CN103970975B - Electrocardiogram (ECG) data processing method and system - Google Patents
Electrocardiogram (ECG) data processing method and system Download PDFInfo
- Publication number
- CN103970975B CN103970975B CN201310041662.9A CN201310041662A CN103970975B CN 103970975 B CN103970975 B CN 103970975B CN 201310041662 A CN201310041662 A CN 201310041662A CN 103970975 B CN103970975 B CN 103970975B
- Authority
- CN
- China
- Prior art keywords
- ecg
- electrocardiogram
- cluster
- data
- population
- 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 provides a kind of electrocardiogram (ECG) data processing method, the method includes：Acquire electrocardiogram (ECG) data；It handles the electrocardiogram (ECG) data and obtains corresponding feature vector；Cluster analysis is carried out to the electrocardiogram (ECG) data by described eigenvector and obtains cluster population；The corresponding result of the electrocardiogram (ECG) data is obtained according to the cluster matched template of population.The system comprises：Cardiac electric recording apparatus, for acquiring electrocardiogram (ECG) data；Characteristic processing device obtains corresponding feature vector for handling the electrocardiogram (ECG) data；Clustering apparatus obtains cluster population for passing through described eigenvector to electrocardiogram (ECG) data progress cluster analysis；Cluster kind cluster treating device, for obtaining the corresponding result of the electrocardiogram (ECG) data according to the cluster matched template of population.The accuracy of electrocardiogram (ECG) data analysis can be improved using the present invention.
Description
Technical field
The present invention relates to computer technology, more particularly to a kind of electrocardiogram (ECG) data processing method and system.
Background technology
Heart disease is one of main disease for threatening human life and health, and usually only early discovery early treatment just has larger
It may cure.Electrocardiogram (ECG) data reflects the function of heart, and the diagnosis of heart disease can be realized using electrocardiogram (ECG) data.Electrocardiogram (ECG) data
Collection mainly realized by ordinary electrocardiogram and Holter two ways, wherein, ordinary electrocardiogram
（Electrocardiogram, abbreviation ECG）Diagnosis is for one of important tool of Diagnosing Cardiac disease, quilt in modern medicine
It is widely used in clinical medicine.
Holter is then that one kind can be recorded continuously and compile analysis human heart in activity and rest state for a long time
The method of lower ECG Change, compared with ordinary electrocardiogram, Holter can continuously record up to 10 in 24 hours
Therefore the electrocardiosignal of ten thousand times or so, can effectively improve and the non-standing rhythm of the heart is played the part of, more particularly to the greatly improving property rhythm of the heart
The recall rate of not normal and of short duration myocardial ischemic attacks.
However, the electrocardiogram (ECG) data that Holter is recorded is that repeated and redundant, data volume are big mostly, electrocardio doctor is relied solely on
Observation electrocardiogram can be diagnosed raw or nursing staff for a long time, and often because of the dispersion of asthenopia or attention
And cause missing inspection.
With the development of computer technology, ECG automatic analysis system is employed to carry out electrocardiogram (ECG) data more and more
Analysis carries out long-time observation to avoid electrocardio doctor or nursing staff to electrocardiogram.
But this traditional ECG automatic analysis system is difficult to realize large-scale clinical practice, ecg wave form individual
Otherness is larger, and single global classification device does not have universality in ECG automatic analysis system, that is to say, that single is complete
Office's grader can not all have the electrocardiogram (ECG) data of first patient consistent general high-accuracy, in turn result in electrocardiogram (ECG) data
Inaccuracy analysis.
Invention content
Based on this, it is necessary to provide a kind of electrocardiogram (ECG) data processing method for the accuracy that can improve electrocardiogram (ECG) data analysis.
In addition, it there is a need to the electrocardiogram (ECG) data processing system that a kind of accuracy that can improve electrocardiogram (ECG) data analysis is provided.
A kind of electrocardiogram (ECG) data processing method, includes the following steps：
Acquire electrocardiogram (ECG) data；
It handles the electrocardiogram (ECG) data and obtains corresponding feature vector；
Cluster analysis is carried out to the electrocardiogram (ECG) data by described eigenvector and obtains cluster population；
The corresponding result of the electrocardiogram (ECG) data is obtained according to the cluster matched template of population.
The step of processing electrocardiogram (ECG) data obtains corresponding feature vector in one of the embodiments, includes：
Pre-process the electrocardiogram (ECG) data；
The calculating of phase feature carrying out the solution of wavelet transformation characteristic of field, RR to the pretreated electrocardiogram (ECG) data respectively
With the calculating of T wave fuzzy characteristics, to obtain corresponding characteristic value；
By the eigenvalue cluster into feature vector.
It is described in one of the embodiments, that the electrocardiogram (ECG) data correspondence is obtained according to the cluster matched template of population
Result the step of include：
Each cluster population with template is matched one by one, obtains the matching between the cluster population and the template
Ratio；
Judge whether the matching ratio is more than threshold value, if so,
It obtains and the corresponding result of the template.
It is described in one of the embodiments, that the electrocardiogram (ECG) data correspondence is obtained according to the cluster matched template of population
Result the step of further include：
If the matching ratio determined between each cluster population and all templates is not more than threshold value, to the cluster kind
Group is screened to obtain representative electrocardio beat, and sends the representative electrocardio beat.
It is described in one of the embodiments, that the cluster population is screened to obtain representative electrocardio beat, concurrently
It is further included after the step of sending the representative electrocardio beat：
It receives according to the representative electrocardio beat feedack；
The cluster population is handled according to the feedack.
Described the step of handling the cluster population according to the feedack, includes in one of the embodiments,：
Judge whether the feedack attaches most importance to and cluster instruction, if so, adjustment clustering parameter, to the cluster population
Corresponding electrocardiogram (ECG) data is clustered again, the cluster population of reunion class is obtained, if it is not, then
The template recorded in the feedack is extracted, and shows the result corresponding to the template of the extraction.
A kind of electrocardiogram (ECG) data processing system, including：
Cardiac electric recording apparatus, for acquiring electrocardiogram (ECG) data；
Characteristic processing device obtains corresponding feature vector for handling the electrocardiogram (ECG) data；
Clustering apparatus obtains cluster population for passing through described eigenvector to electrocardiogram (ECG) data progress cluster analysis；
Cluster kind cluster treating device, it is corresponding for obtaining the electrocardiogram (ECG) data according to the cluster matched template of population
As a result.
The characteristic processing device includes in one of the embodiments,：
Preprocessing module, for pre-processing the electrocardiogram (ECG) data；
Characteristic value computing module is asked for carrying out wavelet transformation characteristic of field to the pretreated electrocardiogram (ECG) data respectively
The calculating of phase feature and the calculating of T wave fuzzy characteristics between solution, RR, to obtain corresponding characteristic value；
Vectorial comprising modules, for by the eigenvalue cluster into feature vector.
The cluster kind cluster treating device includes in one of the embodiments,：
Matching module for each cluster population to be matched one by one with template, obtains described clustering population and described
Matching ratio between template；
Matching judgment module, for judging whether the matching ratio is more than threshold value, if so, notice result obtains mould
Block；
The result acquisition module is for acquisition and the corresponding result of the template.
The cluster kind cluster treating device further includes screening module in one of the embodiments,；
If the matching ratio determined between each cluster population and all templates is not more than threshold value, the screening module
The cluster population is screened to obtain representative electrocardio beat, and sends the representative electrocardio beat.
In one of the embodiments, the system also includes：
Information receiver, for receiving according to the representative electrocardio beat feedack；
Feedback processing device, for handling the cluster population according to the feedack.
The feedback processing device further includes in one of the embodiments,：
Signal judgement module clusters instruction, if so, notice reunion class for judging whether the feedack attaches most importance to
Module, if it is not, then notifying extraction module；
The reunion generic module carries out again the electrocardiogram (ECG) data corresponding to the cluster population for adjusting clustering parameter
Cluster, obtains the cluster population of reunion class；
The extraction module is used to extract the template recorded in the feedack, and show the template institute of the extraction
Corresponding result.
Above-mentioned electrocardiogram (ECG) data processing method and system, handle the electrocardiogram (ECG) data collected, corresponding to obtain
Feature vector carries out cluster analysis to obtain cluster population to electrocardiogram (ECG) data by feature vector, that is, realizes electrocardiogram (ECG) data
Classification, so according to corresponding to the cluster matched template of population obtains the electrocardiogram (ECG) data as a result, compared with prior art, greatly
The electrocardiogram (ECG) data of amount realizes the analysis of electrocardiogram (ECG) data with template matches with gradually refining by classifying, and has fully considered electrocardio number
Individual difference between, and then obtain accurate as a result, greatly increasing the accuracy and universality of electrocardiogram (ECG) data analysis,
It is advantageously implemented prolonged cardiac monitoring.
Description of the drawings
Fig. 1 is the flow chart of the electrocardiogram (ECG) data processing method of one embodiment；
Fig. 2 is that processing electrocardiogram (ECG) data obtains the method flow diagram of corresponding feature vector in Fig. 1；
Fig. 3 is to obtain the method flow diagram of the corresponding result of electrocardiogram (ECG) data according to the cluster matched template of population in Fig. 1；
Fig. 4 is the flow chart of the electrocardiogram (ECG) data processing method of another embodiment；
Fig. 5 is the method flow diagram for handling cluster population in Fig. 4 according to feedack；
Fig. 6 is the structure diagram of the electrocardiogram (ECG) data processing system of one embodiment；
Fig. 7 is the structure diagram of characteristic processing device in Fig. 6；
Fig. 8 is the structure diagram of cluster kind cluster treating device in Fig. 6；
Fig. 9 is the structure diagram of the electrocardiogram (ECG) data processing system of another embodiment；
Figure 10 is the structure diagram of feedback processing device in Fig. 9.
Specific embodiment
As shown in Figure 1, in one embodiment, a kind of electrocardiogram (ECG) data processing method includes the following steps：
Step S110 acquires electrocardiogram (ECG) data.
In the present embodiment, after user completes the connection of cardiac diagnosis lead, the electrocardiogram (ECG) data of user is acquired, and acquire
Electrocardiogram (ECG) data is stored in memory, wherein, the SD which is preferably large capacity, to store and transmit speed fast, small
Card（Secure Digital Memory Card, safe digital card）.
Further, the cardiac electric recording apparatus of electrocardio-data collection is used to implement, for example, portable dynamic electrocardiographic recording
Instrument will use multi-lead structure, e.g., two leads, five leads or 12 lead.In preferred embodiment, will also band logical be used to filter
Wave device and notch filter are filtered electrocardiogram (ECG) data, and then eliminate the interference noise in electrocardiogram (ECG) data.
Step S130, processing electrocardiogram (ECG) data obtain corresponding feature vector.
In the present embodiment, feature vector is briefly demonstrated by the feature of electrocardiogram (ECG) data by accurate, and therefore, it is necessary to electrocardio
Data are handled to obtain corresponding feature vector, accurately to characterize electrocardiogram (ECG) data by feature vector.
Step S150 carries out electrocardiogram (ECG) data by feature vector cluster analysis and obtains cluster population.
In the present embodiment, the Accurate classification of electrocardiogram (ECG) data is realized using feature vector, wherein, used cluster analysis side
Method can be K mean cluster method, K central points clustering method or it is expected maximum（EM）Clustering method etc., and then obtain cluster kind
Group and the cluster population corresponding to species information.Cluster population will include in same category of electrocardiogram (ECG) data.
Step S170 obtains the corresponding result of electrocardiogram (ECG) data according to the cluster matched template of population.
In the present embodiment, multiple template is prestored, which is the electrocardiogram (ECG) data sample extracted in advance, each template
It has corresponded to as a result, the situation corresponding to current electrocardiogram (ECG) data can be known by the result, for example, as a result having recorded to electrocardio
The assessment result and solution of data, so that user is that would know that situation and the alleviation of current cardiac by result
The solution of current cardiac situation.It is obtained from pre-stored multiple template with clustering the matched template of groupy phase,
As shown in Fig. 2, in one embodiment, the detailed process of above-mentioned steps S130 is：
Step S131 pre-processes electrocardiogram (ECG) data.
In the present embodiment, the electrocardiogram (ECG) data of acquisition is read, electrocardiogram (ECG) data is filtered.Specifically, it can be filtered by band logical
Wave device realize, for example, the bandpass filter can be 0.1Hz by cutoff frequency, the high-pass filter that stopband attenuation is 30dB and cut
The low-pass filter composition that only frequency is 35Hz, stopband attenuation is 80dB.
Further, after the filtering of electrocardiogram (ECG) data is completed, will also small wavelength-division be carried out to filtered electrocardiogram (ECG) data
Solution.In a preferred embodiment, the wavelet decomposition carried out is four layers of wavelet decomposition, wherein, morther wavelet uses db4（More Bei Xi
Small echo, N=4）, and using A’trous algorithm（Algorithm a’trous）.
Step S133, phase feature carrying out the solution of wavelet transformation characteristic of field, RR to pretreated electrocardiogram (ECG) data respectively
Calculating and T wave fuzzy characteristics calculating, to obtain corresponding characteristic value.
In the present embodiment, the solution that wavelet transformation characteristic of field is carried out to pretreated electrocardiogram (ECG) data is mainly to solve for electrocardio
The statistical information of 4th scale of data wavelet decomposition, following characteristics value is calculated：
C2x(k)=E { x (n) x (n+k) }
C3x(k, l)=E { x (n) x (n+k) x (n+l) }
C4x(k, l, m)=E { x (n) x (n+k) x (n+l) x (n+m) }
-C2x(k)C2x(m-l)-C2x(l)C2x(m-k)-C2x(m)C2x(l-k)
Wherein, x (n) represents the 4th layer of component of electrocardiogram (ECG) data wavelet decomposition, and expectation is asked in E expressions, k, l, m, during expression
Between postpone.
The calculating of phase feature is carried out according to the R waves position of electrocardiogram (ECG) data between RR, before the computation, should be to pretreatment
Electrocardiogram (ECG) data afterwards carries out R wave detections.Specifically, the R wave energies due to electrocardiogram (ECG) data are concentrated mainly on the sensitivity of wavelet decomposition
On scale, therefore, R waves crest value point will be extracted in third scale, according to the Wavelet Detection principle of singular point, R waves correspond to the
Extreme value pair on three scales, and R waves wave crest point will be corresponding to the zero crossing of extreme value pair.
It detects to have obtained the position of R waves, i.e. extreme value pair on third scale by R waves, at this point, the following formula will be defined：
RR [i]=R [i]-R [i-1]
DiffRR [i]=RR [i]-RR [i-1]
RR [i], RR [i-1], diffRRmean [i, 1], RRmean [i, 5] and RRmean are calculated according to above-mentioned formula
Characteristic values such as [i, 20], and then as the Partial Feature of cluster analysis.
Wherein R [i] represents in electrocardiogram (ECG) data that the time that i-th of R wave occurs, RR [i] represents i-th of R wave and (i-1)-th R
The Time Intervals of wave, diffRR [i] represent the variation of phase between i-th of RR, and diffRRmean [i, k] represents k after i-th of R wavefront
The average value of phase changing value between a RR, RRmean [i, k] represent the average value of phase between k RR after i-th of R wavefront.
Between the calculating of T waves fuzzy characteristics is carried out, T wave detections will be carried out first.The energy of T waves mainly collects in electrocardiogram (ECG) data
In on the 4th scale of wavelet decomposition, therefore, obtained R waves position is detected according to R waves, on the 4th scale of wavelet decomposition
The extreme value pair and zero crossing for corresponding to T waves position and T wave wave crests respectively, example are searched in corresponding search window according to R waves position
Such as, which can be since current QRS wave terminal, until the distance of phase terminates between 0.5 times of RR.
Calculate R waves position and T waves position apart from spacing interval_RT, and detect T wave morphologies, correspond to T waves respectively and fall
It puts, normal T wave and existing inversion also have the inversion of positive T waves, the normal and two-way fuzzy quantity as defined.
Step S135, by eigenvalue cluster into feature vector.
In the present embodiment, it is obtained that phase feature between the obtained characteristic value of wavelet transformation characteristic of field, calculating RR will be solved
Characteristic value and calculate the obtained eigenvalue cluster of T wave fuzzy characteristics into electrocardiogram (ECG) data feature vector.
As shown in figure 3, in one embodiment, the detailed process of above-mentioned steps S170 is：
Each cluster population with template is matched, obtains between cluster population and template by step S171 one by one
With ratio.
In the present embodiment, cluster population obtained to cluster analysis is subjected to template matches one by one, it is each poly- to obtain
The matched template of class population institute.
Specifically, each cluster population is matched one by one with pre-stored multiple template, to obtain and the cluster kind
Matching ratio between group and each template.
Step S173, judges whether matching ratio is more than threshold value, if so, S175 is entered step, if it is not, then, then entering
Step S177.
In the present embodiment, if the matching ratio determined between cluster population and a certain template is more than threshold value, illustrate this
Template is with clustering the matched template of groupy phase, if the matching ratio determined between cluster population and all templates is not big
In threshold value, then illustrate currently identify the matched template of cluster population institute.
Step S175 is obtained and the corresponding result of the template.
In the present embodiment, since each template has corresponding as a result, therefore, determining cluster population and a certain template
Between matching ratio be more than threshold value after, will directly acquire it is corresponding to the template as a result, this result be cluster population
Comprising the situation that is presently in of electrocardiogram (ECG) data.
Step S177 is screened to obtain representative electrocardio beat, and send representative electrocardio beat to cluster population.
In the present embodiment, the center to obtain cluster population is ranked up according to characteristic value, in including for cluster population
Several nearest electrocardio beats of the central point of the selected distance cluster population and in the cluster population in electrocardiogram (ECG) data
Several farthest electrocardio beats of heart point, these electrocardio beats chosen are representative electrocardio beat.
As shown in figure 4, it in one embodiment, is further included after above-mentioned steps S177：
Step S210 is received according to representative electrocardio beat feedack.
In the present embodiment, after the representative electrocardio beat that screening obtains is sent in the form of data packet, distal end
User will receive the data packet, and parse the data packet after by inputting identity and password progress identification
With obtain can not with the cluster population of template matches, at this point, will obtain remote subscriber according to parsing obtain cluster population institute it is defeated
The information entered, and the information is packaged, and feed back the information in a manner of online or is offline by internet, which includes can
To be the instruction of reunion class or the information of a certain template specified for cluster population.
Step S230 handles cluster population according to feedack.
In the present embodiment, according to feedack to reunion class or acquisition can not be carried out with the cluster population of template matches
It is corresponding with specified template as a result, so that electrocardiogram (ECG) data can also obtain accurately in the case where template matches can not be handled
As a result, and then constantly studying positive template matching process.
As shown in figure 5, in one embodiment, the detailed process of above-mentioned steps S230 is：
Step S231 judges whether feedack attaches most importance to cluster instruction, if so, enter step S233, if it is not, then into
Enter step S235.
In the present embodiment, instruction is clustered if determining feedack and attaching most importance to, this can not be gathered with template matches
Class population is clustered again, if determine feedack be not reunion class instruction when, illustrate to record in feedack
Template corresponding to cluster population.
Step S233 adjusts clustering parameter, and the electrocardiogram (ECG) data corresponding to cluster population is clustered again, is met again
The cluster population of class.
In the present embodiment, the clustering parameter adjusted can be population number for increasing cluster etc., so that the reunion on backstage
Class process is more accurate, and after cluster again is completed, the cluster population that counterweight as described above is clustered carries out template matches,
To obtain and cluster the matched template of population, and then obtain corresponding as a result, details are not described herein.
Step S235 extracts the template recorded in feedack, and shows the result corresponding to the template of extraction.
In the present embodiment, the result for clustering population is shown, the information exchange between realization and user.Specifically
, display mode can be the mode of screen.
As shown in fig. 6, in one embodiment, a kind of electrocardiogram (ECG) data processing system includes cardiac electric recording apparatus 110, feature
Processing unit 130, clustering apparatus 150 and cluster kind cluster treating device 170.
Cardiac electric recording apparatus 110, for acquiring electrocardiogram (ECG) data.
In the present embodiment, after user completes the connection of cardiac diagnosis lead, cardiac electric recording apparatus 110 acquires the heart of user
Electric data, and the electrocardiogram (ECG) data acquired is stored in memory, wherein, which is preferably large capacity, stores and transmits speed
Spend fast, small SD card（Secure Digital Memory Card, safe digital card）.
Further, the cardiac electric recording apparatus 110 of electrocardio-data collection is used to implement, for example, portable dynamic electrocardio is remembered
Instrument is recorded, multi-lead structure, e.g., two leads, five leads or 12 lead will be used.In preferred embodiment, cardiac electric recording apparatus
110 include bandpass filter and notch filter, to be filtered to electrocardiogram (ECG) data, and then eliminate the interference in electrocardiogram (ECG) data
Noise.
Characteristic processing device 130 obtains corresponding feature vector for handling electrocardiogram (ECG) data.
In the present embodiment, feature vector is briefly demonstrated by the feature of electrocardiogram (ECG) data by accurate, at feature
Reason device 130 is handled electrocardiogram (ECG) data to obtain corresponding feature vector, accurately to characterize electrocardio by feature vector
Data.
Clustering apparatus 150 obtains cluster population for passing through feature vector to electrocardiogram (ECG) data progress cluster analysis.
In the present embodiment, the application feature vector of clustering apparatus 150 realizes the Accurate classification of electrocardiogram (ECG) data, wherein, it is used
Clustering method can be K mean cluster method, K central points clustering method or it is expected maximum（EM）Clustering method etc., into
And obtain cluster population and the cluster population corresponding to species information.Cluster population will include in the same category of heart
Electric data.
Cluster kind cluster treating device 170, for obtaining the corresponding result of electrocardiogram (ECG) data according to the cluster matched template of population.
In the present embodiment, multiple template is prestored, which is the electrocardiogram (ECG) data sample extracted in advance, each template
It has corresponded to as a result, the situation corresponding to current electrocardiogram (ECG) data can be known by the result, for example, as a result having recorded to electrocardio
The assessment result and solution of data, so that user is that would know that situation and the alleviation of current cardiac by result
The solution of current cardiac situation.It is obtained from pre-stored multiple template with clustering the matched template of groupy phase,
As shown in fig. 7, in one embodiment, features described above processing unit 130 includes preprocessing module 131, characteristic value
Computing module 133 and vectorial comprising modules 135.
Preprocessing module 131, for pre-processing electrocardiogram (ECG) data.
In the present embodiment, preprocessing module 131 reads the electrocardiogram (ECG) data of acquisition, and electrocardiogram (ECG) data is filtered.Specifically,
It can be realized by bandpass filter, for example, the bandpass filter can be 0.1Hz by cutoff frequency, the height that stopband attenuation is 30dB
The low-pass filter composition that bandpass filter and cutoff frequency are 35Hz, stopband attenuation is 80dB.
Further, after the filtering of electrocardiogram (ECG) data is completed, preprocessing module 131 will also be to filtered electrocardio number
According to progress wavelet decomposition.In a preferred embodiment, the wavelet decomposition carried out is four layers of wavelet decomposition, wherein, morther wavelet is adopted
Use db4（More western small echos of shellfish, N=4）, and using A’trous algorithm（Algorithm a’trous）.
Characteristic value computing module 133 is asked for carrying out wavelet transformation characteristic of field to pretreated electrocardiogram (ECG) data respectively
Solution, RR ask the calculating of phase feature and the calculating of T wave fuzzy characteristics, to obtain corresponding characteristic value.
In the present embodiment, characteristic value computing module 133 carries out wavelet transformation characteristic of field to pretreated electrocardiogram (ECG) data
The statistical information for the 4th scale for being mainly to solve for electrocardiogram (ECG) data wavelet decomposition is solved, following characteristics value is calculated：
C2x(k)=E { x (n) x (n+k) }
C3x(k, l)=E { x (n) x (n+k) x (n+l) }
C4x(k, l, m)=E { x (n) x (n+k) x (n+l) x (n+m) }
-C2x(k)C2x(m-l)-C2x(l)C2x(m-k)-C2x(m)C2x(l-k)
Wherein, x (n) represents the 4th layer of component of electrocardiogram (ECG) data wavelet decomposition, and expectation is asked in E expressions, k, l, m, during expression
Between postpone.
The calculating of phase feature is carried out according to the R waves position of electrocardiogram (ECG) data between RR in characteristic value computing module 133, is being counted
Before calculation, characteristic value computing module 133 should carry out R wave detections to pretreated electrocardiogram (ECG) data.Specifically, due to electrocardio number
According to R wave energies be concentrated mainly on the sensitive scale of wavelet decomposition, therefore, characteristic value computing module 133 will be in third scale
R waves crest value point is extracted, according to the Wavelet Detection principle of singular point, R waves correspond to the extreme value pair on third scale, and R wave waves
Peak dot will be corresponding to the zero crossing of extreme value pair.
Characteristic value computing module 133 detects to have obtained the position of R waves, i.e. extreme value pair on third scale by R waves, this
When, the following formula will be defined：
RR [i]=R [i]-R [i-1]
DiffRR [i]=RR [i]-RR [i-1]
RR [i], RR [i-1], diffRRmean [i, 1], RRmean [i, 5] and RRmean are calculated according to above-mentioned formula
Characteristic values such as [i, 20], and then as the Partial Feature of cluster analysis.
Wherein R [i] represents in electrocardiogram (ECG) data that the time that i-th of R wave occurs, RR [i] represents i-th of R wave and (i-1)-th R
The Time Intervals of wave, diffRR [i] represent the variation of phase between i-th of RR, and diffRRmean [i, k] represents k after i-th of R wavefront
The average value of phase changing value between a RR, RRmean [i, k] represent the average value of phase between k RR after i-th of R wavefront.
Between the calculating of T waves fuzzy characteristics is carried out, vectorial comprising modules 135 will carry out T wave detections first.In electrocardiogram (ECG) data
The energy of T waves is focused primarily upon on the 4th scale of wavelet decomposition, therefore, obtained by vectorial comprising modules 135 are detected according to R waves
R waves position, searched in corresponding search window according to R waves position on the 4th scale of wavelet decomposition and correspond to T waves position respectively
The extreme value pair and zero crossing with T wave wave crests are put, for example, the search window can be since the current QRS wave terminal, until 0.5 times
The distance of phase terminates between RR.
Vectorial comprising modules 135 calculate R waves position and T waves position apart from spacing interval_RT, and detect T waveforms
State, correspond to respectively T waves be inverted, normal T wave and existing inversion also have positive T waves inversion, it is normal and it is two-way be definition
Fuzzy quantity.
Vectorial comprising modules 135, for by eigenvalue cluster into feature vector.
In the present embodiment, vectorial comprising modules 135 will be solved between the obtained characteristic value of wavelet transformation characteristic of field, calculating RR
The obtained characteristic value of phase feature and calculate the obtained eigenvalue cluster of T wave fuzzy characteristics into electrocardiogram (ECG) data feature vector.
As shown in figure 8, in one embodiment, above-mentioned cluster kind cluster treating device 170 includes matching module 171, matching
Judgment module 173, result acquisition module 175 and screening module 177.
Matching module 171, for each cluster population to be matched one by one with template, obtain clustering population and template it
Between matching ratio.
In the present embodiment, cluster population obtained to cluster analysis is carried out template matches by matching module 171 one by one, with
Obtain the matched template of each cluster population institute.
Specifically, matching module 171 matches each cluster population with pre-stored multiple template one by one, to obtain
With the matching ratio between the cluster population and each template.
Matching judgment module 173, for judging whether matching ratio is more than threshold value, if so, notice result acquisition module
175, if it is not, then notifying screening module 177.
In the present embodiment, if the matching ratio that matching judgment module 173 is determined between cluster population and a certain template is big
In threshold value, then it is with clustering the matched template of groupy phase, if matching judgment module 173 determines cluster population to illustrate the template
Matching ratio between all templates is not more than threshold value, then illustrates currently identify the matched mould of cluster population institute
Plate.
As a result acquisition module 175, for obtaining and the corresponding result of the template.
In the present embodiment, since each template has corresponding as a result, therefore, determining cluster population and a certain template
Between matching ratio be more than threshold value after, as a result acquisition module 175 will directly acquire corresponding to the template as a result, this knot
Fruit is to cluster the situation that the electrocardiogram (ECG) data that population is included is presently in.
Screening module 177, for being screened to obtain representative electrocardio beat, and send representative electrocardio to cluster population
Beat.
In the present embodiment, screening module 177 is ranked up the center to obtain cluster population according to characteristic value, in cluster kind
Several nearest electrocardio beats of the central point of the selected distance cluster population and poly- apart from this in the electrocardiogram (ECG) data included of group
Several farthest electrocardio beats of the central point of class population, these electrocardio beats chosen are representative electrocardio beat.
As shown in figure 9, in one embodiment, above-mentioned electrocardiogram (ECG) data processing system further includes 210 He of information receiver
Feedback processing device 230.
Information receiver 210, for receiving according to representative electrocardio beat feedack.
In the present embodiment, above-mentioned electrocardiogram (ECG) data processing system further includes auxiliary device, the representative electrocardio that screening is obtained
Beat is sent out in the form of data packet to the auxiliary device of distal end, and remote subscriber in auxiliary device by passing through input at this time
After identity and password carry out identification, the data packet will be received, and parses the data packet so that obtain can not be with mould
The matched cluster population of plate, at this point, the information that acquisition remote subscriber is inputted according to the cluster population that parsing obtains, and be packaged
The information, and the information is fed back in a manner of online or is offline by internet, which includes being the instruction of reunion class,
Can also be the information that a certain template is specified for cluster population.
Feedback processing device 230, for handling cluster population according to feedack.
In the present embodiment, feedback processing device 230 according to feedack to can not with the cluster populations of template matches into
Row reunion class obtains corresponding with the template specified as a result, so that the situation that electrocardiogram (ECG) data can not be handled in template matches
Under can also obtain accurately as a result, and then constantly studying positive template matching process.
As shown in Figure 10, in one embodiment, above-mentioned feedback processing device 230 includes signal judgement module 231, meets again
Generic module 233 and extraction module 235.
Signal judgement module 231 clusters instruction, if so, notice reunion class for judging whether feedack attaches most importance to
Module 233, if it is not, then notifying extraction module 235.
In the present embodiment, if signal judgement module 231 determine feedack attach most importance to cluster instruction, can not to this
Clustered again with the cluster population of template matches, if determine feedack be not reunion class instruction when, illustrate instead
The template corresponding to cluster population is had recorded in the information of feedback.
Reunion generic module 233 for adjusting clustering parameter, gathers the electrocardiogram (ECG) data corresponding to cluster population again
Class obtains the cluster population of reunion class.
In the present embodiment, the clustering parameter that reunion generic module 233 is adjusted can be population number for increasing cluster etc., so that
Backstage reunion class process it is more accurate, after cluster again is completed, cluster population that counterweight as described above is clustered
Template matches are carried out, to obtain and cluster the matched template of population, and then obtain corresponding as a result, details are not described herein.
Extraction module 235 for extracting the template recorded in feedack, and shows the knot corresponding to the template of extraction
Fruit.
In the present embodiment, the result for clustering population is shown, the information exchange between realization and user.Specifically
, display mode can be the mode of screen.
In above-mentioned electrocardiogram (ECG) data processing system, characteristic processing device, clustering apparatus, cluster kind cluster treating device, information connect
Receiving apparatus and feedback processing device constitute user terminal ecg analysis system, run in the computer of user, and auxiliary with distal end
Device is helped to realize information exchange, i.e. a plurality of clients ecg analysis system is interacted with auxiliary device.
Above-mentioned electrocardiogram (ECG) data processing method and system, handle the electrocardiogram (ECG) data collected, corresponding to obtain
Feature vector carries out cluster analysis to obtain cluster population to electrocardiogram (ECG) data by feature vector, that is, realizes electrocardiogram (ECG) data
Classification, so according to corresponding to the cluster matched template of population obtains the electrocardiogram (ECG) data as a result, compared with prior art, greatly
The electrocardiogram (ECG) data of amount realizes the analysis of electrocardiogram (ECG) data, and then accurately tied with gradually refining by classification and template matches
Fruit greatly increases the accuracy and universality of electrocardiogram (ECG) data analysis, is advantageously implemented prolonged cardiac monitoring.
One of ordinary skill in the art will appreciate that realizing all or part of flow in above-described embodiment method, being can be with
Relevant hardware is instructed to complete by computer program, the program can be stored in a computer read/write memory medium
In, the program is when being executed, it may include such as the flow of the embodiment of above-mentioned each method.Wherein, the storage medium can be magnetic
Dish, CD, read-only memory（Read-Only Memory, ROM）Or random access memory（Random Access
Memory, RAM）Deng.
Embodiment described above only expresses the several embodiments of the present invention, and description is more specific and detailed, but simultaneously
Cannot the limitation to the scope of the claims of the present invention therefore be interpreted as.It should be pointed out that for those of ordinary skill in the art
For, without departing from the inventive concept of the premise, various modifications and improvements can be made, these belong to the guarantor of the present invention
Protect range.Therefore, the protection domain of patent of the present invention should be determined by the appended claims.
Claims (2)
1. a kind of electrocardiogram (ECG) data processing system, which is characterized in that including：
Cardiac electric recording apparatus, for acquiring electrocardiogram (ECG) data；
Characteristic processing device obtains corresponding feature vector for handling the electrocardiogram (ECG) data；
Clustering apparatus obtains cluster population for passing through described eigenvector to electrocardiogram (ECG) data progress cluster analysis；
Cluster kind cluster treating device, for obtaining the corresponding knot of the electrocardiogram (ECG) data according to the cluster matched template of population
Fruit；
The cluster kind cluster treating device includes：
Matching module for each cluster population to be matched one by one with template, obtains the cluster population and the template
Between matching ratio；
Matching judgment module, for judging whether the matching ratio is more than threshold value；
Screening module, if not being more than threshold value, institute for determining the matching ratio between each cluster population and all templates
It states screening module the cluster population is screened to obtain representative electrocardio beat, and sends the representative electrocardio beat；
The system also includes：
Information receiver, for receiving according to the representative electrocardio beat feedack；
Feedback processing device, for handling the cluster population according to the feedack；
The feedback processing device further includes：
Signal judgement module clusters instruction, if so, notice reunion class mould for judging whether the feedack attaches most importance to
Block, if it is not, then notifying extraction module；
The reunion generic module gathers the electrocardiogram (ECG) data corresponding to the cluster population for adjusting clustering parameter again
Class obtains the cluster population of reunion class；
The extraction module is used to extract the template recorded in the feedack, and show corresponding to the template of the extraction
Result；
The matching judgment module is additionally operable to when the matching ratio is more than threshold value, then notify result acquisition module；The knot
Fruit acquisition module is for acquisition and the corresponding result of the template.
2. electrocardiogram (ECG) data processing system according to claim 1, which is characterized in that the characteristic processing device includes：
Preprocessing module, for pre-processing the electrocardiogram (ECG) data；
Characteristic value computing module, for the pretreated electrocardiogram (ECG) data is carried out respectively the solution of wavelet transformation characteristic of field,
The calculating of phase feature and the calculating of T wave fuzzy characteristics between RR, to obtain corresponding characteristic value；
Vectorial comprising modules, for by the eigenvalue cluster into feature vector.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310041662.9A CN103970975B (en) | 2013-02-02 | 2013-02-02 | Electrocardiogram (ECG) data processing method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310041662.9A CN103970975B (en) | 2013-02-02 | 2013-02-02 | Electrocardiogram (ECG) data processing method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103970975A CN103970975A (en) | 2014-08-06 |
| CN103970975B true CN103970975B (en) | 2018-07-06 |
Family
ID=51240465
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310041662.9A Active CN103970975B (en) | 2013-02-02 | 2013-02-02 | Electrocardiogram (ECG) data processing method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103970975B (en) |
Families Citing this family (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104398254B (en) * | 2014-11-14 | 2017-02-22 | 中国科学院深圳先进技术研究院 | Electrocardiogram analyzing system, electrocardiogram analyzing equipment and electrocardiogram predication model acquisition equipment |
| 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 |
| CN105748066B (en) * | 2016-03-03 | 2018-03-16 | 深圳竹信科技有限公司 | A kind of extracting method and device of electro-cardiologic signal waveforms characteristic point |
| CN106618552B (en) * | 2016-12-30 | 2019-08-09 | 清华大学 | A kind of detection model acquisition methods and device |
| CN107092789A (en) * | 2017-04-18 | 2017-08-25 | 成都琅瑞医疗技术股份有限公司 | Electrocardiogram (ECG) data analysis tool system and application method |
| CN107622259B (en) * | 2017-10-20 | 2019-12-24 | 山东大学 | T wave detection method, electrocardiogram data analysis method and device |
| CN109171706B (en) * | 2018-09-30 | 2020-12-29 | 南京信息工程大学 | Electrocardiosignal denoising method and system based on classification matching and fractional order diffusion |
| CN112635047B (en) * | 2020-09-22 | 2021-09-07 | 广东工业大学 | Robust electrocardio R peak detection method |
| CN112270359A (en) * | 2020-10-29 | 2021-01-26 | 山西三友和智慧信息技术股份有限公司 | One-dimensional sequence ascending clustering method and system |
| CN112773382A (en) * | 2021-01-20 | 2021-05-11 | 钛虎机器人科技(上海)有限公司 | Myoelectricity sensing method and system with user self-adaption capability |
-
2013
- 2013-02-02 CN CN201310041662.9A patent/CN103970975B/en active Active
Non-Patent Citations (7)
| Title |
|---|
| Personalized long-term ECG classification: A systematic approach;Serkan Kiranyaz et al;《Expert Systems with Applications》;20110430;第38卷(第4期);参见图1，正文第3222页第2栏第7段，第3223页第1栏第2段、第2栏第2-3段、第3224页第1栏第1段， * |
| 一种在线动态心电信号聚类分析策略;于天;《中国优秀硕士学位论文全文数据库 信息科技辑》;20120215;第2012年卷(第2期);I138-1917 * |
| 动态心电图波形改进分类策略研究;曹国超;《中国优秀硕士学位论文全文数据库 信息科技辑》;20100715;第2010年卷(第7期);I138-691 * |
| 基于模板匹配和特征识别相结合的心室期前收缩波形分类算法;段会龙等;《航天医学与医学工程》;20020430;第15卷(第2期);98-102 * |
| 引入异常心电节律分析的心拍分类算法研究;汤征;《中国优秀硕士学位论文全文数据库 医药卫生科技辑》;20081115;第2008年卷(第11期);E062-2 * |
| 心电信号自动分析关键技术研究;季虎;《中国博士学位论文全文数据库 信息科技辑》;20080715;第2008年卷(第7期);I140-17 * |
| 心电自动分析系统;王鸿鹏等;《北京生物医学工程》;20070831;第26卷(第4期);386-394 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103970975A (en) | 2014-08-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103970975B (en) | Electrocardiogram (ECG) data processing method and system | |
| EP3698707B1 (en) | Electrocardiogram information dynamic monitoring system, computer program and computer readable storage medium | |
| US11350868B2 (en) | Electrocardiogram information processing method and electrocardiogram workstation system | |
| CN107951485B (en) | Ambulatory ECG analysis method and apparatus based on artificial intelligence self study | |
| US11783924B2 (en) | ECG information processing method and ECG workstation | |
| Diker et al. | Intelligent system based on Genetic Algorithm and support vector machine for detection of myocardial infarction from ECG signals | |
| Mei et al. | Automatic atrial fibrillation detection based on heart rate variability and spectral features | |
| CN105550653B (en) | Electrocardio lead intelligent selection method and system | |
| CN106066933B (en) | The method for obtaining atrial fibrillation identification artificial neural network weight value matrix | |
| CN113095302B (en) | Depth model for arrhythmia classification, method and device using same | |
| Fang et al. | Dual-channel neural network for atrial fibrillation detection from a single lead ECG wave | |
| CN109288515B (en) | Periodicity monitoring method and device based on premature beat signal in wearable electrocardiosignal | |
| WO2015171804A1 (en) | Detecting artifacts in a signal | |
| Zhao et al. | PVC recognition for wearable ECGs using modified frequency slice wavelet transform and convolutional neural network | |
| CN110811673A (en) | Heart sound analysis system based on probabilistic neural network model | |
| CN114098754B (en) | Atrial fibrillation signal preprocessing method, detection system, equipment and storage medium | |
| CN111345815B (en) | Method, device, equipment and storage medium for detecting QRS wave in electrocardiosignal | |
| Haiying et al. | Real-time automatic ecg diagnosis method dedicated to pervasive cardiac care | |
| Sanamdikar et al. | Using the GAN method, analysis several characteristics of the ECG signal in order to detect cardiac arrhythmia | |
| Zhou et al. | End-Edge Coordinated Multiview Deep Learning for Time-Efficient Atrial Fibrillation Detection | |
| CN113768516B (en) | Electrocardiogram abnormality degree detection method and system based on artificial intelligence | |
| AKS et al. | Estimation of Heart Rate using PPG Sensor at the Suprasternal Notch | |
| Nanarkar et al. | A Survey on Classification and identification of Arrhythmia using Machine Learning techniques | |
| CN118380021B (en) | Stethoscope audio recognition method, system and device based on deep neural network | |
| CN114224352B (en) | Lead falling detection system and method of electrocardiograph based on RBF neural network |
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 |