CN108281184A - Myocardial ischemia based on machine learning examines method, storage medium and electronic equipment in advance - Google Patents
Myocardial ischemia based on machine learning examines method, storage medium and electronic equipment in advance Download PDFInfo
- Publication number
- CN108281184A CN108281184A CN201810100231.8A CN201810100231A CN108281184A CN 108281184 A CN108281184 A CN 108281184A CN 201810100231 A CN201810100231 A CN 201810100231A CN 108281184 A CN108281184 A CN 108281184A
- Authority
- CN
- China
- Prior art keywords
- myocardial ischemia
- machine learning
- advance
- characteristic parameter
- data
- 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
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
Abstract
The present invention provides a kind of myocardial ischemia based on machine learning and examines method, storage medium and electronic equipment in advance, the method includes：It obtains magnetocardiogram data and extracts the data set of magnetocardiogram data；Feature extraction is carried out to data set and forms multiple independent characteristic parameter collection；Each training pattern configured for each characteristic parameter collection is trained respectively according to each characteristic parameter, forms multiple graders；The output result of multiple graders is subjected to integrated study combination, exports final prediction result；Multiple independent characteristic parameter collection include：Time domain charactreristic parameter collection, frequency domain character parameter set and information theory characteristic parameter collection；Support vector machines training pattern is respectively adopted to temporal signatures parameter set and frequency domain character parameter set to be trained, information theory characteristic parameter collection is trained using limit Gradient Propulsion model.The present invention improves the robustness of disaggregated model, reduces the risk of over-fitting by the assembled classification magnetocardiogram data of a variety of machine learning models, to improve the classification accuracy to myocardial ischemia patient.
Description
Technical field
The present invention relates to field of intelligent control technology, more particularly to a kind of medical instrument field of intelligent control technology, specifically
Method, system, storage medium and electronic equipment are examined in advance for the myocardial ischemia based on machine learning.
Background technology
Human life activity's under cover abundant electromagnetic information behind.Magnetocardiogram is a kind of by detecting the work of human heart electricity
The space magnetic field of movable property life and carry out the novel heart disease diagnosis method of imaging analysis.It is similar with traditional electrocardiogram, heart magnetic
Figure reflects the bioelectrical activity of heart, is a kind of functional imaging method.Due to completely it is noninvasive, radiationless, contactless, by body fluid
And the influence of bone etc. is smaller, and ring whirl electric current can be responded, magnetocardiogram includes the electricity that many conventional ECGs can not embody
Physiologic information shows better sensitivity and early diagnosis ability.Clinical studies show, magnetocardiogram are lacked in coronary heart disease, cardiac muscle
Blood etc. has good application potential, thus has high clinical research and application value.
Magnetocardiogram depends directly on data to the sensitivity and reliability of the disease forecastings such as myocardial ischemia and understands.The existing heart
Magnetic chart deciphering method depends on manual sort mostly, less efficient, and accuracy is also influenced by analysis personnel's experience, is constrained significantly
The clinical applicability of magnetocardiogram.Once it researched and proposed and automatic disaggregated model is carried out to magnetocardiogram using machine learning method, these
Model is to be based on single model such as support vector machines mostly, and the methods of direct core method or neural network realize classification feature, point
Class accuracy about 78% to 83% or so, is less than the precision of manual sort.
Invention content
The technical problem to be solved in the present invention is to provide the myocardial ischemia based on machine learning of the quickness and high efficiency side of examining in advance
Method, modeling, and corresponding storage medium and electronic equipment, the classification for solving existing magnetocardiogram rely on manually and divide
The low problem of class accuracy.
In order to solve the above technical problems, the present invention provides a kind of myocardial ischemia based on machine learning examines method in advance, including：
It obtains magnetocardiogram data and extracts the data set of the magnetocardiogram data；Data set progress feature extraction is formed multiple only
Vertical characteristic parameter collection；According to each characteristic parameter respectively to being carried out for each training pattern of each characteristic parameter collection configuration
Training, forms multiple graders；The output result of multiple graders is subjected to integrated study combination, exports final prediction
As a result.
It is described to be based on machine learning before the data set for extracting the magnetocardiogram data in one embodiment of the invention
Myocardial ischemia the method for examining further includes being filtered noise reduction to the magnetocardiogram data of acquisition in advance.
In one embodiment of the invention, the data set of the extraction magnetocardiogram data includes：Mark the heart magnetic
The QRS-T waves of diagram data；Average value processing is carried out to several QRS-T waves and obtains mean value oscillogram；In the mean value oscillogram subscript
Remember T wave wave band datas, forms the data set of the magnetocardiogram data.
In one embodiment of the invention, the multiple independent characteristic parameter collection includes：Time domain charactreristic parameter collection, frequency domain
Characteristic parameter collection and information theory characteristic parameter collection.
In one embodiment of the invention, the time domain charactreristic parameter collection and the frequency domain character parameter set are respectively adopted
Support vector machines training pattern is trained, and is instructed using limit Gradient Propulsion model to described information opinion characteristic parameter collection
Practice.
In one embodiment of the invention, the multiple independent characteristic parameter collection is divided into for each trained mould
The training set that type is trained and the test set for being tested each training pattern.
In one embodiment of the invention, test verification is carried out to each training pattern using ten folding cross-validation methods.
In one embodiment of the invention, the magnetocardiogram data include that the magnetocardiogram data of normal person and myocardial ischemia are suffered from
The magnetocardiogram data of person.
In order to solve the above technical problems, the present invention provides a kind of storage medium, wherein it is stored with computer program, the meter
When calculation machine program is by processor load and execution, realize that as above any myocardial ischemia based on machine learning examines method in advance.
In order to solve the above technical problems, the present invention provides a kind of electronic equipment, including：Processor and memory；Wherein,
The memory is for storing computer program；The processor is for computer program described in load and execution, so that the electricity
Sub- equipment executes as above any myocardial ischemia based on machine learning and examines method in advance.
As described above, the myocardial ischemia based on machine learning of the present invention examines method, storage medium and electronic equipment in advance, have
There is following advantageous effect：
The present invention improves the robust of disaggregated model by the assembled classification magnetocardiogram data of a variety of machine learning models
Property, the risk of over-fitting is reduced, to improve the classification accuracy to myocardial ischemia patient.
Description of the drawings
Fig. 1 is shown as the flow signal that the myocardial ischemia based on machine learning in one embodiment of the invention examines method in advance
Figure.
Fig. 2 is shown as the learning process stream that the myocardial ischemia based on machine learning in one embodiment of the invention examines method in advance
Journey block diagram.
Fig. 3 is shown as the application example that the myocardial ischemia based on machine learning in one embodiment of the invention examines method in advance
Figure.
Fig. 4 is shown as the myocardial ischemia based on machine learning in one embodiment of the invention and examines in advance in method using different instructions
Practice the classification results comparison diagram of model.
Specific implementation mode
Illustrate that embodiments of the present invention, those skilled in the art can be by this specification below by way of specific specific example
Disclosed content understands other advantages and effect of the present invention easily.The present invention can also pass through in addition different specific realities
The mode of applying is embodied or practiced, the various details in this specification can also be based on different viewpoints with application, without departing from
Various modifications or alterations are carried out under the spirit of the present invention.It should be noted that in the absence of conflict, following embodiment and implementation
Feature in example can be combined with each other.
It should be noted that the diagram provided in following embodiment only illustrates the basic structure of the present invention in a schematic way
Think, component count, shape and size when only display is with related component in the present invention rather than according to actual implementation in schema then
Draw, when actual implementation kenel, quantity and the ratio of each component can be a kind of random change, and its assembly layout kenel
It is likely more complexity.
For the automatic classification realized to magnetocardiogram, classification accuracy is improved, the present embodiment provides one kind being based on engineering
The myocardial ischemia of habit examines method, storage medium and electronic equipment in advance, and the classification for solving existing magnetocardiogram relies on manually and divides
The low problem of class accuracy.
A kind of myocardial ischemia based on machine learning provided in this embodiment is examined method, storage medium and electronic equipment and is adopted in advance
It is trained using independent parameter with multiple machine learning models, then is integrated result by integrated learning approach, improve cardiac muscle
Ischemic precision of prediction.(347 normal persons and 227 Patients with Myocardial Ischemia) are verified through a large amount of clinical datas, the present embodiment can be with
Reach accuracy 94.03%, precision ratio 86.56% and recall ratio 97.78%, reaches manual sort's precision.
The myocardial ischemia based on machine learning of the present invention described in detail below is examined method, storage medium and electronics and is set in advance
Standby principle and embodiment, make those skilled in the art do not need creative work be appreciated that the present invention based on engineering
The myocardial ischemia of habit examines method, storage medium and electronic equipment in advance.
As shown in Figure 1, the present embodiment provides a kind of myocardial ischemias based on machine learning to examine method in advance, it is described to be based on machine
The method of examining includes the following steps the myocardial ischemia of study in advance：
Step is S110, obtains magnetocardiogram data and extracts the data set of the magnetocardiogram data；
Step is S120, and carrying out feature extraction to the data set forms multiple independent characteristic parameter collection；
Step is S130, according to each characteristic parameter respectively to each training pattern for each characteristic parameter collection configuration
It is trained, forms multiple graders；
Step is S140, and the output result of multiple graders is carried out integrated study combination, exports final prediction
As a result.
The step S110 in method to step S140 is examined in advance to the myocardial ischemia based on machine learning of the present embodiment below
It is described in detail.
Step is S110, obtains magnetocardiogram data and extracts the data set of the magnetocardiogram data.
Wherein, the magnetocardiogram data include the magnetocardiogram data of normal person and the magnetocardiogram data of Patients with Myocardial Ischemia.
In this present embodiment, as shown in Fig. 2, including the pretreatment to original magnetocardiogram data：Extract the magnetocardiogram number
According to data set before, the method for examining further includes magnetocardiogram data to acquisition to the myocardial ischemia based on machine learning in advance
It is filtered noise reduction.
Original magnetocardiogram data are pre-processed, the original filtered noise reduction of magnetocardiogram data, removes ambient noise and its
He interferes.
In this present embodiment, the data set of the extraction magnetocardiogram data includes：Mark the magnetocardiogram data
QRS-T waves；Average value processing is carried out to several QRS-T waves and obtains mean value oscillogram；In label T wave wave bands in the mean value oscillogram
Data form the data set of the magnetocardiogram data.
I.e. in this present embodiment, pretreatment stage, original magnetocardiogram signal through noise reduction, go the processing such as interference, mean value, obtain
QRS-T waves signal after mean value.On mean value signal, T wave wave bands are intercepted.
That is, to data markers QRS-T waves after noise reduction, average value processing is carried out to several QRS-T waves, obtains mean value letter
Number, mean value oscillogram is obtained, the correspondence wave band of QRS wave and T waves is marked in mean value oscillogram.
Step is S120, and carrying out feature extraction to the data set forms multiple independent characteristic parameter collection.
In this present embodiment, the multiple independent characteristic parameter collection includes but not limited to：Time domain charactreristic parameter collection, frequency domain
Characteristic parameter collection and information theory characteristic parameter collection.
After pre-processing, into characteristic parameter extraction stage, based on T wave band signals extraction time domain, frequency domain, information theory neck
The data set is divided into several independent collection, the input parameter as machine learning model by the characteristic parameters such as domain.
For example, being directed to the data set, time domain, frequency domain are extracted respectively to the T wave wave bands of normal person and myocardial ischemia patient
With the characteristic parameter collection of information theory, time domain charactreristic parameter collection, frequency domain character parameter set and information theory characteristic parameter collection are formed.
Step is S130, according to each characteristic parameter respectively to each training pattern for each characteristic parameter collection configuration
It is trained, forms multiple graders.
For different characteristic parameter collection, it is trained using corresponding machine learning training pattern.I.e. with N number of independent
This N number of training pattern is trained on N number of independent characteristic parameter collection by machine learning classification training pattern respectively.
For example, in this present embodiment, branch is respectively adopted to the time domain charactreristic parameter collection and the frequency domain character parameter set
It holds vector machine training pattern to be trained, described information opinion characteristic parameter collection is trained using limit Gradient Propulsion model.
Three training patterns are independent from each other, and parameter is also adjusted according to corresponding characteristic parameter collection respectively
It is whole.
Wherein, according to the quantity N that can adjust characteristic parameter collection and training pattern the characteristics of the data set, training pattern
It is required that between each characteristic parameter collection independently of each other.
In this present embodiment, the multiple independent characteristic parameter collection is divided into for being instructed to each training pattern
Experienced training set and the test set for being tested each training pattern.The data set is divided into training set and survey
Examination collection is trained adjusting training model using training set data, the performance of training pattern is assessed using test set data test.
In this present embodiment, using but be not limited to ten folding cross-validation methods test verification carried out to each training pattern.
For example, in training process, three graders use ten folding cross-validation methods respectively, improve grader robustness, and then improve and divide
The stability of class device.
Step is S140, and the output result of multiple graders is carried out integrated study combination, exports final prediction
As a result.
The prediction result of N number of grader is integrated by integrated study method, the final result predicted.
Integrated study method can be selected according to the characteristics of data set, including but not limited to：Ballot method, mean value
Method, stacking, mixing method etc..The prediction result that different classifications device is combined for example, by using averaging method, avoids overfitting problem.
Using a clinical magnetocardiogram data set including 347 normal persons and 227 Patients with Myocardial Ischemia as application example.
In 227 patients, including 112 unstable angina patients, 47 Acute Myocardial Infarction Patients.227 people are preced with
Arteries and veins radiography, wherein 68 people's coronary artery chocking-up degrees are less than 35%, 159 people's coronary artery chocking-up degrees are higher than 35%.It establishes and is suitable for the number
It is as shown in Figure 3 according to the training pattern of collection.
It is extracted 164 characteristic parameters altogether for data sets, wherein time domain charactreristic parameter collection includes time domain charactreristic parameter 18
A, frequency domain character parameter set includes frequency domain character parameter 108, and information theory characteristic parameter collection includes information theory characteristic parameter 38
It is a.Different machines learning model classification results are more as shown in Figure 4.Model 5 is training pattern used by the present embodiment, point
Class shows (accuracy rate, precision ratio, recall ratio and F1 values) and is better than single machine learning model.
Relative to diagnostic method of the tradition based on time domain charactreristic parameter, pass through the assembled classification heart of a variety of machine learning models
Magnetic chart data improve the robustness of disaggregated model, reduce the risk of over-fitting, divide myocardial ischemia patient to improve
Class accuracy.
The present embodiment provides a kind of storage medium and a kind of electronic equipment, since the technical characteristic in previous embodiment 1 can
To be applied to storage medium embodiment, electronic equipment embodiment, thus it is no longer repeated.
The storage medium includes：The various media that can store program code such as ROM, RAM, magnetic disc or CD,
In be stored with computer program, the computer program when by processor load and execution, realize previous embodiment in be based on machine
The myocardial ischemia of study examines all or part of step of method in advance.
The electronic equipment is the equipment for including processor (CPU/MCU/SOC), memory (ROM/RAM), such as：It is desk-top
Machine, portable computer, smart mobile phone etc..Particularly, it is stored with computer program in the memory, the processor is in load and execution
When the computer program, the myocardial ischemia based on machine learning examines all or part of step of method in advance in realization previous embodiment
Suddenly.
In conclusion the present invention improves classification mould by the assembled classification magnetocardiogram data of a variety of machine learning models
The robustness of type reduces the risk of over-fitting, to improve the classification accuracy to myocardial ischemia patient.So the present invention has
Effect overcomes various shortcoming in the prior art and has high industrial utilization.
The principle of the present invention and its effect is only illustrated in the above embodiment, and is not intended to limit the present invention.This hair
It is bright to be improved under the premise of without prejudice to overall thought there are many more aspect, it all can be for those skilled in the art
Without prejudice under the spirit and scope of the present invention, can carry out modifications and changes to above-described embodiment.Therefore, technical field such as
Middle all equivalent modifications for being completed without departing from the spirit and technical ideas disclosed in the present invention of tool usually intellectual or
Change, should be covered by the claim of the present invention.
Claims (10)
1. a kind of myocardial ischemia based on machine learning examines method in advance, which is characterized in that including：
It obtains magnetocardiogram data and extracts the data set of the magnetocardiogram data；
Feature extraction is carried out to the data set and forms multiple independent characteristic parameter collection；
According to each characteristic parameter respectively to being trained for each training pattern of each characteristic parameter collection configuration, formed more
A grader；
The output result of multiple graders is subjected to integrated study combination, exports final prediction result.
2. the myocardial ischemia according to claim 1 based on machine learning examines method in advance, which is characterized in that extract the heart
Before the data set of magnetic chart data, the method for examining further includes the heart magnetic to acquisition to the myocardial ischemia based on machine learning in advance
Diagram data is filtered noise reduction.
3. the myocardial ischemia according to claim 1 or 2 based on machine learning examines method in advance, which is characterized in that described to carry
The data set of the magnetocardiogram data is taken to include：
Mark the QRS-T waves of the magnetocardiogram data；
Average value processing is carried out to several QRS-T waves and obtains mean value oscillogram；
In marking T wave wave band datas in the mean value oscillogram, the data set of the magnetocardiogram data is formed.
4. the myocardial ischemia according to claim 1 based on machine learning examines method in advance, which is characterized in that the multiple only
Vertical characteristic parameter collection includes：Time domain charactreristic parameter collection, frequency domain character parameter set and information theory characteristic parameter collection.
5. the myocardial ischemia according to claim 4 based on machine learning examines method in advance, which is characterized in that the time domain
Characteristic parameter collection and the frequency domain character parameter set are respectively adopted support vector machines training pattern and are trained, to described information opinion
Characteristic parameter collection is trained using limit Gradient Propulsion model.
6. the myocardial ischemia based on machine learning examines method in advance according to claim 1 or 5, which is characterized in that will be described
Multiple independent characteristic parameter collection are divided into the training set for being trained to each training pattern and are used for each instruction
Practice the test set that model is tested.
7. the myocardial ischemia according to claim 6 based on machine learning examines method in advance, which is characterized in that handed over using ten foldings
Fork proof method carries out test verification to each training pattern.
8. the myocardial ischemia according to claim 1 based on machine learning examines method in advance, which is characterized in that the magnetocardiogram
Data include the magnetocardiogram data of normal person and the magnetocardiogram data of Patients with Myocardial Ischemia.
9. a kind of storage medium, wherein being stored with computer program, which is characterized in that the computer program is loaded by processor
When execution, realize that the myocardial ischemia based on machine learning as described in any in claim 1 to 8 examines method in advance.
10. a kind of electronic equipment, which is characterized in that including：Processor and memory；Wherein,
The memory is for storing computer program；
The processor is for computer program described in load and execution, so that the electronic equipment is executed as in claim 1 to 8
Any myocardial ischemia based on machine learning examines method in advance.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810100231.8A CN108281184A (en) | 2018-02-01 | 2018-02-01 | Myocardial ischemia based on machine learning examines method, storage medium and electronic equipment in advance |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810100231.8A CN108281184A (en) | 2018-02-01 | 2018-02-01 | Myocardial ischemia based on machine learning examines method, storage medium and electronic equipment in advance |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108281184A true CN108281184A (en) | 2018-07-13 |
Family
ID=62807395
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810100231.8A Pending CN108281184A (en) | 2018-02-01 | 2018-02-01 | Myocardial ischemia based on machine learning examines method, storage medium and electronic equipment in advance |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108281184A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110946573A (en) * | 2019-11-01 | 2020-04-03 | 东软集团股份有限公司 | Cardiac arrest detection device, detection model training device, method and equipment |
| CN115985505A (en) * | 2023-01-19 | 2023-04-18 | 北京未磁科技有限公司 | Multidimensional fusion myocardial ischemia auxiliary diagnosis model and construction method thereof |
| CN116189902A (en) * | 2023-01-19 | 2023-05-30 | 北京未磁科技有限公司 | Myocardial ischemia prediction model based on magnetocardiogram video data and construction method thereof |
| CN117084684A (en) * | 2023-10-19 | 2023-11-21 | 山东大学齐鲁医院 | Characteristic parameter extraction method and system based on electrocardio current density map extension field |
| CN117100276A (en) * | 2023-10-23 | 2023-11-24 | 山东大学齐鲁医院 | Cardiac function detection system, computer storage medium and terminal |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1816303A (en) * | 2003-07-01 | 2006-08-09 | 卡迪尔马格成像公司 | Use of machine learning for classification of magneto cardiograms |
| CN104665849A (en) * | 2014-12-11 | 2015-06-03 | 西南交通大学 | Multi-physiological signal multi-model interaction-based high-speed railway dispatcher stress detecting method |
| US20170032221A1 (en) * | 2015-07-29 | 2017-02-02 | Htc Corporation | Method, electronic apparatus, and computer readable medium of constructing classifier for disease detection |
| CN107301221A (en) * | 2017-06-16 | 2017-10-27 | 华南理工大学 | A kind of data digging method of multiple features dimension heap fusion |
-
2018
- 2018-02-01 CN CN201810100231.8A patent/CN108281184A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1816303A (en) * | 2003-07-01 | 2006-08-09 | 卡迪尔马格成像公司 | Use of machine learning for classification of magneto cardiograms |
| CN104665849A (en) * | 2014-12-11 | 2015-06-03 | 西南交通大学 | Multi-physiological signal multi-model interaction-based high-speed railway dispatcher stress detecting method |
| US20170032221A1 (en) * | 2015-07-29 | 2017-02-02 | Htc Corporation | Method, electronic apparatus, and computer readable medium of constructing classifier for disease detection |
| CN107301221A (en) * | 2017-06-16 | 2017-10-27 | 华南理工大学 | A kind of data digging method of multiple features dimension heap fusion |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110946573A (en) * | 2019-11-01 | 2020-04-03 | 东软集团股份有限公司 | Cardiac arrest detection device, detection model training device, method and equipment |
| CN115985505A (en) * | 2023-01-19 | 2023-04-18 | 北京未磁科技有限公司 | Multidimensional fusion myocardial ischemia auxiliary diagnosis model and construction method thereof |
| CN116189902A (en) * | 2023-01-19 | 2023-05-30 | 北京未磁科技有限公司 | Myocardial ischemia prediction model based on magnetocardiogram video data and construction method thereof |
| CN115985505B (en) * | 2023-01-19 | 2023-12-12 | 北京未磁科技有限公司 | Multidimensional fusion myocardial ischemia auxiliary diagnosis model and construction method thereof |
| CN116189902B (en) * | 2023-01-19 | 2024-01-02 | 北京未磁科技有限公司 | Myocardial ischemia prediction model based on magnetocardiogram video data and construction method thereof |
| CN117084684A (en) * | 2023-10-19 | 2023-11-21 | 山东大学齐鲁医院 | Characteristic parameter extraction method and system based on electrocardio current density map extension field |
| CN117084684B (en) * | 2023-10-19 | 2024-02-02 | 山东大学齐鲁医院 | Characteristic parameter extraction method and system based on electrocardio current density map extension field |
| CN117100276A (en) * | 2023-10-23 | 2023-11-24 | 山东大学齐鲁医院 | Cardiac function detection system, computer storage medium and terminal |
| CN117100276B (en) * | 2023-10-23 | 2024-01-12 | 山东大学齐鲁医院 | Cardiac function detection system, computer storage medium and terminal |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108281184A (en) | Myocardial ischemia based on machine learning examines method, storage medium and electronic equipment in advance | |
| Zhao et al. | Noise rejection for wearable ECGs using modified frequency slice wavelet transform and convolutional neural networks | |
| Marinho et al. | A novel electrocardiogram feature extraction approach for cardiac arrhythmia classification | |
| Deng et al. | Extracting cardiac dynamics within ECG signal for human identification and cardiovascular diseases classification | |
| CN106108889B (en) | Electrocardiogram classification method based on deep learning algorithm | |
| Beritelli et al. | A novel training method to preserve generalization of RBPNN classifiers applied to ECG signals diagnosis | |
| WO2019100566A1 (en) | Artificial intelligence self-learning-based static electrocardiography analysis method and apparatus | |
| CN107837082A (en) | Electrocardiogram automatic analysis method and device based on artificial intelligence self study | |
| CN107951485A (en) | Ambulatory ECG analysis method and apparatus based on artificial intelligence self study | |
| Chen et al. | Self-organized neural network for the quality control of 12-lead ECG signals | |
| CN109864714A (en) | A kind of ECG Signal Analysis method based on deep learning | |
| CN110353665A (en) | ECG signal signature waveform detection method is singly led based on improved one-dimensional U-net | |
| CN111481192B (en) | Electrocardiosignal R wave detection method based on improved U-Net | |
| CN110495877A (en) | A kind of Multi resolution feature extraction method and device based on ECG | |
| CN112932498B (en) | T waveform state classification system with generalization capability based on deep learning | |
| Stracina et al. | Golden standard or obsolete method? Review of ECG applications in clinical and experimental context | |
| Zeng et al. | Ensemble learning on heartbeat type classification | |
| CN109044340B (en) | Electrocardiogram data classification method and device, electronic equipment and storage medium | |
| Yang et al. | Automated intra-patient and inter-patient coronary artery disease and congestive heart failure detection using EFAP-Net | |
| CN107944366A (en) | A kind of finger vein identification method and device based on attribute study | |
| Meqdad et al. | Meta structural learning algorithm with interpretable convolutional neural networks for arrhythmia detection of multisession ECG | |
| Subramanyan et al. | A novel deep neural network for detection of Atrial Fibrillation using ECG signals | |
| Irfan et al. | Deep learning based classification for healthcare data analysis system | |
| Jiao et al. | Pulse recognition of cardiovascular disease patients based on one-dimensional convolutional neural network | |
| Liu et al. | Analyze EEG signals with convolutional neural network based on power spectrum feature selection |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180713 |