are removed;

and obtaining the multi-lead electrocardiosignal according to the denoising signal.

In another embodiment, said obtaining said multi-lead electrocardiographic signal according to said denoised signal comprises:

determining the R wave position of the de-noising signal;

The electrocardiosignal identification method based on the grouping convolution neural network can be used for executing the electrocardiosignal identification device based on the grouping convolution neural network provided by any embodiment, and has corresponding functions and beneficial effects.

For specific limitations of the electrocardiosignal identification method based on the grouped convolutional neural network, reference may be made to the above limitations of the electrocardiosignal identification device based on the grouped convolutional neural network, and details thereof are not repeated herein.

EXAMPLE five

Fig. 10 is a flowchart of a training method of a packet convolutional neural network according to a fifth embodiment of the present application. Referring to fig. 10, the training method of the packet convolutional neural network provided in this embodiment specifically includes:

s1010, acquiring an electrocardiosignal training sample aiming at the grouped convolutional neural network;

s1020, performing machine training on the grouped convolutional neural network by adopting the electrocardiosignal training sample to obtain a trained grouped convolutional neural network; the trained packet convolution neural network comprises a pre-trained convolution block and a packet convolution block; the trained grouped convolutional neural network is used for respectively inputting each group of independent lead electrocardiosignals into the convolution block to obtain convolution characteristics of the corresponding group of independent lead electrocardiosignals; the independent lead electrocardiosignals are obtained by splitting multi-lead electrocardiosignals; the system is also used for grouping the convolution characteristics of each group of independent lead electrocardiosignals to obtain the grouping characteristics of the group of independent lead electrocardiosignals, and inputting the grouping characteristics into the grouping convolution block to obtain a plurality of grouping convolution characteristics of the group of independent lead electrocardiosignals; the system is also used for combining a plurality of grouped convolution characteristics of each group of independent lead electrocardiosignals to obtain the electrocardiosignal combination characteristics of the group of independent lead electrocardiosignals; the system is also used for carrying out full-connection processing on the electrocardiosignal combination characteristics of each group of independent lead electrocardiosignals to obtain the myocardial infarction abnormal probability; and the multi-lead electrocardiosignal is judged to be the myocardial infarction signal according to the myocardial infarction abnormal probability.

The training method of the packet convolutional neural network provided by the above can be used for executing the training device of the packet convolutional neural network provided by any of the above embodiments, and has corresponding functions and advantages.

For specific limitations of the training method of the grouped convolutional neural network, reference may be made to the above limitations of the electrocardiographic signal recognition apparatus of the grouped convolutional neural network, and details are not repeated here.

It should be understood that, although the steps in the flowcharts of fig. 8 and 10 are shown in order as indicated by the arrows, the steps are not necessarily performed in order as indicated by the arrows. The steps are not performed in the exact order shown and described, and may be performed in other orders, unless explicitly stated otherwise. Moreover, at least some of the steps in fig. 8 and 10 may include multiple sub-steps or multiple stages that are not necessarily performed at the same time, but may be performed at different times, and the order of performing the sub-steps or stages is not necessarily sequential, but may be performed alternately or alternately with other steps or at least some of the sub-steps or stages of other steps.

EXAMPLE six

Fig. 9 is a schematic structural diagram of an electronic device according to a sixth embodiment of the present application. As shown in the figure, the electronic device includes: a processor 40, a memory 41, a display screen 42 with touch functionality, an input device 43, an output device 44, and a communication device 45. The number of the processors 40 in the electronic device may be one or more, and one processor 40 is illustrated as an example. The number of the memory 41 in the electronic device may be one or more, and one memory 41 is taken as an example in the figure. The processor 40, the memory 41, the display 42, the input device 43, the output device 44 and the communication device 45 of the electronic device may be connected by a bus or other means, and the bus connection is taken as an example in the figure. In an embodiment, the electronic device may be a computer, a mobile phone, a tablet, a projector, or an interactive smart tablet. In the embodiment, an electronic device is taken as an example of an interactive smart tablet to describe.

The memory 41 is a computer-readable storage medium, and can be used for storing software programs, computer-executable programs, and modules, such as program instructions/modules corresponding to the electrocardiosignal identification method according to any embodiment of the present application. The memory 41 may mainly include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program required for at least one function; the storage data area may store data created according to use of the device, and the like. Further, the memory 41 may include high speed random access memory, and may also include non-volatile memory, such as at least one magnetic disk storage device, flash memory device, or other non-volatile solid state storage device. In some examples, memory 41 may further include memory located remotely from processor 40, which may be connected to the device over a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.

The display screen 42 is a display screen 42 with a touch function, which may be a capacitive screen, an electromagnetic screen, or an infrared screen. In general, the display screen 42 is used for displaying data according to instructions from the processor 40, and is also used for receiving touch operations applied to the display screen 42 and sending corresponding signals to the processor 40 or other devices. Optionally, when the display screen 42 is an infrared screen, the display screen further includes an infrared touch frame, and the infrared touch frame is disposed around the display screen 42, and may also be configured to receive an infrared signal and send the infrared signal to the processor 40 or other devices.

The communication device 45 is used for establishing communication connection with other devices, and may be a wired communication device and/or a wireless communication device.

The input means 43 may be used for receiving input numeric or character information and generating key signal inputs related to user settings and function control of the electronic device, and may be a camera for acquiring images and a sound pickup device for acquiring audio data. The output device 44 may include an audio device such as a speaker. It should be noted that the specific composition of the input device 43 and the output device 44 can be set according to actual conditions.

The processor 40 executes various functional applications and data processing of the device by running software programs, instructions and modules stored in the memory 41, namely, implements the electrocardiosignal identification method based on the packet convolutional neural network.

Specifically, in an embodiment, the packet convolutional neural network includes a volume block and a packet volume block, and when the processor 40 executes one or more programs stored in the memory 41, the following operations are specifically implemented:

acquiring multi-lead electrocardiosignals, and splitting the multi-lead electrocardiosignals to obtain independent lead electrocardiosignals;

inputting the independent lead electrocardiosignals into the convolution blocks respectively to obtain convolution characteristics;

grouping the convolution characteristics to obtain grouping characteristics, and inputting the grouping characteristics into the grouping convolution block to obtain grouping convolution characteristics;

combining the grouped convolution characteristics to obtain electrocardiosignal combination characteristics, and carrying out full-connection processing on the electrocardiosignal combination characteristics to obtain myocardial infarction abnormal probability;

On the basis of the above embodiment, the grouping the convolution features to obtain grouping features includes:

and equally grouping the convolution characteristics according to the grouping convolution group number to obtain the grouping characteristics.

On the basis of the above embodiment, the block convolution block includes a block convolution layer and a block maximum pooling layer, and the inputting the block feature into the block convolution block to obtain the block convolution feature includes:

and performing maximum pooling on the output characteristics of the packet convolution layer through the packet maximum pooling layer to obtain the packet convolution characteristics.

On the basis of the above embodiment, the convolution block of the grouped convolutional neural network includes a first convolution block and a second convolution block, and the separately inputting the independent lead electrocardiographic signals into the convolution blocks to obtain convolution characteristics includes:

On the basis of the above embodiment, the first convolution block includes a first convolution layer and a first maximum pooling layer, and the convolving and pooling the independent lead electrocardiographic signal by the first convolution block to obtain the convolution block output characteristics includes:

On the basis of the foregoing embodiment, the convolving block includes a second convolution layer and a second maximum pooling layer, and the convolving and pooling the output bits of the convolving block by the second convolution block to obtain the convolution characteristic includes:

On the basis of the above embodiment, the acquiring a multi-lead electrocardiographic signal includes:

receiving an original signal;

On the basis of the above embodiment, obtaining the multi-lead electrocardiographic signal according to the denoising signal includes:

determining the R wave position of the de-noising signal;

Specifically, in the embodiment, when the processor 40 executes one or more programs stored in the memory 41, the following operations are further implemented:

EXAMPLE seven

A seventh embodiment of the present application further provides a storage medium containing computer-executable instructions, which when executed by a computer processor, are configured to perform a method for cardiac electrical signal identification based on a packet convolutional neural network, the packet convolutional neural network including pre-trained volume blocks and packet volume blocks, the method including:

The computer-executable instructions, when executed by a computer processor, are further for performing a method for packet convolutional neural network-based cardiac electrical signal identification, the method comprising:

Of course, the storage medium containing the computer-executable instructions provided in the embodiments of the present application is not limited to the operations of the grouped convolutional neural network based electrocardiograph signal identification method described above, and may also perform related operations in the grouped convolutional neural network based electrocardiograph signal identification method provided in any embodiment of the present application, and has corresponding functions and advantages.

It should be noted that the terms "first \ second \ third" related to the embodiments of the present invention are merely used for distinguishing similar objects, and do not represent a specific ordering for the objects, and it should be understood that "first \ second \ third" may exchange a specific order or sequence order if allowed. It should be understood that the terms first, second, and third, as used herein, are interchangeable under appropriate circumstances such that the embodiments of the invention described herein are capable of operation in other sequences than those illustrated or otherwise described herein.

It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, storage, database, or other medium used in the embodiments provided herein may include non-volatile and/or volatile memory, among others. Non-volatile memory can include read-only memory (ROM), Programmable ROM (PROM), Electrically Programmable ROM (EPROM), Electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDRSDRAM), Enhanced SDRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), Rambus Direct RAM (RDRAM), direct bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

The technical features of the above embodiments can be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the above embodiments are not described, but should be considered as the scope of the present specification as long as there is no contradiction between the combinations of the technical features.

The above-mentioned embodiments only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present patent shall be subject to the appended claims.

Claims

1. An apparatus for recognizing a cardiac signal based on a block convolutional neural network, the block convolutional neural network including a pre-trained convolutional block and a block convolutional block, the apparatus comprising:

a feature grouping module, configured to group the convolution features of each group of independent lead electrocardiographic signals to obtain grouping features of the group of independent lead electrocardiographic signals, where the feature grouping module includes: the group number obtaining submodule is used for obtaining the group convolution group number of the grouping convolution block; the grouping submodule is used for uniformly grouping the convolution characteristics of each group of independent lead electrocardiosignals according to the grouping convolution group number to obtain the grouping characteristics of the group of independent lead electrocardiosignals; inputting the grouping features into the grouping volume block, performing one-dimensional convolution on the grouping features through the grouping volume block, pooling the one-dimensional convolved features to reduce the dimension of the high-dimensional features into low-dimensional features and remove redundant features to obtain a plurality of grouping convolution features of the independent lead electrocardiosignals;

2. The apparatus of claim 1, wherein the packet volume block comprises a packet volume layer and a packet max pooling layer, and wherein the feature grouping module comprises:

and the pooling sub-module is used for performing maximal pooling on the output characteristics of the grouped convolution layer through the grouped maximal pooling layer to obtain a plurality of grouped convolution characteristics of the group of independent lead electrocardiosignals so as to reduce the parameter processing amount of the grouped convolution blocks.

3. The apparatus of claim 1, wherein the convolution block of the block convolutional neural network comprises a first convolution block and a second convolution block, and wherein the convolution module comprises:

the first convolution sub-module is used for performing convolution and pooling on the independent lead electrocardiosignals through the first convolution block to obtain convolution block output characteristics so as to reduce parameter processing amount of the second convolution block;

and the second convolution submodule is used for performing convolution and pooling on the output bits of the convolution block through the second convolution block to obtain the convolution characteristics so as to avoid characteristic loss caused by overlarge dimensionality reduction gradient.

4. The apparatus of claim 3, wherein the first convolution block comprises a first convolution layer and a first max-pooling layer, and wherein the first convolution sub-module is specifically configured to:

5. The apparatus of claim 3, wherein the second convolution block comprises a second convolution layer and a second max-pooling layer, and wherein the second convolution sub-module is specifically configured to:

6. The apparatus of claim 1, wherein the signal acquisition module comprises:

7. The apparatus of claim 6, wherein the multi-lead acquisition sub-module is specifically configured to:

determining the R wave position of the de-noising signal;

8. An apparatus for training a packet convolutional neural network, the apparatus comprising:

the machine training module is used for performing machine training on the packet convolutional neural network by adopting the electrocardiosignal training sample to obtain a trained packet convolutional neural network; the trained packet convolution neural network comprises a pre-trained convolution block and a packet convolution block; the trained grouped convolutional neural network is used for respectively inputting each group of independent lead electrocardiosignals into the convolution block to obtain convolution characteristics of the corresponding group of independent lead electrocardiosignals; the independent lead electrocardiosignals are obtained by splitting multi-lead electrocardiosignals; the system is further used for grouping the convolution characteristics of each group of independent lead electrocardiosignals to obtain the grouping characteristics of the group of independent lead electrocardiosignals, wherein the number of the grouping convolution groups of the grouping convolution blocks is obtained; uniformly grouping the convolution characteristics of each group of independent lead electrocardiosignals according to the grouping convolution group number to obtain the grouping characteristics of the group of independent lead electrocardiosignals; inputting the grouping features into the grouping volume block, performing one-dimensional convolution on the grouping features through the grouping volume block, pooling the one-dimensional convolved features to reduce the dimension of the high-dimensional features into low-dimensional features and remove redundant features to obtain a plurality of grouping convolution features of the independent lead electrocardiosignals; the system is also used for combining a plurality of grouped convolution characteristics of each group of independent lead electrocardiosignals to obtain the electrocardiosignal combination characteristics of the group of independent lead electrocardiosignals; the system is also used for carrying out full-connection processing on the electrocardiosignal combination characteristics of each group of independent lead electrocardiosignals to obtain the myocardial infarction abnormal probability; and the multi-lead electrocardiosignal is judged to be the myocardial infarction signal according to the myocardial infarction abnormal probability.

9. An electronic device, comprising: a memory, one or more processors;

the memory for storing one or more programs;

when executed by the one or more processors, cause the one or more processors to perform a method for cardiac electrical signal recognition based on a packet convolutional neural network, wherein the packet convolutional neural network comprises pre-trained volume blocks and packet volume blocks, the method comprising the steps of;

grouping the convolution characteristics of each group of independent lead electrocardiosignals to obtain the grouping characteristics of the group of independent lead electrocardiosignals, wherein the grouping convolution group number of the grouping convolution blocks is obtained; uniformly grouping the convolution characteristics of each group of independent lead electrocardiosignals according to the grouping convolution group number to obtain the grouping characteristics of the group of independent lead electrocardiosignals; inputting the grouping features into the grouping volume block, performing one-dimensional convolution on the grouping features through the grouping volume block, pooling the one-dimensional convolved features to reduce the dimension of the high-dimensional features into low-dimensional features and remove redundant features to obtain a plurality of grouping convolution features of the independent lead electrocardiosignals;

10. A storage medium containing computer-executable instructions, which when executed by a computer processor, perform a method for cardiac electrical signal identification based on a packet convolutional neural network, wherein the packet convolutional neural network comprises pre-trained volume blocks and packet volume blocks, the method comprising the steps of;