CN109157211A - A kind of portable cardiac on-line intelligence monitoring diagnosis system design method - Google Patents
A kind of portable cardiac on-line intelligence monitoring diagnosis system design method Download PDFInfo
- Publication number
- CN109157211A CN109157211A CN201810921525.7A CN201810921525A CN109157211A CN 109157211 A CN109157211 A CN 109157211A CN 201810921525 A CN201810921525 A CN 201810921525A CN 109157211 A CN109157211 A CN 109157211A
- Authority
- CN
- China
- Prior art keywords
- sample
- scatter plot
- rdr
- parameter
- component analysis
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
- A61B5/316—Modalities, i.e. specific diagnostic methods
- A61B5/318—Heart-related electrical modalities, e.g. electrocardiography [ECG]
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7225—Details of analog processing, e.g. isolation amplifier, gain or sensitivity adjustment, filtering, baseline or drift compensation
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7235—Details of waveform analysis
- A61B5/7264—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems
Landscapes
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Animal Behavior & Ethology (AREA)
- Veterinary Medicine (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Heart & Thoracic Surgery (AREA)
- Medical Informatics (AREA)
- Molecular Biology (AREA)
- Surgery (AREA)
- Artificial Intelligence (AREA)
- General Health & Medical Sciences (AREA)
- Public Health (AREA)
- Pathology (AREA)
- Signal Processing (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physiology (AREA)
- Psychiatry (AREA)
- Power Engineering (AREA)
- Cardiology (AREA)
- Evolutionary Computation (AREA)
- Fuzzy Systems (AREA)
- Mathematical Physics (AREA)
- Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)
Abstract
The invention discloses a kind of portable cardiac on-line intelligence monitoring diagnosis system design methods.Including being filtered denoising to electrocardiosignal, R crest value position is extracted;RdR scatter plot is drawn using heartbeat interval;RdR scatter plot is zoomed in and out, grayscale image is changed into, image data is normalized;The scatter plot sample of acquisition is marked;Sample is sampled, randomly selects 85% data as training sample;Parameter is set;Approximate base, characteristic value, feature vector are solved to every class sample respectively;The SPE of each test sample is calculated separately, value and the SPE difference reckling of certain class sample calculate accuracy rate compared with concrete class for the prediction classification of test sample;Disaggregated model parameter is obtained, and the disaggregated model of acquisition is used in cardiac diagnosis system.The present invention carries out cardioelectric monitor diagnosis using the method for integrating sparse kernel principle component analysis, has satisfactory diagnosis recognition effect.
Description
Technical field
The present invention relates to a kind of portable cardiac on-line intelligence monitoring diagnosis system design methods, belong to intelligent medical technology
Field.
Background technique
Cardiovascular disease is unhealthful important killer, in June, 2017 " Chinese cardiovascular disease report 2016 " publication.Report
It points out: currently, cardiovascular death accounts for the first place of the total cause of death of urban and rural residents, and 10 years cardiovascular disease numbers of patients from now on
It will rapid growth.In addition, will be about 23,300,000 people according to the report of the World Health Organization to the year two thousand thirty and died of angiocarpy
Disease.In face of this trend, the early diagnosis and prevention of cardiovascular disease are particularly important.
Traditional diagnostic method is patient to hospital, and doctor carries out ECG examination to it, and then provides diagnostic result, is appointed
It is engaged in heavy and doctor is needed to have clinical experience abundant and professional knowledge;And rare medical resource, it is difficult to meet substantial amounts
The requirement of PATIENT POPULATION.In order to improve medical efficiency, convenience and agility, there is automated diagnostic technology, assist doctor
It is diagnosed.
Heart rate variability refers to the time-variance number between cardiac cycle, and research object is cardiac cycle rather than heart rate.
The heart rate of people be not it is unalterable, there is small time differences between heartbeat twice, calculate the difference of cardiac cycle,
It can be appreciated that heart rate variability (Heart rate variability, HRV).
Heart rate variability can assess the influence of cardiac sympathetic nerve and parasympathetic nerve to cardiovascular activity, contain the heart
Bulk information in terms of blood vessel.Clinical research shows that the reduction of heart rate variability is the hearts such as myocardial infarction, hypertension, angina pectoris
The symptom of vascular morbidity.Therefore, the research of heart rate variability, in evaluation systema cariovasculare functional, prediction cardiovascular disease
Breaking-out, and have great importance for the early diagnosis of cardiovascular disease.
Poincare scatter plot is a kind of important research method of heart rate variability: being existed by using continuous heartbeat interval
Graphing in rectangular coordinate system reflects the variation of adjacent heartbeat interval, can show the feature of heartbeat interval;Poincare dissipates
For point diagram there are many form, including comet formation, sector etc., different shapes reflects different heart states.
Although Poincare scatter plot is a kind of effective heart rate variance analyzing method, it can not be embodied at any time
Between the trend that changes, its heart rate variability cannot be embodied well for certain cardiovascular diseases, physical condition etc.
Matter.Then, some scholars propose improvement strategy, i.e. first order difference plot, are drawn by the difference of adjacent heartbeat interval
Scatter plot.However, this method is lost the absolute value information of original heartbeat interval again;Therefore, and scholar combines the two
Get up, propose a kind of RdR scatter plot, come with this while reflecting heartbeat interval and its variation.
Currently, many for the heart rate variability analysis of different cardiovascular diseases;But there is no come from according to scatter plot
Different cardiovascular diseases is distinguished in dynamic identification, realizes electrocardio automated diagnostic.
Summary of the invention
In view of the above existing problems in the prior art, the present invention provides a kind of portable cardiac on-line intelligence monitoring, diagnosing systems
Unite design method, by integrate sparse kernel principle component analysis method can automatic recognition classification scatter plot, to heart rate variability make
Analysis provides to realize automated diagnostic, the waste alleviated medical resource in short supply, reduce medical resource, improving medical efficiency
Basis.
To achieve the goals above, the present invention provides a kind of portable cardiac on-line intelligence monitoring diagnosis system design side
Method, specific steps are as follows:
Step 1) acquires electrocardiosignal, is filtered denoising to electrocardiosignal, extracts R crest value position;
Step 2) draws electrocardio RdR scatter plot using heartbeat interval;
Step 3) zooms in and out electrocardio RdR scatter plot, zooms to same specification, changes into grayscale image, and to image data
It is normalized, to reduce calculation amount;
The scatter plot sample of acquisition is marked in step 4);
Step 5) samples sample, randomly selects 85% data as training sample;
Parameter is arranged in step 6), and parameter includes the value of the error parameter of approximate base, gaussian kernel function parameter and control limit;
Step 7) solves approximate base, characteristic value, feature vector to every class sample respectively;
Step 8) calculates separately the SPE of each test sample, and the SPE difference reckling of value and certain class sample is test specimens
This prediction classification calculates accuracy rate compared with concrete class, the step 9) if meeting the requirements, otherwise return step 6) again
Parameter training is set;
Step 9) obtains disaggregated model parameter, and the disaggregated model of acquisition is used in cardiac diagnosis system.
Further, using integrating sparse kernel principle component analysis method for the Classification and Identification of the electrocardio RdR scatter plot.
RdR scatter plot is a kind of heart rate variance analyzing method, can embody its trend changed over time.
Further, sparse kernel principle component analysis method will be integrated and is used for Classification and Identification to the electrocardio RdR scatter plot, first
Foundation integrates sparse kernel principle component analysis model, then selects Con trolling index by square prediction error method.
Integrating sparse kernel principle component analysis is a kind of unsupervised machine learning algorithm, integrates the base of sparse kernel principle component analysis
This method is as follows:
Principal component analysis is a kind of typical unsupervised algorithm, is usually used in solving the linear problem of luv space, and in order to
The nonlinear problem of luv space is solved with linear method in feature space, B.Scholkopf et al. proposes core principal component
It analyzes (Kernel Principal Component Analysis, KPCA).It defines from luv space RnTo feature space F's
Nonlinear Mapping:If given sample X={ x1,…,xN},xi∈Rn, then pass throughMapping can obtain one
Group vectorAssuming that this group of vector meetsThen the Correlation Matrix in feature space is
If this group of vectorIt can then enableKnown toMeet
Condition, instead of in formulaThen KPCA problem, which can be converted to, seeks Correlation Matrix in feature spaceEigenvalue λ, that is, feature
Vector
Wherein,It is the linear combination of sample, enablesɑ=[ɑ1,…,ɑN]T, thenWhenWhen cannot explicitly obtaining, kernel function is introduced, ifFirstly the need of calculating:
K=ΦTΦ;
Wherein, matrix K is the matrix of NxN, also referred to as nuclear matrix.Then problem is converted are as follows:
K ɑ=N λ ɑ;
Wherein, ɑ=[ɑ1,…,ɑN].WhenThe process of centralization can the operation directly on K:
Wherein,MeetIt is 1 matrix of a NxN.Assuming that obtaining
1 >=λ of eigenvalue λ 2 >=... λ n and its corresponding feature vector ɑ1,ɑ2,…,ɑN,K-th of feature in feature space to
Amountɑi,kIt indicatesK-th of feature vector i-th of value, byIt normalizesK-th of the feature vector of variable x after normalizationDirection is projected as k-th of principal component,
Formula are as follows:
The present invention selects a kind of common kernel function, and Radial basis kernel function is calculated.When selection Radial basis kernel function
When, apparent overfitting problem is had, mainly since the corresponding feature space of Radial basis kernel function is wirelessly tieed up, is passed through
The number for the principal component that this method obtains is unrelated with the dimension of given sample, and related with sample size.It is asked to solve this
Topic, introduces rarefaction method, as sparse kernel principle component analysis (Sparse Kernel Principal Component
Analysis,SKPCA)。
In kernel principle component analysis, feature vector can use sampleIt is expressed as
That is feature vectorIt is the linear combination of whole samples, this provides a kind of thinking for the rarefaction of feature vector.It is logical
Approximate base derivation algorithm is crossed, to askApproximate base, to obtainRarefaction method.
What approximate base solved comprises the concrete steps that:
A. set is establishedFor approximate Maximum independent group, XA=φ;
B. for k=2 ..., N'sMinimizing value;
C. if obtaining minimum value≤ε, corresponding element is added to XAIt in the middle, is otherwise Xl.ε is linearly related
Truncated error asks the base of infinite dimensional space almost without sparsity in limited sample, therefore selects approximate calculation.
D. return step B, until completing all calculating, calculating terminates.
Wherein, the solution procedure in step B is as follows:
By Lagrange conditionObtain -2K0λ=0+2K, wherein K0=(k (x1,xk),…,k(xl,xk))T, the side of being K
Battle array, Kij=k (xi,xj).When kernel function is gaussian radial basis function, matrix K positive definite obtains λmin=K-1K0As f's (λ)
Minimum point.
Assuming that the one group of approximation base acquired isWithIndicate approximate base
The base vector of composition, feature vector can be expressed asProblem changes into:
Both sides are same to be multipliedNote?
Known to the problem of being eigen vector, i.e.,It enablesThen
It derives:
Wherein, K (m :) indicates the m row of nuclear matrix K, and K (:, n) indicates the n-th column, 1N=[1 ..., 1] indicate 1 row N column
Row vector.Problem has also been converted into (KI)-1Ksα=λ α is typical eigen vector problem.
Although rarefaction can control the learning ability of learning machine, overfitting is prevented, rarefaction method still may
Certain key properties of data set can be lost, that is, owe problem concerning study.Therefore, it is necessary to introduce integrated approach, referred to as integrate sparse
Kernel principle component analysis (Integration Sparse Kernel Principal Component Analysis, ISKPCA).
Since kernel principle component analysis is a kind of unsupervised learning, cannot be instructed from outside, i.e., it cannot be to certain study
As a result it is awarded or is punished, so, the method for simple average is selected herein.The integrated approach specific steps process of SKPCA is such as
Under:
A., duplicate number Re is set;
B. nuclear matrix K=Φ is calculatedTΦ;
C. sample set is calculatedThe approximate base Φ of Re groupI 1,…,ΦI Re(solving kth group approximation base
Φ I kWhen, previous group approximation base ΦI k-1Remaining vector is the prioritized vector of kth group);
D. according to above-mentioned SKPCA method, to each group of approximate base ΦI kSeek eigenvalue λk, feature vector ɑk, obtain altogether
Re group eigenvalue λ1,…,λReWith feature vector ɑ1,…,ɑRe;
E. to preceding n eigenvalue λ1≥…≥λnWith corresponding feature vector, integration characteristic value is sought respectivelyIntegration characteristic vectorWherein λi kIt is kth group approximation base ΦI kThe subspace opened is
I-th acquired the characteristic value of solution space,It is corresponding feature vector;
F. to integration characteristic valueWith integrated feature vectorIt is normalized to get sparse kernel principle component analysis is integrated
Characteristic value and feature vector.
The Classification and Identification for sparse kernel principle component analysis method will be integrated being used for electrocardio RdR scatter plot, needs first to establish integrated
Sparse kernel principle component analysis model, then selects Con trolling index.The present invention selects a kind of common Con trolling index, i.e., square pre-
It surveys error (Squared Prediction Error, SPE).
The specific method of square prediction error:
For i-th of sample Xi, it is assumed that acquiring its preceding n principal component by SKPCA is t1,…,tn, corresponding characteristic value
For λ1,…,λn.WithIndicate the reconstruct vector of the N number of principal component of feature spaceSimilarly
The then SPE of sample X is defined as:
Wherein,
Compared with prior art, beneficial effects of the present invention are as follows:
The present invention is using sparse kernel principle component analysis method is integrated, by calculating sample data and using kernel principle component analysis
Difference between mapping data carrys out the maximum correlation between data, and judges electrocardiogram (ECG) data classification with this, research from
Dynamic identification classification RdR scatter plot.Performed an analysis using the method for the present invention to heart rate variability, can for realize automated diagnostic,
Alleviate medical resource in short supply, the waste for reducing medical resource, some bases of medical efficiency offer are provided.
Detailed description of the invention
Fig. 1 is specific method flow chart of the present invention;
Fig. 2 is system structure diagram;
Fig. 3 is RdR scatter plot.
Fig. 4 is to randomly choose test sample and all kinds of SPE differences;
Fig. 5 is the partial test result figure with test data testing classification device performance.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, it is obtained by those of ordinary skill in the art without making creative efforts it is all its
His embodiment, shall fall within the protection scope of the present invention.
As shown in Figure 1, a kind of portable cardiac on-line intelligence monitoring diagnosis system design method provided by the invention, specifically
Step are as follows:
Step 1) acquires electrocardiosignal, is filtered denoising to electrocardiosignal, extracts R crest value position;
Step 2) draws electrocardio RdR scatter plot using heartbeat interval;
Step 3) zooms in and out electrocardio RdR scatter plot, zooms to same specification, changes into grayscale image, and to image data
It is normalized, to reduce calculation amount;
The scatter plot sample of acquisition is marked in step 4);
Step 5) samples sample, randomly selects 85% data as training sample;
Parameter is arranged in step 6), and parameter includes the value of the error parameter of approximate base, gaussian kernel function parameter and control limit;
Step 7) solves approximate base, characteristic value, feature vector to every class sample respectively;
Step 8) calculates separately the SPE of each test sample, and the SPE difference reckling of value and certain class sample is test specimens
This prediction classification calculates accuracy rate compared with concrete class, the step 9) if meeting the requirements, otherwise return step 6) again
Parameter training is set;
Step 9) obtains disaggregated model parameter, and the disaggregated model of acquisition is used in cardiac diagnosis system.
Further, using integrating sparse kernel principle component analysis method for the Classification and Identification of the electrocardio RdR scatter plot.
RdR scatter plot is a kind of heart rate variance analyzing method, can embody its trend changed over time.
Further, sparse kernel principle component analysis method will be integrated and is used for Classification and Identification to the electrocardio RdR scatter plot, first
Foundation integrates sparse kernel principle component analysis model, then selects Con trolling index by square prediction error method.
Integrating sparse kernel principle component analysis is a kind of unsupervised machine learning algorithm, integrates the base of sparse kernel principle component analysis
This method is as follows:
Principal component analysis is a kind of typical unsupervised algorithm, is usually used in solving the linear problem of luv space, and in order to
The nonlinear problem of luv space is solved with linear method in feature space, B.Scholkopf et al. proposes core principal component
It analyzes (Kernel Principal Component Analysis, KPCA).It defines from luv space RnTo feature space F's
Nonlinear Mapping:If given sample X={ x1,…,xN},xi∈Rn, then pass throughMapping can obtain one group
VectorAssuming that this group of vector meetsThen the Correlation Matrix in feature space is
If this group of vectorIt can then enableKnown toMeet
Condition, instead of in formulaThen KPCA problem, which can be converted to, seeks Correlation Matrix in feature spaceEigenvalue λ, that is, feature
Vector
Wherein,It is the linear combination of sample, enablesɑ=[ɑ1,…,ɑN]T, thenWhenWhen cannot explicitly obtaining, kernel function is introduced, ifFirstly the need of calculating:
K=ΦTΦ;
Wherein, matrix K is the matrix of NxN, also referred to as nuclear matrix.Then problem is converted are as follows:
K ɑ=N λ ɑ;
Wherein, ɑ=[ɑ1,…,ɑN].WhenThe process of centralization can the operation directly on K:
Wherein,MeetIt is 1 matrix of a NxN.Assuming that obtaining
1 >=λ of eigenvalue λ 2 >=... λ n and its corresponding feature vector ɑ1,ɑ2,…,ɑN,K-th of feature in feature space to
Amountɑi,kIt indicatesK-th of feature vector i-th of value, byIt normalizesK-th of the feature vector of variable x after normalizationDirection is projected as k-th of master point
Amount, formula are as follows:
The present invention selects a kind of common kernel function, and Radial basis kernel function is calculated.When selection Radial basis kernel function
When, apparent overfitting problem is had, mainly since the corresponding feature space of Radial basis kernel function is wirelessly tieed up, is passed through
The number for the principal component that this method obtains is unrelated with the dimension of given sample, and related with sample size.It is asked to solve this
Topic, introduces rarefaction method, as sparse kernel principle component analysis (Sparse Kernel Principal Component
Analysis,SKPCA)。
In kernel principle component analysis, feature vector can use sampleIt is expressed as?
That is feature vectorIt is the linear combination of whole samples, this provides a kind of thinking for the rarefaction of feature vector.Pass through
Approximate base derivation algorithm, to askApproximate base, to obtainRarefaction method.
What approximate base solved comprises the concrete steps that:
A. set is establishedFor approximate Maximum independent group, XA=φ;
B. for k=2 ..., N'sMinimizing value;
Wherein,
C. if obtaining minimum value≤ε, corresponding element is added to XAIt in the middle, is otherwise Xl.ε is linearly related
Truncated error asks the base of infinite dimensional space almost without sparsity in limited sample, therefore selects approximate calculation.
D. return step B, until completing all calculating, calculating terminates.
Wherein, the solution procedure in step B is as follows:
It enables
By Lagrange conditionObtain -2K0λ=0+2K, wherein K0=(k (x1,xk),…,k(xl,xk))T, the side of being K
Battle array, Kij=k (xi,xj).When kernel function is gaussian radial basis function, matrix K positive definite obtains λmin=K-1K0As f's (λ)
Minimum point.
Assuming that the one group of approximation base acquired isWithIndicate approximate base
The base vector of composition, feature vector can be expressed asProblem changes into:
Both sides are same to be multipliedNote?
Known to the problem of being eigen vector, i.e.,It enablesThen
It derives:
Wherein, K (m :) indicates the m row of nuclear matrix K, and K (:, n) indicates the n-th column, 1N=[1 ..., 1] indicate 1 row N column
Row vector.Problem has also been converted into (KI)-1Ksα=λ α is typical eigen vector problem.
Although rarefaction can control the learning ability of learning machine, overfitting is prevented, rarefaction method still may
Certain key properties of data set can be lost, that is, owe problem concerning study.Therefore, it is necessary to introduce integrated approach, referred to as integrate sparse
Kernel principle component analysis (Integration Sparse Kernel Principal Component Analysis, ISKPCA).
Since kernel principle component analysis is a kind of unsupervised learning, cannot be instructed from outside, i.e., it cannot be to certain study
As a result it is awarded or is punished, so, the method for simple average is selected herein.The integrated approach specific steps process of SKPCA is such as
Under:
A., duplicate number Re is set;
B. nuclear matrix K=Φ is calculatedTΦ;
C. sample set is calculatedThe approximate base Φ of Re groupI 1,…,ΦI Re(solving kth group approximation base
Φ I kWhen, previous group approximation base ΦI k-1Remaining vector is the prioritized vector of kth group);
D. according to above-mentioned SKPCA method, to each group of approximate base ΦI kSeek eigenvalue λk, feature vector ɑk, obtain altogether
Re group eigenvalue λ1,…,λReWith feature vector ɑ1,…,ɑRe;
E. to preceding n eigenvalue λ1≥…≥λnWith corresponding feature vector, integration characteristic value is sought respectivelyIntegration characteristic vectorWherein λi kIt is kth group approximation base ΦI kThe subspace opened is
I-th acquired the characteristic value of solution space,It is corresponding feature vector;
F. to integration characteristic valueWith integrated feature vectorIt is normalized to get sparse kernel principle component analysis is integrated
Characteristic value and feature vector.
The Classification and Identification for sparse kernel principle component analysis method will be integrated being used for electrocardio RdR scatter plot, needs first to establish integrated
Sparse kernel principle component analysis model, then selects Con trolling index.The present invention selects a kind of common Con trolling index, i.e., square pre-
It surveys error (Squared Prediction Error, SPE).
The specific method of square prediction error:
For i-th of sample Xi, it is assumed that acquiring its preceding n principal component by SKPCA is t1,…,tn, corresponding characteristic value
For λ1,…,λn.WithIndicate the reconstruct vector of the N number of principal component of feature spaceSimilarly
The then SPE of sample X is defined as:
Wherein,
Using the electrocardiogram (ECG) data in MIT-BIH database, to make schematically analysis with the data of this database convenient for narration
It introduces.
Fig. 2 is system structure diagram, the hardware of the system mainly by Acquisition Circuit, microcontroller circuit, storage circuit,
RAM circuit, jtag circuit, telecommunication circuit, circuit for alarming, LED operation indicating circuit, electric power management circuit and Cloud Server etc.
Composition.In order to reduce the complexity of system, this system selects B/S (browser/server) mode in remote monitoring terminal part
To develop.This system connects into an entirety by internet, and to meet portable feature, is designed using single lead.Acquisition
Circuit is responsible for acquiring electrocardiosignal, carries out data filtering on the microprocessor, then the information such as LCD display waveform, heart rate, can
Data are uploaded to Cloud Server using telecommunication circuit, received server-side to data handles data, and obtains
Diagnostic analysis report, according to diagnostic analysis as a result, judging whether to need to send the operation such as warning information.Doctor can also pass through
Browser checks user's electrocardiogram (ECG) data, and provides diagnostic comments.
Fig. 3 is the RdR scatter plot schematic diagram that MIT-BIH electrocardiogram (ECG) data is drawn by MATLAB.RdR scatter plot can have
Effect reflection heartbeat interval changes with time trend, contains a large amount of clinical information, can show different heartbeat features.Experiment
It has been shown that, carries out identification classification to it using the method for the present invention, effect is preferable.
Fig. 4 is that certain a kind of electrocardiogram (ECG) data sample is randomly choosed from test sample, if figure is to choose concrete class to be
It, is calculated the difference of its SPE by some test sample collections of " 3 " with each classification respectively, as a result such as figure (4), it can be seen that its
With the SPE difference minimum that classification is " 3 ", classify all correct;
Fig. 5 is that method, the result display diagram of part classifying test display according to the present invention, and accuracy rate is higher, in figure, ginseng
Number true is actual sample class, and predicted is the classification of classifier prediction, and the image shown in figure is RdR scatterplot
Figure, it can be seen that classification results are all correct.
In conclusion the present invention by calculating sample data and uses core using sparse kernel principle component analysis method is integrated
Difference between principal component analysis mapping data carrys out the maximum correlation between data, and judges electrocardiogram (ECG) data with this
Classification studies automatic recognition classification RdR scatter plot.It is performed an analysis using the method for the present invention to heart rate variability, can be realization
Automated diagnostic, the waste alleviated medical resource in short supply, reduce medical resource improve some bases of medical efficiency offer.
The above is only a preferred embodiment of the present invention, protection scope of the present invention is not limited merely to above-mentioned implementation
Example, all technical solutions belonged under thinking of the present invention all belong to the scope of protection of the present invention.It should be pointed out that being led for this technology
For the those of ordinary skill in domain, several improvements and modifications without departing from the principles of the present invention, these improvements and modifications
Also it should be regarded as protection scope of the present invention.
Claims (3)
1. a kind of portable cardiac on-line intelligence monitoring diagnosis system design method, which is characterized in that including following main body step:
Step 1) acquires electrocardiosignal, is filtered denoising to electrocardiosignal, extracts R crest value position;
Step 2) draws electrocardio RdR scatter plot using heartbeat interval;
Step 3) zooms in and out electrocardio RdR scatter plot, zooms to same specification, changes into grayscale image, and carry out to image data
Normalized, to reduce calculation amount;
The scatter plot sample of acquisition is marked in step 4);
Step 5) samples sample, randomly selects 85% data as training sample;
Parameter is arranged in step 6), and parameter includes the value of the error parameter of approximate base, gaussian kernel function parameter and control limit;
Step 7) solves approximate base, characteristic value, feature vector to every class sample respectively;
Step 8) calculates separately the SPE of each test sample, and the SPE difference reckling of value and certain class sample is test sample
Predict classification, compared with concrete class, calculate accuracy rate, the step 9) if meeting the requirements, otherwise return step 6) it resets
Parameter training;
Step 9) obtains disaggregated model parameter, and the disaggregated model of acquisition is used in cardiac diagnosis system.
2. a kind of portable cardiac on-line intelligence monitoring diagnosis system design method according to claim 1, feature exist
In,
Using integrating sparse kernel principle component analysis method for the Classification and Identification of the electrocardio RdR scatter plot.
3. a kind of portable cardiac on-line intelligence monitoring diagnosis system design method according to claim 2, feature exist
In,
Sparse kernel principle component analysis method will be integrated and be used for Classification and Identification to the electrocardio RdR scatter plot, first establish integrate it is dilute
Kernel principle component analysis model is dredged, Con trolling index is then selected by square prediction error method.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810921525.7A CN109157211A (en) | 2018-08-14 | 2018-08-14 | A kind of portable cardiac on-line intelligence monitoring diagnosis system design method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810921525.7A CN109157211A (en) | 2018-08-14 | 2018-08-14 | A kind of portable cardiac on-line intelligence monitoring diagnosis system design method |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109157211A true CN109157211A (en) | 2019-01-08 |
Family
ID=64895481
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810921525.7A Pending CN109157211A (en) | 2018-08-14 | 2018-08-14 | A kind of portable cardiac on-line intelligence monitoring diagnosis system design method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109157211A (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110403600A (en) * | 2019-07-26 | 2019-11-05 | 武汉海星通技术股份有限公司 | Paroxysmal Atrial Fibrillation intelligent analysis method and system based on differential time scatter plot |
CN111920397A (en) * | 2020-08-07 | 2020-11-13 | 江苏师范大学 | Arteriosclerosis degree detection method based on sparse least square support vector machine |
CN114496209A (en) * | 2022-02-18 | 2022-05-13 | 青岛市中心血站 | Blood donation intelligent decision method and system |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106073755A (en) * | 2016-05-27 | 2016-11-09 | 成都信汇聚源科技有限公司 | The implementation method that in a kind of miniature holter devices, atrial fibrillation identifies automatically |
CN106725426A (en) * | 2016-12-14 | 2017-05-31 | 深圳先进技术研究院 | A kind of method and system of electrocardiosignal classification |
CN107292292A (en) * | 2017-07-20 | 2017-10-24 | 浙江好络维医疗技术有限公司 | A kind of QRS complex sorting technique based on SVMs |
CN206792400U (en) * | 2017-01-16 | 2017-12-26 | 吉林东华原医疗设备有限责任公司 | HRV detection means |
-
2018
- 2018-08-14 CN CN201810921525.7A patent/CN109157211A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106073755A (en) * | 2016-05-27 | 2016-11-09 | 成都信汇聚源科技有限公司 | The implementation method that in a kind of miniature holter devices, atrial fibrillation identifies automatically |
CN106725426A (en) * | 2016-12-14 | 2017-05-31 | 深圳先进技术研究院 | A kind of method and system of electrocardiosignal classification |
CN206792400U (en) * | 2017-01-16 | 2017-12-26 | 吉林东华原医疗设备有限责任公司 | HRV detection means |
CN107292292A (en) * | 2017-07-20 | 2017-10-24 | 浙江好络维医疗技术有限公司 | A kind of QRS complex sorting technique based on SVMs |
Non-Patent Citations (2)
Title |
---|
甘良志: "核学习算法与集成方法研究", 《中国博士学位论文全文数据库(信息科技辑)》 * |
陆宏伟等: "基于RdR新型散点图心率变异性研究", 《生物医学工程学杂志》 * |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110403600A (en) * | 2019-07-26 | 2019-11-05 | 武汉海星通技术股份有限公司 | Paroxysmal Atrial Fibrillation intelligent analysis method and system based on differential time scatter plot |
CN110403600B (en) * | 2019-07-26 | 2022-02-08 | 武汉海星通技术股份有限公司 | Intelligent analysis method and system for paroxysmal atrial fibrillation based on difference time scatter diagram |
CN111920397A (en) * | 2020-08-07 | 2020-11-13 | 江苏师范大学 | Arteriosclerosis degree detection method based on sparse least square support vector machine |
CN114496209A (en) * | 2022-02-18 | 2022-05-13 | 青岛市中心血站 | Blood donation intelligent decision method and system |
CN114496209B (en) * | 2022-02-18 | 2022-09-27 | 青岛市中心血站 | Intelligent decision-making method and system for blood donation |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Rai et al. | Hybrid CNN-LSTM deep learning model and ensemble technique for automatic detection of myocardial infarction using big ECG data | |
Acharya et al. | Automated identification of normal and diabetes heart rate signals using nonlinear measures | |
Sridhar et al. | Accurate detection of myocardial infarction using non linear features with ECG signals | |
CN110090012A (en) | A kind of human body diseases detection method and testing product based on machine learning | |
US20100217144A1 (en) | Diagnostic and predictive system and methodology using multiple parameter electrocardiography superscores | |
Zhang et al. | A multi-dimensional association information analysis approach to automated detection and localization of myocardial infarction | |
CN112365978A (en) | Method and device for establishing early risk assessment model of tachycardia event | |
Sridhar et al. | Automated diagnosis of Coronary Artery Disease using nonlinear features extracted from ECG signals | |
Nascimento et al. | Heart arrhythmia classification based on statistical moments and structural co-occurrence | |
Pt et al. | Automated diagnosis of diabetes using heart rate variability signals | |
CN109157211A (en) | A kind of portable cardiac on-line intelligence monitoring diagnosis system design method | |
Swapna et al. | Diabetes detection using ecg signals: An overview | |
Cheng et al. | Atrial fibrillation identification with PPG signals using a combination of time-frequency analysis and deep learning | |
Nasimov et al. | A new approach to classifying myocardial infarction and cardiomyopathy using deep learning | |
Shen et al. | Risk prediction for cardiovascular disease using ECG data in the China Kadoorie Biobank | |
Zhang et al. | Cardiac arrhythmia classification by time–frequency features inputted to the designed convolutional neural networks | |
Zhang et al. | Application of deep neural network for congestive heart failure detection using ECG signals | |
Huang et al. | A multiview feature fusion model for heartbeat classification | |
Jiang et al. | Visualization deep learning model for automatic arrhythmias classification | |
Nath et al. | Quantum annealing for automated feature selection in stress detection | |
TWI688371B (en) | Intelligent device for atrial fibrillation signal pattern acquisition and auxiliary diagnosis | |
WO2023129752A1 (en) | Assessment of hemodynamics parameters | |
Hong et al. | Gated temporal convolutional neural network and expert features for diagnosing and explaining physiological time series: a case study on heart rates | |
TWI732489B (en) | Method and system for quickly detecting abnormal concentration of potassium ion in blood from electrocardiogram | |
CN108922619A (en) | A kind of RdR scatter plot recognition methods based on sparse core leading role |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
WD01 | Invention patent application deemed withdrawn after publication | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20190108 |