CN113288161A - Single-lead ECG signal classification method and system based on improved residual error network - Google Patents
Single-lead ECG signal classification method and system based on improved residual error network Download PDFInfo
- Publication number
- CN113288161A CN113288161A CN202110614917.0A CN202110614917A CN113288161A CN 113288161 A CN113288161 A CN 113288161A CN 202110614917 A CN202110614917 A CN 202110614917A CN 113288161 A CN113288161 A CN 113288161A
- Authority
- CN
- China
- Prior art keywords
- residual error
- error network
- ecg signal
- data
- layer
- 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
- 238000000034 method Methods 0.000 title claims abstract description 36
- 206010003119 arrhythmia Diseases 0.000 claims abstract description 30
- 230000006793 arrhythmia Effects 0.000 claims abstract description 30
- 238000012545 processing Methods 0.000 claims abstract description 20
- 238000012360 testing method Methods 0.000 claims abstract description 14
- 238000000605 extraction Methods 0.000 claims abstract description 11
- 230000002159 abnormal effect Effects 0.000 claims description 18
- 238000012549 training Methods 0.000 claims description 18
- 238000001514 detection method Methods 0.000 claims description 17
- 238000002372 labelling Methods 0.000 claims description 16
- 238000011176 pooling Methods 0.000 claims description 16
- 239000012634 fragment Substances 0.000 claims description 13
- 230000008569 process Effects 0.000 claims description 13
- 208000037170 Delayed Emergence from Anesthesia Diseases 0.000 claims description 11
- 230000006870 function Effects 0.000 claims description 10
- 238000005070 sampling Methods 0.000 claims description 10
- 230000004913 activation Effects 0.000 claims description 8
- 230000008859 change Effects 0.000 claims description 6
- 238000010606 normalization Methods 0.000 claims description 6
- 230000011218 segmentation Effects 0.000 claims description 4
- 238000011478 gradient descent method Methods 0.000 claims description 2
- 238000012937 correction Methods 0.000 claims 1
- 230000004927 fusion Effects 0.000 abstract description 3
- 238000010586 diagram Methods 0.000 description 4
- 230000033764 rhythmic process Effects 0.000 description 4
- 206010003658 Atrial Fibrillation Diseases 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 3
- 230000007547 defect Effects 0.000 description 3
- 238000003745 diagnosis Methods 0.000 description 3
- 239000011159 matrix material Substances 0.000 description 2
- 238000012216 screening Methods 0.000 description 2
- 208000024172 Cardiovascular disease Diseases 0.000 description 1
- 230000005856 abnormality Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000013527 convolutional neural network Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 208000019622 heart disease Diseases 0.000 description 1
- 230000010365 information processing Effects 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000008447 perception Effects 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 230000002265 prevention Effects 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 230000010349 pulsation Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000010998 test method Methods 0.000 description 1
- 230000002861 ventricular Effects 0.000 description 1
Images
Classifications
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Data Mining & Analysis (AREA)
- Artificial Intelligence (AREA)
- General Health & Medical Sciences (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Molecular Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computing Systems (AREA)
- Software Systems (AREA)
- Computational Linguistics (AREA)
- Mathematical Physics (AREA)
- Physiology (AREA)
- Pathology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Psychiatry (AREA)
- Signal Processing (AREA)
- Probability & Statistics with Applications (AREA)
- Heart & Thoracic Surgery (AREA)
- Medical Informatics (AREA)
- Surgery (AREA)
- Animal Behavior & Ethology (AREA)
- Public Health (AREA)
- Veterinary Medicine (AREA)
- Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)
Abstract
A single lead ECG signal classification method and system based on improved residual error network, taking 5 seconds as window length, segmenting original ECG signal; and processing the signals by using an improved residual error network, inputting the segmented data into the residual error network, and outputting a processed output result as an identification result of the corresponding ECG signal. According to the method, the original electrocardiosignals do not need to be subjected to heart beat extraction processing; can realize the classification and identification of 5 types of arrhythmia, namely normal (N), supraventricular (S), venticular (V), fusion (F) and unknown (Q); the comprehensive accuracy rate of the test on the MIT-BIH arrhythmia database reaches more than 99 percent.
Description
Technical Field
The invention relates to the technical field of information processing, in particular to a single-lead ECG signal classification method and system based on an improved residual error network.
Background
Currently, the detection and diagnosis of arrhythmia mainly uses ECG, i.e. electrocardio signals. The traditional manual diagnosis method has various defects: the workload is large. The influence of experience and level of a reader is large, and errors are easy to occur. Because of the above problems, many automated detection algorithms are proposed, but the following two problems generally exist in the existing algorithms: the electrocardiosignals need to be separately photographed, namely P wave, QRS wave complex and the like are identified, and various abnormal detection can be carried out on the basis. And the error inevitably exists in the process of separately shooting the electrocardiosignal, and once the error occurs, the error can bring influence to the subsequent arrhythmia detection. Fewer arrhythmia types can be identified, more commonly including: distinguishing between normal sinus rhythm and arrhythmia; distinguishing between normal sinus rhythm and certain types of arrhythmia (e.g., atrial fibrillation); in addition to the above problems, how to use less lead data to realize detection and how to improve the recognition accuracy rate are also challenges faced by various existing detection algorithms.
In the prior art, a method for extracting characteristic parameters of arrhythmia, a device for identifying arrhythmia and a computer readable medium (publication number: CN108852347A) need parameter and characteristic extraction.
In the prior art, the arrhythmia identification and classification method (publication number: CN108647584A) based on sparse representation and neural network needs parameter extraction and complex preprocessing (needs to perform sub-beat on the original electrocardiogram, needs to perform dimensionality reduction and the like), and can distinguish 6 arrhythmia types.
In the prior art documents, a convolutional Neural Network is used in Robust ECG Signal Classification for Detection of Atrial Fibrillation Using a Novel Neural Network, but only four types of normal, Atrial Fibrillation, noise and the like can be identified, and the comprehensive accuracy can only reach 82%.
In the prior art document, Cardiologist-Level Arrhythmia Detection with conditional neural networks are used and identified in a sequence manner, but only 12 Arrhythmia types can be identified, and the comprehensive accuracy is lower than 80%.
Disclosure of Invention
Aiming at the defects in the prior art, the invention solves the main problems as follows:
(1) the traditional manual diagnosis method has the defects of large workload, large influence by experience and level of a reader and easy error occurrence.
(2) The existing various automatic processing methods have the following problems:
the electrocardiosignal needs to be subjected to sub-beat, namely P wave, QRS wave group and the like are identified, and errors can be introduced in the sub-beat process and influence is caused on the final identification effect;
feature extraction needs to be performed manually, such as: the process has large workload on one hand, and different feature extraction methods also have great influence on the final identification result on the other hand;
some identification methods need to use all data of 12-lead electrocardiogram, thereby invisibly improving the difficulty of identification and detection.
Therefore, the single-lead ECG signal classification method and system based on the improved residual error network are provided, less lead data are used, even only the single-lead data are used, the segmented original ECG signals are directly used as sequences to be processed on the premise of not performing sub-shooting and artificial feature extraction, and efficient and accurate identification of various arrhythmia types is realized.
In order to achieve the purpose, the invention adopts the following technical scheme:
a single lead ECG signal classification method based on an improved residual error network, comprising the steps of:
s1, segmenting a raw ECG signal: setting the length of a time window, segmenting an original ECG signal, acquiring voltage values of a plurality of points in each time period according to a sampling rate, and labeling segmented data;
s2, processing the signals by using a residual error network: and directly taking the segmented voltage data as the input of a residual error network, and outputting the classification result of the original ECG signal after the processing of the residual error network.
In order to optimize the technical scheme, the specific measures adopted further comprise:
further, the sampling rate is 360 times/second;
the time of the time window length is any one of 2 seconds, 4 seconds and 5 seconds;
and directly taking the segmented voltage data as the input of the residual error network, processing the voltage data in a sequence mode, and not performing heartbeat extraction or artificial feature extraction.
Further, a single lead ECG arrhythmia detection classification system based on an improved residual error network, comprising:
a raw ECG signal segmentation module; setting the length of a time window, segmenting an original ECG signal, acquiring voltage values of a plurality of points in each time period according to the sampling rate, and labeling segmented data by a labeling unit;
a signal processing module; and directly taking the segmented voltage data as the input of a residual error network, and outputting the classification result of the original ECG signal after the processing of the residual error network.
Furthermore, the number of the residual error networks is 3, and a single residual error network sequentially comprises a convolution layer and 8 groups of residual error blocks; each group of residual error blocks comprises two convolution layers, input voltage data are input into a leveling layer and a full connection layer after being processed by three independent residual error networks, wherein the leveling layer converts multidimensional data into one-dimensional data, the full connection layer classifies the arrhythmia data of the one-dimensional data, the size of a convolution kernel in each residual error network can be adjusted through design, accordingly, perception of the residual error network on different sizes is increased, and feature diversity can be increased due to the fact that the sizes of the convolution kernels are different to reflect distinctiveness.
Further, the air conditioner is provided with a fan,
before each convolution layer, respectively adopting batch processing normalization and modified linear activation;
each group of residual blocks comprises a discarding layer, and a certain proportion of connections are discarded randomly;
a pooling layer is arranged in each group of residual blocks;
the full connection layer comprises a dense layer and a Softmax activation function; a discarding layer is arranged in front of the full connection layer, wherein the dense layer is used for carrying out nonlinear change on the features acquired in front of the full connection layer, extracting the correlation among the features, mapping the features to an output space, and finally classifying the features by using softmax;
wherein the first pooling layer in the residual block in each residual network uses a pooling window with the size of 1, and the remaining pooling layers in the seven residual blocks sequentially use pooling windows with the sizes of 2, 1, 2, 1, 2, 1, 2;
convolution layers in three residual error networks respectively use convolution kernels with the sizes of 16, 32 and 64, the step length is 1, and edges are filled.
Further, the specific content of the labeling by the labeling unit is as follows:
if all heartbeats in a segment are normal, the segment is normal;
if normal and abnormal heart beats exist in one segment at the same time, the segment is abnormal;
if multiple types of abnormal heartbeats exist in one fragment at the same time, taking the most abnormal type in the fragment as the fragment type;
if there are multiple types and the same number of abnormal heartbeats in a fragment, the first occurring abnormal type is taken as the fragment type.
Further, the labeling unit rejects the label classified by non-pulsation in the labeling process;
dividing the labeled data into a training set and a test set according to the proportion of 70% of the training set and 30% of the test set, and sending the training set data and the test set data into a residual error network for training and testing;
in the residual error network training process, the loss function uses a cross entropy loss function, the optimizer uses a random gradient descent method, the initial value of the learning rate is 0.1, and exponential decay change is adopted in the follow-up process.
Further, a terminal carrying a controller of a single lead ECG signal classification method based on an improved residual error network according to any one of claims 1-2.
Further, a computer-readable storage medium comprising instructions which, when executed on a computer, cause the computer to perform a single lead ECG signal classification method based on a modified residual error network according to any one of claims 1-2.
The invention has the beneficial effects that:
arrhythmia recognition is realized through electrocardiosignals; in the identification process, the original electrocardiosignals do not need to be subjected to sub-shooting and artificial feature extraction; the identification can be completed only by single-lead electrocardiosignals, and the requirements on data acquisition equipment and acquisition operation are low; can identify 5 types of heart rhythms, such as Normal, Supraventricular, Ventricular, Fusion and Unknown, and has wide application range; meanwhile, the method has high identification accuracy, and the comprehensive accuracy of the test on the MIT-BIH arrhythmia database reaches more than 98 percent.
Meanwhile, the invention is helpful to reduce the difficulty and threshold of arrhythmia detection, such as: the arrhythmia screening can be realized through the intelligent watch, so that the popularization degree and the detection accuracy of early screening of various heart diseases are improved undoubtedly, and the early prevention of cardiovascular diseases is facilitated.
Drawings
FIG. 1 is a block diagram of the overall logic of the present invention.
Fig. 2 is a schematic diagram of an arrhythmia classification system of the present invention.
Fig. 3 is a schematic diagram of the residual error network structure of the present invention.
FIG. 4 is a graph of the distribution of the types of data in the training set of the present invention.
FIG. 5 is a graph of the distribution of the various types of data in the test set of the present invention.
FIG. 6 is a graph of model loss change (50 rounds) during the training process of the present invention.
Fig. 7 is a confusion matrix diagram of the present invention.
Detailed Description
The present invention will now be described in further detail with reference to the accompanying drawings.
As shown in fig. 1, a single lead ECG signal classification method based on improved residual error network includes:
s101: segmentation of the raw ECG signal: the raw ECG signal is segmented with 5 seconds as the window length.
Time window lengths of 2 seconds and 4 seconds can also be used, but experimental results show that the impact on the accuracy is not great. In addition, the time window may contain a different number of ECG signal data points depending on the sampling rate, where the MIT-BIH arrhythmia data sampling rate is 360Hz, i.e., 360 times/second.
S102: processing the signal using a residual network: the segmented data is input into a residual error network, and the result output by the residual error network after processing is the identification result corresponding to the original ECG signal.
As shown in fig. 2, a single lead ECG arrhythmia detection classification system based on a modified residual error network. The method comprises the following steps:
and a raw ECG signal segmentation module which segments the raw ECG signal with a window length of 5 seconds.
A signal processing module; and processing the signals by using an improved residual error network, inputting the segmented data into the residual error network, and outputting a result which is an identification result of the corresponding original ECG signal by using the processed residual error network.
The structure of the residual error network provided by the invention is shown in fig. 3, each single residual error network comprises 8 residual error blocks, and each residual error block has 2 convolutional layers; before each convolution layer, a batch normalization layer is adopted for batch processing normalization and a pre-activation layer is adopted for correcting linear activation; a pooling layer Maxpooling is arranged in each group of residual blocks to be used as a jump connection; a full link layer is shared and a drop layer Dropout is added before the full link layer.
In an embodiment of the present invention, the residual error network further includes:
batch Normalization: and (4) batch normalization layer for accelerating data convergence.
MaxPooling 1D: one-dimensional maximum pooling layer.
Conv 1D: a one-dimensional convolutional layer.
Pre-Activation: pre-activation layer, here all ReLu, i.e. modified linear elements, are chosen as activation functions.
Flatten: and the leveling layer converts the multidimensional data into one-dimensional data.
Dropout: layers are discarded and a certain proportion of connections are discarded randomly to prevent overfitting.
Full Connection: and (4) fully connecting the layers.
Softmax: softmax activates the function.
The dotted line represents the Skip Connection, i.e. the Skip Connection, used to construct the residual unit.
The parameters of each module of the single-lead ECG arrhythmia detection and classification system based on the residual error network are set as follows:
MaxPooling 1D: the MaxPooling1D unit in the first residual block uses a pooling window with the size of 1, the remaining seven residual blocks use pooling windows with the sizes of 2, 1, 2, 1, 2, 1 and 2 respectively, and the rest settings in the pooling layer except the pooling window are default settings.
Conv 1D: the three residual networks use convolution kernels of size 16, 32, 64, respectively, step size 1, edge-filling.
Dropout: the coefficient is set to 0.5, i.e. 25% of the connections are randomly dropped.
The invention is further described below with reference to the test procedures and results.
The raw data is 48 segments of two-lead data of 30 minutes with a sampling rate of 360Hz, i.e. 360 samples per second.
Segmenting data according to windows of 1800 sampling points by using II leads, simultaneously relieving the influence caused by unbalanced categories by mutual overlapping of the segments, and finally labeling the segmented data.
The labeling scheme is as follows: when all heartbeats in a segment are normal, the segment is normal; normal and abnormal heartbeats exist in one segment at the same time, and the segment is abnormal; multiple types of abnormal heartbeats exist in one fragment at the same time, and the most abnormal type in the fragment is taken as the fragment type; there are multiple types and the same number of abnormal heartbeats in a segment, and the type of the abnormality which appears first is taken as the segment type.
In the process of labeling, labels of non-beat classification are also removed, such as: waveform generation, P-wave peak, rhythm variation, etc. After processing, 5 types of 34945 data and labels thereof are obtained. The quantity of each type of data is as follows: normal (N)11781, supraventricular (S)6326, Ventriventricular (V)7352, fusion (F)6178, Unknow (Q) 3308. The data are divided into a training set and a test set according to the proportion of 70% of the training set and 30% of the test set, and the distribution of each type of the training set and the test set is shown in fig. 4 and 5.
Training set data was input into the modified residual network of fig. 3 for training. The loss function uses a cross entropy loss function, the optimizer uses a random gradient to decrease, the initial value of the learning rate is 0.1, and exponential decay change is adopted subsequently. Higher recognition accuracy can be realized after about 50 rounds, loss change in the training process is shown in figure 6, and a corresponding confusion matrix is shown in figure 7.
Among others, it is another object of the invention to provide a terminal carrying a controller implementing a single lead ECG signal classification method based on an improved residual error network.
It is, inter alia, another object of the invention to provide a computer readable storage medium comprising instructions which, when run on a computer, cause the computer to perform a single lead ECG signal classification method based on an improved residual network.
It should be noted that the terms "upper", "lower", "left", "right", "front", "back", etc. used in the present invention are for clarity of description only, and are not intended to limit the scope of the present invention, and the relative relationship between the terms and the terms is not limited by the technical contents of the essential changes.
The above is only a preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above-mentioned embodiments, and all technical solutions belonging to the idea of the present invention belong to the protection scope of the present invention. It should be noted that modifications and embellishments within the scope of the invention may be made by those skilled in the art without departing from the principle of the invention.
Claims (9)
1. A single lead ECG signal classification method based on an improved residual error network, comprising the steps of:
s1, segmenting a raw ECG signal: setting the length of a time window, segmenting an original ECG signal, acquiring voltage values of a plurality of points in each time period according to a sampling rate, and labeling segmented data;
s2, processing the signals by using a residual error network: and directly taking the segmented voltage data as the input of a residual error network, and outputting the classification result of the original ECG signal after the processing of the residual error network.
2. The single lead ECG signal classification method based on improved residual error network according to claim 1,
the sampling rate is 360 times/second;
the time of the time window length is any one of 2 seconds, 4 seconds and 5 seconds;
and directly taking the segmented voltage data as the input of the residual error network, processing the voltage data in a sequence mode, and not performing heartbeat extraction or artificial feature extraction.
3. A single lead ECG arrhythmia detection classification system based on an improved residual error network, comprising:
a raw ECG signal segmentation module; setting the length of a time window, segmenting an original ECG signal, acquiring voltage values of a plurality of points in each time period according to the sampling rate, and labeling segmented data by a labeling unit;
a signal processing module; and directly taking the segmented voltage data as the input of a residual error network, and outputting the classification result of the original ECG signal after the processing of the residual error network.
4. The improved residual error network-based single-lead ECG arrhythmia detection and classification system as claimed in claim 3, wherein the number of the residual error networks is 3, and the single residual error network comprises convolutional layer, 8 sets of residual error blocks; each group of residual error blocks comprises two convolution layers, input voltage data are input into a leveling layer and a full connection layer after being processed by three independent residual error networks, wherein the leveling layer converts multidimensional data into one-dimensional data, and the full connection layer classifies the one-dimensional data into arrhythmia data.
5. The improved residual error network-based single lead ECG arrhythmia detection classification system according to claim 4,
before each convolution layer, adopting batch processing normalization and correction linear activation;
each group of residual blocks comprises a discarding layer, and a certain proportion of connections are discarded randomly;
a pooling layer is arranged in each group of residual blocks;
the full connection layer comprises a dense layer and a Softmax activation function; a discarding layer is arranged in front of the full connecting layer;
wherein the first pooling layer in the residual block in each residual network uses a pooling window with the size of 1, and the remaining pooling layers in the seven residual blocks sequentially use pooling windows with the sizes of 2, 1, 2, 1, 2, 1, 2;
convolution layers in three residual error networks respectively use convolution kernels with the sizes of 16, 32 and 64, the step length is 1, and edges are filled.
6. The system for detecting and classifying single-lead ECG arrhythmia based on improved residual error network as claimed in claim 3, wherein the labeling unit labels the following specific contents:
if all heartbeats in a segment are normal, the segment is normal;
if normal and abnormal heart beats exist in one segment at the same time, the segment is abnormal;
if multiple types of abnormal heartbeats exist in one fragment at the same time, taking the most abnormal type in the fragment as the fragment type;
if there are multiple types and the same number of abnormal heartbeats in a fragment, the first occurring abnormal type is taken as the fragment type.
7. The system for detecting and classifying single-lead ECG arrhythmia based on improved residual error network as claimed in claim 6, wherein the labeling unit is used for rejecting the label of non-beat classification in the labeling process;
dividing the labeled data into a training set and a test set according to the proportion of 70% of the training set and 30% of the test set, and sending the training set data and the test set data into a residual error network for training and testing;
in the residual error network training process, the loss function uses a cross entropy loss function, the optimizer uses a random gradient descent method, the initial value of the learning rate is 0.1, and exponential decay change is adopted in the follow-up process.
8. A terminal characterized in that the terminal carries a controller of a single lead ECG signal classification method based on an improved residual error network as claimed in any one of claims 1-2.
9. A computer-readable storage medium comprising instructions which, when executed on a computer, cause the computer to perform a method of single lead ECG signal classification based on a modified residual error network according to any one of claims 1-2.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110614917.0A CN113288161A (en) | 2021-06-02 | 2021-06-02 | Single-lead ECG signal classification method and system based on improved residual error network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110614917.0A CN113288161A (en) | 2021-06-02 | 2021-06-02 | Single-lead ECG signal classification method and system based on improved residual error network |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113288161A true CN113288161A (en) | 2021-08-24 |
Family
ID=77327095
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110614917.0A Pending CN113288161A (en) | 2021-06-02 | 2021-06-02 | Single-lead ECG signal classification method and system based on improved residual error network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113288161A (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109063552A (en) * | 2018-06-22 | 2018-12-21 | 深圳大学 | A kind of multi-lead electrocardiosignal classification method and system |
| CN109303560A (en) * | 2018-11-01 | 2019-02-05 | 杭州质子科技有限公司 | A kind of atrial fibrillation recognition methods of electrocardiosignal in short-term based on convolution residual error network and transfer learning |
| CN109998532A (en) * | 2019-06-04 | 2019-07-12 | 广州视源电子科技股份有限公司 | Electrocardiosignal recognition methods and device based on the more topology convergence networks of multi-lead |
| CN110046604A (en) * | 2019-04-25 | 2019-07-23 | 成都信息工程大学 | A kind of single lead ECG arrhythmia detection classification method based on residual error network |
| CN111329445A (en) * | 2020-02-20 | 2020-06-26 | 广东工业大学 | Atrial fibrillation identification method based on group convolution residual error network and long-term and short-term memory network |
-
2021
- 2021-06-02 CN CN202110614917.0A patent/CN113288161A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109063552A (en) * | 2018-06-22 | 2018-12-21 | 深圳大学 | A kind of multi-lead electrocardiosignal classification method and system |
| CN109303560A (en) * | 2018-11-01 | 2019-02-05 | 杭州质子科技有限公司 | A kind of atrial fibrillation recognition methods of electrocardiosignal in short-term based on convolution residual error network and transfer learning |
| CN110046604A (en) * | 2019-04-25 | 2019-07-23 | 成都信息工程大学 | A kind of single lead ECG arrhythmia detection classification method based on residual error network |
| CN109998532A (en) * | 2019-06-04 | 2019-07-12 | 广州视源电子科技股份有限公司 | Electrocardiosignal recognition methods and device based on the more topology convergence networks of multi-lead |
| CN111329445A (en) * | 2020-02-20 | 2020-06-26 | 广东工业大学 | Atrial fibrillation identification method based on group convolution residual error network and long-term and short-term memory network |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110046604B (en) | Single-lead ECG arrhythmia detection and classification method based on residual error network | |
| CN109620152B (en) | MutifacolLoss-densenert-based electrocardiosignal classification method | |
| EP4092570A1 (en) | Heartbeat classification method for multi-tag ecg signal labeling, and device | |
| CN108511055B (en) | Ventricular premature beat recognition system and method based on classifier fusion and diagnosis rules | |
| CN111626114B (en) | Electrocardiosignal arrhythmia classification system based on convolutional neural network | |
| CN110974214A (en) | Automatic electrocardiogram classification method, system and equipment based on deep learning | |
| CN110638430B (en) | Method for building cascade neural network ECG signal arrhythmia classification model | |
| CN113229825A (en) | Deep neural network-based multi-label multi-lead electrocardiogram classification method | |
| CN113995419B (en) | Atrial fibrillation risk prediction system based on heartbeat rhythm signal and application thereof | |
| CN111956208B (en) | ECG signal classification method based on ultra-lightweight convolutional neural network | |
| CN110367969A (en) | A kind of improved electrocardiosignal K-Means Cluster | |
| CN110558975B (en) | Electrocardiosignal classification method and system | |
| CN113057648A (en) | ECG signal classification method based on composite LSTM structure | |
| CN111904411A (en) | Multi-lead heartbeat signal classification method and device based on multi-scale feature extraction | |
| CN113288157A (en) | Arrhythmia classification method based on depth separable convolution and improved loss function | |
| CN113855042B (en) | Sequence labeling-based multi-lead electrocardiosignal classification method integrating depth and medical features | |
| CN112957052B (en) | Multi-lead electrocardiosignal classification method based on NLF-CNN lead fusion depth network | |
| CN112450944A (en) | Label correlation guide feature fusion electrocardiogram multi-classification prediction system and method | |
| CN113288161A (en) | Single-lead ECG signal classification method and system based on improved residual error network | |
| CN113876331B (en) | Electrocardiosignal-oriented semi-supervised atrial fibrillation automatic detection system | |
| CN114224355A (en) | Electrocardiosignal classification training method, classification method, device and storage medium | |
| CN114366116B (en) | Parameter acquisition method based on Mask R-CNN network and electrocardiogram | |
| CN111261283A (en) | Electrocardiosignal deep neural network modeling method based on pyramid type convolution layer | |
| CN115414049A (en) | Wearable electrocardiogram real-time diagnosis system based on deep neural network | |
| CN113679393B (en) | ECG data feature generation model based on contrast predictive coding |
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: 20210824 |