CN108491769A - Atrial fibrillation sorting technique based on phase between RR and multiple characteristic values - Google Patents
Atrial fibrillation sorting technique based on phase between RR and multiple characteristic values Download PDFInfo
- Publication number
- CN108491769A CN108491769A CN201810190669.XA CN201810190669A CN108491769A CN 108491769 A CN108491769 A CN 108491769A CN 201810190669 A CN201810190669 A CN 201810190669A CN 108491769 A CN108491769 A CN 108491769A
- Authority
- CN
- China
- Prior art keywords
- phase
- atrial fibrillation
- feature
- characteristic values
- multiple characteristic
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/02—Preprocessing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/08—Feature extraction
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/12—Classification; Matching
- G06F2218/16—Classification; Matching by matching signal segments
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Signal Processing (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)
Abstract
The present invention is based on the atrial fibrillation sorting techniques of phase between RR and multiple characteristic values, extract the validity feature of atrial fibrillation signal, belong to field of signal processing.Its feature is to include the following steps：1）The acquisition and segmentation of phase between progress electrocardiosignal RR；2）The feature vector of phase between extraction characterization atrial fibrillation Duan Yufei atrial fibrillation sections RR；3）Training sorter model is for predicting input sample label.Compared with existing electrocardiosignal recognition methods, from multi-angle, the phase between RR carries out signature analysis --- dispersion degree, distributional pattern, complexity to the present invention, and more effective parameter is selected to carry out characteristic present in each angle analysis, realized in MIT BIH atrial fibrillation databases 95.81% sensitivity, 96.48% specificity and 96.09% accuracy rate, the specificity that 95.16% is realized in MIT BIH sinus rhythm databases, to real-time analyzing processing atrial fibrillation important in inhibiting.
Description
Technical field
The present invention is based on the atrial fibrillation sorting techniques of phase between RR and multiple characteristic values, extract the validity feature of atrial fibrillation signal, belong to
Field of signal processing.
Background technology
In global medical health care system, the detection and management of patients with atrial fibrillation just become the most important thing.Since atrial fibrillation can make
Patient occurs stroke risk and increases by 500, so the detection and treatment of early stage are particularly important.
The phase is the time being separated by between R waves in two QRS waves between RR, can reflect heart rate to a certain extent.Atrial fibrillation electrocardio
One prominent features of figure are that the phase is absolutely irregular between RR, and R-wave amplitude is larger, compared to the detection positioning to P, f wave, between RR
The detection positioning of phase is easier, is accurate.
The feature that effectively reflection disease is extracted from electrocardiosignal, is the process of a feature selecting.In practical application,
Statistical analysis method be only data are qualitatively analyzed, and nonlinear theory can extraction system differentiation information, it is fixed
The complexity of the description system of amount.So the nonlinear parameter analytic approach of phase between RR can be regarded as to one kind of traditional statistical method
Supplement can obtain the multi-angle distribution spy of phase between RR by the statistics measure feature of phase and nonlinear parameter feature between combination RR
Sign, to improve the accuracy of atrial fibrillation detection.The present invention extracts the statistical nature of phase between RR --- steady coefficient of variation parameter, skewness
Parameter and nonlinear parameter feature --- LZ complexities composition characteristic vector carry out atrial fibrillation classification and Detection.
Invention content
The present invention is based on the atrial fibrillation detection method of phase between RR and multiple characteristic values, this method extracts phase validity feature between RR, should
Method is intended to realize the precise classification of atrial fibrillation using the multi-angle distribution characteristics of phase between RR.
The technical solution adopted in the present invention is as follows：
Atrial fibrillation detection method based on phase between RR and multiple characteristic values, is as follows：
Step 1：Pretreatment：The acquisition and segmentation of phase between progress electrocardiosignal RR；
Step 2：Feature extraction：The feature vector of phase between extraction characterization atrial fibrillation Duan Yufei atrial fibrillation sections RR；
Step 3：Classification：Training sorter model is for predicting input sample label；
Advantageous effect：Categorizing system of the present invention describes dispersion degree, the skewness parameter of phase between RR using steady coefficient of variation parameter
Distribution shape, the Lempel-Ziv of phase between description RR（LZ）Complexity parameter describes the complexity of phase between RR, then by three kinds of spies
The feature vector input SVM classifier of value indicative composition carries out the classification of atrial fibrillation and non-atrial fibrillation.The experiment is in MIT-BIH atrial fibrillation data
Realized in library 95.81% sensitivity, 96.48% specificity and 96.09% accuracy rate, in MIT-BIH sinus rhythm data
The specificity that 95.16% is realized in library is suitble to apply to tele-medicine and atrial fibrillation signal identifies in real time, to the small of intelligent medical
Type important in inhibiting.
Description of the drawings
Fig. 1 is present system schematic diagram
Fig. 2 is phase and corresponding classification mark figure between RR of the present invention
Fig. 3 is that phase length is 64 points of steady coefficient of variation, skewness, LZ complexity characteristics value graphics between RR.
Specific implementation mode
The present invention is described in further detail With reference to embodiment：
1. pretreatment
By preceding 12 electrocardiographic recordings in MIT-BIH atrial fibrillation databases（Smaller 12 of recording mechanism）Classification mould is trained as training set
Type, rear 13 electrocardiographic recordings（Larger 13 of recording mechanism）The inspection of disaggregated model accuracy is carried out as test set.To one
When the whole cardiac RR intervals of a patient carry out segmentation with concrete class mark, it is one section to take 32,64 and 128 heartbeats respectively,
Also include non-atrial fibrillation heartbeat, for the phase between these RR from Fig. 2 it can be found that including not only atrial fibrillation heartbeat in phase section between some RR
Section, if atrial fibrillation heartbeat quantity is more than the half of this section of total heartbeat quantity, it is atrial fibrillation section to mark these sections, is otherwise labeled as non-
Atrial fibrillation section.MIT-BIH databases are segmented, mark arrangement, table 1 has counted different section lengths, the room in each database
It quivers and non-atrial fibrillation hop count mesh.
2. feature extraction
We are extracted three characteristic values of phase between RR：（1）The steady coefficient of variation,（2）Skewness parameter,（3）Lempel-Ziv is multiple
Miscellaneous degree.
The steady coefficient of variation：
In robust statistical techniques, the steady coefficient of variation（RCV）Be expressed as normalization RR between phase quartile spacing（IQR）Between RR
Interim digit（）Ratio, calculation formula is：
Expression in above formula is normalization IQR, is allowed to the standard deviation equal to normal distribution
Skewness：
Skewness can assess the skew direction showed in data distribution form and degree, can to the asymmetry of data distribution into
Row description.The skewness calculation formula of one group of RR interval series is：
In formula, the phase between i-th of RR is indicated,Phase average value between expression RR, N is phase length between RR, and SD is standard deviation
Lempel-Ziv complexities：
Lempel and Ziv proposes a kind of product complexity theory, and the complexity of a certain sequence of length is considered as in the sequence
The speed that new model is generated with being incremented by for sequence length.Its specific algorithm is as follows：
Step 1：First by known arrayReconstruct, calculates the average value of the sequence, will meet in sequencePart all sets, meetPart all resets, former sequence is reconstructed into one by such method and is only wrapped
Include the symbol sebolic addressing of " 0 " and " 1 "。
Step 2：To gained above（" 0 ", " 1 "）An already present substring in sequence, add behind
A upper character, or add a string characters, obtain a symbol sebolic addressing；Character string SQ removal last character is obtained into SQv, that is, then check Q whether with SQv
In oneself substring that occurred it is identical, that is, check whether there are some,, make
；If in the presence of Q is just added to subsequent mode and referred to as " is replicated ", at this point, Q is extended, that is, increases k, continuously repeats above
The step of, until Q is not belonging to the substring occurred in SQv；And by the Q substrings for being not belonging to occur in SQv
Process referred to as " is inserted into ", after " " symbol is placed on Q at " insertion "；Then by alphabet all conducts before " "
S recycles above step.
Step 3：As above, a sequence is broken by " " is divided into multiple string segments, and the character segment number defined is " multiple
Miscellaneous degree "。
Step 4：According to the research of Lempel and Ziv, the sequence of a ∈ { " 0 ", " 1 " }, complexity can approach one
Value：
What is indicated is the progressive form of sequence, can be usedIt willIt is normalized, obtains " relative complexity "
Estimate：
From calculating process it is found that LZ complexity metrics are the speed for generating new model in a sequence,Different value is taken,
The speed that new model is generated in sequence is different, which is demonstrated by different characteristics；Or it can think from another angleBe calculate a time series with random sequence similar degree.
3. classification
After through the above steps, feature vector is input in grader, atrial fibrillation classification results can be obtained.The method of the present invention is logical
It crosses and the classifying quality of MIT-BIH-AF databases and MIT-BIH-NSR databases is verified, table 2 has counted phase segment length between RR
When degree is 32,64,128, for the classifying quality of phase length between different RR, wherein when phase length is 64 between taking RR, obtain
The classifying quality arrived is best, reached in MIT-BIH-AF database test sets 95.81% sensitivity, 96.48% it is special
Property and 96.09% accuracy rate, 95.16% specificity has been reached in MIT-BIH-NSR databases.
Claims (4)
1. the atrial fibrillation sorting technique based on phase between RR and multiple characteristic values, characterized in that including following steps：
Step 1, pretreatment：The acquisition and segmentation of phase between progress electrocardiosignal RR：
Segmentation is carried out according to MIT-BIH atrial fibrillation database cardiac RR intervals to mark with concrete class；For phase section between these RR, if
When atrial fibrillation heartbeat quantity is more than the half of this section of total heartbeat quantity, then it is atrial fibrillation section to mark these sections, is otherwise labeled as non-atrial fibrillation
Section；
Step 2, feature extraction：
The statistical nature of phase between extraction RR --- the steady coefficient of variation, skewness and nonlinear parameter feature --- LZ complexities, group
At feature vector；Dispersion degree, the skewness parameter that wherein steady coefficient of variation parameter describes the phase between RR describe the distribution of phase between RR
Shape, Lempel-Ziv（LZ）Complexity parameter describes the complexity of phase between RR；
Classify in step 3, feature vector input grader：
By extracted in step 2 three kinds of eigenvalue clusters at feature vector, then in conjunction with the class of atrial fibrillation in MIT-BIH databases
Type marks, and training set feature vector is marked one piece of input SVM classifier with atrial fibrillation to be trained, finally by the feature of test set
Vector inputs above-mentioned trained grader, classifies to test set；
Wherein, present invention employs radial basis function（radial basis kernel,RBF）The SVM of core.
2. the atrial fibrillation sorting technique according to claim 1 based on phase between RR and multiple characteristic values, it is characterised in that：Step 1
It is middle according to electrocardiosignal characteristic, obtain the phase between RR, be then segmented and marked.
3. the atrial fibrillation sorting technique according to claim 1 based on phase between RR and multiple characteristic values：To electrocardiosignal in step 2
RR between the phase carry out feature extraction, refer to respectively between RR the phase carry out operation, obtain steady coefficient of variation parameter value, skewness parameter
Value, Lempel-Ziv（LZ）Complexity parameter value.
4. the atrial fibrillation sorting technique according to claim 1 based on phase between RR and multiple characteristic values：In step 3 classification refer to by
The three kinds of characteristic values extracted in step 2 are mixed, composition characteristic vector, are then inputted grader and are classified.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810190669.XA CN108491769A (en) | 2018-03-08 | 2018-03-08 | Atrial fibrillation sorting technique based on phase between RR and multiple characteristic values |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810190669.XA CN108491769A (en) | 2018-03-08 | 2018-03-08 | Atrial fibrillation sorting technique based on phase between RR and multiple characteristic values |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108491769A true CN108491769A (en) | 2018-09-04 |
Family
ID=63338119
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810190669.XA Pending CN108491769A (en) | 2018-03-08 | 2018-03-08 | Atrial fibrillation sorting technique based on phase between RR and multiple characteristic values |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108491769A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109303561A (en) * | 2018-11-01 | 2019-02-05 | 杭州质子科技有限公司 | It is a kind of to clap the recognition methods clapped with the abnormal heart based on the artifact heart of misclassification and supervised learning |
| CN111067508A (en) * | 2019-12-31 | 2020-04-28 | 深圳安视睿信息技术股份有限公司 | Non-intervention monitoring and evaluating method for hypertension in non-clinical environment |
| WO2022073220A1 (en) * | 2020-10-10 | 2022-04-14 | 上海市第一人民医院 | Atrial fibrillation detection device and method, and system and storage medium |
Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101630364A (en) * | 2009-08-20 | 2010-01-20 | 天津大学 | Method for gait information processing and identity identification based on fusion feature |
| CN101807245A (en) * | 2010-03-02 | 2010-08-18 | 天津大学 | Artificial neural network-based multi-source gait feature extraction and identification method |
| CN101865859A (en) * | 2009-04-17 | 2010-10-20 | 华为技术有限公司 | Detection method and device for image scratch |
| CN103630912A (en) * | 2013-11-26 | 2014-03-12 | 中国科学院嘉兴微电子与***工程中心 | Detection method of stillness of satellite receiver |
| CN104161509A (en) * | 2014-08-08 | 2014-11-26 | 申岱 | Heart rate variability analyzing method based on amplitude spectrum and instruments |
| CN104410762A (en) * | 2014-11-18 | 2015-03-11 | 沈阳工业大学 | Steady echo cancellation method in hand free cell phone conversation system |
| CN104665803A (en) * | 2014-12-10 | 2015-06-03 | 上海理工大学 | Atrial fibrillation detecting system based on intelligent platform |
| CN105205315A (en) * | 2015-09-08 | 2015-12-30 | 山东大学 | Biological signal quantization level representing method based on ELZ encoding algorithm |
| CN105228508A (en) * | 2013-03-08 | 2016-01-06 | 新加坡健康服务有限公司 | A kind of system and method measured for the risk score of classifying |
| CN105320969A (en) * | 2015-11-20 | 2016-02-10 | 北京理工大学 | A heart rate variability feature classification method based on multi-scale Renyi entropy |
| CN105411565A (en) * | 2015-11-20 | 2016-03-23 | 北京理工大学 | Heart rate variability feature classification method based on generalized scale wavelet entropy |
| 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 |
| CN107358196A (en) * | 2017-07-12 | 2017-11-17 | 北京卫嘉高科信息技术有限公司 | A kind of sorting technique of heart beat type, device and electrocardiogram equipment |
-
2018
- 2018-03-08 CN CN201810190669.XA patent/CN108491769A/en active Pending
Patent Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101865859A (en) * | 2009-04-17 | 2010-10-20 | 华为技术有限公司 | Detection method and device for image scratch |
| CN101630364A (en) * | 2009-08-20 | 2010-01-20 | 天津大学 | Method for gait information processing and identity identification based on fusion feature |
| CN101807245A (en) * | 2010-03-02 | 2010-08-18 | 天津大学 | Artificial neural network-based multi-source gait feature extraction and identification method |
| CN105228508A (en) * | 2013-03-08 | 2016-01-06 | 新加坡健康服务有限公司 | A kind of system and method measured for the risk score of classifying |
| CN103630912A (en) * | 2013-11-26 | 2014-03-12 | 中国科学院嘉兴微电子与***工程中心 | Detection method of stillness of satellite receiver |
| CN104161509A (en) * | 2014-08-08 | 2014-11-26 | 申岱 | Heart rate variability analyzing method based on amplitude spectrum and instruments |
| CN104410762A (en) * | 2014-11-18 | 2015-03-11 | 沈阳工业大学 | Steady echo cancellation method in hand free cell phone conversation system |
| CN104665803A (en) * | 2014-12-10 | 2015-06-03 | 上海理工大学 | Atrial fibrillation detecting system based on intelligent platform |
| CN105205315A (en) * | 2015-09-08 | 2015-12-30 | 山东大学 | Biological signal quantization level representing method based on ELZ encoding algorithm |
| CN105320969A (en) * | 2015-11-20 | 2016-02-10 | 北京理工大学 | A heart rate variability feature classification method based on multi-scale Renyi entropy |
| CN105411565A (en) * | 2015-11-20 | 2016-03-23 | 北京理工大学 | Heart rate variability feature classification method based on generalized scale wavelet entropy |
| 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 |
| CN107358196A (en) * | 2017-07-12 | 2017-11-17 | 北京卫嘉高科信息技术有限公司 | A kind of sorting technique of heart beat type, device and electrocardiogram equipment |
Non-Patent Citations (2)
| Title |
|---|
| 邹芃等: ""稳健统计方法应用实例分析"", 《科技传播》 * |
| 陈志博等: ""基于RR间期和多特征值的房颤自动检测分类"", 《生物医学工程学杂志》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109303561A (en) * | 2018-11-01 | 2019-02-05 | 杭州质子科技有限公司 | It is a kind of to clap the recognition methods clapped with the abnormal heart based on the artifact heart of misclassification and supervised learning |
| CN111067508A (en) * | 2019-12-31 | 2020-04-28 | 深圳安视睿信息技术股份有限公司 | Non-intervention monitoring and evaluating method for hypertension in non-clinical environment |
| CN111067508B (en) * | 2019-12-31 | 2022-09-27 | 深圳安视睿信息技术股份有限公司 | Non-intervention monitoring and evaluating method for hypertension in non-clinical environment |
| WO2022073220A1 (en) * | 2020-10-10 | 2022-04-14 | 上海市第一人民医院 | Atrial fibrillation detection device and method, and system and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Kumari et al. | An automated detection of heart arrhythmias using machine learning technique: SVM | |
| Sellami et al. | A robust deep convolutional neural network with batch-weighted loss for heartbeat classification | |
| Lynn et al. | A deep bidirectional GRU network model for biometric electrocardiogram classification based on recurrent neural networks | |
| de Albuquerque et al. | Robust automated cardiac arrhythmia detection in ECG beat signals | |
| Limam et al. | Atrial fibrillation detection and ECG classification based on convolutional recurrent neural network | |
| CN109620152B (en) | MutifacolLoss-densenert-based electrocardiosignal classification method | |
| CN104367317B (en) | Electrocardiogram electrocardiosignal classification method with multi-scale characteristics combined | |
| CN108596142B (en) | PCANet-based electrocardiogram feature extraction method | |
| WO2020006939A1 (en) | Electrocardiogram generation and classification method based on generative adversarial network | |
| CN106778685A (en) | Electrocardiogram image-recognizing method, device and service terminal | |
| Ahmed et al. | An investigative study on motifs extracted features on real time big-data signals | |
| Kaya et al. | Feature selection using genetic algorithms for premature ventricular contraction classification | |
| CN108491769A (en) | Atrial fibrillation sorting technique based on phase between RR and multiple characteristic values | |
| Zhang et al. | [Retracted] An ECG Heartbeat Classification Method Based on Deep Convolutional Neural Network | |
| Perumal et al. | Lung cancer detection and classification on CT scan images using enhanced artificial bee colony optimization | |
| CN109620210A (en) | A kind of electrocardiosignal classification method of the CNN based on from coding mode in conjunction with GRU | |
| Jeong et al. | Convolutional neural network for classification of eight types of arrhythmia using 2D time–frequency feature map from standard 12-lead electrocardiogram | |
| CN109124620A (en) | A kind of atrial fibrillation detection method, device and equipment | |
| CN111956214A (en) | QRS wave automatic detection method based on U-net end-to-end neural network model | |
| CN113288157A (en) | Arrhythmia classification method based on depth separable convolution and improved loss function | |
| Zhao et al. | An explainable attention-based TCN heartbeats classification model for arrhythmia detection | |
| Dong et al. | An improved YOLOv5 network for lung nodule detection | |
| Sridevi et al. | Quanvolution neural network to recognize arrhythmia from 2D scaleogram features of ECG signals | |
| Rahuja et al. | A deep neural network approach to automatic multi-class classification of electrocardiogram signals | |
| Ahmed et al. | Improving prediction of plant disease using k-efficient clustering and classification algorithms |
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 |
Application publication date: 20180904 |
|
| WD01 | Invention patent application deemed withdrawn after publication |