CN116616784A - Electrocardiogram classification method, device and storage medium based on deep learning - Google Patents
Electrocardiogram classification method, device and storage medium based on deep learning Download PDFInfo
- Publication number
- CN116616784A CN116616784A CN202310701079.XA CN202310701079A CN116616784A CN 116616784 A CN116616784 A CN 116616784A CN 202310701079 A CN202310701079 A CN 202310701079A CN 116616784 A CN116616784 A CN 116616784A
- Authority
- CN
- China
- Prior art keywords
- data
- electrocardiogram
- plane data
- identified
- classification
- 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 39
- 238000013135 deep learning Methods 0.000 title claims abstract description 31
- 238000000605 extraction Methods 0.000 claims abstract description 40
- 230000004927 fusion Effects 0.000 claims abstract description 40
- 238000012545 processing Methods 0.000 claims abstract description 39
- 230000006870 function Effects 0.000 claims abstract description 24
- 238000007781 pre-processing Methods 0.000 claims abstract description 12
- 238000003491 array Methods 0.000 claims description 22
- 230000005856 abnormality Effects 0.000 claims description 9
- 238000005070 sampling Methods 0.000 claims description 9
- 238000002565 electrocardiography Methods 0.000 claims description 5
- 238000007499 fusion processing Methods 0.000 claims description 3
- 238000013507 mapping Methods 0.000 claims description 3
- 230000002159 abnormal effect Effects 0.000 abstract description 12
- 238000010801 machine learning Methods 0.000 abstract 1
- 208000037170 Delayed Emergence from Anesthesia Diseases 0.000 description 6
- 238000010586 diagram Methods 0.000 description 4
- 230000008569 process Effects 0.000 description 4
- 230000000694 effects Effects 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 238000004590 computer program Methods 0.000 description 2
- 238000003745 diagnosis Methods 0.000 description 2
- 239000004973 liquid crystal related substance Substances 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 230000009467 reduction Effects 0.000 description 2
- 230000006793 arrhythmia Effects 0.000 description 1
- 206010003119 arrhythmia Diseases 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000006835 compression Effects 0.000 description 1
- 238000007906 compression Methods 0.000 description 1
- 230000008094 contradictory effect Effects 0.000 description 1
- 238000013527 convolutional neural network Methods 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 208000019622 heart disease Diseases 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 208000010125 myocardial infarction Diseases 0.000 description 1
- 208000031225 myocardial ischemia Diseases 0.000 description 1
- 230000003334 potential effect Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 230000001360 synchronised effect Effects 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Classifications
-
- 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
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
- A61B5/316—Modalities, i.e. specific diagnostic methods
- A61B5/318—Heart-related electrical modalities, e.g. electrocardiography [ECG]
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
- A61B5/316—Modalities, i.e. specific diagnostic methods
- A61B5/318—Heart-related electrical modalities, e.g. electrocardiography [ECG]
- A61B5/346—Analysis of electrocardiograms
-
- 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
- A61B5/7203—Signal processing specially adapted for physiological signals or for diagnostic purposes for noise prevention, reduction or removal
-
- 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
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7235—Details of waveform analysis
- A61B5/7264—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems
- A61B5/7267—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems involving training the classification device
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/10—Pre-processing; Data cleansing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
- G06F18/251—Fusion techniques of input or preprocessed data
-
- 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
- G06N3/0464—Convolutional networks [CNN, ConvNet]
-
- 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
- G06F2218/04—Denoising
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- Molecular Biology (AREA)
- Biophysics (AREA)
- Evolutionary Computation (AREA)
- Biomedical Technology (AREA)
- General Health & Medical Sciences (AREA)
- Surgery (AREA)
- Public Health (AREA)
- General Engineering & Computer Science (AREA)
- Veterinary Medicine (AREA)
- General Physics & Mathematics (AREA)
- Animal Behavior & Ethology (AREA)
- Medical Informatics (AREA)
- Heart & Thoracic Surgery (AREA)
- Pathology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Signal Processing (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Physiology (AREA)
- Psychiatry (AREA)
- Cardiology (AREA)
- Mathematical Physics (AREA)
- Probability & Statistics with Applications (AREA)
- Computing Systems (AREA)
- Computational Linguistics (AREA)
- Software Systems (AREA)
- Fuzzy Systems (AREA)
- Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)
Abstract
The invention relates to the field of signal processing and medical treatment, and discloses an electrocardiogram classification method based on deep learning, which comprises the following steps: preprocessing an electrocardiogram to be identified to obtain frontal plane data of 6*n lead signals and transverse plane data of 6*n lead signals; processing the frontal plane data and the transverse plane data respectively by using a fusion model to obtain first heart level plane data and second heart level plane data, and splicing the first heart level plane data and the second heart level plane data to obtain spliced data; extracting features of the spliced data through a feature extraction model to obtain feature data; and classifying the extracted characteristic data by using a classification function to obtain a classification result of the electrocardiogram to be identified. The invention can be used for classifying the electrocardiosignals normally and abnormally by utilizing machine learning, and can be used for intelligently classifying the electrocardiosignals of patients in the medical field, so that doctors have more time to analyze the abnormal electrocardiosignals.
Description
Technical Field
The present invention relates to the field of signal processing, and in particular, to an electrocardiographic classification method, device and storage medium based on deep learning.
Background
Along with the continuous improvement of the living standard of people, the health of people is more and more emphasized. Electrocardiography is an important means of clinical medical diagnosis of heart diseases, and can assist medical staff in diagnosis.
Currently, in the medical field, various types of electrocardiographs have been widely used in clinic.
However, the current electrocardiographic acquisition requires manual judgment of an electrocardiograph to judge whether the electrocardiograph of the patient is normal or abnormal, and manual checking of whether the electrocardiograph is abnormal is very time-consuming and inefficient, so that a doctor takes much time to judge whether the electrocardiograph of the patient is abnormal, and the doctor is separated from the effort of analyzing the cause of the electrocardiograph abnormality of the patient.
Therefore, the electrocardiograph has a problem that whether the electrocardiogram is abnormal or not cannot be judged intelligently.
Disclosure of Invention
In view of the foregoing, it is necessary to provide an electrocardiographic classification method based on deep learning, which aims to solve the technical problem that whether an electrocardiograph is abnormal or not cannot be intelligently judged in the prior art.
The invention provides an electrocardiogram classification method based on deep learning, which comprises the following steps:
preprocessing an electrocardiogram to be identified to obtain electrocardiosignal data, and dividing the electrocardiosignal data into 12 x n two-dimensional arrays, wherein the two-dimensional arrays comprise frontal plane data of 6*n lead signals and transverse plane data of 6*n lead signals;
inputting the two-dimensional array into a preset fusion model, processing the frontal plane data to obtain first heart level plane data, processing the transverse plane data to obtain second heart level plane data, and splicing the first heart level plane data and the second heart level plane data to obtain spliced data;
performing feature extraction on the spliced data by using a preset feature extraction model to obtain extracted feature data;
and classifying the extracted characteristic data by using a preset classification function to obtain a classification result of the electrocardiogram to be identified.
Optionally, the preprocessing the electrocardiogram to be identified to obtain electrocardiographic signal data includes:
denoising the electrocardiogram to be identified by using a preset first filter;
and carrying out downsampling treatment on the denoised electrocardiograph data by using a preset second filter to obtain electrocardiograph signal data with a preset sampling rate.
Optionally, the processing the frontal plane data to obtain first heart level plane data includes:
and carrying out data signal fusion on 6*n lead signals of the frontal plane data by utilizing convolution of the fusion model to obtain first heart level plane data with data dimension of 1x n 1x a, wherein a represents the channel number of the convolution kernel, and n1 is the data size obtained after convolution of the convolution kernel.
Optionally, the processing the cross plane data to obtain second heart level plane data includes:
and carrying out data signal fusion on 6*n lead signals of the frontal plane data by utilizing the convolution of the fusion model to obtain second heart level plane data with the data dimension of 1x n 1x a, wherein a represents the channel number of the convolution kernel, and n1 is the data size obtained after the convolution of the convolution kernel.
Optionally, the first heart level surface data and the second heart level surface data are spliced to obtain spliced data; comprising the following steps:
and connecting the first heart level surface data with the data dimension of 1x n 1x a and the second heart level surface data with the data dimension of 1x n 1x a after fusion processing of the fusion model to obtain spliced data with the data dimension of 1x n 1x 2a, wherein n1 is the data size obtained after convolution of the convolution kernel of the fusion model, and a is the channel number representing the convolution kernel.
Optionally, the classifying the extracted feature data by using a preset classification function to obtain a classification result of the electrocardiogram to be identified includes:
classifying the electrocardiogram result to be identified by using a softmax function, mapping the extracted characteristic data into a value of (0, 1) by the softmax function, outputting 0 or 1,0 representing the electrocardiogram abnormality to be identified, and 1 representing the electrocardiogram abnormality to be identified.
Optionally, the feature extraction is performed on the spliced data by using a preset feature extraction model, so as to obtain extracted feature data, which includes:
carrying out convolution processing on the spliced data for a preset number of times by utilizing the characteristic extraction model;
and performing feature extraction on the spliced data after the convolution processing by using the full connection layer of the feature extraction model to obtain the feature data.
In order to solve the above problems, the present invention also provides an electrocardiographic signal image classifying apparatus, the apparatus comprising:
the data processing module is used for preprocessing an electrocardiogram to be identified to obtain electrocardiosignal data, dividing the electrocardiosignal data into 12-n two-dimensional arrays, wherein the two-dimensional arrays comprise frontal surface data of 6*n lead signals and transverse surface data of 6*n lead signals;
the data layer fusion module is used for inputting the two-dimensional array into a preset fusion model, processing the frontal plane data to obtain first heart level plane data, processing the transverse plane data to obtain second heart level plane data, and splicing the first heart level plane data and the second heart level plane data to obtain spliced data;
the feature extraction module is used for carrying out feature extraction on the spliced data by utilizing a preset feature extraction model to obtain extracted feature data;
and the classification module is used for classifying the extracted characteristic data by utilizing a preset classification function to obtain the classification result of the electrocardiogram to be identified.
In order to solve the above-mentioned problems, the present invention also provides an electronic apparatus including:
at least one processor; the method comprises the steps of,
a memory communicatively coupled to the at least one processor; wherein,,
the memory stores a deep learning based electrocardiogram classification program executable by the at least one processor, the deep learning based electrocardiogram classification program being executable by the at least one processor to enable the at least one processor to perform a deep learning based electrocardiogram classification method as described above.
In order to solve the above-mentioned problems, the present invention also provides a computer-readable storage medium having stored thereon a deep learning-based electrocardiogram classification program executable by one or more processors to implement the above-mentioned deep learning-based electrocardiogram classification method.
Compared with the prior art, the method captures the synchronous electrocardiosignal characteristics of different leads compared with the prior art, then fuses and finally classifies.
The method comprises the steps of preprocessing an electrocardiogram to be identified to obtain electrocardiosignal data, dividing the electrocardiosignal data into a two-dimensional array of 12 x n, wherein the two-dimensional array comprises frontal surface data of 6*n lead signals and transverse surface data of 6*n lead signals, inputting the two-dimensional array into a preset fusion model, processing the frontal surface data to obtain first heart level surface data, processing the transverse surface data to obtain second heart level surface data, and splicing the first heart level surface data and the second heart level surface data to obtain spliced data, so that the characteristics of the electrocardiogram to be identified are fused together, the characteristics of the spliced data are conveniently extracted by using a preset characteristic extraction model, the extracted characteristic data is obtained, the extracted characteristic data is classified by using a preset classification function, the classification result of the electrocardiogram to be identified is obtained, and the extracted characteristic data can be classified.
Therefore, the technical scheme of the invention respectively fuses the electrocardiographic activity information collected by different planes, reflects the heart activity original information and is more beneficial to the following feature extraction and classification work. And the multi-lead electrocardiosignals are fused through the convolutional neural network, so that fusion signals of different planes are obtained, then feature extraction is carried out, and abnormal electrocardiosignals are reclassified, and in the medical field, whether the electrocardiosignals are abnormal can be intelligently judged, so that the electrocardiosignal classification efficiency of the electrocardiosignals is improved, and doctors can have more time to analyze the abnormal electrocardiosignals.
Drawings
Fig. 1 is a schematic flow chart of an electrocardiographic classification method based on deep learning according to an embodiment of the present invention;
FIG. 2 is a schematic block diagram of an electrocardiographic classifying device based on deep learning according to an embodiment of the present invention;
fig. 3 is a schematic structural diagram of an electronic device for implementing an electrocardiographic classification method based on deep learning according to an embodiment of the present invention;
FIG. 4a shows the lead positions of a conventional ECG12 leads;
fig. 4b shows a schematic position of 6 frontal leads I, II, III, aVR, aVL, aVF of the 12 leads;
FIG. 4c shows a schematic of the positions of 6 transversal leads V1, V2, V3, V4, V5, V6 of the 12 leads;
the achievement of the objects, functional features and advantages of the present invention will be further described with reference to the accompanying drawings, in conjunction with the embodiments.
Detailed Description
The present invention will be described in further detail with reference to the drawings and examples, in order to make the objects, technical solutions and advantages of the present invention more apparent. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
It should be noted that the description of "first", "second", etc. in this disclosure is for descriptive purposes only and is not to be construed as indicating or implying a relative importance or implying an indication of the number of technical features being indicated. Thus, a feature defining "a first" or "a second" may explicitly or implicitly include at least one such feature. In addition, the technical solutions of the embodiments may be combined with each other, but it is necessary to base that the technical solutions can be realized by those skilled in the art, and when the technical solutions are contradictory or cannot be realized, the combination of the technical solutions should be considered to be absent and not within the scope of protection claimed in the present invention.
Referring to fig. 1, a flowchart of an electrocardiographic classification method based on deep learning according to an embodiment of the present invention is shown. The method is performed by an electronic device.
In this embodiment, an electrocardiographic classification method based on deep learning includes:
s1, preprocessing an electrocardiogram to be identified to obtain electrocardiosignal data, and dividing the electrocardiosignal data into 12 x n two-dimensional arrays, wherein the two-dimensional arrays comprise frontal surface data of 6*n lead signals and transverse surface data of 6*n lead signals.
In this embodiment, an electrocardiogram is a main way to record and monitor the heart potential activity signal, and further diagnose abnormal conditions of the heart, including arrhythmia, ischemic heart disease, myocardial infarction, etc., by observing the law of heart beat. The electrocardiographic signal data are electrocardiographic 12-lead data, and the electrocardiographic 12-lead data include 6 limb lead data, namely frontal face data (I, II, III, aVR, aVL, aVF) and 6 chest lead data, namely transverse face data (V1-V6). The limb leads including standard bipolar leads (i, ii, and iii) and compression leads (aVR, aVL, and aVF) are shown in fig. 4:
the 12 lead data are time sequence signals, and reflect the electrocardio activity information of the heart at different angles. A two-dimensional array is essentially an array with an array as array elements, i.e., an array of arrays. The preprocessing of the electrocardiogram to be identified involves denoising and dimension reduction of the electrocardiogram data to be identified. The first filter is a noise reduction filter, and in this embodiment, an EMI filter may be used, and denoising processing is performed on electrocardiographic data to be identified by using the first filter; the first filter is a downsampling filter, in this embodiment, a downsampling digital filter may be used, and the second filter that can downsample is used to downsample the electrocardiographic data to be identified, so that the sampling rate of 12-lead extraction is 250HZ, and the sampling values are 2000, and the sampling value and the sampling rate are not limited. The final data is processed into a two-dimensional array of 12 x 2000, each row representing one lead of electrocardiographic signal data, reducing the data sample and thus the time for computer data computation.
In one embodiment, the electrocardiographic data to be identified may be denoised by a binning method, and further description is omitted. In the step S1, the electrocardiogram to be identified is divided into 12-n two-dimensional arrays after denoising and downsampling, the two-dimensional arrays are directly input into a preset fusion model to be fused, in the medical field, the electrocardiogram of a patient consulted by a hospital is acquired through an electrocardiogram instrument, and then the acquired electrocardiogram is subjected to complaint operation, so that the 12-n two-dimensional arrays of the electrocardiogram of the patient consulted can be obtained.
S2, inputting the two-dimensional array into a preset fusion model, processing the frontal plane data to obtain first heart level plane data, processing the transverse plane data to obtain second heart level plane data, and splicing the first heart level plane data and the second heart level plane data to obtain spliced data.
In this embodiment, the electrocardiographic sampling signals of 12 lead data in electrocardiographic data to be identified are processed into two-dimensional arrays, for example, 12 lead data are processed into 12 x 2000 two-dimensional arrays, and according to the frontal plane data and the leads of the transverse plane data, the 12 lead data are respectively divided into 6 lead data electrocardiographic sampling signals as a group, so as to obtain two 6 x 2000 two-dimensional arrays, wherein one 6 x 2000 two-dimensional array is frontal plane data, and the other 6 x 2000 two-dimensional array is transverse plane data.
And finally, simultaneously passing the two-dimensional arrays of 6 x 2000 into a preset fusion model respectively.
The method comprises the steps that a preset fusion model is a convolution neural network with a convolution kernel 6*N (N can be formulated according to requirements and is not limited in more than needed), then signal fusion of a data layer is carried out on a two-dimensional array of 6 x 2000 of frontal plane data, and convolution is carried out on first lead data of 6 x 2000 of frontal plane data by using the convolution of the fusion model to obtain first heart level plane data; meanwhile, a preset fusion model is a convolution neural network with a convolution kernel of 6*N, and a 6 x 2000 two-dimensional array of transverse plane data is convolved to obtain second heart level plane data.
The data dimension of the first heart level surface data and the second heart level surface data is 1×n1a, wherein n1 is the data size obtained after convolution, and a is a channel representing a convolution kernel.
The first heart level surface data and the second heart level surface data are respectively obtained by fusing fusion models of the same convolution kernel, the dimension of the data is 1x n 1x a, namely (height, width) two dimensions are the same, the first heart level surface data and the second heart level surface data are directly input into a computer, the computer can connect the data with the same dimension (height and width of the data) to generate three-dimensional data, spliced data with the data dimension of 1x n 1x 2a is obtained, n1 is the data size obtained after convolution, and 2a is the channel number representing the convolution kernel.
In the step S2, the frontal plane data and the transverse plane data can be respectively fused together through a fusion model to obtain original characteristics and spliced data, so that the characteristics of the electrocardiograms to be identified are fused together, the subsequent characteristic extraction is convenient, and the characteristic information contained in the electrocardiograms of the patient to be consulted can be obtained in the medical field.
And S3, performing feature extraction on the spliced data by using a preset feature extraction model to obtain extracted feature data.
In this embodiment, spliced data with a data dimension of 1×n12a is input into a preset feature extraction model, the preset feature extraction model uses a convolution residual network model to refer to 17 layers of convolution layers, wherein 4 residual blocks including two layers of residual learning units are provided, and an eighteenth layer is a full-connection layer.
First, the spliced data is input to the first layer as an input value X1, the first layer is subjected to convolution of 1X50X4, the output characteristic is represented by F1 (X), after the second layer 1X50X4 is subjected to convolution with the third layer 1X50X8, the third layer output characteristic is represented by F3 (X), the F3 (X) is input as a residual block, the residual block is divided into two branches, one branch directly reaches the input position of the sixth layer after passing through shortcut connections (shortcut connection), the other branch F3 (X) continues to perform convolution output F4 (X) of the fourth layer 1X50X8, F4 (X) is output F5 (X) through convolution of the fifth layer 1X50X8, and at this time, the input value x6=f5 (X) +f3 (X) of the sixth layer.
Then, X6 is subjected to convolution of a sixth layer 1X50X8, the output F6 (X) is input to a seventh layer to be subjected to convolution of 1X50X16, and the output characteristic is represented by F7 (X).
The residual block is input to F7 (X), one is directly reached to the tenth layer input position after shortcut connections (shortcut connection), the other branch F7 (X) continues to perform the convolution of the eighth layer 1X50X16, F8 (X) is output through the convolution of the ninth layer 1X50X16, and the tenth layer input value x10=f9 (X) +f7 (X) at this time.
X10 is convolved with a tenth layer of 1X50X16, the output F10 (X) is input to an eleventh layer to convolve with a 1X50X32, and the output characteristic is denoted as F11 (X).
The residual block is input to F11 (X), one is directly reached to the fourteenth layer input position after shortcut connections (shortcut connection), the other branch F11 (X) continues to perform the convolution of the twelfth layer 1X50X32, F12 (X) is output through the convolution of the thirteenth layer 1X50X32, and the fourteenth layer input value x14=f13 (X) +f11 (X) at this time.
X14 is convolved with 1X50X16 in the fourteenth layer, F14 (X) output is input to the fifteenth layer for convolving with 1X50X32, and the output characteristic is denoted by F15 (X).
The residual block is input to F15 (X), one is directly reached to the input position of the eighteenth layer after shortcut connections (shortcut connection), the other branch F15 (X) continues to perform convolution of the sixteenth layer 1X50X32, F16 (X) is output through convolution of the seventeenth layer 1X50X32, and the input value x18=f17 (X) +f15 (X) of the eighteenth layer at this time.
X18 is input to the fully connected layer.
The step S3 may be to extract features of the electrocardiogram to be identified, to obtain feature data of two different planes.
S4, classifying the extracted characteristic data by using a preset classification function to obtain a classification result of the electrocardiogram to be identified.
The Softmax is a classification function known in the art, the step can classify the electrocardiogram result to be identified by using the Softmax function, the extracted characteristic data is mapped into a value output of 0 or 1,0 (0, 1) by the Softmax function, the value output of 0 or 1,0 represents the electrocardiogram abnormality to be identified, and 1 represents the electrocardiogram abnormality to be identified, and the description is omitted again.
The step S4 can classify the extracted characteristic data, and in the medical field, the electrocardiograph of the patient consulted by the hospital can be intelligently classified, and whether the electrocardiograph of the patient consulted by the hospital is abnormal or not can be known without manual judgment, so that doctors have more time to analyze the abnormal electrocardiograph.
Fig. 2 is a schematic block diagram of an image classification device based on electrocardiographic signals according to an embodiment of the present invention.
The image classification device 100 based on electrocardiosignals can be installed in electronic equipment. Depending on the implementation, the apparatus 100 for classifying images based on electrocardiographic signals may include a data processing module 110, a data layer fusion module 120, a feature extraction module 130, and a classification module 140. The module of the invention, which may also be referred to as a unit, refers to a series of computer program segments, which are stored in the memory of the electronic device, capable of being executed by the processor of the electronic device and of performing a fixed function.
In the present embodiment, the functions concerning the respective modules/units are as follows:
the data processing module 110 is configured to pre-process an electrocardiogram to be identified to obtain electrocardiographic signal data, and divide the electrocardiographic signal data into 12×n two-dimensional arrays, where the two-dimensional arrays include frontal plane data of 6*n lead signals and transverse plane data of 6*n lead signals;
the data layer fusion module 120 is configured to input the two-dimensional array into a preset fusion model, process the frontal plane data to obtain first heart level plane data, process the transverse plane data to obtain second heart level plane data, and splice the first heart level plane data and the second heart level plane data to obtain spliced data;
the feature extraction module 130 is configured to perform feature extraction on the spliced data by using a preset feature extraction model, so as to obtain extracted feature data;
and the classification module 140 is configured to classify the extracted feature data by using a preset classification function, so as to obtain a classification result of the electrocardiogram to be identified.
In one embodiment, the preprocessing the electrocardiogram to be identified to obtain electrocardiographic signal data includes:
denoising the electrocardiogram to be identified by using a preset first filter;
and carrying out downsampling treatment on the denoised electrocardiograph data by using a preset second filter to obtain electrocardiograph signal data with a preset sampling rate.
In one embodiment, the processing the frontal plane data to obtain first heart level plane data includes:
and carrying out data signal fusion on 6*n lead signals of the frontal plane data by utilizing convolution of the fusion model to obtain first heart level plane data with data dimension of 1x n 1x a, wherein a represents the channel number of the convolution kernel, and n1 is the data size obtained after convolution of the convolution kernel.
In one embodiment, the processing the cross plane data to obtain second heart level plane data includes:
and carrying out data signal fusion on 6*n lead signals of the frontal plane data by utilizing the convolution of the fusion model to obtain second heart level plane data with the data dimension of 1x n 1x a, wherein a represents the channel number of the convolution kernel, and n1 is the data size obtained after the convolution of the convolution kernel.
In one embodiment, the first heart level surface data and the second heart level surface data are spliced to obtain spliced data; comprising the following steps:
and connecting the first heart level surface data with the data dimension of 1x n 1x a and the second heart level surface data with the data dimension of 1x n 1x a after fusion processing of the fusion model to obtain spliced data with the data dimension of 1x n 1x 2a, wherein n1 is the data size obtained after convolution of the convolution kernel of the fusion model, and a is the channel number representing the convolution kernel.
In one embodiment, the classifying the extracted feature data by using a preset classification function to obtain a classification result of the electrocardiogram to be identified includes:
classifying the electrocardiogram result to be identified by using a softmax function, mapping the extracted characteristic data into a value of (0, 1) by the softmax function, outputting 0 or 1,0 representing the electrocardiogram abnormality to be identified, and 1 representing the electrocardiogram abnormality to be identified.
In one embodiment, the feature extraction of the spliced data by using a preset feature extraction model, to obtain extracted feature data, includes:
carrying out convolution processing on the spliced data for a preset number of times by utilizing the characteristic extraction model;
and performing feature extraction on the spliced data after the convolution processing by using the full connection layer of the feature extraction model to obtain the feature data.
Fig. 3 is a schematic structural diagram of an electronic device for implementing an electrocardiographic classification method based on deep learning according to an embodiment of the present invention.
In the present embodiment, the electronic device 1 includes, but is not limited to, a memory 11, a processor 12, and a network interface 13, which are communicably connected to each other via a system bus, and the memory 11 stores therein an electrocardiographic-signal-based image classification program 10, the electrocardiographic-signal-based image classification program 10 being executable by the processor 12. Fig. 3 shows only an electronic device 1 with components 11-13 and one sort of classification procedure 10 based on electrocardiographic signals, it being understood by a person skilled in the art that the structure shown in fig. 3 does not constitute a limitation of the electronic device 1, and may include fewer or more components than shown, or may combine certain components, or a different arrangement of components.
Wherein the storage 11 comprises a memory and at least one type of readable storage medium. The memory provides a buffer for the operation of the electronic device 1; the readable storage medium may be a non-volatile storage medium such as flash memory, hard disk, multimedia card, card memory (e.g., SD or DX memory, etc.), random Access Memory (RAM), static Random Access Memory (SRAM), read Only Memory (ROM), electrically Erasable Programmable Read Only Memory (EEPROM), programmable Read Only Memory (PROM), magnetic memory, magnetic disk, optical disk, etc. In some embodiments, the readable storage medium may be an internal storage unit of the electronic device 1; in other embodiments, the nonvolatile storage medium may also be an external storage device of the electronic device 1, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card) or the like, which are provided on the electronic device 1. In this embodiment, the readable storage medium of the memory 11 is generally used for storing an operating system and various application software installed in the electronic device 1, for example, storing a code or the like based on the electrocardiographic signal image classifying program 10 in one embodiment of the present invention. Further, the memory 11 may be used to temporarily store various types of data that have been output or are to be output.
Processor 12 may be a central processing unit (Central Processing Unit, CPU), controller, microcontroller, microprocessor, or other data processing chip in some embodiments. The processor 12 is typically used to control the overall operation of the electronic device 1, such as performing control and processing related to data interaction or communication with other devices, etc. In this embodiment, the processor 12 is configured to execute a program code or process data stored in the memory 11, for example, to execute an image classification program 10 based on electrocardiographic signals.
The network interface 13 may comprise a wireless network interface or a wired network interface, the network interface 13 being used for establishing a communication connection between the electronic device 1 and a terminal (not shown).
Optionally, the electronic device 1 may further comprise a user interface, which may comprise a Display (Display), an input unit such as a Keyboard (Keyboard), and optionally a standard wired interface, a wireless interface. Alternatively, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an OLED (Organic Light-Emitting Diode) touch, or the like. The display may also be referred to as a display screen or display unit, as appropriate, for displaying information processed in the electronic device 1 and for displaying a visual user interface.
It should be understood that the embodiments described are for illustrative purposes only and are not limited to this configuration in the scope of the patent application.
A deep learning based electrocardiogram classification program 10 stored in the memory 11 of the electronic device 1 is a combination of instructions that, when executed in the processor 12, may implement:
preprocessing an electrocardiogram to be identified to obtain electrocardiosignal data, and dividing the electrocardiosignal data into 12 x n two-dimensional arrays, wherein the two-dimensional arrays comprise frontal plane data of 6*n lead signals and transverse plane data of 6*n lead signals;
inputting the two-dimensional array into a preset fusion model, processing the frontal plane data to obtain first heart level plane data, and processing the transverse plane data to obtain second heart level plane data; connecting the first heart level surface data and the second heart level surface data to obtain spliced data;
performing feature extraction on the spliced data through a preset feature extraction model;
and classifying the extracted characteristic data by using a preset classification function to obtain a classification result of the electrocardiogram to be identified.
Specifically, the processor 12 may refer to the description of the related steps in the corresponding embodiment of fig. 1 for the specific implementation method of the electrocardiographic classification procedure 10 based on deep learning, which is not described herein.
Further, the modules/units integrated in the electronic device 1 may be stored in a computer readable storage medium if implemented in the form of software functional units and sold or used as separate products. The computer readable medium may be nonvolatile or nonvolatile. The computer readable medium may include: any entity or device capable of carrying the computer program code, a recording medium, a U disk, a removable hard disk, a magnetic disk, an optical disk, a computer Memory, a Read-Only Memory (ROM).
The computer readable storage medium stores an electrocardiogram classification program 10 based on deep learning, where the electrocardiogram classification program 10 based on deep learning can be executed by one or more processors, and the specific implementation of the computer readable storage medium is basically the same as that of the above embodiments of an electrocardiogram classification method based on deep learning, and will not be described herein.
In the several embodiments provided in the present invention, it should be understood that the disclosed apparatus, device and method may be implemented in other manners. For example, the above-described apparatus embodiments are merely illustrative, and for example, the division of the modules is merely a logical function division, and there may be other manners of division when actually implemented.
The modules described as separate components may or may not be physically separate, and components shown as modules may or may not be physical units, may be located in one place, or may be distributed over multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
In addition, each functional module in the embodiments of the present invention may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit. The integrated units can be realized in a form of hardware or a form of hardware and a form of software functional modules.
It will be evident to those skilled in the art that the invention is not limited to the details of the foregoing illustrative embodiments, and that the present invention may be embodied in other specific forms without departing from the spirit or essential characteristics thereof.
The present embodiments are, therefore, to be considered in all respects as illustrative and not restrictive, the scope of the invention being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein. Any reference signs in the claims shall not be construed as limiting the claim concerned.
Furthermore, it is evident that the word "comprising" does not exclude other elements or steps, and that the singular does not exclude a plurality. A plurality of units or means recited in the system claims can also be implemented by means of software or hardware by means of one unit or means. The terms second, etc. are used to denote a name, but not any particular order.
Finally, it should be noted that the above-mentioned embodiments are merely for illustrating the technical solution of the present invention and not for limiting the same, and although the present invention has been described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that modifications and equivalents may be made to the technical solution of the present invention without departing from the spirit and scope of the technical solution of the present invention.
Claims (10)
1. An electrocardiographic classification method based on deep learning, the method comprising:
preprocessing an electrocardiogram to be identified to obtain electrocardiosignal data, and dividing the electrocardiosignal data into 12 x n two-dimensional arrays, wherein the two-dimensional arrays comprise frontal plane data of 6*n lead signals and transverse plane data of 6*n lead signals;
inputting the two-dimensional array into a preset fusion model, processing the frontal plane data to obtain first heart level plane data, processing the transverse plane data to obtain second heart level plane data, and splicing the first heart level plane data and the second heart level plane data to obtain spliced data;
performing feature extraction on the spliced data by using a preset feature extraction model to obtain extracted feature data;
and classifying the extracted characteristic data by using a preset classification function to obtain a classification result of the electrocardiogram to be identified.
2. The method for classifying electrocardiography based on deep learning according to claim 1, wherein the preprocessing of electrocardiography to be identified to obtain electrocardiographic signal data includes:
denoising the electrocardiogram to be identified by using a preset first filter;
and carrying out downsampling treatment on the denoised electrocardiograph data by using a preset second filter to obtain electrocardiograph signal data with a preset sampling rate.
3. The deep learning-based electrocardiogram classification method according to claim 1, wherein the processing the frontal plane data to obtain first heart level plane data comprises:
and carrying out data signal fusion on 6*n lead signals of the frontal plane data by utilizing convolution of the fusion model to obtain first heart level plane data with data dimension of 1x n 1x a, wherein a represents the channel number of the convolution kernel, and n1 is the data size obtained after convolution of the convolution kernel.
4. The deep learning-based electrocardiogram classification method of claim 1 wherein said processing said cross-plane data to obtain second heart-level plane data comprises:
and carrying out data signal fusion on 6*n lead signals of the frontal plane data by utilizing the convolution of the fusion model to obtain second heart level plane data with the data dimension of 1x n 1x a, wherein a represents the channel number of the convolution kernel, and n1 is the data size obtained after the convolution of the convolution kernel.
5. The method for classifying electrocardiography based on deep learning according to claim 1, wherein the first level surface data and the second level surface data are spliced to obtain spliced data; comprising the following steps:
and connecting the first heart level surface data with the data dimension of 1x n 1x a and the second heart level surface data with the data dimension of 1x n 1x a after fusion processing of the fusion model to obtain spliced data with the data dimension of 1x n 1x 2a, wherein n1 is the data size obtained after convolution of the convolution kernel of the fusion model, and a is the channel number representing the convolution kernel.
6. The deep learning-based electrocardiogram classification method according to claim 1, wherein classifying the extracted feature data by using a predetermined classification function to obtain the classification result of the electrocardiogram to be identified comprises:
classifying the electrocardiogram result to be identified by using a softmax function, mapping the extracted characteristic data into a value of (0, 1) by the softmax function, outputting 0 or 1,0 representing the electrocardiogram abnormality to be identified, and 1 representing the electrocardiogram abnormality to be identified.
7. The method for classifying electrocardiography based on deep learning according to claim 1, wherein the feature extraction of the spliced data using a preset feature extraction model to obtain extracted feature data comprises:
carrying out convolution processing on the spliced data for a preset number of times by utilizing the characteristic extraction model;
and performing feature extraction on the spliced data after the convolution processing by using the full connection layer of the feature extraction model to obtain the feature data.
8. An electrocardiographic signal-based image classification device, the device comprising:
the data processing module is used for preprocessing an electrocardiogram to be identified to obtain electrocardiosignal data, dividing the electrocardiosignal data into 12-n two-dimensional arrays, wherein the two-dimensional arrays comprise frontal surface data of 6*n lead signals and transverse surface data of 6*n lead signals;
the data layer fusion module is used for inputting the two-dimensional array into a preset fusion model, processing the frontal plane data to obtain first heart level plane data, and processing the transverse plane data to obtain second heart level plane data; connecting the first heart level surface data and the second heart level surface data to obtain spliced data;
the feature extraction module is used for extracting features of the spliced data through a preset feature extraction model;
and the classification module is used for classifying the extracted characteristic data by utilizing a preset classification function to obtain the classification result of the electrocardiogram to be identified.
9. An electronic device, the electronic device comprising:
at least one processor; the method comprises the steps of,
a memory communicatively coupled to the at least one processor; wherein,,
the memory stores an electrocardiographic-signal-based image classification program executable by the at least one processor to enable the at least one processor to perform a deep learning-based electrocardiographic classification method of any one of claims 1-7.
10. A computer-readable storage medium having stored thereon an electrocardiographic-signal-based image classification program executable by one or more processors to implement a deep learning-based electrocardiographic classification method of any one of claims 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310701079.XA CN116616784A (en) | 2023-06-13 | 2023-06-13 | Electrocardiogram classification method, device and storage medium based on deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310701079.XA CN116616784A (en) | 2023-06-13 | 2023-06-13 | Electrocardiogram classification method, device and storage medium based on deep learning |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116616784A true CN116616784A (en) | 2023-08-22 |
Family
ID=87613438
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310701079.XA Pending CN116616784A (en) | 2023-06-13 | 2023-06-13 | Electrocardiogram classification method, device and storage medium based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116616784A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117257322A (en) * | 2023-09-21 | 2023-12-22 | 齐鲁工业大学(山东省科学院) | Multi-label electrocardiosignal classification method based on dual-branch network |
-
2023
- 2023-06-13 CN CN202310701079.XA patent/CN116616784A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117257322A (en) * | 2023-09-21 | 2023-12-22 | 齐鲁工业大学(山东省科学院) | Multi-label electrocardiosignal classification method based on dual-branch network |
| CN117257322B (en) * | 2023-09-21 | 2024-04-19 | 齐鲁工业大学(山东省科学院) | Multi-label electrocardiosignal classification method based on dual-branch network |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2023071268A1 (en) | Method and apparatus for analyzing high-frequency qrs-complex data | |
| CN114901143A (en) | System and method for reduced-lead electrocardiography diagnosis using deep neural networks and rule-based systems | |
| EP3593722A1 (en) | Method and system for identification of cerebrovascular abnormalities | |
| CN106073760A (en) | 18 Lead ambulatory electrocardiograms based on wilson's lead2s analyze method and system | |
| JP2007531580A (en) | Computer method and system for reading and analyzing ECG signals | |
| CN116616784A (en) | Electrocardiogram classification method, device and storage medium based on deep learning | |
| Bond et al. | Methods for presenting and visualising electrocardiographic data: From temporal signals to spatial imaging | |
| Smith et al. | The reconstruction of a 12-lead electrocardiogram from a reduced lead set using a focus time-delay neural network | |
| CN112869753B (en) | Analysis method, equipment, medium and electrocardiograph for QRST waveform of electrocardiogram | |
| CN114048343A (en) | Classification platform covering medical image information of patients in whole disease course | |
| CN110432894B (en) | Electrocardiogram key point labeling method and electronic equipment | |
| Jacono et al. | Data acquisition and complex systems analysis in critical care: developing the intensive care unit of the future | |
| CN114901145A (en) | System and method for electrocardiographic diagnosis using deep neural networks and rule-based systems | |
| Rueda et al. | Compelling new electrocardiographic markers for automatic diagnosis | |
| CN115337018B (en) | Electrocardiogram signal classification method and system based on overall dynamic characteristics | |
| CN104545882B (en) | Electrocardiogram measuring device and synthesis electrocardio drawing generating method | |
| CN114999638B (en) | Big data visualization processing method and system for medical diagnosis based on artificial intelligence | |
| CN111128321A (en) | Information display method and system, device, electronic equipment and readable medium | |
| Hu et al. | Low‐cost wireless sensor networks for remote cardiac patients monitoring applications | |
| CN113116358B (en) | Electrocardiogram display method and device, terminal equipment and storage medium | |
| CN110916649B (en) | Processing device, processing method and detection device for long-range electrocardiogram scatter diagram | |
| Zhang et al. | Multi-scale and multi-channel information fusion for exercise electrocardiogram feature extraction and classification | |
| Kim et al. | SE-ResNet-ViT Hybrid Model for Noise Classification in Adhesive Patch-type Wearable Electrocardiographs | |
| Popa et al. | Medical data storage, visualization and interpretation: a case study using a proprietary ECG XML format | |
| Fraser et al. | Time-delay lifts for physiological signal exploration: An application to ECG analysis |
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 |