CN104102915B - Personal identification method based on ECG multi-template matching under a kind of anomalous ecg state - Google Patents
Personal identification method based on ECG multi-template matching under a kind of anomalous ecg state Download PDFInfo
- Publication number
- CN104102915B CN104102915B CN201410313480.7A CN201410313480A CN104102915B CN 104102915 B CN104102915 B CN 104102915B CN 201410313480 A CN201410313480 A CN 201410313480A CN 104102915 B CN104102915 B CN 104102915B
- Authority
- CN
- China
- Prior art keywords
- ecg
- template
- data
- user
- matching
- 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.)
- Active
Links
Landscapes
- Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)
Abstract
The present invention relates to the multi-template matching personal identification methods based on ECG under a kind of anomalous ecg state, belong to biological characteristics identity recognizing technology field, the electrocardiogram (ECG) data of user to be identified is compared with the data for registering user in template library, obtains identification result.The key technology of this method includes: ECG signal processing, for eliminating noise jamming；Electrocardiosignal decomposes, and isolates the ecg wave form in each period；Standardization standardizes on time and amplitude scale respectively；Feature extraction extracts feature using wavelet transformation, and ISODATA algorithm carries out clustering, and then constructs ECG template library；Correlation analysis calculates the correlation of ECG test data and each template, selects best match template, final to obtain identification result.Multi-template matching personal identification method proposed by the present invention identifies identity using the electrocardiosignal accumulate in human body, and the ECG data under abnormality is taken into account.
Description
Technical field
The present invention relates to the multi-template matching bodies that ECG (Electrocardiogram) is based under a kind of anomalous ecg state
Part recognition methods, belongs to biological characteristics identity recognizing technology field, identifies identity using the electrocardiosignal accumulate in human body, not only needle
To healthy population, and it is also applied for the individual with arrhythmia cordis.
Background technique
In modern society, identification has important application in many fields.Along with today's society to security requirement
Continuous enhancing, the drawbacks of traditional identity identification method, gradually shows, for example certificate is easily lost, password is easy to crack.?
Under this background, the identity recognizing technology based on biological characteristic gradually causes the extensive concern of people, has become a hot topic of research it
One.
Identification based on biological characteristic refers to the physiological characteristic or behavioral trait using human body, to identify personal identification
A technology.Biological characteristic or behavioral trait for identification need to meet generality, uniqueness, stability and can measure
The characteristics such as property.The advantages that by feat of unique portability and reliability, fingerprint recognition, recognition of face and speech recognition etc. are raw
Object characteristic identity identification technology achieves fast development, and is applied widely.
PQRST wave morphology in ECG keeps relative stability within one period, even if easily leading to the heart in pressure and movement etc.
Under conditions of rate variation, QRS complex group is still stable.In addition, receptor type, age, gender, cardiac position, size, dissection knot
Structure, chest configuration and heart physiological characteristic etc. influence, and ECG signal varies with each individual.Between the same normal heartbeat of individual, the rhythm of the heart of the same race
There is very big similitude between not normal heartbeat, internal diversity is less than the difference between individual heartbeat, and therefore, ECG can be used as one kind
Biological characteristic is used for identification.
Compared with traditional biometric identity recognition methods, identification is carried out by means of the ECG signal accumulate in human body
Gradually by the concern of scholar, it has many particular advantages: 1. antifalsifications, and ECG signal is a kind of from the heart of user
Live biometric eliminates the hidden danger imitated or stolen easily compared with the features such as fingerprint, face and voice；2. easily place
Reason, ECG is one-dimensional signal, and data volume is small, and processing is simple, saves memory space.
Many scholars inquire into ECG signal for identification, and the research work of early stage is mainly around Healthy People
Group carries out, and identifies identity according to normal ECG signal, and achieve very high recognition accuracy.However, having in actual life
The individual of arrhythmia cordis is generally existing, and effect of the above method under anomalous ecg state is simultaneously bad；It is recent some to grind
Study carefully work to take into account abnormal ECG signal, many methods pertain only to a certain or several types of arrhythmia, in practical application
With significant limitation.
Summary of the invention
The present invention comprehensive considers normally to propose that a kind of accuracy rate is high and is suitable for anomalous ecg state with abnormal ECG signal
Identification algorithm.Overall procedure is as shown in Figure 1, key technology includes ECG signal processing, signal decomposition and standardization
Processing, ecg characteristics are extracted, ECG data cluster, ECG template data library constructs and identification.It particularly may be divided into following five
Module:
1. ECG signal processing module
1.1 obtain the multistage electrocardiogram (ECG) data under user's stable state with certain time length, are stored in the form of data file,
Corresponding header file includes the identity information and electrocardiogram (ECG) data format of the user；
1.2 pairs of every section of electrocardiogram (ECG) datas pre-process, and are realized by designing suitable filter, filter out Hz noise, base
The noises such as line drift and myoelectricity interference.
2. signal decomposition and standardized module
The ECG data of 2.1 couples of users detects R crest location；
2.2 use the method based on R crest location and adjacent R R interphase, divide continuous ecg wave form, that is, will weekly
Phase ecg wave form is gradually separated；
2.3 standardizations make all ecg wave forms of same individual length having the same in time scale, in width
Being worth on scale has equal maximum voltage value.
3. electrocardiosignal characteristic extracting module
3.1, using R crest location and average RR interphase information, extract every QRS complex for clapping ecg wave form；
3.2 extract the feature of ECG signal based on wavelet transformation (Wavelet Transform)；
3.3 drop feature space using principal component analysis (Principal Component Analysis, PCA)
Dimension.
4. ECG template data library constructs module
4.1 use ISODATA (Iterative Self-Organization Data Analysis Techniques)
Algorithm clusters the ECG data of same user；
4.2 pairs of every class ECG datas, using corresponding QRS complex as ECG template to be selected；
4.3 pairs of every class templates to be selected choose suitable number of typical template as such data using dependent thresholds method
Matching template；
4.4 couples of all registration users do identical processing, to construct ECG template data library.
5. template matching and identification module
One section of electrocardiogram (ECG) data with certain time length under 5.1 acquisition user's stable states to be identified, execution data prediction,
The operations such as signal decomposition, standardization and QRS complex extraction；
5.2 use dependent thresholds method, choose suitable number and can characterize the QRS complex of the user as ECG survey
Try data；
5.3 match each test data with all templates in ECG template library；
5.4 related coefficients between test data and template data find test data as template matching criterion
Best match template comprehensively considers the template matching results of same user's whole ECG test data to be identified, determines the user
Final identity.
The present invention has the advantages that
1. algorithm is standardized each cycle ecg wave form isolated, thus eliminate well due to pressure
Changes in heart rate caused by the external factors such as power, movement is reflected in the inconsistency in ecg wave form time scale；
2. mature R wave detection algorithm when extracting QRS complex, is utilized in algorithm, increase extraction reliability and
Precision；And when extracting feature by wavelet transformation, without detecting P wave, Q involves T wave position, and significantly reduces algorithm
Time complexity；
3. algorithm not only utilizes normal ECG data when constructing electro-cardiologic template library, and for the use with arrhythmia cordis
Family, takes into account exception ECG data, uniformly uses ISODATA algorithm to cluster normal and all kinds of arrhythmia cordis data, both mentioned
High identification precision, has also been enlarged application range；
4., in order to eliminate influence of user's electrocardio singular value to algorithm, increasing system in template matching and identification
Robustness is chosen multiple test datas and is differentiated, the best match template and correspondence of each test data of user are comprehensively considered
Related coefficient, obtain final identification result.
Detailed description of the invention
Fig. 1 is technical solution of the present invention overview flow chart；
Fig. 2 is technical solution of the present invention modular structure schematic diagram；
Fig. 3 is ECG signal processing module flow diagram；
Fig. 4 is signal decomposition and standardized module flow chart；
Fig. 5 is electrocardiosignal characteristic extracting module flow chart；
Fig. 6 is that ECG template data library constructs module flow diagram；
Fig. 7 is template matching and identification module flow chart.
Specific embodiment
To keep implementation steps of the invention, effect and advantage apparent, below in conjunction with attached drawing to implementation of the invention
Mode is described in further detail.
The present invention relates to the multi-template matching personal identification methods based on ECG under a kind of anomalous ecg state, referring to fig. 2,
This method comprises:
201: ECG signal processing module；
202: signal decomposition and standardized module；
203: electrocardiosignal characteristic extracting module；
204: ECG template data library constructs module；
205: template matching and identification module.
Wherein, ECG signal processing module is used to obtain the raw ECG data of user, and storage in the data file, passes through
The noises such as the myoelectricity interference in Hz noise, baseline drift and the collection process in signal are eliminated in filtering processing.Electrocardiosignal is pre-
The flow chart of processing module is specifically included referring to Fig. 3:
301: obtaining the raw ECG data of user.The present invention is with MIT-BIH arrhythmia cordis database (MIT-BIH
Arrhythmia Database, MITDB) in electrocardiogram (ECG) data as experimental data, choose wherein 44 electrocardiogram (ECG) data files,
The duration of each file about 30 minutes, each user was characterized respectively, is denoted as UseriI=1,2 ..., 44 makees for first 20 minutes
For training dataset, it is used as test data set within latter 10 minutes.
302: data filtering processing.Human body electrocardio is small-signal, and vulnerable to various noise jammings in collection process,
Therefore, filtering processing is essential step.
The filtering algorithm that the present invention uses mainly completes the work of following several respects: 1. pairs of original ECG signals carry out
Value processing；2. application moving average filter eliminates high-frequency noise interference；3. eliminating baseline drift, external signal is mainly considered
The low frequency signal in source；4. cutoff frequency is used to be higher than the noise of 30Hz for the Butterworth filter rejection frequency of 30Hz.Through such as
Various main noise interference in ECG signal can be effectively removed in upper four steps processing.
Signal decomposition and standardization module.It will according to certain segmentation rule for pretreated ECG data
User's each cycle ecg wave form is separated from continuous electrocardiosignal, and does standardization.Signal decomposition and standardization mould
The flow chart of block referring to fig. 4, specifically includes:
The detection of 401:R crest location.For the feature for extracting ECG signal, R crest location need to be detected, for this purpose, the present invention uses
The QRS wave detector of entitled ECGPUWAVE realizes, the output of the detector be one include R wave crest sampling point position text
Part.
402: electrocardiosignal decomposes.In order to by the ECG data of user cluster, and in turn construct electro-cardiologic template library, need by
The ECG signal of each cardiac cycle is separated from continuous electrocardiogram (ECG) data record.The present invention, which uses, is based on R crest location and phase
The method of adjacent RR interphase is realized.The sample frequency of electrocardiogram (ECG) data is 360Hz in MITDB, to user UseriElectrocardio in training set
Data, can be calculated periodicity be M, sampling number n, then it is a length of when the segment data
Average RR interphase is
Electrocardiosignal decomposition follows following principle:
1. pair current ecg wave form, R crest location is tR-Position remembers with the time interval of previous adjacent R wave crest
For tRR_Pre is denoted as t with the time interval of the latter adjacent R wave crestRR_next；
2. access time section is [tR_position-0.4*tRR_Pre, tR_position+0.6*tRR_Next] one section
The current ecg wave form of data characterization；
3. the electrocardiogram (ECG) data of couple all users does same treatment, realize that electrocardiosignal decomposes.
403: time scale standardization.Heart rate is easily affected by the external environment, and makes to eliminate changes in heart rate to arithmetic accuracy
At adverse effect, need to do standardization in time scale to each ecg wave form, specific implementation is as follows:
On the basis of current period ecg-r wave peak position, and with reference to time value t between the average RR of the userRRIt is adjusted.
1. if the time interval length of current period ecg wave form is greater than average RR interphase, i.e. 0.4*tRR_pre+0.6*
tRR_Next > tRR, then region constriction is carried out, region constriction coefficient is
2. if the time interval length of current period ecg wave form is less than average RR interphase, i.e. 0.4*tRR-pre+0.6*
tRR_Next < tRR, then interval extension is carried out, interval extension coefficient is
3. pair all ecg wave forms do above-mentioned region constriction or extension process, standardized ECG letter in time scale is obtained
Number.
404: amplitude scale calibration.In order to eliminate the variation of amplitude caused by measuring instrument, need to each ecg wave form
It is standardized on amplitude scale, specific implementation is as follows:
1. recording the amplitude at all ECG signal R crest locations of same user, averaged
2. if the amplitude at current ECG signal R crest location is greater than average value, i.e.,Then by amplitude
Compression, the compressed coefficient are
3. if the amplitude at current ECG signal R crest location is less than average value, i.e.,Then by amplitude
It stretches, drawing coefficient is
405: electrocardiogram (ECG) data storage.To user UseriThe electrocardiogram (ECG) data of i=1,2 ..., 44, through signal decomposition and standard
Change processing, is saved with two-dimensional matrix, is denoted as ECGi mn, i=1,2 ..., 44, wherein m is ECG number, and n is the sampling of ECG data
Points.
Electrocardiosignal characteristic extracting module.For ECG standardized data, QRS complex is on the one hand extracted, as user's
On the other hand template set to be selected extracts primitive character with wavelet transformation, principal component analysis carries out dimensionality reduction to feature space, that is,
Redundancy feature is eliminated, the time complexity of algorithm can be reduced.The flow chart of electrocardiosignal characteristic extracting module is referring to Fig. 5, specifically
Include:
501: extracting QRS complex.QRS complex is extracted on the basis of R crest location, for standardized ECG data,
R crest location and RR interphase average value t need to only be utilizedRRIt extracts.If current ECG signal R crest location is tR-
Position, with 0.15*tRRDuration intercept one section of electrocardiogram (ECG) data respectively forwardly and backward as QRS complex.For user
UseriI=1,2 ..., 44, the QRS complex extracted with two-dimensional matrix storage, is denoted as QRSi mn, i=1,2 ..., 44,
Middle m is QRS wave shape number, and n is the sampling number of each QRS wave shape.
502: wavelet transformation extracts feature.Small echo change is carried out to ECG signal using the wavelet function of entitled Daubechies
It changes, Daubechies small echo abbreviation dbN, wherein N is wavelet-order, the Support of wavelet function ψ (t) and scaling function φ (t)
It is 2N-1.ψ (t) can be found out by φ (t), be φ (2t) weighted shift and, be shown below
N value is different, weight gkAlso different.φ (t) limited length, supporting domain are [0,2N-1], and therefore, obtained ψ (t) is
Finite support, supporting domain are [1-N, N].
It is theoretical and experience have shown that, db3 small echo is similar to ecg wave form, meets the principle of similarity of selection, therefore, selects
For db3 small echo as wavelet basis, it is 5 supporting domain that it, which has length, shorter bearing length can be effectively reduced algorithm when
Between complexity, be conducive to the feature extraction of ECG signal.
Since different user coefficient of wavelet decomposition waveform becomes apparent than time domain waveform difference, and each heart bat of same user is small
Wave Decomposition coefficient different wave shape is small, more stable, ECG time domain waveform is carried out 6 grades of wavelet decompositions, db3 is as wavelet basis, after transformation
Wavelet coefficient form feature vector, be denoted as x=[x1, x2..., xp]。
503:PCA carries out Feature Dimension Reduction.Principal component analysis is a kind of data analysing method that K.Pearson is proposed, purpose
It is that one group of new feature arranged from big to small by importance is calculated from primitive character, they are linear groups of primitive character
It closes, and irrelevant.
If new feature is yi, i=1,2 ..., p are the linear combination of above-mentioned primitive character
For unified yiScale, the mould of linear combination coefficient might as well be set as 1, i.e.,
αi Tαi=1
It is expressed as follows with matrix
Y=ATx
Wherein, y is new feature yiThe feature vector of composition, A are eigentransformation matrixes.Need to solve optimal orthogonal transformation
Matrix A makes new feature yiVariance reach extreme value.
The covariance matrix of x is set as ∑, is estimated with training sample.
μ=E { x }
∑=E { (x- μ) (x- μ)T}
Covariance matrix ∑ shares p characteristic value λi, i=1,2 ..., p (including may it is equal and may for 0 it is intrinsic
Value), it sorts from large to small as λ1≥λ2≥…≥λp。
PCA is to indicate data with less principal component as a kind of feature extracting method, k principal component before taking, then
The variance that they represent data account for the ratio of population variance as
In the present invention, aforementioned proportion is set as 90%, the k value of above formula can be calculated accordingly, and then realize Feature Dimension Reduction.
504: final characteristic storage in two-dimensional matrix, being denoted as Fi mk, i=1,2 ..., 44, wherein m is the heart of user
Electrical waveform number, k are the principal component number used.
ECG template data library constructs module.It is clustered using ECG data of the ISODATA algorithm to user, is judged in classification
Reasonable classification results are obtained under criterion.By dependent thresholds method, typical ECG signal is selected in each categorical data, in turn
The template for generating the category does same treatment, building registration user's electro-cardiologic template to the electrocardiogram (ECG) data of each classification of all users
Database.ECG template data library constructs the flow chart of module referring to Fig. 6, specifically includes:
601:ISODATA algorithm cluster.ISODATA(Iterative Self-Organizing Data Analysis
Techniques, iteration self-organizing data analysis technique) it can be regarded as a kind of improved C means clustering algorithm.The algorithm is
All kinds of mean value of meter grate, can be improved operation efficiency, in addition, drawing in cluster process in this way after whole samples are adjusted
Enter the judge criterion to classification, whereby can automatically by certain categories combinations or division, to obtain more reasonable cluster result,
Also the limitation of given class number in advance is breached to a certain extent.
For user UseriIf forming sample set by N number of ECG data, is indicated, be denoted as with respective feature vectorAfter ISODATA algorithm cluster, c cluster centre is obtained, m is usedj, j=1,2 ..., c are indicated.
602: dependent thresholds method selects template.In order to reduce the time complexity of algorithm, using dependent thresholds method in user
Every class ECG data in select typical QRS complex as such matching template, the specific implementation process is as follows:
1. user UseriClassification ГjMean value mjFor
Wherein, NjIt is the number of samples of j-th of cluster.
2. in classification ГjMiddle selection K away from class center mjThe corresponding QRS complex of nearest sample as template to be selected,
It is arranged successively from small to large by distance, is denoted as QRSi st, i=1,2 ..., 44s=1,2 ..., K.
3. QRS might as well be selectedi 1tAs reference templates, the related coefficient with residue K-1 templates to be selected is calculated
Wherein, Cov (QRSi 1t, QRSi st) it is template QRS to be selectedi 1tAnd QRSi stCovariance, D (QRSi 1t) and D
(QRSi st) it is QRS respectivelyi 1tAnd QRSi stVariance.
Cov(QRSi 1t, QRSi st)=E { [QRSi 1t-E(QRSi 1t)][QRSi st-E(QRSi st)]}
D(QRSi 1t)=E | QRSi 1t-E(QRSi 1t)|2}
D(QRSi st)=E | QRSi st-E(QRSi st)|2}
And E (QRSi 1t) and E (QRSi st) it is QRS respectivelyi 1tAnd QRSi stMean value.
603: building electro-cardiologic template library.In order to construct ECG template data library, classification Γ is setjStencil-chosen threshold value be
Thj 1, as the thresholding for selecting typical template from template to be selected, using the average value of related coefficient as threshold value Thj 1
Set such ГjTemplate-matching threshold be Thj 2, as the criterion of ECG test data template matching success or not,
Using the minimum value of related coefficient as threshold value Thj 2
Above-mentioned processing is done to the ECG data of all users, the multiple template for characterizing every class electrocardiogram (ECG) data is obtained, is denoted as
Tempi jk, i=1,2 ..., 44, wherein j is UseriCluster numbers, and k be jth class template number.
Template matching and identification module.User Useri10 minutes electrocardiogram (ECG) datas are made after i=1,2 ..., 44
ECG standardized data is obtained, still after data preprocessing module and signal decomposition and standardized module processing for test data
R crest location information extraction QRS complex is so utilized, ECG test data is selected by dependent thresholds method, in conjunction with the electrocardio of building
Template database does correlation analysis, and then obtains the best match template of each test data, if corresponding related coefficient
Greater than matching threshold, then shows successful match, be otherwise considered as invalid data, refused.It is completed when all test datas match
Afterwards, then comprehensively consider the matching results of all test datas of same user, provide final identification result.Template matching
And the flow chart of identification module is specifically included referring to Fig. 7:
701: extracting QRS complex.For user UserlL=1,2 ..., 44, after choosing corresponding electrocardiogram (ECG) data file
10 minutes data obtain ECG after data preprocessing module and signal decomposition and standardized module processing as test data
Standardized data.
ECG signal QRS complex is extracted on the basis of R crest location, since electrocardiogram (ECG) data has been standardized processing,
R crest location and RR interphase average value t need to only be utilizedRRIt extracts.If current ECG signal R crest location is tR_
Position, with 0.15*tRRDuration respectively forwardly and backward interception one piece of data as QRS complex.For user Userl
The QRS complex of the electrocardiogram (ECG) data of l=1,2 ..., 44, extraction is stored with two-dimensional matrix, is denoted as QRSl uv, l=1,2 ...,
44, wherein u is QRS wave shape number, and v is the sampling number of QRS wave shape.
702: dependent thresholds method selects test data.For the QRS complex of user, randomly select K waveform be used as to
Test data is selected, QRS is denoted asl uv, l=1,2 ..., 44u=1,2 ..., K, for the similitude and test data for guaranteeing waveform
Typicalness obtains final ECG test data using correlation coefficient threshold method.QRS might as well be chosenl 1vAs reference data, calculate
With the correlation coefficient r of remaining K-1 datau(QRSl 1v, QRSl uv), u=2,3 ..., K are selected using average value as test data
The threshold value Th selectedl
Above-mentioned processing is done to the ECG data of all users, the ECG test data of characterization user identity is obtained, is denoted as
Testl uv, l=1,2 ..., 44, wherein u is UserlQRS complex number.
703: correlation analysis.So far, the ECG template data Temp of 44 users is obtainedi jk, i=1,2 ..., 44 and survey
Try data Testl uv, l=1,2 ..., 44.
U-th of electrocardio test data Test of user to be identified for ll u, calculate and all kinds of moulds of all registration users
The related coefficient of whole template datas, i.e. r in plateI, j, k(Testl u, Tempi jk)。
704: obtaining best match template.Using the best match template of Correlation Coefficient Criteria selection test data, first
The maximum template of related coefficient, i.e. max are selected in the similar electrocardiogram (ECG) data of userk{rI, j, k(Testl u, Tempi jk), then
It is selected in the every class electrocardiogram (ECG) data of user, i.e. maxj{maxk{rI, j, k(Testl u, Tempi jk), finally in the heart of all users
Best match template, i.e. max are selected in electric datai{maxj{maxk{rI, j, k(Testl u, Tempi jk)}}}.So far, process is above-mentioned
The processing of three step maximizings finds best match template for each test data of user to be identified.
705: comprehensively considering template matching situations and obtain identification result.It is searched for each test data of user to be identified
After rope to best match template, if corresponding related coefficient is greater than default template-matching threshold, show successful match, otherwise
Algorithm is refused using the test data as invalid data.
User to be identified has multiple ECG test datas, it is therefore desirable to comprehensively consider the template of same user's total data
Match condition, and the weight of matching result is directly proportional to corresponding correlation coefficient value, obtains final identification knot accordingly
Fruit.To user UserlThe test data Test of l=1,2 ..., 44l u, corresponding best match template is denoted as Tempi j, i=1,
2 ..., 44.
If it is considered that single matching of the ECG test data to template, is denoted as
If it is considered that matching of the single ECG test data to registration user, is denoted as
If it is considered that matching of the user to be identified to registration user, then be denoted as
Finally, algorithm exports final identification result.
Claims (6)
1. the multi-template matching personal identification method under a kind of anomalous ecg state based on ECG, which is characterized in that including walking as follows
It is rapid:
Step 1: raw ECG data obtains, pretreatment；The raw ECG data acquisition, which refers to, not only utilizes normal ECG data,
And for the user with arrhythmia cordis, exception ECG data is taken into account；
Step 2: electrocardiosignal decomposes and standardization；
Step 3: wavelet transformation extracts the characteristic feature of electrocardiosignal, and principal component analysis carries out dimensionality reduction to feature space, obtains table
Levy the feature vector of ECG signal；
Step 4: ECG data being clustered using ISODATA algorithm, and then constructs registration user ECG template library；
Step 5: the ECG test data of user to be identified being matched with the electro-cardiologic template in template library one by one, using correlation
Coefficient obtains identification result as similarity criteria；
The step 4 includes the following steps:
The ECG data of same user is clustered using ISODATA algorithm, obtains c class data；
Using QRS complex as the template to be selected of every class ECG data；
Using dependent thresholds method, suitable number of typical QRS complex is chosen from every class template to be selected, as such ECG number
According to matching template；
Set classification ΓjStencil-chosen threshold value be Thj 1, as the thresholding for selecting typical template from template to be selected, correlation
The average value of coefficient is as threshold value Thj 1
Set classification ΓjTemplate-matching threshold be Thj 2, as the criterion of ECG test data template matching success or not, phase
The minimum value of relationship number is as threshold value Thj 2
Same treatment is done to each registration user, to construct ECG template data library；Wherein, each user of i characterization, i=1,
2…；J is user UseriCluster numbers；K is the number of template to be selected；T indicates data duration；QRSi 1tFor benchmark template；
QRSi stFor template to be selected；rsIndicate the related coefficient of reference templates and residue K-1 templates to be selected；S indicate except reference templates it
The serial number of remaining K-1 outer templates to be selected, s=2,3 ..., K.
2. the multi-template matching personal identification method under anomalous ecg state according to claim 1 based on ECG, feature
It is, the pretreatment in the step 1 comprises the following steps:
1.1 obtain the multistage electrocardiogram (ECG) data under user's stable state with certain time length；
1.2 pairs of every section of electrocardiogram (ECG) datas pre-process, and design filter, filter out Hz noise, baseline drift and myoelectricity interference.
3. the multi-template matching personal identification method under anomalous ecg state according to claim 1 based on ECG, feature
It is, the step 2 comprises the following steps:
The ECG data of 2.1 couples of users detects R crest location；
2.2 use the method based on R crest location and adjacent R R interphase, isolate the waveform of each cardiac electrical cycle；
2.3 standardizations, with identical length in time scale, on amplitude scale with it is equal most
Big voltage value.
4. the multi-template matching personal identification method under anomalous ecg state according to claim 1 based on ECG, feature
It is, the step 3 comprises the following steps:
3.1, using R crest location and the information of RR interphase, extract every QRS complex for clapping ecg wave form；
3.2 extract the primitive character of ecg wave form based on wavelet transformation；
3.3 carry out dimensionality reduction to ecg characteristics space using principal component analysis, obtain the feature vector that can characterize ecg wave form.
5. the multi-template matching personal identification method under anomalous ecg state according to claim 1 based on ECG, feature
It is, the step 5 comprises the following steps:
5.1 obtain lower section of the user's stable state to be identified electrocardiogram (ECG) data with certain time length, complete data prediction, electrocardio
Signal decomposition, standardization and QRS complex extract；
5.2 use dependent thresholds method, choose the QRS complex that can characterize user to be identified as ECG test data；
5.3 match each test data of user to be identified with all templates in ECG template library；
5.4 use related coefficient as template matching similarity criteria, best match template are found, to user UserlL=1,
2 ... test data Testl u, corresponding best match template is denoted as Tempi j, i=1,2 ...；It is complete to comprehensively consider same user
The template matching situations of portion's data:
Consider matching of the single ECG test data to template, is denoted as
Consider matching of the single ECG test data to registration user, is denoted as
Consider matching of the user to be identified to registration user, is then denoted as
The weight of matching result is directly proportional to corresponding correlation coefficient value, obtains final identification result accordingly；Wherein,
U is user UserlQRS complex number；J is user UseriCluster numbers；K is the template number of jth class；TestlIt is l
The electrocardio test data of user to be identified；Tempi jkFor the multiple template for characterizing every class electrocardiogram (ECG) data；TempiFor registration user's
Template.
6. a kind of computer readable storage medium is stored with the computer program being used in combination with calculating equipment, the calculating
Machine program is executed by processor to realize any one of claim 1 to 5 the method.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410313480.7A CN104102915B (en) | 2014-07-01 | 2014-07-01 | Personal identification method based on ECG multi-template matching under a kind of anomalous ecg state |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410313480.7A CN104102915B (en) | 2014-07-01 | 2014-07-01 | Personal identification method based on ECG multi-template matching under a kind of anomalous ecg state |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104102915A CN104102915A (en) | 2014-10-15 |
| CN104102915B true CN104102915B (en) | 2019-02-22 |
Family
ID=51671055
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410313480.7A Active CN104102915B (en) | 2014-07-01 | 2014-07-01 | Personal identification method based on ECG multi-template matching under a kind of anomalous ecg state |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104102915B (en) |
Families Citing this family (51)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104573458B (en) * | 2014-12-30 | 2017-05-31 | 深圳先进技术研究院 | A kind of personal identification method based on electrocardiosignal, apparatus and system |
| JP6655794B2 (en) * | 2015-03-17 | 2020-02-26 | パナソニックＩｐマネジメント株式会社 | Personal authentication device, personal authentication method, and program |
| CN106137184B (en) * | 2015-04-08 | 2019-02-12 | 四川锦江电子科技有限公司 | Electrocardiosignal QRS complex detection method based on wavelet transformation |
| WO2016168979A1 (en) | 2015-04-20 | 2016-10-27 | 深圳市长桑技术有限公司 | Vital sign analysis method and system |
| WO2016168980A1 (en) * | 2015-04-20 | 2016-10-27 | 深圳市长桑技术有限公司 | Physiological sign information acquisition method and system |
| CN104834921B (en) * | 2015-05-22 | 2018-01-16 | 清华大学深圳研究生院 | Electrocardio just/abnormal big data processing method and processing device |
| US9824287B2 (en) * | 2015-09-29 | 2017-11-21 | Huami Inc. | Method, apparatus and system for biometric identification |
| CN105654046B (en) * | 2015-12-29 | 2019-01-18 | 中国科学院深圳先进技术研究院 | Electrocardiosignal personal identification method and device |
| US10123741B2 (en) * | 2016-11-30 | 2018-11-13 | Huami Inc. | Cardiac condition detection |
| WO2018000396A1 (en) | 2016-06-30 | 2018-01-04 | 华为技术有限公司 | Authentication method and communication terminal |
| CN106295300B (en) * | 2016-08-15 | 2019-04-26 | 深圳先进技术研究院 | A kind of wearable intelligent equipment identity identifying method and device |
| US9642577B1 (en) | 2016-08-16 | 2017-05-09 | American Reliance, Inc. | Methods and systems for disease analysis based on transformations of diagnostic signals |
| CN107872427B (en) * | 2016-09-26 | 2020-04-21 | 清华大学深圳研究生院 | Cloud computing-based electrocardiogram identity recognition system and method |
| DE102016011700A1 (en) * | 2016-09-28 | 2018-03-29 | Personal Medsystems Gmbh | Monitoring of biosignals, in particular electrocardiograms |
| CN106418757B (en) * | 2016-09-29 | 2018-06-12 | 华中科技大学 | A kind of ECG authentications wisdom clothing |
| CN106618552B (en) * | 2016-12-30 | 2019-08-09 | 清华大学 | A kind of detection model acquisition methods and device |
| CN108304746B (en) * | 2017-01-11 | 2022-09-13 | 北京三星通信技术研究有限公司 | Method and equipment for updating authentication reference information for electrocardio identity authentication |
| CN108392176A (en) * | 2017-02-08 | 2018-08-14 | 上海跃扬医疗科技有限公司 | A kind of Sleep architecture detection method based on the acquisition of heart impact signal |
| CN108460318B (en) * | 2017-02-22 | 2020-06-16 | 清华大学深圳研究生院 | Identity authentication/identification method and equipment based on electrocardiosignals |
| WO2018152711A1 (en) * | 2017-02-22 | 2018-08-30 | 清华大学深圳研究生院 | Electrocardiographic authentication-based door control system and authentication method therefor |
| CN106897693B (en) * | 2017-02-23 | 2021-01-08 | 深圳先进技术研究院 | Identity recognition method, device and system |
| CN108701219B (en) * | 2017-03-14 | 2021-05-04 | 华为技术有限公司 | Method and device for processing waveform signal |
| CN107174236B (en) * | 2017-06-21 | 2020-03-10 | 广东工业大学 | Electrocardiosignal denoising method and device based on optimization theory |
| WO2019006631A1 (en) * | 2017-07-03 | 2019-01-10 | 深圳市汇顶科技股份有限公司 | Quality evaluation method and apparatus, model establishment method and module, and wearable device |
| CN107647860B (en) * | 2017-08-31 | 2021-01-15 | 深圳和而泰智能控制股份有限公司 | Heart rate detection method and device, electronic equipment and storage medium |
| GB2566101A (en) * | 2017-09-05 | 2019-03-06 | B Secur Ltd | Wearable authentication device |
| CN109840451A (en) * | 2017-11-28 | 2019-06-04 | 深圳市岩尚科技有限公司 | A kind of wearable ring of intelligence payment and its method of payment based on electrocardio identification |
| CN108154085B (en) * | 2017-12-06 | 2022-02-18 | 北京顺源开华科技有限公司 | Method and device for identity recognition based on electrocardiogram data and electronic equipment |
| CN108596142B (en) * | 2018-05-09 | 2022-01-11 | 吉林大学 | PCANet-based electrocardiogram feature extraction method |
| EP3866681A1 (en) * | 2018-10-18 | 2021-08-25 | Deep Science, LLC | Systems and methods for detecting physiological information using multi-modal sensors |
| CN109637610A (en) * | 2018-11-19 | 2019-04-16 | 深圳市理邦精密仪器股份有限公司 | Configuration method, terminal device and the medium of electrocardio report |
| CN109875545A (en) * | 2018-12-17 | 2019-06-14 | 浙江好络维医疗技术有限公司 | A kind of 12 lead electrocardiogram method for reconstructing of individuation based on CNN |
| CN109887594A (en) * | 2018-12-17 | 2019-06-14 | 浙江好络维医疗技术有限公司 | A kind of multi-lead arrhythmia cordis intelligent diagnosing method based on MODWT and TDCNN |
| CN110123304B (en) * | 2019-01-22 | 2021-08-27 | 东南大学 | Dynamic electrocardio noise filtering method based on multi-template matching and correlation coefficient matrix |
| CN109674464B (en) * | 2019-01-29 | 2021-04-30 | 郑州大学 | Multi-lead electrocardiosignal composite feature extraction method and corresponding monitoring system |
| CN109875570B (en) * | 2019-01-30 | 2020-08-18 | 华南理工大学 | Effective method for recognizing electrocardiosignal identities before and after movement |
| CN109919050B (en) * | 2019-02-21 | 2021-07-20 | 天津惊帆科技有限公司 | Identity recognition method and device |
| CN110169766A (en) * | 2019-04-30 | 2019-08-27 | 厦门大学 | A kind of cardiogram wave detection method, apparatus based on wavelet transformation, terminal device |
| CN110420022B (en) * | 2019-07-29 | 2020-12-11 | 浙江大学 | P wave detection method based on dual-density wavelet transform |
| CN114126487A (en) * | 2019-12-02 | 2022-03-01 | 深圳迈瑞生物医疗电子股份有限公司 | Monitoring method, monitoring device, monitoring equipment and computer readable storage medium |
| CN111582138A (en) * | 2020-04-30 | 2020-08-25 | 山东大学 | Electrocardio identity recognition method and system based on frequency domain cepstrum coefficient characteristics |
| CN111797816B (en) * | 2020-07-21 | 2024-05-14 | 迪姆软件(北京)有限公司 | Electrocardiogram analysis method and analysis device |
| CN112016405B (en) * | 2020-08-07 | 2022-07-15 | 浙江大学 | Electrocardiogram identity authentication method based on wearable equipment |
| CN112257518A (en) * | 2020-09-30 | 2021-01-22 | 西安交通大学第二附属医院 | ECG identity recognition method based on WT and WOA-PNN algorithm |
| CN112232253B (en) * | 2020-10-26 | 2021-11-09 | 南通大学 | Fetal electrocardiogram abnormity detection method driven by rough and reasonable granularity |
| CN112686094B (en) * | 2020-12-03 | 2021-08-10 | 华中师范大学 | Non-contact identity recognition method and system based on millimeter wave radar |
| CN113468988B (en) * | 2021-06-18 | 2024-04-05 | 南京润楠医疗电子研究院有限公司 | Identification method under multiple pressure states based on ECG (electronic pulse-taking) signals |
| CN114027853B (en) * | 2021-12-16 | 2022-09-27 | 安徽心之声医疗科技有限公司 | QRS complex detection method, device, medium and equipment based on feature template matching |
| CN115512427B (en) * | 2022-11-04 | 2023-04-25 | 北京城建设计发展集团股份有限公司 | User face registration method and system combined with matched biopsy |
| CN115840906B (en) * | 2023-02-13 | 2023-05-30 | 博睿康科技（常州）股份有限公司 | Action potential decomposition method, decomposition model, and signal analysis device |
| CN116028914B (en) * | 2023-03-27 | 2023-06-16 | 深圳市魔样科技有限公司 | Intelligent finger ring identity authentication method and system |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101773394A (en) * | 2010-01-06 | 2010-07-14 | 中国航天员科研训练中心 | Identification method and identification system using identification method |
| 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 |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101268938A (en) * | 2008-04-30 | 2008-09-24 | 董军 | Method and apparatus for electrocardiogram recognition and specification |
-
2014
- 2014-07-01 CN CN201410313480.7A patent/CN104102915B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101773394A (en) * | 2010-01-06 | 2010-07-14 | 中国航天员科研训练中心 | Identification method and identification system using identification method |
| 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 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104102915A (en) | 2014-10-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104102915B (en) | Personal identification method based on ECG multi-template matching under a kind of anomalous ecg state | |
| Fratini et al. | Individual identification via electrocardiogram analysis | |
| CN106473750B (en) | Personal identification method based on photoplethysmographic optimal period waveform | |
| Sasikala et al. | Identification of individuals using electrocardiogram | |
| CN103714281A (en) | Identity recognition method based on electrocardiosignals | |
| Dar et al. | ECG biometric identification for general population using multiresolution analysis of DWT based features | |
| WO2018152711A1 (en) | Electrocardiographic authentication-based door control system and authentication method therefor | |
| Belgacem et al. | Person identification system based on electrocardiogram signal using LabVIEW | |
| Hadiyoso et al. | One-lead electrocardiogram for biometric authentication using time series analysis and Support Vector Machine | |
| Chiu et al. | Discrete wavelet transform applied on personal identity verification with ECG signal | |
| Kuila et al. | Feature extraction of electrocardiogram signal using machine learning classification | |
| Sellami et al. | ECG as a biometric for individual's identification | |
| Tan et al. | ECG biometric identification using wavelet analysis coupled with probabilistic random forest | |
| Gao et al. | Real time ECG characteristic point detection with randomly selected signal pair difference (RSSPD) feature and random forest classifier | |
| CN111783715A (en) | Identity recognition method based on pulse signal feature extraction | |
| CN114638272A (en) | Identity recognition method and device based on fingertip pulse wave signals | |
| Li et al. | A novel abnormal ECG beats detection method | |
| Fattah et al. | An approach for human identification based on time and frequency domain features extracted from ECG signals | |
| Elshahed | Personal identity verification based ECG biometric using non-fiducial features | |
| Bassiouni et al. | A study on the Intelligent Techniques of the ECG-based Biometric Systems | |
| Biran et al. | Automatic qrs detection and segmentation using short time fourier transform and feature fusion | |
| Jia et al. | Decision level fusion for pulse signal classification using multiple features | |
| Loong et al. | Effects of diseased ECG on the robustness of ECG biometric systems | |
| Wang et al. | Research on ECG biometric in cardiac irregularity conditions | |
| Saeed et al. | A Slantlet based statistical features extraction for classification of normal, arrhythmia, and congestive heart failure in electrocardiogram |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |