CN104398252A - Electrocardiogram signal processing method and device - Google Patents

Electrocardiogram signal processing method and device Download PDF

Info

Publication number
CN104398252A
CN104398252A CN201410617357.4A CN201410617357A CN104398252A CN 104398252 A CN104398252 A CN 104398252A CN 201410617357 A CN201410617357 A CN 201410617357A CN 104398252 A CN104398252 A CN 104398252A
Authority
CN
China
Prior art keywords
electrocardiosignal
sample data
feature
signal
training
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN201410617357.4A
Other languages
Chinese (zh)
Inventor
周丰丰
赵苗苗
刘记奎
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shenzhen Institute of Advanced Technology of CAS
Original Assignee
Shenzhen Institute of Advanced Technology of CAS
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Shenzhen Institute of Advanced Technology of CAS filed Critical Shenzhen Institute of Advanced Technology of CAS
Priority to CN201410617357.4A priority Critical patent/CN104398252A/en
Publication of CN104398252A publication Critical patent/CN104398252A/en
Pending legal-status Critical Current

Links

Classifications

    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
    • A61B5/316—Modalities, i.e. specific diagnostic methods
    • A61B5/318—Heart-related electrical modalities, e.g. electrocardiography [ECG]
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B1/00—Instruments for performing medical examinations of the interior of cavities or tubes of the body by visual or photographical inspection, e.g. endoscopes; Illuminating arrangements therefor
    • A61B1/00002—Operational features of endoscopes
    • A61B1/00004—Operational features of endoscopes characterised by electronic signal processing
    • A61B1/00009—Operational features of endoscopes characterised by electronic signal processing of image signals during a use of endoscope
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
    • A61B5/316—Modalities, i.e. specific diagnostic methods
    • A61B5/318—Heart-related electrical modalities, e.g. electrocardiography [ECG]
    • A61B5/346—Analysis of electrocardiograms
    • A61B5/349—Detecting specific parameters of the electrocardiograph cycle
    • A61B5/352—Detecting R peaks, e.g. for synchronising diagnostic apparatus; Estimating R-R interval
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A61B5/7203—Signal processing specially adapted for physiological signals or for diagnostic purposes for noise prevention, reduction or removal
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A61B5/7235—Details of waveform analysis
    • A61B5/725—Details of waveform analysis using specific filters therefor, e.g. Kalman or adaptive filters

Abstract

The embodiment of the invention discloses an electrocardiogram signal processing method and device. The method comprises the steps of denoising sample data including a reference characteristic type electrocardiogram signal to obtain a filter electrocardiogram signal, wherein the sample data includes sample data for training and sample data for test, extracting characteristics of the filter electrocardiogram signal on the basis of a multi-electrocardiogram period fusion method to obtain a characteristic electrocardiogram signal, carrying out data normalization treatment on the characteristic electrocardiogram signal to obtain a standard electrocardiogram signal, training a BP neural network according to the standard electrocardiogram signal corresponding to the sample data for training to obtain a trained BP neural network, injecting the standard electrocardiogram signal corresponding to the sample data for test into the trained BP neural network, and acquiring a classification result of a standard signal corresponding to the sample data for test. By adopting the electrocardiogram signal processing method and device disclosed by the embodiment, the comprehensiveness and accuracy for extracting and classifying the characteristic electrocardiogram signals can be improved.

Description

A kind of ECG's data compression method and device
Technical field
The present invention relates to signal processing technology field, particularly relate to a kind of ECG's data compression method and device.
Background technology
Electrocardiosignal is the general performance of cardiac electrical activity at body surface.The object of ECG's data compression infers the state of cardiovascular system, and make auxiliary medical decision according to it.
ECG's data compression method comprises two parts, the feature extraction of electrocardiosignal and classifier design.The feature extraction of electrocardiosignal mainly contains Wavelet Transform and time parameters estimation method, but above method only carries out feature extraction for key character wave band in a cardiac electrical cycle, make to mainly contain support vector machine, linear judgement and Self-organizing Maps in classifier design, in above ECG's data compression method, feature electrocardiosignal is extracted and the comprehensive and accuracy of classification all has much room for improvement.
Summary of the invention
The embodiment of the present invention provides a kind of ECG's data compression method and device, can improve comprehensive, accuracy that feature electrocardiosignal is extracted and classified.
Embodiments provide a kind of ECG's data compression method, it can comprise:
To the sample data comprising fixed reference feature classification electrocardiosignal except process of making an uproar obtains filtering electrocardiosignal, described sample data comprises training sample data and test sample data;
Based on multiple cardiac cycles fusion method, feature extraction is carried out to described filtering electrocardiosignal and obtain feature electrocardiosignal;
Data normalization process is carried out to described feature electrocardiosignal and obtains standard cardioelectric signal;
The standard cardioelectric signal training BP neutral net corresponding according to training sample data, obtains training rear BP neutral net;
By BP neutral net after training described in standard cardioelectric signal injection corresponding for test sample data, obtain the classification results of standard signal corresponding to described test sample data.
Embodiments provide a kind of electrocardiogram signal processing device, it can comprise:
Signal denoising unit, for the sample data comprising fixed reference feature classification electrocardiosignal except process of making an uproar obtains filtering electrocardiosignal, described sample data comprises training sample data and test sample data;
Feature extraction unit, carries out feature extraction based on multiple cardiac cycles fusion method to described filtering electrocardiosignal and obtains feature electrocardiosignal;
Normalized unit, obtains standard cardioelectric signal for carrying out data normalization process to described feature electrocardiosignal;
Training unit, for the standard cardioelectric signal training BP neutral net corresponding according to training sample data, obtains training rear BP neutral net;
Taxon, for by BP neutral net after training described in standard cardioelectric signal injection corresponding for test sample data, obtains the classification results of standard signal corresponding to described test sample data.
Therefore, the embodiment of the present invention obtains standard cardioelectric signal to after the denoising of electrocardiosignal sample data, multicycle fusion feature extraction, normalized, the standard cardioelectric signal training BP neutral net using training sample data corresponding, and classify with the standard cardioelectric signal that BP neutral net after training is corresponding to test sample data, the method can improve comprehensive, the accuracy that feature electrocardiosignal is extracted and classified.
Accompanying drawing explanation
In order to be illustrated more clearly in the embodiment of the present invention or technical scheme of the prior art, be briefly described to the accompanying drawing used required in embodiment or description of the prior art below, apparently, accompanying drawing in the following describes is only some embodiments of the present invention, for those of ordinary skill in the art, under the prerequisite not paying creative work, other accompanying drawing can also be obtained according to these accompanying drawings.
Fig. 1 is the schematic flow sheet of the first embodiment of a kind of ECG's data compression method provided by the invention;
Fig. 2 is the schematic flow sheet of the second embodiment of a kind of ECG's data compression method provided by the invention;
Fig. 3 is the schematic flow sheet of the 3rd embodiment of a kind of ECG's data compression method provided by the invention;
Fig. 4 is the schematic flow sheet of the 4th embodiment of a kind of ECG's data compression method provided by the invention;
Fig. 5 is the structural representation of the embodiment of a kind of electrocardiogram signal processing device provided by the invention.
Detailed description of the invention
Below in conjunction with the accompanying drawing in the embodiment of the present invention, be clearly and completely described the technical scheme in the embodiment of the present invention, obviously, described embodiment is only the present invention's part embodiment, instead of whole embodiments.Based on the embodiment in the present invention, those of ordinary skill in the art, not making the every other embodiment obtained under creative work prerequisite, belong to the scope of protection of the invention.
See Fig. 1 to Fig. 4, it is the first to the 3rd embodiment schematic flow sheet of the ECG's data compression method that the embodiment of the present invention provides.ECG's data compression method described in the present embodiment, comprises step:
Step S101, to the sample data comprising fixed reference feature classification electrocardiosignal except process of making an uproar obtains filtering electrocardiosignal, described sample data comprises training sample data and test sample data;
In specific embodiment, as shown in Figure 2, embodiment of the present invention step S101 to described training electrocardiosignal sample data except make an uproar process specifically can be realized by following steps:
S1011, extracts the high-frequency interferencing signal in described sample data;
S1012, the opening operation in employing morphology function and closed operation function extract the baseline drift signal in described sample data;
S1013, by the high-frequency interferencing signal in sample data described in sef-adapting filter filtering and described baseline drift signal, obtains filtering electrocardiosignal.
In specific implementation, described sef-adapting filter can irrigate husband's band filter for the Bart of 0.5-45Hz, and described Bart irrigates husband's band filter and Wavelet Transformation Algorithm can be adopted to remove baseline drift.
Step S102, carries out feature extraction based on multiple cardiac cycles fusion method to described filtering electrocardiosignal and obtains feature electrocardiosignal;
In the embodiment that some are feasible, as shown in Figure 3, embodiment of the present invention step S102 is carried out feature extraction based on multiple cardiac cycles fusion method to described filtering electrocardiosignal and obtains feature electrocardiosignal and specifically can be realized by following steps:
S1021, injects autoregressive moving-average model arma modeling by described filtering electrocardiosignal, the electrocardiosignal set of eigenvectors that it is feature with ARMA coefficient that described autoregressive moving-average model arma modeling generates;
In specific implementation, the single channel common version of described arma modeling is
y ( k ) + a 1 y ( k - 1 ) + . . . + a i y ( k - 1 ) + . . . + a p y ( k - p ) = e ( k ) + b 1 e ( k - 1 ) + . . . + b j e ( k - j ) + . . . + b q e ( k - q ) - - - ( 1 )
In formula, described y (k) is described electrocardiosignal set of eigenvectors; a iand b i(i=1,2 ... p; J=1,2 ... q) AR coefficient and MA coefficient is respectively; P and q is respectively AR order and MA order; E (k) is ARMA forecast error.
S1022, obtains described electrocardiosignal set of eigenvectors, carries out whitening processing, obtain feature electrocardiosignal to described ecg characteristics vector set;
Step S103, carries out data normalization process to described feature electrocardiosignal and obtains standard cardioelectric signal;
In the embodiment that some are feasible, as shown in Figure 4, embodiment of the present invention step S103 is carried out data normalization process to described feature electrocardiosignal and specifically can be realized by following steps:
S1031, determines an electrocardiosignal of leading in described feature electrocardiosignal;
S1032, the R wave-wave peak dot of the electrocardiosignal sequence at electrocardiosignal of leading described in detection place, with the wave crest point of described R ripple for demarcation line, resolves into multicycle electrocardiosignal group by described feature electrocardiosignal;
S1033, to described multicycle electrocardiosignal group, on transverse axis time shaft and longitudinal axis voltage axis, carry out minimax normalized obtains standard cardioelectric signal simultaneously.
Step S104, the standard cardioelectric signal training BP neutral net corresponding according to training sample data, obtains training rear BP neutral net.
In the embodiment that some are feasible, the standard cardioelectric signal training BP neutral net corresponding according to training sample data described by embodiment of the present invention step S104, obtains training rear BP neutral net can be realized by following steps:
S1041, netinit, determines network input layer nodes n, node in hidden layer l according to system input and output electrocardiosignal sequence (X, Y), output layer nodes m, initializes input layer, connection weights ω between hidden layer and output layer neuron ijand ω jk, initialize hidden layer threshold value a, output layer threshold value b, given learning rate and neuron excitation function.
S1042, hidden layer exports and calculates.According to input vector X, between input layer with hidden layer, be connected weights ω ijand hidden layer threshold value a, calculate hidden layer and export H.
H j = f ( Σ i = 1 n ω ij x i - a j ) , j = 1,2 , . . . , l - - - ( 2 )
In formula (2), l is node in hidden layer; F is hidden layer excitation function, and this function has multiple representation form, and selected function is herein:
f ( x ) = 1 1 + e - x
S1043, output layer exports and calculates.Export H according to hidden layer, connect weights ω jkwith output layer threshold value b, calculate the neural neural network prediction of BP and export O;
O k = Σ j = 1 l H j ω jk - b k , k = 1,2 , . . . , m
S1044, Error Calculation.O and desired output Y, computing network forecast error e is exported according to neural network forecast;
e k＝Y k-O kk＝1,2,…,m
S1045, right value update.Upgrade network according to neural network forecast error e and connect weights ω ijand ω jk, computing formula is
ω ij = ω ij + η H j ( 1 - H j ) x ( i ) Σ k = 1 m ω jk e k , i = 1,2 , . . . , n ; j = 1,2 , . . . , l - - - ( 3 - 1 )
ω jk＝ω jk+ηH je kj＝1,2,…,l；k＝1,2,…,m (3-2)
In formula (3-1), (3-2), η is learning rate, and computing formula is
η(t)＝η max-t(η max-η min)/t max(4)
In formula (4), η maxwith η minmaximum learning rate and minimum learning rate respectively; t maxbe maximum iteration time and current iteration number of times with t.
In specific implementation, the method for additional momentum can be adopted to calculate connection weights, the connection weights computing formula of band additional momentum is
ω(k)＝ω(k-1)+△ω(k)+a[ω(k-1)-ω(k-2)] (5)
In formula (5), ω (k), ω (k-1), ω (k-2) is the weights in k, k-1, k-2 moment respectively; A is momentum learning rate.
S1046, threshold value upgrades.Upgrade network node threshold value a and b according to neural network forecast error e, computing formula is
a j = a j + η H j ( 1 - H j ) Σ k = 1 m ω jk e k , j = 1,2 , . . . , l
b k＝b k+e kk＝1,2,…,m
S1047, whether evaluation algorithm iteration terminates, if do not terminate, returns step S1042.
S105, by BP neutral net after training described in standard cardioelectric signal injection corresponding for test sample data, obtains the classification results of standard signal corresponding to described test sample data.
In specific embodiment, electrocardiogram acquisition equipment is used to gather the characteristic signal of 4 class experiment samples, totally 2000 sample datas (derive from america's MIT/BIH (arrhythmia database data) standard database, sample frequency 250Hz, precision 12, the electrocardiosignal sample data that wherein health characteristics index is corresponding is 500 examples, electrocardiosignal sample data corresponding to myocardial infarction characteristic index is 500 examples, electrocardiosignal sample data corresponding to ventricular arrhythmia characteristic index is 500 examples, electrocardiosignal sample data corresponding to bundle branch block characteristic index is 500 examples), different types of electrocardiosignal uses 1 (corresponding health characteristics index) respectively, 2 (corresponding myocardial infarction characteristic indexs), 3 (respective chamber/chambers arrhythmia characteristic indexs), 4 (respective chamber/chambers arrhythmia characteristic index) identifies, according to described step S101, step S102, step S103 processes described electrocardiosignal, obtain 4 class standard electrocardiosignaies, and be stored in data1 respectively, data2, data3, in data4 database file, wherein often organizing data is 25 dimensions, first dimension is classification logotype, rear 24 dimensions are feature electrocardiosignaies, 125 sample datas in the every category feature electrocardiosignal of random selecting also merge rear as training data, using after 375 sample datas remaining in every category feature electrocardiosignal merge as test data, the classification accuracy rate obtaining BP neutral net is as shown in the table:
Because described standard cardioelectric signal has 24 dimensions, 4 classes, preferably, the input layer of described BP neutral net has 24 nodes, and hidden layer has 25 nodes, and output layer has 4 nodes.
The embodiment of the present invention obtains standard cardioelectric signal to after the denoising of electrocardiosignal sample data, multicycle fusion feature extraction, normalized, the standard cardioelectric signal training BP neutral net using training sample data corresponding, and classify with the standard cardioelectric signal that BP neutral net after training is corresponding to test sample data, the method can improve comprehensive, the accuracy that feature electrocardiosignal is extracted and classified.
See Fig. 5, it is the structural representation of a kind of electrocardiogram signal processing device that the embodiment of the present invention provides.Electrocardiogram signal processing device described in the present embodiment, comprising:
Signal denoising unit 501, for obtaining training electrocardiosignal sample data, to described training electrocardiosignal sample data except process of making an uproar obtains filtering electrocardiosignal;
In specific implementation, described signal denoising unit 501, specifically for:
Extract the high-frequency interferencing signal in described sample data;
Opening operation in employing morphology function and closed operation function extract the baseline drift signal in described sample data;
By the high-frequency interferencing signal in sample data described in sef-adapting filter filtering and described baseline drift signal, obtain filtering electrocardiosignal.
Feature extraction unit 502, obtains feature electrocardiosignal for carrying out feature extraction based on multiple cardiac cycles fusion method to described filtering electrocardiosignal;
In specific implementation, described feature extraction unit 502, specifically for:
Described filtering electrocardiosignal is injected autoregressive moving-average model arma modeling, the electrocardiosignal set of eigenvectors that it is feature with ARMA coefficient that described autoregressive moving-average model arma modeling generates;
Obtain described electrocardiosignal set of eigenvectors, whitening processing is carried out to described electrocardiosignal set of eigenvectors, obtains feature electrocardiosignal.
Normalized unit 503, obtains standard cardioelectric signal for carrying out data normalization process to described feature electrocardiosignal;
In specific implementation, described normalized unit 503, specifically for:
Determine an electrocardiosignal of leading in described feature electrocardiosignal;
The R wave-wave peak dot of the electrocardiosignal sequence at electrocardiosignal of leading described in detection place, with the wave crest point of described R ripple for demarcation line, resolves into multicycle electrocardiosignal group by described feature electrocardiosignal;
To described multicycle electrocardiosignal group, on transverse axis time shaft and longitudinal axis voltage axis, carry out minimax normalized obtains standard cardioelectric signal simultaneously.
Training unit 504, for the standard cardioelectric signal training BP neutral net corresponding according to training sample data, obtains training rear BP neutral net.
Taxon 505, for by BP neutral net after training described in standard cardioelectric signal injection corresponding for test sample data, obtains the classification results of standard signal corresponding to described test sample data.
In specific embodiment, electrocardiogram acquisition equipment is used to gather the characteristic signal of 4 class experiment samples, totally 2000 sample datas (derive from america's MIT/BIH (arrhythmia database data) standard database, sample frequency 250Hz, precision 12, the electrocardiosignal sample data that wherein health characteristics index is corresponding is 500 examples, electrocardiosignal sample data corresponding to myocardial infarction characteristic index is 500 examples, electrocardiosignal sample data corresponding to ventricular arrhythmia characteristic index is 500 examples, electrocardiosignal sample data corresponding to bundle branch block characteristic index is 500 examples), different types of electrocardiosignal uses 1 (corresponding health characteristics index) respectively, 2 (corresponding myocardial infarction characteristic indexs), 3 (respective chamber/chambers arrhythmia characteristic indexs), 4 (respective chamber/chambers arrhythmia characteristic index) identifies, according to described step S101, step S102, step S103 processes described electrocardiosignal, obtain 4 class standard electrocardiosignaies, and be stored in data1 respectively, data2, data3, in data4 database file, wherein often organizing data is 25 dimensions, first dimension is classification logotype, rear 24 dimensions are feature electrocardiosignaies, 125 sample datas in the every category feature electrocardiosignal of random selecting also merge rear as training data, using after 375 sample datas remaining in every category feature electrocardiosignal merge as test data, after adding up described training, the classification accuracy rate of BP neutral net to test data is as shown in the table:
Because described standard cardioelectric signal has 24 dimensions, 4 classes, preferably, the input layer of described BP neutral net has 24 nodes, and hidden layer has 25 nodes, and output layer has 4 nodes.
The embodiment of the present invention obtains standard cardioelectric signal to after the denoising of electrocardiosignal sample data, multicycle fusion feature extraction, normalized, the standard cardioelectric signal training BP neutral net using training sample data corresponding, and classify with the standard cardioelectric signal that BP neutral net after training is corresponding to test sample data, the method can improve comprehensive, the accuracy that feature electrocardiosignal is extracted and classified.
One of ordinary skill in the art will appreciate that all or part of flow process realized in above-described embodiment method, that the hardware that can carry out instruction relevant by computer program has come, described program can be stored in a computer read/write memory medium, this program, when performing, can comprise the flow process of the embodiment as above-mentioned each side method.Wherein, described storage medium can be magnetic disc, CD, read-only store-memory body (Read-Only Memory, ROM) or random store-memory body (Random Access Memory, RAM) etc.
Above disclosedly be only present pre-ferred embodiments, certainly can not limit the interest field of the present invention with this, therefore according to the equivalent variations that the claims in the present invention are done, still belong to the scope that the present invention is contained.

Claims (10)

1. an ECG's data compression method, is characterized in that, comprises the following steps:
To the sample data comprising fixed reference feature classification electrocardiosignal except process of making an uproar obtains filtering electrocardiosignal, described sample data comprises training sample data and test sample data;
Based on multiple cardiac cycles fusion method, feature extraction is carried out to described filtering electrocardiosignal and obtain feature electrocardiosignal;
Data normalization process is carried out to described feature electrocardiosignal and obtains standard cardioelectric signal;
The standard cardioelectric signal training BP neutral net corresponding according to training sample data, obtains training rear BP neutral net;
By BP neutral net after training described in standard cardioelectric signal injection corresponding for test sample data, obtain the classification results of standard signal corresponding to described test sample data.
2. method according to claim 1, is characterized in that,
Described fixed reference feature classification comprises health characteristics classification and myocardial infarction feature classification and ventricular arrhythmia feature classification and respective chamber/chambers arrhythmia classification.
3. method according to claim 1 and 2, is characterized in that, the described sample data to comprising fixed reference feature classification electrocardiosignal comprises except process of making an uproar obtains filtering electrocardiosignal number:
Extract the high-frequency interferencing signal in described sample data;
Opening operation in employing morphology function and closed operation function extract the baseline drift signal in described sample data;
By the high-frequency interferencing signal in sample data described in sef-adapting filter filtering and described baseline drift signal, obtain filtering electrocardiosignal.
4. method according to claim 1, is characterized in that, describedly carries out feature extraction based on multiple cardiac cycles fusion method to described filtering electrocardiosignal and obtains feature electrocardiosignal and comprise:
Described filtering electrocardiosignal is injected autoregressive moving-average model arma modeling, the electrocardiosignal set of eigenvectors that it is feature with ARMA coefficient that described autoregressive moving-average model arma modeling generates;
Obtain described electrocardiosignal set of eigenvectors, whitening processing is carried out to described electrocardiosignal set of eigenvectors, obtains feature electrocardiosignal.
5. method according to claim 1, is characterized in that, describedly carries out data normalization process to described feature electrocardiosignal and obtains standard cardioelectric signal and comprise:
Determine an electrocardiosignal of leading in described feature electrocardiosignal;
The R wave-wave peak dot of the electrocardiosignal sequence at electrocardiosignal of leading described in detection place, with the wave crest point of described R ripple for demarcation line, resolves into multicycle electrocardiosignal group by described feature electrocardiosignal;
To described multicycle electrocardiosignal group, on transverse axis time shaft and longitudinal axis voltage axis, carry out minimax normalized obtains standard cardioelectric signal simultaneously.
6. an electrocardiogram signal processing device, is characterized in that, described device comprises:
Signal denoising unit, for the sample data comprising fixed reference feature classification electrocardiosignal except process of making an uproar obtains filtering electrocardiosignal, described sample data comprises training sample data and test sample data;
Feature extraction unit, carries out feature extraction based on multiple cardiac cycles fusion method to described filtering electrocardiosignal and obtains feature electrocardiosignal;
Normalized unit, obtains standard cardioelectric signal for carrying out data normalization process to described feature electrocardiosignal;
Training unit, for the standard cardioelectric signal training BP neutral net corresponding according to training sample data, obtains training rear BP neutral net;
Taxon, for by BP neutral net after training described in standard cardioelectric signal injection corresponding for test sample data, obtains the classification results of standard signal corresponding to described test sample data.
7. device according to claim 6, is characterized in that, in described signal denoising unit,
Described fixed reference feature classification in described training sample data comprises health characteristics classification and myocardial infarction feature classification and ventricular arrhythmia feature classification and respective chamber/chambers arrhythmia classification.
8. the device according to claim 6 or 7, is characterized in that, described signal denoising unit, specifically for,
Extract the high-frequency interferencing signal in described sample data;
Opening operation in employing morphology function and closed operation function extract the baseline drift signal in described sample data;
By the high-frequency interferencing signal in sample data described in sef-adapting filter filtering and described baseline drift signal, obtain filtering electrocardiosignal.
9. device according to claim 6, is characterized in that, described feature extraction unit, specifically for,
Described filtering electrocardiosignal is injected autoregressive moving-average model arma modeling, the electrocardiosignal set of eigenvectors that it is feature with ARMA coefficient that described autoregressive moving-average model arma modeling generates;
Obtain described electrocardiosignal set of eigenvectors, whitening processing is carried out to described electrocardiosignal set of eigenvectors, obtains feature electrocardiosignal.
10. device according to claim 6, is characterized in that, described normalized unit, specifically for,
Determine an electrocardiosignal of leading in described feature electrocardiosignal;
The R wave-wave peak dot of the electrocardiosignal sequence at electrocardiosignal of leading described in detection place, with the wave crest point of described R ripple for demarcation line, resolves into multicycle electrocardiosignal group by described feature electrocardiosignal;
To described multicycle electrocardiosignal group, on transverse axis time shaft and longitudinal axis voltage axis, carry out minimax normalized obtains standard cardioelectric signal simultaneously.
CN201410617357.4A 2014-11-05 2014-11-05 Electrocardiogram signal processing method and device Pending CN104398252A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201410617357.4A CN104398252A (en) 2014-11-05 2014-11-05 Electrocardiogram signal processing method and device

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201410617357.4A CN104398252A (en) 2014-11-05 2014-11-05 Electrocardiogram signal processing method and device

Publications (1)

Publication Number Publication Date
CN104398252A true CN104398252A (en) 2015-03-11

Family

ID=52635968

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201410617357.4A Pending CN104398252A (en) 2014-11-05 2014-11-05 Electrocardiogram signal processing method and device

Country Status (1)

Country Link
CN (1) CN104398252A (en)

Cited By (21)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105105743A (en) * 2015-08-21 2015-12-02 山东省计算中心（国家超级计算济南中心） Intelligent electrocardiogram diagnosis method based on deep neural network
CN105550653A (en) * 2015-12-14 2016-05-04 广州视源电子科技股份有限公司 ECG lead intelligent selection method and system
CN106344004A (en) * 2016-09-28 2017-01-25 清华大学 Electrocardiosignal feature point detecting method and device
CN106510716A (en) * 2016-11-07 2017-03-22 广州视源电子科技股份有限公司 Method and device for calculating respiratory rate based on electrocardiosignals
CN106539586A (en) * 2016-11-07 2017-03-29 广州视源电子科技股份有限公司 A kind of breathing rate computational methods and device
CN106580324A (en) * 2016-11-07 2017-04-26 广州视源电子科技股份有限公司 Method and device for extracting respiratory signal
WO2017124946A1 (en) * 2016-01-22 2017-07-27 周常安 Dynamic cardiovascular activity monitoring method and system, and wearable monitoring device
CN107736888A (en) * 2017-11-27 2018-02-27 中山大学 A kind of method and system that real-time cardiac monitoring is carried out using wearable ecg equipment
CN108403107A (en) * 2018-02-06 2018-08-17 北京大学深圳研究生院 A kind of arrhythmia cordis method of discrimination and system
CN108836312A (en) * 2018-07-13 2018-11-20 希蓝科技（北京）有限公司 A kind of method and system of the novel progress clutter rejecting based on artificial intelligence
CN108836316A (en) * 2018-05-02 2018-11-20 湖北工业大学 A kind of R wave of electrocardiosignal extracting method based on BP neural network
CN109009047A (en) * 2018-05-24 2018-12-18 东软集团股份有限公司 A kind of arrhythmia detection method, device and equipment
CN109864714A (en) * 2019-04-04 2019-06-11 北京邮电大学 A kind of ECG Signal Analysis method based on deep learning
CN109998495A (en) * 2019-05-23 2019-07-12 河南工业大学 A kind of electrocardiosignal classification method based on particle group optimizing BP neural network
CN110037681A (en) * 2019-03-15 2019-07-23 上海数创医疗科技有限公司 Method and device based on the identification rhythm of the heart type for improving self-encoding encoder network
CN110245562A (en) * 2019-05-13 2019-09-17 中国水产科学研究院东海水产研究所 Ocean based on deep learning produces malicious microalgae type automatic identifying method
WO2019205067A1 (en) * 2018-04-27 2019-10-31 Vita-Course Technologies Co., Ltd. Systems and methods for determining an arrhythmia type
CN110858309A (en) * 2018-08-22 2020-03-03 中国信息通信研究院 Multi-reference time clock weighting synthesis method
CN113261972A (en) * 2020-02-17 2021-08-17 华为技术有限公司 Electrocardio detection device, circuit and method
CN113509186A (en) * 2021-06-30 2021-10-19 重庆理工大学 ECG classification system and method based on deep convolutional neural network
CN113768517A (en) * 2021-09-28 2021-12-10 彩之物科技(深圳)有限公司 Intelligent early warning system and early warning method for heart health quality

Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20030204146A1 (en) * 2002-04-29 2003-10-30 Carlson Dan B. Apparatus and methods for analysis of cardiac device stored episodes containing sensed signals and waveforms
CN1539372A (en) * 2003-10-24 2004-10-27 南京大学 Method and device for early diagnosis of heart disease basaed on high-frequency waveform of cardiograph
CN1951320A (en) * 2006-11-16 2007-04-25 上海交通大学 Abnormal electrocardiogram recognition method based on ultra-complete characteristics
CN101658425A (en) * 2009-09-11 2010-03-03 西安电子科技大学 Device and method for detecting attention focusing degree based on analysis of heart rate variability
CN102129509A (en) * 2011-02-22 2011-07-20 郑州大学 Coronary heart disease self-diagnosis system based on electrocardiographic monitoring and back-propagation neural network
CN102779234A (en) * 2012-06-28 2012-11-14 中国科学院苏州纳米技术与纳米仿生研究所 Electrocardiogram classified processing method and device
CN103345600A (en) * 2013-06-24 2013-10-09 中国科学院深圳先进技术研究院 Electrocardiosignal data processing method
US8655435B2 (en) * 2012-02-08 2014-02-18 Samsung Electronics Co., Ltd. Apparatus and method for generating atrial fibrillation prediction model, and apparatus and method for predicting atrial fibrillation
CN103610457A (en) * 2013-11-19 2014-03-05 深圳先进技术研究院 Electrocardiosignal processing method and electrocardiosignal processing system
CN103927556A (en) * 2014-05-07 2014-07-16 天津工业大学 ECG signal classifying method based on wavelet packet and approximate entropy

Patent Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20030204146A1 (en) * 2002-04-29 2003-10-30 Carlson Dan B. Apparatus and methods for analysis of cardiac device stored episodes containing sensed signals and waveforms
CN1539372A (en) * 2003-10-24 2004-10-27 南京大学 Method and device for early diagnosis of heart disease basaed on high-frequency waveform of cardiograph
CN1951320A (en) * 2006-11-16 2007-04-25 上海交通大学 Abnormal electrocardiogram recognition method based on ultra-complete characteristics
CN101658425A (en) * 2009-09-11 2010-03-03 西安电子科技大学 Device and method for detecting attention focusing degree based on analysis of heart rate variability
CN102129509A (en) * 2011-02-22 2011-07-20 郑州大学 Coronary heart disease self-diagnosis system based on electrocardiographic monitoring and back-propagation neural network
US8655435B2 (en) * 2012-02-08 2014-02-18 Samsung Electronics Co., Ltd. Apparatus and method for generating atrial fibrillation prediction model, and apparatus and method for predicting atrial fibrillation
CN102779234A (en) * 2012-06-28 2012-11-14 中国科学院苏州纳米技术与纳米仿生研究所 Electrocardiogram classified processing method and device
CN103345600A (en) * 2013-06-24 2013-10-09 中国科学院深圳先进技术研究院 Electrocardiosignal data processing method
CN103610457A (en) * 2013-11-19 2014-03-05 深圳先进技术研究院 Electrocardiosignal processing method and electrocardiosignal processing system
CN103927556A (en) * 2014-05-07 2014-07-16 天津工业大学 ECG signal classifying method based on wavelet packet and approximate entropy

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
葛丁飞等: "心电信号多周期融合特征提取和分类研究", 《中国生物医学工程学报》 *

Cited By (33)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105105743B (en) * 2015-08-21 2017-03-29 山东省计算中心（国家超级计算济南中心） Electrocardiogram intelligent processing method based on deep neural network
CN105105743A (en) * 2015-08-21 2015-12-02 山东省计算中心（国家超级计算济南中心） Intelligent electrocardiogram diagnosis method based on deep neural network
WO2017101529A1 (en) * 2015-12-14 2017-06-22 广州视源电子科技股份有限公司 Electrocardio lead intelligent selection method and system
CN105550653A (en) * 2015-12-14 2016-05-04 广州视源电子科技股份有限公司 ECG lead intelligent selection method and system
CN105550653B (en) * 2015-12-14 2018-11-27 广州视源电子科技股份有限公司 Cardiac diagnosis lead intelligent selecting method and system
WO2017124946A1 (en) * 2016-01-22 2017-07-27 周常安 Dynamic cardiovascular activity monitoring method and system, and wearable monitoring device
CN106344004A (en) * 2016-09-28 2017-01-25 清华大学 Electrocardiosignal feature point detecting method and device
CN106580324A (en) * 2016-11-07 2017-04-26 广州视源电子科技股份有限公司 Method and device for extracting respiratory signal
CN106580324B (en) * 2016-11-07 2019-07-16 广州视源电子科技股份有限公司 A kind of breath signal extracting method and device
CN106510716B (en) * 2016-11-07 2019-09-20 广州视源电子科技股份有限公司 A kind of method and device calculating respiratory rate based on electrocardiosignal
CN106539586A (en) * 2016-11-07 2017-03-29 广州视源电子科技股份有限公司 A kind of breathing rate computational methods and device
CN106539586B (en) * 2016-11-07 2019-07-16 广州视源电子科技股份有限公司 A kind of respiratory rate calculation method and device
CN106510716A (en) * 2016-11-07 2017-03-22 广州视源电子科技股份有限公司 Method and device for calculating respiratory rate based on electrocardiosignals
CN107736888A (en) * 2017-11-27 2018-02-27 中山大学 A kind of method and system that real-time cardiac monitoring is carried out using wearable ecg equipment
CN108403107A (en) * 2018-02-06 2018-08-17 北京大学深圳研究生院 A kind of arrhythmia cordis method of discrimination and system
CN108403107B (en) * 2018-02-06 2019-12-31 北京大学深圳研究生院 Arrhythmia discrimination method and system
US11534099B2 (en) 2018-04-27 2022-12-27 Vita-Course Technologies Co., Ltd. Systems and methods for determining an arrhythmia type
CN112312836A (en) * 2018-04-27 2021-02-02 深圳市长桑技术有限公司 System and method for determining type of arrhythmia
WO2019205067A1 (en) * 2018-04-27 2019-10-31 Vita-Course Technologies Co., Ltd. Systems and methods for determining an arrhythmia type
CN108836316A (en) * 2018-05-02 2018-11-20 湖北工业大学 A kind of R wave of electrocardiosignal extracting method based on BP neural network
CN109009047B (en) * 2018-05-24 2021-10-26 东软集团股份有限公司 Arrhythmia detection method, device and equipment
CN109009047A (en) * 2018-05-24 2018-12-18 东软集团股份有限公司 A kind of arrhythmia detection method, device and equipment
CN108836312A (en) * 2018-07-13 2018-11-20 希蓝科技（北京）有限公司 A kind of method and system of the novel progress clutter rejecting based on artificial intelligence
CN110858309A (en) * 2018-08-22 2020-03-03 中国信息通信研究院 Multi-reference time clock weighting synthesis method
CN110858309B (en) * 2018-08-22 2023-04-18 中国信息通信研究院 Multi-reference time clock weighting synthesis method
CN110037681A (en) * 2019-03-15 2019-07-23 上海数创医疗科技有限公司 Method and device based on the identification rhythm of the heart type for improving self-encoding encoder network
CN109864714A (en) * 2019-04-04 2019-06-11 北京邮电大学 A kind of ECG Signal Analysis method based on deep learning
CN110245562A (en) * 2019-05-13 2019-09-17 中国水产科学研究院东海水产研究所 Ocean based on deep learning produces malicious microalgae type automatic identifying method
CN109998495A (en) * 2019-05-23 2019-07-12 河南工业大学 A kind of electrocardiosignal classification method based on particle group optimizing BP neural network
CN113261972A (en) * 2020-02-17 2021-08-17 华为技术有限公司 Electrocardio detection device, circuit and method
CN113509186A (en) * 2021-06-30 2021-10-19 重庆理工大学 ECG classification system and method based on deep convolutional neural network
CN113768517A (en) * 2021-09-28 2021-12-10 彩之物科技(深圳)有限公司 Intelligent early warning system and early warning method for heart health quality
CN113768517B (en) * 2021-09-28 2024-03-15 彩之物科技(深圳)有限公司 Intelligent early warning system and early warning method for heart health quality

Similar Documents

Publication Publication Date Title
CN104398252A (en) Electrocardiogram signal processing method and device
CN107822622B (en) Electrocardiogram diagnosis method and system based on deep convolutional neural network
JP7065185B2 (en) Electrocardiographic signal detection
CN109645980A (en) A kind of rhythm abnormality classification method based on depth migration study
KR102077605B1 (en) Method and apparatus for an automatic artifact removal of EEG based on a deep learning algorithm
Liu et al. Applying topological persistence in convolutional neural network for music audio signals
Kumar et al. Towards adaptive classification using Riemannian geometry approaches in brain-computer interfaces
CN105212922A (en) The method and system that R wave of electrocardiosignal detects automatically are realized towards FPGA
CN104809434A (en) Sleep staging method based on single-channel electroencephalogram signal ocular artifact removal
CN110367967A (en) A kind of pocket lightweight human brain condition detection method based on data fusion
Kang et al. A method of denoising multi-channel EEG signals fast based on PCA and DEBSS algorithm
CN109480824A (en) Processing method, device and the server of EGC waveform data
CN104409080A (en) Voice end node detection method and device
KR102501837B1 (en) Signal feature extraction method and signal feature extraction apparatus
CN110399846A (en) A kind of gesture identification method based on multichannel electromyography signal correlation
CN108968941A (en) A kind of arrhythmia detection method, apparatus and terminal
CN109124620A (en) A kind of atrial fibrillation detection method, device and equipment
CN104035563A (en) W-PCA (wavelet transform-principal component analysis) and non-supervision GHSOM (growing hierarchical self-organizing map) based electrocardiographic signal identification method
Duan et al. EEG feature selection method based on decision tree
CN109567789A (en) Division processing method, device and the readable storage medium storing program for executing of ECG data
Zhang et al. Classification of EEG signals based on AR model and approximate entropy
CN110558975A (en) Electrocardiosignal classification method and system based on LRF-ELM and BLSTM
Avramidis et al. Multiscale fractal analysis on EEG signals for music-induced emotion recognition
CN111067505A (en) Arrhythmia detection method, device and storage medium
Li et al. A novel motor imagery EEG recognition method based on deep learning

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
RJ01 Rejection of invention patent application after publication

Application publication date: 20150311