CN116712056B - Characteristic image generation and identification method, equipment and storage medium for electrocardiogram data - Google Patents
Characteristic image generation and identification method, equipment and storage medium for electrocardiogram data Download PDFInfo
- Publication number
- CN116712056B CN116712056B CN202310983456.3A CN202310983456A CN116712056B CN 116712056 B CN116712056 B CN 116712056B CN 202310983456 A CN202310983456 A CN 202310983456A CN 116712056 B CN116712056 B CN 116712056B
- Authority
- CN
- China
- Prior art keywords
- data
- order
- standard
- electrocardiogram
- electrocardio
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 64
- 230000009467 reduction Effects 0.000 claims abstract description 19
- 238000005259 measurement Methods 0.000 claims abstract description 8
- 238000012545 processing Methods 0.000 claims description 24
- 239000011159 matrix material Substances 0.000 claims description 15
- 238000011176 pooling Methods 0.000 claims description 13
- 238000004590 computer program Methods 0.000 claims description 10
- 230000009466 transformation Effects 0.000 claims description 10
- NAWXUBYGYWOOIX-SFHVURJKSA-N (2s)-2-[[4-[2-(2,4-diaminoquinazolin-6-yl)ethyl]benzoyl]amino]-4-methylidenepentanedioic acid Chemical compound C1=CC2=NC(N)=NC(N)=C2C=C1CCC1=CC=C(C(=O)N[C@@H](CC(=C)C(O)=O)C(O)=O)C=C1 NAWXUBYGYWOOIX-SFHVURJKSA-N 0.000 claims description 9
- 238000013507 mapping Methods 0.000 claims description 9
- 238000007781 pre-processing Methods 0.000 claims description 9
- 238000001914 filtration Methods 0.000 claims description 8
- 230000003044 adaptive effect Effects 0.000 claims description 7
- 238000013461 design Methods 0.000 claims description 7
- 238000000605 extraction Methods 0.000 claims description 7
- 208000000418 Premature Cardiac Complexes Diseases 0.000 claims description 6
- 230000006870 function Effects 0.000 claims description 6
- 238000005070 sampling Methods 0.000 claims description 6
- 230000004927 fusion Effects 0.000 claims description 3
- 238000010606 normalization Methods 0.000 claims description 3
- 241000764238 Isis Species 0.000 claims description 2
- 238000006243 chemical reaction Methods 0.000 abstract description 4
- 238000005086 pumping Methods 0.000 abstract description 4
- 238000004891 communication Methods 0.000 description 8
- 238000013135 deep learning Methods 0.000 description 7
- 230000009471 action Effects 0.000 description 4
- 238000013527 convolutional neural network Methods 0.000 description 4
- 230000008569 process Effects 0.000 description 4
- 238000012549 training Methods 0.000 description 4
- 238000011426 transformation method Methods 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000002093 peripheral effect Effects 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000003759 clinical diagnosis Methods 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000000802 evaporation-induced self-assembly Methods 0.000 description 1
- 208000019622 heart disease Diseases 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000004118 muscle contraction Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 230000001575 pathological effect Effects 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/02—Detecting, measuring or recording pulse, heart rate, blood pressure or blood flow; Combined pulse/heart-rate/blood pressure determination; Evaluating a cardiovascular condition not otherwise provided for, e.g. using combinations of techniques provided for in this group with electrocardiography or electroauscultation; Heart catheters for measuring blood pressure
- A61B5/024—Detecting, measuring or recording pulse rate or heart rate
- A61B5/0245—Detecting, measuring or recording pulse rate or heart rate by using sensing means generating electric signals, i.e. ECG signals
-
- 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
- A61B5/24—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
- A61B5/349—Detecting specific parameters of the electrocardiograph cycle
-
- 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
- A61B5/7253—Details of waveform analysis characterised by using transforms
- A61B5/7257—Details of waveform analysis characterised by using transforms using Fourier transforms
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
Landscapes
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Engineering & Computer Science (AREA)
- Cardiology (AREA)
- Physics & Mathematics (AREA)
- Heart & Thoracic Surgery (AREA)
- Public Health (AREA)
- Biophysics (AREA)
- Pathology (AREA)
- Biomedical Technology (AREA)
- Veterinary Medicine (AREA)
- Medical Informatics (AREA)
- Molecular Biology (AREA)
- Surgery (AREA)
- Animal Behavior & Ethology (AREA)
- General Health & Medical Sciences (AREA)
- Signal Processing (AREA)
- Physiology (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Psychiatry (AREA)
- Mathematical Physics (AREA)
- Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)
Abstract
The invention relates to a characteristic image generation and identification method, equipment and a storage medium of electrocardiogram data, wherein the method comprises the steps of carrying out noise reduction treatment on electrocardiogram data x (t) obtained based on measurement of electrocardiogram equipment, and intercepting the noise-reduced electrocardiogram data to obtain a complete electrocardiogram period sample; performing fast Fourier transform on the standard electrocardiographic data to obtain a standard electrocardiographic frequency signal thereof, performing 1/2 frequency-pumping treatment and 1/3 frequency-pumping treatment on the standard electrocardiographic frequency signal, performing fast Fourier inverse transform, and performing expansion treatment to generate 270 second-order standard electrocardiographic data and three-order standard electrocardiographic data; then performing fast Fourier transform to generate a multi-order rainbow code; finally, a C-ViT-3 model is designed to realize the classification and identification of the multi-order rainbow codes, namely the identification of electrocardiogram data. The invention can complete the conversion of one-dimensional data of electrocardiogram data into image data of two-dimensional multi-order rainbow codes, and can realize the classification and identification of electric signals through a C-ViT-3 model, and the identification accuracy reaches 91.75%.
Description
Technical Field
The invention relates to the technical field of heartbeat signal image generation and heartbeat signal image recognition, in particular to a characteristic image generation and recognition method, equipment and a storage medium of electrocardiogram data.
Background
Electrocardiographic data (ECG) contains a large amount of regular, physiological and pathological information of heart activity, and is of great importance for clinical diagnosis of heart diseases. At present, the analysis of electrocardiographic data mainly comprises the step of diagnosing diseases by analyzing electrocardiographic features by doctors, wherein an electrocardiographic expression form electrocardiograph is a one-dimensional waveform. However, in electrocardiographic data acquisition, the electrocardiographic data is easily affected by site environment, limb actions of human body and muscle contraction to contain noise, and in order to filter noise interference in electrocardiographic data, a filter is usually designed to realize the noise filtering process of electrocardiographic data.
Most of the electrocardiographic data anomalies are sudden and occasional, and it is necessary to monitor electrocardiographic data for a long period of time to obtain portions of electrocardiographic data anomalies. However, the doctor is required to screen out the abnormal parts of the electrocardiographic data from a large amount of electrocardiographic data, the workload is huge, and the knowledge and clinical experience mastered by the doctor are required to be high. With the development of large data deep learning technology, an electrocardiogram data recognition method based on deep learning image classification recognition is paid attention to widely. Compared with the method for identifying the electrocardiogram data by manually identifying the electrocardiogram data, the method for identifying the electrocardiogram data by adopting deep learning image classification has no strict prior knowledge requirement, has extremely low manual workload, and can be completed by a computer for the most of work. However, the electrocardiographic data recognition method using deep learning image classification recognition requires that one-dimensional electrocardiographic data be converted into two-dimensional image data, and that a model for image recognition be designed to perform electrocardiographic data recognition.
Disclosure of Invention
The invention provides a characteristic image generation and identification method, equipment and a storage medium of electrocardiogram data, which can at least solve one of the technical problems in the background technology.
In order to achieve the above purpose, the present invention adopts the following technical scheme:
a characteristic image generating and recognizing method of electrocardiogram data includes the following steps,
step A: for electrocardiographic data obtained based on electrocardiographic device measurementsThe unit is millivolts; />The time is in seconds, a band-pass filter and a wave trap are designed to conduct noise reduction treatment, and an adaptive interception method is adopted to intercept the electrocardiograph data after noise reduction, so that a complete electrocardiograph period sample is obtained;
and (B) step (B): the intercepted signals are predicted to form standard electrocardiogram data with a sample point of 270, the standard electrocardiogram data is subjected to fast Fourier transform to obtain standard electrocardio frequency signals, and then the standard electrocardio frequency signals are subjected toFrequency-extracting processing and->Performing frequency extraction processing to obtain a second-order electrocardio-frequency signal and a third-order electrocardio-frequency signal;
step C: performing Fourier inverse transformation on the second-order electrocardio frequency signal and the third-order electrocardio frequency signal to obtain second-order standard electrocardio data and third-order standard electrocardio data, performing fast Fourier transformation on the second-order standard electrocardio data and the third-order standard electrocardio data to obtain second-order standard electrocardio frequency signal and third-order standard electrocardio frequency signal, and constructing a multi-order rainbow code according to the standard electrocardio data, the second-order standard electrocardio frequency signal, the third-order standard electrocardio frequency signal and the third-order standard electrocardio frequency signal;
Step D: the CNN feature extractor and the feature connector are designed to improve the Vision Transformer model, namely, the Vision Transformer model (C-ViT-3 model) is improved, the C-ViT-3 model is trained, and the multi-order rainbow codes are input into the trained model to realize the identification of electrocardiographic data.
Further said step a: for electrocardiographic data obtained based on electrocardiographic device measurementsThe unit is millivolts;the method comprises the following steps of designing a band-pass filter and a trap filter for noise reduction treatment for time in seconds, and intercepting the electrocardiograph data after noise reduction by adopting a self-adaptive intercepting method to obtain a complete electrocardiograph period sample, wherein the method comprises the following specific steps of:
s101: electrocardiogram data to be obtained based on electrocardiographic device measurementsExpressed in the form of a sequence, i.e. an electrocardiographic data sequence +.>A Butterworth band-pass filter pair is adopted>Noise reduction is carried out to obtain->Then adopting a second order stuffing wave device pair +.>Noise reduction is carried out to obtain noise reduction electrocardiogram data>；
wherein ,for the sampling amount +.>For electrocardiographic data->Is used for the sampling frequency of (a). />、And->All are sequence signals in millivolts, which are a representation of a signal, and are not specific to a particular signal. Butterworth band-pass filter is a method for filtering fixed frequency, parameters of Butterworth band-pass filter are set, and cutoff frequency is high >Low cut-off frequency->. A method for filtering fixed frequency of second order stuffing device, setting parameters of second order stuffing device, and filtering constant>Stuffing frequency->。
S102: for noise reduction of electrocardiogram dataPerforming adaptive interception to obtain truncated electrocardiographic data +.>；
wherein ,for the window function +.>、/>The start and end of the window function, respectively. Interception of electrocardiographic data->To intercept electrocardiographic data, millivolts are measured.
Further, in the step S101、/>Is to reduce the noise of the electrocardiogram data +.>Calculating the data points (1 st R peak to 2 nd R peak) based on 3R peaks>Data points contained in the 2 nd to 3 rd R peaks +.>。/>To take the forward number of the data point of the 2 nd R peak +.>Data points of>To take the data point of the 2 nd R peak to be backward by +.>If->And->The decimal fraction is rounded up.
Further, the step B: the intercepted signals are predicted to form standard electrocardiogram data with a sample point of 270, the standard electrocardiogram data is subjected to fast Fourier transform to obtain standard electrocardio frequency signals, and then the standard electrocardio frequency signals are subjected toFrequency-extracting processing and->The frequency extraction process obtains a second-order electrocardio frequency signal and a third-order electrocardio frequency signal, which comprises the following specific steps:
S103: for intercepting electrocardiogram dataNormalizing and preprocessing to obtain preprocessed electrocardiogram data +.>；
wherein ,is to->Normalized electrocardiographic data is performed. />Is pre-processing electrocardiogram data, < >>Is->Data point position of R peak of ∈ ->The number of data points, which is the number of R peaks backward, ">Is the number of data points forward of the R peak.
S104: for preprocessing electrocardiogram dataData prediction is performed, and standard electrocardiogram data +.>The number of data points of (2) is 270;
wherein ,representing solution->Is->The normalized autoregressive parameters corresponding to the order model,，/>for normalizing autoregressive parameters, < >>。/>Representing solving the inverse of the corresponding sequence, +.>Is predictive initial data, ++>。/>Is predictive data +.>，/>Representing the number of data points to predict for generation. />Represents backward predicted electrocardiographic data, +.>Representing forward predicted electrocardiographic data.Standard electrocardiographic data obtained after prediction contains 270 data points.
S105: for standard electrocardiogram dataPerforming fast Fourier transform to obtain standard electrocardio frequency signalsAnd standard electrocardio frequency signals are subjected to +.>Frequency-extracting processing and->Frequency extraction processing to obtain a second-order electrocardio frequency signal +.>And third-order electrocardio frequency signal- >；
Among them, the fast fourier transform is a mathematical transform method.，/>In hertz, containing 270 data points. />In hertz, containing 135 data points. />In hertz, containing 90 data points.
Further, in the step S104Is the selection rule and execution rule of->When in use, let->=/>Executing backward prediction formulas (5) and (6), formulas (7) and (8), and making +.>=Executing a forward prediction formula, a formula (7) and a formula (9) of the formula (5) and the formula (6), and finally executing a formula (10); when->First let->=/>Forward prediction equations, equation (7), equation (9) of equation (5), equation (6) are performed, and let +.>For null sequences, finally executing formula (10); when->When in use, let->=/>Executing backward prediction formulas (5) and (6), formulas (7) and (8), and making +.>Finally, executing the formula (10); when->Direct order->、/>For null sequences, equation (10) is performed.
Further, the step C: performing Fourier inverse transformation on the second-order electrocardio frequency signal and the third-order electrocardio frequency signal to obtain second-order standard electrocardio data and third-order standard electrocardio data, performing fast Fourier transformation on the second-order standard electrocardio data and the third-order standard electrocardio data to obtain second-order standard electrocardio frequency signal and third-order standard electrocardio frequency signal, and constructing a multi-order rainbow code by the standard electrocardio data, the second-order standard electrocardio frequency signal, the third-order standard electrocardio frequency signal and the third-order standard electrocardio frequency signal, wherein the method comprises the following specific steps of:
S106: for second-order electrocardio frequency signalsAnd third-order electrocardio frequency signal->Performing inverse fast Fourier transform, and performing 2-fold expansion and 3-fold expansion to obtain second-order standard electrocardiogram data>And third-order standard electrocardiographic data->；
wherein ,is a second-order electrocardio frequency signal +.>The second-order electrocardiogram data obtained by performing the inverse fast fourier transform contains 135 data points. />Is a third-order electrocardio frequency signal +.>The third-order electrocardiogram data obtained by performing the inverse fast fourier transform contains 90 data points. />Is to->Second order standard obtained by 2 times expansionQuasi-electrocardiographic data, comprising 270 data points. />Is to->Three-order standard electrocardiographic data obtained by 3-fold expansion contains 270 data points.
S107: for second-order standard electrocardiogram dataAnd third-order standard electrocardiographic data->Performing fast Fourier transform to obtain second-order standard electrocardio frequency signal +.>And third-order standard electrocardio frequency signal +.>And is about->、、/>Normalization treatment is carried out to obtain->、/>、/>；
wherein ：is a second-order standard electrocardio-frequency signal, has the unit of hertz and comprises270 data points. />Is a third-order standard electrocardio-frequency signal, has a unit of hertz and comprises 270 data points. />、/>、/>The normalized electrocardiographic frequency signal contains 270 data points.
S108: will be、/>、/>、/>、/>Constructing a multi-order rainbow code matrix>And linearly map the feature map matrix to the RGB image matrix with length and width +.>I.e. a multi-level rainbow code;
wherein the linear mapping is a linear transformation method. The RGB image matrix refers to a 3-channel color image standard matrix.
Further, the step D: the design of a CNN feature extractor and a feature connector improves a Vision Transformer model, namely improves a Vision Transformer model (C-ViT-3 model) and trains a C-ViT-3 model, and inputs a multi-order rainbow code into the trained model to realize identification of electrocardiographic data, which comprises the following specific steps:
s109: designing CNN feature extractor structure to input dataIs output by a structure composed of 3 convolution layers, 3 pooling layers, 3 BN operations and 3 Reul operations>Image data of (a)；
Wherein, 3 convolution layers, 3 pooling layers, 3 BN operations and 3 Reul operations are all a mathematical transformation method, and different combination modes and different parameters can obtain different output results.Is represented as a size ofIs a color image of (a) a color image of (b). />Is of the size +.>Gray image of (c) a gray image of (c).
S110: the input data of the design feature connector isFor->Generating image data by performing cubic spline interpolation>Will->Linear mapping to RGB image generating image data>。
wherein ,is->Is described. />Is->Is described. Cubic spline interpolation is a mathematical interpolation method that can expand the data. The linear mapping to RGB image is an image mapping method, which maps +.>Grey image data of (2) becomes->Is described.
S111: design Vision Transformer module, toThe input Vision Transformer module obtains a recognition result F;
among them, the Vision Transformer module is the ViT-B16 model among Vision Transformer models, which is a deep learning neural network. The recognition result F is a classification label, and one contains four labels of 0, 1, 2 and 3, and only one of the four labels is given by the recognition result each time, and represents one category of electrocardiographic data.
S112: inputting the multi-order rainbow code generated by the electrocardiogram data into a trained C-ViT-3 model to obtain the identification result of the electrocardiogram data;
wherein the C-ViT-3 model refers to all the structures contained in step S109 to step S111. The trained C-ViT-3 model is a model obtained by training the C-ViT-3 model through a large number of multi-order rainbow codes, and is a self-learning method, and the learnable network parameters can be automatically optimized to obtain an optimal learnable network parameter. And inputting the multi-order rainbow codes participating in training into a trained C-ViT-3 model to obtain a corresponding identification result. The four labels of 0, 1, 2 and 3 of the identification result F respectively correspond to fusion heart beat, non-ectopic heart beat, unknown heart beat and ectopic heart beat of the electrocardiogram data.
Further, in the step S109, the combination mode and the parameter are set such that the input of the layer 1 convolution layer isIs set to a convolution kernel size +.>The step length is 1, the depth is 8, and BN operation and Reul operation are carried out to output ∈>Characteristic data of->. Layer 1 pooling layer input is +.>The convolution kernel size is set to +.>Step size of 2, output->Characteristic data of->. Layer 2 rollThe laminated input is +.>The convolution kernel size is set to +.>The step length is 1, the depth is 16, and BN operation and Reul operation are carried out to output ∈>Characteristic data of->. The layer 2 pooling layer input is +.>The convolution kernel size is set to +.>Step size of 2, output->Characteristic data of->. Layer 3 convolutional layer input is +.>The convolution kernel size is set to +.>The step length is 1, the depth is 1, and BN operation and Reul operation are carried out to outputCharacteristic data of->. The layer 3 pooling layer input is +.>The convolution kernel size is set to +.>Step length of 1, outputImage data of->。
In yet another aspect, the invention also discloses a computer readable storage medium storing a computer program which, when executed by a processor, causes the processor to perform the steps of the method as described above.
In yet another aspect, the invention also discloses a computer device comprising a memory and a processor, the memory storing a computer program which, when executed by the processor, causes the processor to perform the steps of the method as above.
According to the technical scheme, the characteristic image generation and identification method of the electrocardiogram data comprises the steps of designing a Butterworth band-pass filter and a second-order stuffing wave device for noise reduction treatment on electrocardiogram data x (t) (with the unit of millivolts; t is time and the unit of seconds) obtained based on measurement of electrocardiogram equipment, and intercepting the noise-reduced electrocardiogram data by adopting an adaptive interception method to obtain a complete electrocardiogram period sample; secondly, standard electrocardiogram data with a sample point of 270 is obtained through prediction processing, standard electrocardiogram data is subjected to fast Fourier transform to obtain a standard electrocardiogram frequency signal, 1/2 frequency-extracting processing and 1/3 frequency-extracting processing are carried out on the standard electrocardiogram frequency signal, fast Fourier transform is carried out, and expansion processing is carried out to generate second-order standard electrocardiogram data with the sample point of 270 and third-order standard electrocardiogram data; then, performing fast Fourier transformation on the second-order standard electrocardiograph data and the third-order standard electrocardiograph data to obtain a second-order standard electrocardiograph frequency signal and a third-order standard electrocardiograph frequency signal, and generating a multi-order rainbow code by the standard electrocardiograph data, the second-order standard electrocardiograph data, the third-order standard electrocardiograph data, the standard electrocardiograph frequency signal, the second-order standard electrocardiograph frequency signal and the third-order standard electrocardiograph frequency signal; finally, a C-ViT-3 model of a CNN feature extractor, a feature connector and a Vision Transformer structure is designed to realize the classification and identification of multi-order rainbow codes, namely the identification of electrocardiogram data. The invention can complete the conversion of one-dimensional data of electrocardiogram data into image data of two-dimensional multi-order rainbow codes, and can realize the classification and identification of electric signals through a C-ViT-3 model, and the identification accuracy reaches 91.75%.
In general, the characteristic image generation and identification method of the electrocardiogram data can realize the conversion from one-dimensional data to two-dimensional data of the electrocardiogram data, and generate a multi-order rainbow code characteristic image; the intelligent classification and identification of the multi-order rainbow code features are realized by the big data deep learning technology, namely the intelligent classification and identification of the electrocardiogram data is realized, the identification accuracy rate reaches 91.75% at present, and the method has good practical application value.
Drawings
FIG. 1 is a flow chart of a method for generating and identifying multi-level rainbow codes of electrocardiographic data according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of a generated multi-level rainbow code according to an embodiment of the present invention;
FIG. 3 is a schematic diagram of a network structure of the C-ViT-3 model according to an embodiment of the invention.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is apparent that the described embodiments are some embodiments of the present invention, but not all embodiments of the present invention.
As shown in fig. 1, the characteristic image generating and identifying method of electrocardiographic data according to the present embodiment specifically includes the following steps,
Step A: for electrocardiographic data obtained based on electrocardiographic device measurementsThe unit is millivolts; />The band-pass filter and the trap filter are designed for time in secondsCarrying out noise reduction treatment, and intercepting the electrocardiographic data after noise reduction by adopting a self-adaptive intercepting method to obtain a complete electrocardiographic cycle sample;
s101: electrocardiogram data to be obtained based on electrocardiographic device measurementsExpressed in the form of a sequence, i.e. an electrocardiographic data sequence +.>A Butterworth band-pass filter pair is adopted>Noise reduction is carried out to obtain->Then adopting a second order stuffing wave device pair +.>Noise reduction is carried out to obtain noise reduction electrocardiogram data>；
wherein ,for the sampling amount +.>For electrocardiographic data->Is used for the sampling frequency of (a). />、And->All are sequence signals in millivolts, which are a representation of a signal, and are not specific to a particular signal. Butterworth band-pass filter is a method for filtering fixed frequency, parameters of Butterworth band-pass filter are set, and cutoff frequency is high>Low cut-off frequency->. A method for filtering fixed frequency of second order stuffing device, setting parameters of second order stuffing device, and filtering constant>Stuffing frequency->。
S102: for noise reduction of electrocardiogram dataPerforming adaptive interception to obtain truncated electrocardiographic data +. >；
wherein ,for the window function +.>、/>The start and end of the window function, respectively. Interception of electrocardiographic data->To intercept electrocardiographic data, millivolts are measured. />、/>Is to reduce the noise of the electrocardiogram data +.>Calculating the data points (1 st R peak to 2 nd R peak) based on 3R peaks>Data points contained by the 2 nd R peak to the 3 rd R peak。/>To take the forward number of the data point of the 2 nd R peak +.>Data points of>To take the data point of the 2 nd R peak to be backward by +.>If->And->The decimal fraction is rounded up.
And (B) step (B): the intercepted signals are predicted to form standard electrocardiogram data with a sample point of 270, the standard electrocardiogram data is subjected to fast Fourier transform to obtain standard electrocardio frequency signals, and then the standard electrocardio frequency signals are subjected toFrequency-extracting processing and->Performing frequency extraction processing to obtain a second-order electrocardio-frequency signal and a third-order electrocardio-frequency signal;
s103: for intercepting electrocardiogram dataNormalizing and preprocessing to obtain preprocessed electrocardiogram data +.>；
wherein ,is to->Normalized electrocardiographic data is performed. />Is pre-processing electrocardiogram data, < >>Is->Data point position of R peak of ∈ ->The number of data points, which is the number of R peaks backward, " >Is the number of data points forward of the R peak.
S104: for preprocessing electrocardiogram dataData prediction is performed, and standard electrocardiogram data +.>The number of data points of (2) is 270;
wherein ,representing solution->Is->The normalized autoregressive parameters corresponding to the order model,，/>for normalizing autoregressive parameters, < >>。/>Representing solving the inverse of the corresponding sequence, +.>Is predictive initial data, ++>。/>Is predictive data +.>，/>Representing the number of data points to predict for generation. />Represents backward predicted electrocardiographic data, +.>Representing forward predicted electrocardiographic data.Standard electrocardiographic data obtained after prediction contains 270 data points. />Is the selection rule and execution rule of->When in use, let->=/>Executing backward prediction formulas (5) and (6), formulas (7) and (8), and making +.>=/>Executing a forward prediction formula, a formula (7) and a formula (9) of the formula (5) and the formula (6), and finally executing a formula (10); when->First let->=/>Executing forward prediction formulas of formulas (5) and (6), formulas (7) and (9),let->For null sequences, finally executing formula (10); when (when)When in use, let->=/>Executing backward prediction formulas (5) and (6), formulas (7) and (8), and making +. >Finally, executing the formula (10); when->Direct order->、/>For null sequences, equation (10) is performed.
S105: for standard electrocardiogram dataPerforming fast Fourier transform to obtain standard electrocardio frequency signalsAnd standard electrocardio frequency signals are subjected to +.>Frequency-extracting processing and->Frequency extraction processing to obtain a second-order electrocardio frequency signal +.>And third-order electrocardio frequency signal->；
Among them, the fast fourier transform is a mathematical transform method.，/>In hertz, containing 270 data points. />In hertz, containing 135 data points. />In hertz, containing 90 data points.
Step C: performing Fourier inverse transformation on the second-order electrocardio frequency signal and the third-order electrocardio frequency signal to obtain second-order standard electrocardio data and third-order standard electrocardio data, performing fast Fourier transformation on the second-order standard electrocardio data and the third-order standard electrocardio data to obtain second-order standard electrocardio frequency signal and third-order standard electrocardio frequency signal, and constructing a multi-order rainbow code according to the standard electrocardio data, the second-order standard electrocardio frequency signal, the third-order standard electrocardio frequency signal and the third-order standard electrocardio frequency signal;
s106: for second-order electrocardio frequency signals And third-order electrocardio frequency signal->Performing inverse fast Fourier transform, and performing 2-fold expansion and 3-fold expansion to obtain second-order standard electrocardiogram data>And third-order standard electrocardiographic data->；
wherein ,is a second-order electrocardio frequency signal +.>The second-order electrocardiogram data obtained by performing the inverse fast fourier transform contains 135 data points. />Is a third-order electrocardio frequency signal +.>The third-order electrocardiogram data obtained by performing the inverse fast fourier transform contains 90 data points. />Is to->The second order standard electrocardiographic data obtained by performing 2-fold expansion contains 270 data points. />Is to->Three-order standard electrocardiographic data obtained by 3-fold expansion contains 270 data points.
S107: for second-order standard electrocardiogram dataAnd third-order standard electrocardiographic data->Performing fast Fourier transform to obtain second-order standard electrocardio frequency signal +.>And third-order standard electrocardio frequency signal +.>And is about->、、/>Normalization treatment is carried out to obtain->、/>、/>；
wherein ：is a second order standard electrocardiographic frequency signal in hertz containing 270 data points. />Is a third-order standard electrocardio-frequency signal, has a unit of hertz and comprises 270 data points. />、/>、/>Normalized toComprises 270 data points.
S108: will be、/>、/>、/>、/>Constructing a multi-order rainbow code matrix>And linearly map the feature map matrix to the RGB image matrix with length and width +.>I.e. a multi-level rainbow code;
wherein the linear mapping is a linear transformation method. The RGB image matrix refers to a 3-channel color image standard matrix.
As shown in fig. 2, a multi-level rainbow code pattern is generated in which the first-level rainbow code is composed of standard electrocardiographic data, the second-level rainbow code is composed of second-level standard electrocardiographic data, the third-level rainbow code is composed of third-level standard electrocardiographic data, the fourth-level rainbow code is composed of standard electrocardiographic frequency signals, the fifth-level rainbow code is composed of second-level standard electrocardiographic frequency signals, and the sixth-level rainbow code is composed of third-level standard electrocardiographic frequency signals, thus forming the multi-level rainbow code.
Step D: the CNN feature extractor and the feature connector are designed to improve the Vision Transformer model, namely, the Vision Transformer model (C-ViT-3 model) is improved, the C-ViT-3 model is trained, and the multi-order rainbow codes are input into the trained model to realize the identification of electrocardiographic data.
As shown in FIG. 3, which is a network structure diagram of the C-ViT-3 model, the network structure diagram is composed of a CNN feature extractor, a feature connector and a Vision Transformer module, and the Vision Transformer structure in the figure is a Vision Transformer model.
S109: designing CNN feature extractor structure to input dataIs output by a structure composed of 3 convolution layers, 3 pooling layers, 3 BN operations and 3 Reul operations>Image data of (a)；
Wherein, 3 convolution layers, 3 pooling layers, 3 BN operations and 3 Reul operations are all a mathematical transformation method, and different combination modes and different parameters can obtain different output results.Is represented as a size ofIs a color image of (a) a color image of (b). />Is of the size +.>Gray image of (c) a gray image of (c). Layer 1 convolutional layer input is +.>Is set to a convolution kernel size +.>Step size of 1Depth of 8, performing BN operation and Reul operation output->Characteristic data of->. Layer 1 pooling layer input is +.>The convolution kernel size is set to +.>Step size of 2, output->Characteristic data of->. The layer 2 convolutional layer input is +.>The convolution kernel size is set to +.>The step length is 1, the depth is 16, and BN operation and Reul operation are carried out to output ∈>Characteristic data of->. The layer 2 pooling layer input is +.>The convolution kernel size is set to +.>Step size of 2, output->Characteristic data of->. Layer 3 convolutional layer input is +.>The convolution kernel size is set to +.>Step length of 1 and depth of 1, performing BN operation and Reul operation to output ∈1- >Characteristic data of->. The layer 3 pooling layer input is +.>The convolution kernel size is set to +.>Step size 1, output->Image data of->。
S110: the input data of the design feature connector isFor->Generating image data by performing cubic spline interpolation>Will->Linear mapping to RGB image generating image data>。
wherein ,is->Is described. />Is->Is described. Cubic spline interpolation is a mathematical interpolation method that can expand the data. The linear mapping to RGB image is an image mapping method, which maps +.>Grey image data of (2) becomes->Is described.
S111: design Vision Transformer module, toThe input Vision Transformer module obtains a recognition result F;
among them, the Vision Transformer module is the ViT-B16 model among Vision Transformer models, which is a deep learning neural network. The recognition result F is a classification label, and one contains four labels of 0, 1, 2 and 3, and only one of the four labels is given by the recognition result each time, and represents one category of electrocardiographic data.
S112: inputting the multi-order rainbow code generated by the electrocardiogram data into a trained C-ViT-3 model to obtain the identification result of the electrocardiogram data;
Wherein the C-ViT-3 model refers to all the structures contained in step S109 to step S111. The trained C-ViT-3 model is a model obtained by training the C-ViT-3 model through a large number of multi-order rainbow codes, and is a self-learning method, and the learnable network parameters can be automatically optimized to obtain an optimal learnable network parameter. And inputting the multi-order rainbow codes participating in training into a trained C-ViT-3 model to obtain a corresponding identification result. The four labels of 0, 1, 2 and 3 of the identification result F respectively correspond to fusion heart beat, non-ectopic heart beat, unknown heart beat and ectopic heart beat of the electrocardiogram data.
In summary, the method for generating and intelligently identifying the characteristic image of the electrocardiogram data according to the embodiment of the invention comprises the following steps: (1) For the electrocardiogram data x which is obtained based on the measurement of the electrocardiogram equipmentt) (in millivolts;ttime is in seconds), a Butterworth band-pass filter and a second-order stuffing wave device are designed for noise reduction treatment, and an adaptive interception method is adopted for intercepting electrocardiograph data after noise reduction, so that a complete electrocardiograph period sample is obtained; (2) Obtaining standard electrocardiogram data with a sample point of 270 through prediction processing, performing fast Fourier transform on the standard electrocardiogram data to obtain a standard electrocardiogram frequency signal of the standard electrocardiogram data, performing 1/2 frequency-pumping processing and 1/3 frequency-pumping processing on the standard electrocardiogram frequency signal, performing fast Fourier inverse transform, and performing expansion processing to generate second-order standard electrocardiogram data and third-order standard electrocardiogram data with the sample point of 270; (3) Performing fast Fourier transformation on the second-order standard electrocardiograph data and the third-order standard electrocardiograph data to obtain a second-order standard electrocardiograph frequency signal and a third-order standard electrocardiograph frequency signal, and generating a multi-order rainbow code by the standard electrocardiograph data, the second-order standard electrocardiograph data, the third-order standard electrocardiograph data, the standard electrocardiograph frequency signal, the second-order standard electrocardiograph frequency signal and the third-order standard electrocardiograph frequency signal; (4) A C-ViT-3 model of a CNN feature extractor, a feature connector and a Vision Transformer structure is designed to realize classification recognition of multi-order rainbow codes, namely recognition of electrocardiogram data. The method can reliably remove noise components in the signals, effectively reduce interference of the noise components on the true value of the electrocardiogram data, finish conversion of one-dimensional data of the electrocardiogram data into image data of two-dimensional multi-order rainbow codes, realize classification and identification of the electric signals through a C-ViT-3 model, and achieve an identification accuracy rate of 91.75 percent.
In yet another aspect, the application also discloses a computer readable storage medium storing a computer program which, when executed by a processor, causes the processor to perform the steps of the method as described above.
In yet another aspect, the application also discloses a computer device comprising a memory and a processor, the memory storing a computer program which, when executed by the processor, causes the processor to perform the steps of the method as above.
In yet another embodiment of the present application, a computer program product containing instructions that, when run on a computer, cause the computer to perform the method of generating and identifying characteristic images of electrocardiographic data of any one of the above embodiments is also provided.
It may be understood that the system provided by the embodiment of the present application corresponds to the method provided by the embodiment of the present application, and explanation, examples and beneficial effects of the related content may refer to corresponding parts in the above method.
The embodiment of the application also provides an electronic device, which comprises a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory are communicated with each other through the communication bus,
A memory for storing a computer program;
and the processor is used for realizing the characteristic image generation and identification method of the electrocardiogram data when executing the program stored in the memory.
The communication bus mentioned by the above electronic device may be a peripheral component interconnect standard (english: peripheral Component Interconnect, abbreviated: PCI) bus or an extended industry standard architecture (english: extended Industry Standard Architecture, abbreviated: EISA) bus, or the like. The communication bus may be classified as an address bus, a data bus, a control bus, or the like.
The communication interface is used for communication between the electronic device and other devices.
The Memory may include random access Memory (Random Access Memory, abbreviated as RAM) or nonvolatile Memory (NVM), such as at least one disk Memory. Optionally, the memory may also be at least one memory device located remotely from the aforementioned processor.
The processor may be a general-purpose processor, including a central processing unit (Central Processing Unit, CPU for short), a network processor (Network Processor, NP for short), etc.; it may also be a digital signal processor (English: digital Signal Processing; DSP; for short), an application specific integrated circuit (English: application Specific Integrated Circuit; ASIC; for short), a Field programmable gate array (English: field-Programmable Gate Array; FPGA; for short), or other programmable logic device, discrete gate or transistor logic device, discrete hardware components.
In the above embodiments, it may be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented in software, may be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When loaded and executed on a computer, produces a flow or function in accordance with embodiments of the present application, in whole or in part. The computer may be a general purpose computer, a special purpose computer, a computer network, or other programmable apparatus. The computer instructions may be stored in or transmitted from one computer-readable storage medium to another, for example, by wired (e.g., coaxial cable, optical fiber, digital Subscriber Line (DSL)), or wireless (e.g., infrared, wireless, microwave, etc.). The computer readable storage medium may be any available medium that can be accessed by a computer or a data storage device such as a server, data center, etc. that contains an integration of one or more available media. The usable medium may be a magnetic medium (e.g., floppy Disk, hard Disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid State Disk (SSD)), etc.
It is noted that relational terms such as first and second, and the like are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Moreover, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.
In this specification, each embodiment is described in a related manner, and identical and similar parts of each embodiment are all referred to each other, and each embodiment mainly describes differences from other embodiments. In particular, for system embodiments, since they are substantially similar to method embodiments, the description is relatively simple, as relevant to see a section of the description of method embodiments.
The above embodiments are only for illustrating the technical solution of the present invention, and are not limiting; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims (10)
1. A characteristic image generating and identifying method of electrocardiogram data is characterized by comprising the following steps,
step A: for electrocardiographic data obtained based on electrocardiographic device measurementsThe unit is millivolts; />The time is in seconds, a band-pass filter and a wave trap are designed to conduct noise reduction treatment, and an adaptive interception method is adopted to intercept the electrocardiograph data after noise reduction, so that a complete electrocardiograph period sample is obtained;
and (B) step (B): the intercepted signals are predicted to form standard electrocardiogram data with a sample point of 270, the standard electrocardiogram data is subjected to fast Fourier transform to obtain standard electrocardio frequency signals, and then the standard electrocardio frequency signals are subjected to Frequency-extracting processing and->Performing frequency extraction processing to obtain a second-order electrocardio-frequency signal and a third-order electrocardio-frequency signal;
step C: performing Fourier inverse transformation on the second-order electrocardio frequency signal and the third-order electrocardio frequency signal to obtain second-order standard electrocardio data and third-order standard electrocardio data, performing fast Fourier transformation on the second-order standard electrocardio data and the third-order standard electrocardio data to obtain second-order standard electrocardio frequency signal and third-order standard electrocardio frequency signal, and constructing a multi-order rainbow code according to the standard electrocardio data, the second-order standard electrocardio frequency signal, the third-order standard electrocardio frequency signal and the third-order standard electrocardio frequency signal;
step D: the CNN feature extractor and the feature connector are designed to improve the Vision Transformer model, namely the Vision Transformer model is improved, the C-ViT-3 model is trained, and the multi-order rainbow codes are input into the trained model to realize the identification of electrocardiogram data.
2. The characteristic image generating and recognizing method of electrocardiographic data according to claim 1, characterized in that: the step A specifically comprises the following steps of,
s101: electrocardiogram data to be obtained based on electrocardiographic device measurements Expressed in the form of a sequence, i.e. an electrocardiographic data sequence +.>A Butterworth band-pass filter pair is adopted>Noise reduction is carried out to obtain->Then the second order stuffing wave device pair is adoptedNoise reduction is carried out to obtain noise reduction electrocardiogram data>；
wherein ,for the sampling amount +.>For electrocardiographic data->Is a sampling frequency of (a); />、/>And (3) withAre all serial signals in millivolts; setting parameters of Butterworth band-pass filter, high cut-off frequencyLow cut-off frequency->The method comprises the steps of carrying out a first treatment on the surface of the Setting parameters of a second order stuffing machine, and filtering constant +.>Stuffing frequency->；
S102: for noise reduction of electrocardiogram dataPerforming adaptive interception to obtain truncated electrocardiographic data +.>；
wherein ,for the window function +.> 、/>Respectively window boxCounting a starting point and a finishing point; interception of electrocardiographic data->To intercept electrocardiographic data, millivolts are measured.
3. The characteristic image generating and recognizing method of electrocardiographic data according to claim 2, characterized in that: in the step S101 、/>Is to reduce the noise of the electrocardiogram data +.>Calculating the data points (1 st R peak to 2 nd R peak) based on 3R peaks>Data points contained in the 2 nd to 3 rd R peaks +.>；/>To take the forward number of the data point of the 2 nd R peak +.>Data points of>To take the data point of the 2 nd R peak to be backward by +. >If->And->The decimal fraction is rounded up.
4. The characteristic image generating and recognizing method for electrocardiographic data according to claim 3, characterized in that: the step B comprises the following steps of, in particular,
s103: for intercepting electrocardiogram dataNormalizing and preprocessing to obtain preprocessed electrocardiogram data +.>；
wherein ,is to->Carrying out normalized electrocardiogram data; />Is pre-processing electrocardiogram data, < >>Is thatData point position of R peak of ∈ ->The number of data points, which is the number of R peaks backward, ">Is the number of data points forward of the R peak;
s104: for preprocessing electrocardiogram dataData prediction is carried out, and standard electrocardiogram data obtained after prediction is setThe number of data points of (2) is 270;
wherein ,representing solution->Is->Normalized autoregressive parameters corresponding to the order model, +.>，For normalizing autoregressive parameters, < >>；/>Representing solving the inverse of the corresponding sequence, +.>Is predictive initial data, ++>；/>Is predictive data +.>，/>Representing a number of data points to be predictively generated; />Represents backward predicted electrocardiographic data, +.>Electrocardiographic data representing forward predictions; />Standard electrocardiogram data obtained after prediction comprises 270 data points;
s105: for standard electrocardiogram data Performing fast Fourier transform to obtain standard electrocardio frequency signal +.>And standard electrocardio frequency signals are subjected to +.>Frequency-extracting processing and->Frequency extraction processing to obtain a second-order electrocardio frequency signal +.>And third-order heartElectric frequency signal->；
wherein ,，/>in hertz, containing 270 data points; />In hertz, containing 135 data points; />In hertz, containing 90 data points.
5. The method for generating and recognizing a characteristic image of electrocardiographic data according to claim 4, characterized in that: in the step S104Is the selection rule and execution rule of->When in use, let->=/>Executing backward prediction formulas (5) and (6), formulas (7) and (8), and making +.>=/>Executing a forward prediction formula, a formula (7) and a formula (9) of the formula (5) and the formula (6), and finally executing a formula (10);
when (when)First let->=/>Forward prediction equations, equation (7), equation (9) of equation (5), equation (6) are performed, and let +.>For null sequences, finally executing formula (10); when->When it is, firstly make=/>Executing backward prediction formulas (5) and (6), formulas (7) and (8), and making +.>Finally, executing the formula (10); when->Direct order->、/>For null sequences, equation (10) is performed.
6. The method for generating and recognizing a characteristic image of electrocardiographic data according to claim 5, characterized in that:
the step C comprises the following steps of, in particular,
s106: for second-order electrocardio frequency signalsAnd third-order electrocardio frequency signal->Performing inverse fast Fourier transform, and performing 2-fold expansion and 3-fold expansion to obtain second-order standard electrocardiogram data>And third-order standard electrocardiogram data；
wherein ,is a second-order electrocardio frequency signal +.>Performing inverse fast fourier transform to obtain second-order electrocardiogram data comprising 135 data points; />Is a third-order electrocardio frequency signal +.>Third-order electrocardiogram obtained by performing inverse fast Fourier transformData, comprising 90 data points; />Is to->Performing 2 times expansion to obtain second-order standard electrocardiographic data comprising 270 data points; />Is to->Three-order standard electrocardiographic data obtained by 3 times expansion comprises 270 data points;
s107: for second-order standard electrocardiogram dataAnd third-order standard electrocardiographic data->Performing fast Fourier transform to obtain second-order standard electrocardio frequency signal +.>And third-order standard electrocardio frequency signal +.>And is about->、/>、Normalization treatment is carried out to obtain->、/>、/>；
wherein ：is a second-order standard electrocardio-frequency signal, has a unit of hertz and comprises 270 data points; / >Is a third-order standard electrocardio-frequency signal, has the unit of hertz and comprises 270 data points; />、/>、/>Normalized electrocardiographic frequency signals, comprising 270 data points;
s108: will be、/>、/>、/>、/>Constructing a multi-order rainbow code matrix>And linearly map the feature map matrix to the RGB image matrix with length and width +.>I.e. a multi-level rainbow code;
wherein the RGB image matrix refers to a 3-channel color image standard matrix.
7. The method for generating and recognizing a characteristic image of electrocardiographic data according to claim 5, characterized in that:
the step D specifically includes the steps of,
s109: designing CNN feature extractor structure to input dataIs output by a structure composed of 3 convolution layers, 3 pooling layers, 3 BN operations and 3 Reul operations>Image data of->；
wherein ,is of the size +.>Is a color image of (1); />Image data representation size of (2)Is->Gray image of (a);
s110: the input data of the design feature connector isFor->Generating image data by performing cubic spline interpolation>Will beLinear mapping to RGB image generating image data>；
wherein ,is->Is a digital image of the image data; />Is->Is a digital image of the image data; will beGrey image data of (2) becomes- >Color image data of (a);
s111: design Vision Transformer module, toThe input Vision Transformer module obtains a recognition result F;
the identification result F is a classification label, and the identification result comprises four labels of 0, 1, 2 and 3, wherein each time, the identification result only gives one of the four labels, and represents one category of electrocardiographic data;
s112: inputting the multi-order rainbow code generated by the electrocardiogram data into a trained C-ViT-3 model to obtain the identification result of the electrocardiogram data;
wherein the C-ViT-3 model refers to all the structures contained in step S109 to step S111; the four labels of 0, 1, 2 and 3 of the identification result F respectively correspond to fusion heart beat, non-ectopic heart beat, unknown heart beat and ectopic heart beat of the electrocardiogram data.
8. The characteristic image generating and recognizing method of electrocardiographic data according to claim 7, characterized in that:
the combination mode and parameters in the step S109 are set as that the input of the layer 1 convolution layer isIs set to a convolution kernel size +.>The step length is 1, the depth is 8, and BN operation and Reul operation are carried out to output ∈>Characteristic data of->；
Layer 1 pooling layer input isThe convolution kernel size is set to +. >Step size of 2, output->Characteristic data of->；
The layer 2 convolutional layer input isThe convolution kernel size is set to +.>The step length is 1, the depth is 16, and BN operation and Reul operation are carried out to output ∈>Characteristic data of->；
Layer 2 pooling layer input isThe convolution kernel size is set to +.>Step size of 2, output->Characteristic data of->；
The layer 3 convolutional layer input isThe convolution kernel size is set to +.>Step length of 1 and depth of 1, performing BN operation and Reul operation to output ∈1->Characteristic data of->；
Layer 3 pooling layer input isThe convolution kernel size is set to +.>Step size 1, output->Image data of (a)。
9. A computer readable storage medium storing a computer program which, when executed by a processor, causes the processor to perform the steps of the method of any one of claims 1 to 8.
10. A computer device comprising a memory and a processor, the memory storing a computer program that, when executed by the processor, causes the processor to perform the steps of the method of any of claims 1 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310983456.3A CN116712056B (en) | 2023-08-07 | 2023-08-07 | Characteristic image generation and identification method, equipment and storage medium for electrocardiogram data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310983456.3A CN116712056B (en) | 2023-08-07 | 2023-08-07 | Characteristic image generation and identification method, equipment and storage medium for electrocardiogram data |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116712056A CN116712056A (en) | 2023-09-08 |
| CN116712056B true CN116712056B (en) | 2023-11-03 |
Family
ID=87868306
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310983456.3A Active CN116712056B (en) | 2023-08-07 | 2023-08-07 | Characteristic image generation and identification method, equipment and storage medium for electrocardiogram data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116712056B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116965822B (en) * | 2023-09-25 | 2023-12-29 | 合肥工业大学 | Magnetocardiogram color space circle diagram generation and wave band time identification method and storage medium |
| CN116982989A (en) * | 2023-09-26 | 2023-11-03 | 深圳新蕊科技有限公司 | Electrocardiosignal R wave monitoring method and wearable device thereof |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| FR2648567A1 (en) * | 1989-05-24 | 1990-12-21 | Inst Nat Sante Rech Med | Method for the digital processing of a signal by reversible transformation into wavelets |
| CN106887225A (en) * | 2017-03-21 | 2017-06-23 | 百度在线网络技术（北京）有限公司 | Acoustic feature extracting method, device and terminal device based on convolutional neural networks |
| CN109350078A (en) * | 2018-10-19 | 2019-02-19 | 天津大学 | Blood sugar concentration detection method based on ultra-wideband microwave inverse Fourier transform |
| CN114145725A (en) * | 2022-02-08 | 2022-03-08 | 广东工业大学 | PPG sampling rate estimation method based on noninvasive continuous blood pressure measurement |
| CN114469124A (en) * | 2022-01-30 | 2022-05-13 | 北京理工大学 | Method for identifying abnormal electrocardiosignals in motion process |
| CN115881276A (en) * | 2023-02-22 | 2023-03-31 | 合肥工业大学 | Time/frequency dual barcode feature image generation method for electrocardiosignal and storage medium |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113539231B (en) * | 2020-12-30 | 2024-06-18 | 腾讯科技（深圳）有限公司 | Audio processing method, vocoder, device, equipment and storage medium |
-
2023
- 2023-08-07 CN CN202310983456.3A patent/CN116712056B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| FR2648567A1 (en) * | 1989-05-24 | 1990-12-21 | Inst Nat Sante Rech Med | Method for the digital processing of a signal by reversible transformation into wavelets |
| CN106887225A (en) * | 2017-03-21 | 2017-06-23 | 百度在线网络技术（北京）有限公司 | Acoustic feature extracting method, device and terminal device based on convolutional neural networks |
| CN109350078A (en) * | 2018-10-19 | 2019-02-19 | 天津大学 | Blood sugar concentration detection method based on ultra-wideband microwave inverse Fourier transform |
| CN114469124A (en) * | 2022-01-30 | 2022-05-13 | 北京理工大学 | Method for identifying abnormal electrocardiosignals in motion process |
| CN114145725A (en) * | 2022-02-08 | 2022-03-08 | 广东工业大学 | PPG sampling rate estimation method based on noninvasive continuous blood pressure measurement |
| CN115881276A (en) * | 2023-02-22 | 2023-03-31 | 合肥工业大学 | Time/frequency dual barcode feature image generation method for electrocardiosignal and storage medium |
Non-Patent Citations (3)
| Title |
|---|
| Irene S Lensen，等.Heart rate fragmentation gives novel insights into non-autonomic mechanisms governing beat-to-beat control of the heart's rhythm.《JRSM CARDIOVASCULAR DISEASE》.2020,(第9期),第1-14页. * |
| Shockwave Signal Downsampling Rate Acquisition Based on Sparse Fourier Transform;BO XU，等;《IEEE》;第10卷;第44076-44087页 * |
| 全卷积循环神经网络的语音情感识别;朱敏，等;《声学技术》;第40卷(第5期);第645-651页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116712056A (en) | 2023-09-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN116712056B (en) | Characteristic image generation and identification method, equipment and storage medium for electrocardiogram data | |
| Aziz et al. | ECG-based machine-learning algorithms for heartbeat classification | |
| US10869610B2 (en) | System and method for identifying cardiac arrhythmias with deep neural networks | |
| CN107516075B (en) | Electrocardiosignal detection method and device and electronic equipment | |
| JP2021526063A (en) | Arrhythmia detection methods, devices, electronic devices and computer storage media | |
| CA3069891A1 (en) | Discovering novel features to use in machine learning techniques, such as machine learning techniques for diagnosing medical conditions | |
| CN111046824A (en) | Time series signal efficient denoising and high-precision reconstruction modeling method and system | |
| CN109124620B (en) | Atrial fibrillation detection method, device and equipment | |
| WO2020228420A1 (en) | Denoising autoencoder training method, electrocardiography signal denoising method, and apparatuses | |
| CN109948396B (en) | Heart beat classification method, heart beat classification device and electronic equipment | |
| WO2022077888A1 (en) | Scatter diagram classification method and apparatus for photoplethysmography signal | |
| CN111714124B (en) | Magnetic resonance film imaging method, device, imaging equipment and storage medium | |
| JP2023104885A (en) | Electrocardiographic heart rate multi-type prediction method based on graph convolution | |
| CN113647905A (en) | Training method for predicting blood pressure by deep neural network, computer device and storage medium | |
| CN114707539A (en) | Hand joint angle estimation method, hand joint angle estimation device, storage medium and equipment | |
| Zeleniy et al. | Image Processing as an Analysis Tool in Medical Research | |
| CN111461158A (en) | Method, apparatus, storage medium, and system for identifying features in ultrasound images | |
| Zhang et al. | Research on lung sound classification model based on dual-channel CNN-LSTM algorithm | |
| Liang et al. | Identification of heart sounds with arrhythmia based on recurrence quantification analysis and Kolmogorov entropy | |
| Banerjee et al. | A new moving horizon estimation based real-time motion artifact removal from wavelet subbands of ECG signal using particle filter | |
| CN116671929A (en) | Method, system, equipment and medium for determining optimal decomposition scale for electrocardiosignal denoising | |
| CN114010227B (en) | Right ventricle characteristic information identification method and device | |
| CN110853012B (en) | Method, apparatus and computer storage medium for obtaining cardiac parameters | |
| CN113349796B (en) | Sleep monitoring method, equipment and storage medium based on multi-source signals | |
| Dias et al. | Cleaning ECG with Deep Learning: A Denoiser Tested in Industrial Settings |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |