CN108186011B - Atrial fibrillation detection method, atrial fibrillation detection device and readable storage medium - Google Patents
Atrial fibrillation detection method, atrial fibrillation detection device and readable storage medium Download PDFInfo
- Publication number
- CN108186011B CN108186011B CN201711334437.9A CN201711334437A CN108186011B CN 108186011 B CN108186011 B CN 108186011B CN 201711334437 A CN201711334437 A CN 201711334437A CN 108186011 B CN108186011 B CN 108186011B
- Authority
- CN
- China
- Prior art keywords
- atrial fibrillation
- recognition model
- electrocardiosignal
- training
- electrocardiosignals
- 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
Images
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7235—Details of waveform analysis
- 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
-
- 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
- A61B5/361—Detecting fibrillation
Landscapes
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Physics & Mathematics (AREA)
- Cardiology (AREA)
- Biophysics (AREA)
- Animal Behavior & Ethology (AREA)
- Veterinary Medicine (AREA)
- Public Health (AREA)
- General Health & Medical Sciences (AREA)
- Surgery (AREA)
- Molecular Biology (AREA)
- Pathology (AREA)
- Biomedical Technology (AREA)
- Heart & Thoracic Surgery (AREA)
- Medical Informatics (AREA)
- Fuzzy Systems (AREA)
- Mathematical Physics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Signal Processing (AREA)
- Psychiatry (AREA)
- Physiology (AREA)
- Evolutionary Computation (AREA)
- Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)
Abstract
The invention discloses an atrial fibrillation detection method based on electrocardiosignals, which comprises the following steps: acquiring an electrocardiosignal for training an atrial fibrillation recognition model, and extracting morphological characteristic parameters and time sequence information of the electrocardiosignal for training the atrial fibrillation recognition model; taking the morphological characteristic parameters and the time sequence information as training samples, and establishing an atrial fibrillation recognition model by adopting a deep learning mode; acquiring an atrial fibrillation detection result of the to-be-detected electrocardiosignal, and analyzing the to-be-detected electrocardiosignal based on the atrial fibrillation identification model to obtain the to-be-detected electrocardiosignal. The invention also discloses an atrial fibrillation detection device and a readable storage medium. The invention can realize the automatic detection of atrial fibrillation and improve the efficiency and the accuracy of the detection of atrial fibrillation.
Description
Technical Field
The invention relates to the field of medical signal processing, in particular to an atrial fibrillation detection method and device and a readable storage medium.
Background
Heart disease is a common cardiovascular disease and one of the most serious diseases in incidence and fatality, so the prevention and diagnosis of heart disease has become an important issue in the medical field. Atrial fibrillation (short for atrial fibrillation) is the most common sustained arrhythmia with the great harm to the heart health and is often accompanied by symptoms such as palpitation, dizziness, chest discomfort, shortness of breath and the like.
At present, the electrocardiogram technology is always used as an important technical means for diagnosing heart diseases due to the advantages of simple and convenient operation method, no invasive damage to users and the like, but the detection result needs doctors to carefully observe the electrocardiogram through naked eyes so as to judge whether the users have atrial fibrillation problems. When doctors face massive electrocardiograms every day, the feasibility and the accuracy of atrial fibrillation detection are easily influenced by eye fatigue and repeated judgment. The existing computer-aided detection of atrial fibrillation mainly judges whether a user has the problem of atrial fibrillation by calculating the absolute value of the variance of the positions of two adjacent R peaks in an electrocardiosignal, but the existing detection algorithm cannot cope with complicated and variable practical conditions, such as race, gender, age, emotion and the like of the user.
Disclosure of Invention
The invention mainly aims to provide an atrial fibrillation detection method and device based on electrocardiosignals and a readable storage medium, and aims to solve the technical problem that the existing method for detecting atrial fibrillation by electrocardiosignals is inaccurate and reasonable.
In order to achieve the purpose, the invention provides an atrial fibrillation detection method based on electrocardiosignals, which comprises the following steps:
acquiring an electrocardiosignal for training an atrial fibrillation recognition model, and extracting morphological characteristic parameters and time sequence information of the electrocardiosignal for training the atrial fibrillation recognition model;
taking the morphological characteristic parameters and the time sequence information as training samples, and establishing an atrial fibrillation recognition model by adopting a deep learning mode;
acquiring an atrial fibrillation detection result of the to-be-detected electrocardiosignal, and analyzing the to-be-detected electrocardiosignal based on the atrial fibrillation identification model to obtain the to-be-detected electrocardiosignal.
Preferably, the acquiring the electrocardiosignals used for training the atrial fibrillation recognition model and extracting morphological characteristic parameters and timing sequence information of the electrocardiosignals used for training the atrial fibrillation recognition model includes:
acquiring atrial fibrillation electrocardiosignals and normal electrocardiosignals for training an atrial fibrillation recognition model;
preprocessing the atrial fibrillation electrocardiosignals and the normal electrocardiosignals to obtain a plurality of electrocardio data sequences;
judging whether each electrocardiogram data sequence meets a preset condition or not;
when a preset condition is met, morphological feature extraction is carried out on the electrocardio data sequence to obtain a plurality of morphological feature parameters;
identifying the time position of each morphological characteristic parameter in the electrocardiogram data sequence to obtain a plurality of time sequence information;
and determining the corresponding relation between each morphological characteristic parameter and each time sequence information in the electrocardiogram data sequence.
Preferably, the establishing an atrial fibrillation recognition model by using the morphological characteristic parameters and the time sequence information as training samples and adopting a deep learning mode comprises:
and taking the morphological characteristic parameters, the time sequence information and the corresponding relation as input quantities, taking the electrocardio conditions of the atrial fibrillation electrocardiosignals and the normal electrocardiosignals as output quantities, and training value samples of the input quantities and the output quantities by adopting a convolutional neural network and a long-time memory neural network to obtain an atrial fibrillation recognition model.
Preferably, after the morphological characteristic parameters and the time sequence information are used as training samples and a deep learning mode is adopted to establish an atrial fibrillation recognition model, the method for detecting atrial fibrillation further comprises the following steps:
and verifying the atrial fibrillation identification model by adopting K-fold cross verification or a confusion matrix to obtain a verification result so that maintenance personnel can maintain the atrial fibrillation identification model according to the verification result.
Preferably, the acquiring waits to detect the electrocardiosignal, and based on the atrial fibrillation recognition model analysis wait to detect the electrocardiosignal, obtain wait to detect the electrocardiosignal after the atrial fibrillation testing result, the atrial fibrillation testing method still includes:
combining the electrocardiosignal to be detected with the electrocardiosignal for training the atrial fibrillation recognition model to update a training sample of the atrial fibrillation recognition model;
and further training the atrial fibrillation recognition model according to the updated training samples.
In addition, to achieve the above object, the present invention also provides an atrial fibrillation detection apparatus, including:
a memory storing an atrial fibrillation detection program;
a processor for executing the atrial fibrillation detection program to perform the following operations:
acquiring an electrocardiosignal for training an atrial fibrillation recognition model, and extracting morphological characteristic parameters and time sequence information of the electrocardiosignal for training the atrial fibrillation recognition model;
taking the morphological characteristic parameters and the time sequence information as training samples, and establishing an atrial fibrillation recognition model by adopting a deep learning mode;
acquiring an atrial fibrillation detection result of the to-be-detected electrocardiosignal, and analyzing the to-be-detected electrocardiosignal based on the atrial fibrillation identification model to obtain the to-be-detected electrocardiosignal.
Preferably, the operation of acquiring the electrocardiosignals for training the atrial fibrillation recognition model and extracting morphological characteristic parameters and timing information of the electrocardiosignals for training the atrial fibrillation recognition model comprises the following steps:
acquiring atrial fibrillation electrocardiosignals and normal electrocardiosignals for training an atrial fibrillation recognition model;
preprocessing the atrial fibrillation electrocardiosignals and the normal electrocardiosignals to obtain a plurality of electrocardio data sequences;
judging whether each electrocardiogram data sequence meets a preset condition or not;
when a preset condition is met, morphological feature extraction is carried out on the electrocardio data sequence to obtain a plurality of morphological feature parameters;
identifying the time position of each morphological characteristic parameter in the electrocardiogram data sequence to obtain a plurality of time sequence information;
and determining the corresponding relation between each morphological characteristic parameter and each time sequence information in the electrocardiogram data sequence.
Preferably, the performing the operation of establishing the atrial fibrillation recognition model by using the morphological characteristic parameters and the time sequence information as training samples and adopting a deep learning mode includes:
and taking the morphological characteristic parameters, the time sequence information and the corresponding relation as input quantities, taking the electrocardio conditions of the atrial fibrillation electrocardiosignals and the normal electrocardiosignals as output quantities, and training value samples of the input quantities and the output quantities by adopting a convolutional neural network and a long-time memory neural network to obtain an atrial fibrillation recognition model.
Preferably, after performing the operation of acquiring the cardiac signal to be detected and analyzing the cardiac signal to be detected based on the atrial fibrillation recognition model to obtain an atrial fibrillation detection result of the cardiac signal to be detected, the processor further performs the following operations:
combining the electrocardiosignal to be detected with the electrocardiosignal for training the atrial fibrillation recognition model to update a training sample of the atrial fibrillation recognition model;
and further training the atrial fibrillation recognition model according to the updated training samples.
In addition, to achieve the above object, the present invention further provides a computer readable storage medium having an atrial fibrillation detection program stored thereon, which when executed by a processor, implements the steps of the method for detecting atrial fibrillation based on cardiac electrical signals as described in any one of the above.
According to the method, firstly, an electrocardiosignal for training an atrial fibrillation recognition model is obtained, morphological characteristic parameters and time sequence information of the electrocardiosignal are extracted, secondly, the extracted morphological characteristic parameters and the extracted time sequence information are used as training samples, an atrial fibrillation recognition model is established in a deep learning mode, finally, an electrocardiosignal to be detected is obtained, the electrocardiosignal to be detected is analyzed according to the atrial fibrillation recognition model, and an atrial fibrillation detection result of the electrocardiosignal to be detected is obtained, so that automatic detection of atrial fibrillation is achieved, and the efficiency and accuracy of atrial fibrillation detection are improved.
Drawings
Fig. 1 is a schematic structural diagram of an operating environment of an atrial fibrillation detection apparatus according to an embodiment of the present invention;
FIG. 2 is a schematic flowchart illustrating an embodiment of a method for detecting atrial fibrillation based on cardiac electrical signals according to the present invention;
FIG. 3 is a detailed flowchart of step S10 in FIG. 2;
fig. 4 is a schematic flow chart of another embodiment of the atrial fibrillation detection method based on the electrocardiosignals.
The implementation, functional features and advantages of the objects of the present invention will be further explained with reference to the accompanying drawings.
Detailed Description
It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
Referring to fig. 1, fig. 1 is a schematic structural diagram of an operating environment of an atrial fibrillation detection apparatus according to an embodiment of the present invention.
As shown in fig. 1, the atrial fibrillation detection apparatus may include: a processor 1001, such as a CPU, a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. Wherein a communication bus 1002 is used to enable connective communication between these components. The user interface 1003 may include a Display (Display), an input unit such as a Keyboard (Keyboard), and the network interface 1004 may optionally include a standard wired interface, a wireless interface (e.g., WI-FI interface). The memory 1005 may be a high-speed RAM memory or a non-volatile memory (e.g., a magnetic disk memory). The memory 1005 may alternatively be a storage device separate from the processor 1001.
Optionally, the atrial fibrillation detection apparatus may further include a tuner demodulator, a return channel, RF (Radio Frequency) circuitry, a sensor, audio circuitry, a speaker, and the like.
Those skilled in the art will appreciate that the hardware configuration of the atrial fibrillation detection apparatus shown in fig. 1 does not constitute a limitation of the atrial fibrillation detection apparatus, and may include more or fewer components than those shown, or some components may be combined, or a different arrangement of components may be used.
As shown in fig. 1, a memory 1005, which is a kind of computer-readable storage medium, may include therein an operating system, a network communication module, a user interface module, and a computer program. The operating system is a program for managing and controlling the atrial fibrillation detection apparatus and software resources, and supports the operation of the atrial fibrillation detection program and other software and/or programs.
In the hardware configuration of the atrial fibrillation detection apparatus shown in fig. 1, the network interface 1004 is mainly used for accessing a network; the user interface 1003 is mainly used for detecting a confirmation instruction, an editing instruction, and the like. And processor 1001 may be configured to call the atrial fibrillation detection program stored in memory 1005 and perform the following operations:
acquiring an electrocardiosignal for training an atrial fibrillation recognition model, and extracting morphological characteristic parameters and time sequence information of the electrocardiosignal for training the atrial fibrillation recognition model;
taking the morphological characteristic parameters and the time sequence information as training samples, and establishing an atrial fibrillation recognition model by adopting a deep learning mode;
acquiring an atrial fibrillation detection result of the to-be-detected electrocardiosignal, and analyzing the to-be-detected electrocardiosignal based on the atrial fibrillation identification model to obtain the to-be-detected electrocardiosignal.
Further, the atrial fibrillation detection apparatus calls the atrial fibrillation detection program stored in the memory 1005 by the processor 1001 to perform the following operations:
acquiring atrial fibrillation electrocardiosignals and normal electrocardiosignals for training an atrial fibrillation recognition model;
preprocessing the atrial fibrillation electrocardiosignals and the normal electrocardiosignals to obtain a plurality of electrocardio data sequences;
judging whether each electrocardiogram data sequence meets a preset condition or not;
when a preset condition is met, morphological feature extraction is carried out on the electrocardio data sequence to obtain a plurality of morphological feature parameters;
identifying the time position of each morphological characteristic parameter in the electrocardiogram data sequence to obtain a plurality of time sequence information;
and determining the corresponding relation between each morphological characteristic parameter and each time sequence information in the electrocardiogram data sequence.
Further, the atrial fibrillation detection apparatus calls the atrial fibrillation detection program stored in the memory 1005 by the processor 1001 to perform the following operations:
and taking the morphological characteristic parameters, the time sequence information and the corresponding relation as input quantities, taking the electrocardio conditions of the atrial fibrillation electrocardiosignals and the normal electrocardiosignals as output quantities, and training value samples of the input quantities and the output quantities by adopting a convolutional neural network and a long-time memory neural network to obtain an atrial fibrillation recognition model.
Further, the atrial fibrillation detection apparatus calls the atrial fibrillation detection program stored in the memory 1005 by the processor 1001 to perform the following operations:
and verifying the atrial fibrillation identification model by adopting K-fold cross verification or a confusion matrix to obtain a verification result so that maintenance personnel can maintain the atrial fibrillation identification model according to the verification result.
Further, the atrial fibrillation detection apparatus calls the atrial fibrillation detection program stored in the memory 1005 by the processor 1001 to perform the following operations:
combining the electrocardiosignal to be detected with the electrocardiosignal for training the atrial fibrillation recognition model to update a training sample of the atrial fibrillation recognition model;
and further training the atrial fibrillation recognition model according to the updated training samples.
Based on the hardware structure of the atrial fibrillation detection device, the embodiments of the atrial fibrillation detection method based on the electrocardiosignals are provided.
Referring to fig. 2, fig. 2 is a schematic flow chart of an embodiment of the atrial fibrillation detection method based on the electrocardiosignals.
In this embodiment, the atrial fibrillation detection method based on the electrocardiosignals includes the following steps:
step S10, acquiring electrocardiosignals for training an atrial fibrillation recognition model, and extracting morphological characteristic parameters and time sequence information of the electrocardiosignals for training the atrial fibrillation recognition model;
in this embodiment, before the heart beats, the cardiac muscle is excited first, and a weak current is generated during the excitation, that is, each cardiac cycle of the heart is accompanied by bioelectrical changes, and the weak current can be transmitted to various parts of the body surface. Because the tissues of each part of the body are different and the distances between each part and the heart are different, different potential changes are shown at each part of the body surface, and the direction, the frequency and the intensity of the surface potential changes generated by the electrical activity in the heart of the human body are regular. Therefore, the electrical signals of different parts of the body surface are dynamically detected by the signal acquisition equipment, amplified by the amplifier and traced by the recorder, and then the electrocardiosignals can be obtained. Meanwhile, the data of the electrocardiosignals change along with time, and the electrocardiogram is used for drawing and presenting the electrocardiosignals and is an intuitive presentation mode of the electrocardiosignals. Specifically, before acquiring the electrocardiosignals, the method further comprises the step of acquiring the electrocardiosignals, wherein the acquired electrocardiosignals can be acquired through preset electrocardiosignal acquisition equipment. The manner of acquiring the electrocardiosignals here may be consistent with conventional acquisition. The heart disease can be diagnosed by the form, amplitude and relative time relationship between the waves of the electrocardiogram waveform. Such as cardiac arrhythmia, myocardial infarction, extra systole, hypertension, ectopic beating of the heart, etc.
In this embodiment, the waveform change occurring in each cardiac cycle of the conventional electrocardiogram has a certain rule, such as the waveform form, amplitude, and relative time relationship between the waves, and thus can be summarized as a peak point feature, an interval feature, a baseline feature, a waveform feature, a heartbeat form, and the like. The form characteristic parameters of the extracted electrocardiosignals preferably comprise interval characteristics, waveform characteristics and heart beat forms which are used as characteristic parameters of atrial fibrillation detection. Because the data of the electrocardiosignals change along with time, the time sequence information of the electrocardiosignals is extracted, and the moving change position of the electrocardiosignals at a certain time point or a certain time period can be more accurately known. Furthermore, because various noises are often doped in the acquisition process of the electrocardiosignals, and the sources of the noises mainly include power frequency interference, baseline drift, myoelectricity interference or other noise interference and the like, preprocessing is required before extraction so as to avoid interfering atrial fibrillation detection.
Step S20, taking morphological characteristic parameters and time sequence information as training samples, and establishing an atrial fibrillation recognition model by adopting a deep learning mode;
in this embodiment, the model is a series of preconditions for solving a learning problem, the atrial fibrillation recognition model is a mathematical model constructed by using a mathematical logic method and a mathematical language, and the deep learning method is to enable a computer to learn new knowledge from existing data, that is, to perform systematic learning according to data extracted from electrocardiographic signals, for example, how to classify characteristic parameters, how to optimize morphological characteristics, and the like. The training process is a process of determining model parameters by using training samples, namely existing data.
And step S30, acquiring the electrocardiosignals to be detected, and analyzing the electrocardiosignals to be detected based on the atrial fibrillation recognition model to obtain atrial fibrillation detection results of the electrocardiosignals to be detected.
In this embodiment, the atrial fibrillation recognition model can be locally deployed or can be deployed on line at the cloud end, and the preferred online deployment at the cloud end can update data in real time, optimizes accuracy and reliability of the atrial fibrillation recognition model, and simultaneously can also be distributed to expand, further satisfies the complicated changeable actual atrial fibrillation detection condition. When the atrial fibrillation recognition model is locally deployed, electrocardiosignals of a user are acquired through the signal acquisition equipment, namely the electrocardiosignals to be detected are analyzed according to the acquired electrocardiosignals to be detected and combined with the atrial fibrillation recognition model, and then the electrocardiosignals to be detected are obtained, so that whether the electrocardiosignals to be detected are normal or atrial fibrillation is obtained.
In this embodiment, the electrocardiographic signal to be detected is used as data to be detected, and also needs to be preprocessed first to remove power frequency interference, baseline drift, other noise interference and the like of the electrocardiographic signal to be detected, so as to improve the signal-to-noise ratio of the electrocardiographic signal, and then morphological characteristic parameters and timing sequence information of the electrocardiographic signal to be detected are extracted and used as input values of the model. And (3) analyzing and comparing morphological characteristic parameters and time sequence information of the electrocardiosignal to be detected through mapping data of the atrial fibrillation identification model, and outputting a result, namely obtaining an atrial fibrillation detection result of the electrocardiosignal to be detected, namely obtaining whether the electrocardio condition of the electrocardiosignal to be detected belongs to atrial fibrillation or normal.
It should be noted that, for further improving the reliability of atrial fibrillation detection, firstly, the atrial fibrillation recognition model is used for preliminarily detecting electrocardiosignals to be detected, and then when the preliminary detection result is a small amount of atrial fibrillation results with complicated conditions, the doctor performs secondary rechecking confirmation, so that the time and energy for the doctor to look up the electrocardiogram can be saved, and the doctor can concentrate on more valuable case judgment.
According to the atrial fibrillation detection method, firstly, an electrocardiosignal used for training an atrial fibrillation recognition model is obtained, morphological characteristic parameters and time sequence information of the electrocardiosignal are extracted, secondly, the extracted morphological characteristic parameters and the extracted time sequence information are used as training samples, an atrial fibrillation recognition model is established in a deep learning mode, finally, an electrocardiosignal to be detected is obtained, the electrocardiosignal to be detected is analyzed according to the atrial fibrillation recognition model, an atrial fibrillation detection result of the electrocardiosignal to be detected is obtained, therefore, automatic detection of atrial fibrillation is achieved, and the efficiency and accuracy of atrial fibrillation detection are improved.
Referring to fig. 3, fig. 3 is a detailed flowchart of step S10 in fig. 2.
Based on the foregoing embodiment, in this embodiment, in step S10, acquiring an electrocardiographic signal used for training an atrial fibrillation recognition model, and extracting morphological characteristic parameters and timing information of the electrocardiographic signal used for training the atrial fibrillation recognition model, the method includes:
step S11, acquiring atrial fibrillation electrocardiosignals and normal electrocardiosignals for training an atrial fibrillation recognition model;
in this embodiment, the atrial fibrillation ecg signals are those diagnosed as having the atrial fibrillation problem, and the normal ecg signals are those diagnosed as healthy without the atrial fibrillation problem. Through two different electrocardiosignals as the basic data of training atrial fibrillation recognition model, can make the computer study of degree of depth for deal with different users, and then distinguish and wait to detect the electrocardiosignal and belong to the atrial fibrillation type. Further, in order to make the output result of the atrial fibrillation identification model more accurate, at the initial modeling stage, a certain proportion of atrial fibrillation electrocardiographic signals and normal electrocardiographic signals, such as a ratio of 500 atrial fibrillation electrocardiographic signals to 1500 normal electrocardiographic signals, needs to be acquired. Meanwhile, in order to further improve the operation efficiency, the electrocardiosignals can be labeled, for example, AF ═ 1 indicates atrial fibrillation, AF ═ 0 indicates normal, and the electrocardiosignals of two different electrocardio conditions can be distinguished through labeling.
Step S12, preprocessing atrial fibrillation electrocardiosignals and normal electrocardiosignals to obtain a plurality of electrocardio data sequences;
in this embodiment, although two electrocardiographic signals of different electrocardiographic conditions are obtained, both electrocardiographic signals need to be preprocessed in various manners, such as conversion, filtering, feature extraction, and other function operations. Because various noises are often doped in the acquisition process of the electrocardiosignals, the sources of the noises mainly include power frequency interference, baseline drift, myoelectric interference or other noise interference and the like. The acquired atrial fibrillation electrocardiosignals and normal electrocardiosignals are respectively preprocessed, power frequency interference, baseline drift, electromyographic interference or other noise interference of the electrocardiosignals are preferably filtered by a digital filtering algorithm to improve the signal to noise ratio of the electrocardiosignals, and then the electrocardiosignals are converted into a plurality of electrocardio data sequences, wherein the electrocardio data sequences at least comprise any one or more of filtering electrocardio data, R wave peak annotation data, premature beat point annotation data and interference segment annotation data. Through the electrocardio data sequence of different grade type, can be more accurate classify, and then improve the ability that the computer discerned atrial fibrillation.
Step S13, judging whether each electrocardiogram data sequence meets the preset conditions;
in this embodiment, the preset conditions are conditions preset by a developer, for example, the duration of each electrocardiographic data sequence is 40 to 240s, the sampling frequency is 512HZ, or at least 20 effective beats are included in a segment of the electrocardiographic data sequence with the largest RR interval fluctuation in continuous effective segments, and the preset conditions are specifically set according to actual conditions.
Step S14, when a preset condition is met, morphological feature extraction is carried out on the electrocardio data sequence to obtain a plurality of morphological feature parameters;
in this embodiment, because a plurality of electrocardiographic data sequences are obtained by preprocessing, whether each electrocardiographic data sequence meets a preset condition or not is judged one by one, and when the preset condition is met, the next operation can be performed, that is, morphological characteristic parameters are extracted, so as to ensure the reliability of atrial fibrillation identification model data. It should be noted that when the electrocardiogram data sequence does not meet the preset condition, the inconsistent electrocardiogram data sequence is removed, so that the operation is reduced, the data redundancy is avoided, and the training of the atrial fibrillation recognition model is influenced.
In this embodiment, a Convolutional Neural Network (CNN) algorithm is widely applied to image recognition and depth feature extraction, that is, various morphological feature parameters can be extracted by performing multiple convolutions and pooling on a electrocardiographic data sequence. The electrocardiogram data sequences are filtering electrocardiogram data, R wave peak annotation data, premature beat point annotation data, interference segment annotation data and the like, each electrocardiogram data sequence corresponds to different morphological characteristic parameters, for example, RR interval and/or R peak position can be extracted from the R wave peak annotation data, ectopic pacing point position can be extracted from the premature beat point annotation data, and the specific morphological characteristic parameters are set and extracted according to actual needs. In particular, the amount of the solvent to be used,
step S15, identifying the time position of each morphological characteristic parameter in the electrocardiogram data sequence to obtain a plurality of time sequence information;
step S16, determining the corresponding relationship between each morphological characteristic parameter and each time series information in the electrocardiographic data sequence.
In this embodiment, the Long Short Term Memory neural network (LSTM) algorithm is widely applied to sequence identification, that is, an electrocardiographic data sequence is identified by using the Long Short Term Memory neural network, specifically, since data is time-varying, each morphological characteristic parameter necessarily has a corresponding time point in the electrocardiographic data sequence, and thus time series information corresponding to each morphological characteristic parameter can be obtained, and further, a corresponding relationship between the morphological characteristic parameter and the time series information in the same electrocardiographic data sequence is determined, for example, a first ectopic pacing point in the sequence corresponds to a certain time point, and a second ectopic pacing point in the sequence corresponds to another time point.
In this embodiment, one electrocardiographic data sequence has a plurality of time series information and a plurality of morphological characteristic parameters, and each morphological characteristic parameter has corresponding time series information. Specifically, atrial fibrillation electrocardiograph signals are processed to obtain a plurality of electrocardiograph data sequences, and meanwhile, normal electrocardiograph signals are processed to obtain a plurality of electrocardiograph data sequences.
Further, in another embodiment of the present invention, morphological characteristic parameters, timing sequence information, and a corresponding relationship are used as input quantities of important data required for modeling, the electrocardiographic states of the atrial fibrillation electrocardiographic signals and the normal electrocardiographic signals are used as output quantities, a convolutional neural network and a long-and-short-term memory neural network are adopted to train value samples of the input quantities and the output quantities, a convolutional neural network structure model and a long-and-short-term memory neural network structure model are constructed, and the long-and-short-term memory neural network structure model and the convolutional neural network structure model are combined to obtain the atrial fibrillation recognition model of the present invention, which can give consideration to both morphological characteristic recognition and timing sequence recognition, and improve the accuracy and reliability of the atrial fibrillation recognition model.
In this embodiment, an atrial fibrillation recognition model is established, preferably, two classes of cross entropy loss (binary cross entropy loss), specifically, probability distributions of two classes in a sample set are p and q, where p is a true distribution and q is a non-true distribution. A cross-entropy cost function is used to measure the similarity between the two probability distributions. For a random variable X, the expectation of the information content of all its possible values (E [ i (X)) ] log (1/p)) is called entropy.
The entropy of the true distribution is:
if the error distribution q is used to represent the data from the true distribution p, then it should be:
since the sample judged to be q is from the true distribution p, the probability in H (p, q) is expected to be p (i). H (p, q) we call the "cross entropy". The model training process is to use the Adam gradient descent method to iteratively reduce the cross entropy of the two classes.
Further, in another embodiment of the invention, the final result evaluation of the atrial fibrillation diagnostic model adopts two ways of K-fold cross validation or confusion matrix to ensure that the overall accuracy reaches more than 95%, and the reproducibility and stability of the atrial fibrillation diagnostic model are verified by sampling for multiple times. Specifically, after the operation of establishing the atrial fibrillation recognition model by taking the timing sequence information and the morphological characteristic parameters as training samples and adopting a deep learning mode is executed, the processor further executes the following operation: and verifying the atrial fibrillation identification model by adopting K-fold cross verification or a confusion matrix to obtain a verification result. The atrial fibrillation recognition model is conveniently maintained by feeding back the verification result to related maintenance personnel, for example, when the total accuracy of the verification result does not reach a preset threshold value such as 90%, the maintenance personnel can analyze data according to the input quantity and the output quantity of the atrial fibrillation recognition model, optimize a calculation method and further improve the atrial fibrillation recognition model; for example, when the total accuracy of the verification result reaches a preset threshold, the system can be put into production and use.
In this embodiment, the first verification method: the K-fold cross validation preferentially randomizes the electrocardiosignals, wherein the electrocardiosignals comprise 228 atrial fibrillation electrocardiosignals and 4064 normal electrocardiosignals, each time, the electrocardiosignals are randomly divided into 80% of training sets and 20% of testing sets, and the effect stability of the model is repeatedly trained and predicted and evaluated. And a second verification method: the confusion matrix is used to compare the classification result with the actual measured value, i.e. the classification result is displayed in a confusion matrix. Each column of the confusion matrix represents a prediction category, the total number of each column representing the number of data predicted for that category; each row represents a true attribution category of data, and the total number of data in each row represents the number of data instances for that category. The values in each column represent the number of outputs for which real data is predicted as that class.
Further, in order to improve accuracy of atrial fibrillation detection, referring to fig. 4, another embodiment of the method for detecting atrial fibrillation according to the present invention is provided based on the foregoing embodiment, in this embodiment, the method for detecting atrial fibrillation further includes:
step S40, combining the electrocardiosignal to be detected and the electrocardiosignal for training the atrial fibrillation recognition model to update a training sample of the atrial fibrillation recognition model;
and step S50, further training the atrial fibrillation recognition model according to the updated training samples.
In this embodiment, because the electrocardiosignal to be detected obtains the detection result through the atrial fibrillation recognition model, preferably, label annotation AF ═ 1 or 0 is added to the electrocardiosignal to be detected, wherein 1 represents atrial fibrillation, and 0 represents normal, so preferably through incremental learning, the electrocardiosignal to be detected that has detected is incorporated into the data set of the electrocardiosignal when the original modeling, utilize the training sample that is continuously newly added, constantly adjust and optimize the atrial fibrillation recognition model, increase the analysis data of the electrocardiosignal, when being helpful to analyze the electrocardiosignal to be detected, improve more typical and high-accuracy analysis basis, and then improve the diagnosis accuracy of atrial fibrillation recognition.
In this embodiment, after the electrocardiographic signal to be detected and the electrocardiographic signal used for training the atrial fibrillation recognition model are combined, the corresponding morphological characteristic parameter and timing sequence information are extracted by the same data processing method as the original electrocardiographic signal, and the new morphological characteristic parameter and timing sequence information are incorporated into the original training sample, that is, the original training sample is updated. By adding the detected electrocardiosignals, real-time data updating is realized, and meanwhile, more data are used for training by the updated training samples, so that an atrial fibrillation recognition model is optimized.
Further optionally, in order to achieve the above object, the present invention further provides a computer-readable storage medium, where the computer-readable storage medium provided in this embodiment stores an atrial fibrillation detection program, which includes acquiring an electrocardiographic signal, extracting morphological feature parameters, extracting timing information, and building a model. The stored atrial fibrillation detection program can be read, interpreted and executed by the processor, so that the steps of the atrial fibrillation detection method based on the electrocardiosignals in any one of the embodiments of the atrial fibrillation detection method are realized.
It should be noted that, in this document, 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 an … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.
The above-mentioned serial numbers of the embodiments of the present invention are merely for description and do not represent the merits of the embodiments.
Through the above description of the embodiments, those skilled in the art will clearly understand that the method of the above embodiments can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware, but in many cases, the former is a better implementation manner. Based on such understanding, the technical solutions of the present invention may be embodied in the form of a software product, which is stored in a readable storage medium (such as ROM/RAM, magnetic disk, optical disk) and includes instructions for enabling a terminal (such as a mobile phone, a computer, a server, an air conditioner, or a network device) to execute the method according to the embodiments of the present invention.
The present invention is described in connection with the accompanying drawings, but the present invention is not limited to the above embodiments, which are only illustrative and not restrictive, and those skilled in the art can make various changes without departing from the spirit and scope of the invention as defined by the appended claims, and all changes that come within the meaning and range of equivalency of the specification and drawings that are obvious from the description and the attached claims are intended to be embraced therein.
The above description is only a preferred embodiment of the present invention, and not intended to limit the scope of the present invention, and all modifications of equivalent structures and equivalent processes, which are made by using the contents of the present specification and the accompanying drawings, or directly or indirectly applied to other related technical fields, are included in the scope of the present invention.
Claims (4)
1. An atrial fibrillation detection apparatus, comprising:
a memory storing an atrial fibrillation detection program;
a processor for executing the atrial fibrillation detection program to perform the following operations:
acquiring an electrocardiosignal for training an atrial fibrillation recognition model, and extracting morphological characteristic parameters and time sequence information of the electrocardiosignal for training the atrial fibrillation recognition model, wherein the morphological characteristic parameters and the time sequence information specifically comprise the following steps: acquiring atrial fibrillation electrocardiosignals and normal electrocardiosignals for training an atrial fibrillation recognition model; preprocessing the atrial fibrillation electrocardiosignals and the normal electrocardiosignals to obtain a plurality of electrocardio data sequences; the electrocardio data sequence at least comprises any one or more of filtered electrocardio data, R wave peak annotation data, premature beat point annotation data and interference segment annotation data; each electrocardio data sequence corresponds to different morphological characteristic parameters; judging whether each electrocardiogram data sequence meets preset conditions one by one; the preset condition is that the duration of each electrocardiogram data sequence is 40-240 s, or the electrocardiogram data sequences comprise at least 20 effective heartbeats in a segment with the largest RR interval fluctuation in continuous effective segments; when a preset condition is met, morphological feature extraction is carried out on the electrocardio data sequence to obtain a plurality of morphological feature parameters; when the electrocardiogram data sequence does not meet the preset condition, rejecting the non-conforming electrocardiogram data sequence; identifying the time position of each morphological characteristic parameter in the electrocardiogram data sequence to obtain a plurality of time sequence information; determining the corresponding relation between each morphological characteristic parameter and each time sequence information in the electrocardiogram data sequence;
taking the morphological characteristic parameters and the time sequence information as training samples, and establishing an atrial fibrillation recognition model by adopting a deep learning mode;
verifying the atrial fibrillation identification model by adopting K-fold cross verification or a confusion matrix to obtain a verification result so that maintenance personnel can maintain the atrial fibrillation identification model according to the verification result;
acquiring an atrial fibrillation detection result of the to-be-detected electrocardiosignal, and analyzing the to-be-detected electrocardiosignal based on the atrial fibrillation identification model to obtain the to-be-detected electrocardiosignal.
2. The apparatus according to claim 1, wherein the performing of the operation of establishing the atrial fibrillation recognition model by using the morphological feature parameters and the timing sequence information as training samples and using a deep learning manner comprises:
and taking the morphological characteristic parameters, the time sequence information and the corresponding relation as input quantities, taking the electrocardio conditions of the atrial fibrillation electrocardiosignals and the normal electrocardiosignals as output quantities, and training value samples of the input quantities and the output quantities by adopting a convolutional neural network and a long-time memory neural network to obtain an atrial fibrillation recognition model.
3. The atrial fibrillation detection apparatus according to any one of claims 1 to 2, wherein after performing the operations of acquiring an atrial fibrillation detection result of an atrial fibrillation detection signal to be detected and analyzing the atrial fibrillation detection signal based on the atrial fibrillation recognition model, said processor further performs the operations of:
combining the electrocardiosignal to be detected with the electrocardiosignal for training the atrial fibrillation recognition model to update a training sample of the atrial fibrillation recognition model;
and further training the atrial fibrillation recognition model according to the updated training samples.
4. A computer-readable storage medium having an atrial fibrillation detection program stored thereon, the atrial fibrillation detection program being executed by a processor to:
acquiring an electrocardiosignal for training an atrial fibrillation recognition model, and extracting morphological characteristic parameters and time sequence information of the electrocardiosignal for training the atrial fibrillation recognition model, wherein the morphological characteristic parameters and the time sequence information specifically comprise the following steps: acquiring atrial fibrillation electrocardiosignals and normal electrocardiosignals for training an atrial fibrillation recognition model; preprocessing the atrial fibrillation electrocardiosignals and the normal electrocardiosignals to obtain a plurality of electrocardio data sequences; the electrocardio data sequence at least comprises any one or more of filtered electrocardio data, R wave peak annotation data, premature beat point annotation data and interference segment annotation data; each electrocardio data sequence corresponds to different morphological characteristic parameters; judging whether each electrocardiogram data sequence meets preset conditions one by one; the preset condition is that the duration of each electrocardiogram data sequence is 40-240 s, or the electrocardiogram data sequences comprise at least 20 effective heartbeats in a segment with the largest RR interval fluctuation in continuous effective segments; when a preset condition is met, morphological feature extraction is carried out on the electrocardio data sequence to obtain a plurality of morphological feature parameters; when the electrocardiogram data sequence does not meet the preset condition, rejecting the non-conforming electrocardiogram data sequence; identifying the time position of each morphological characteristic parameter in the electrocardiogram data sequence to obtain a plurality of time sequence information; determining the corresponding relation between each morphological characteristic parameter and each time sequence information in the electrocardiogram data sequence;
taking the morphological characteristic parameters and the time sequence information as training samples, and establishing an atrial fibrillation recognition model by adopting a deep learning mode;
verifying the atrial fibrillation identification model by adopting K-fold cross verification or a confusion matrix to obtain a verification result so that maintenance personnel can maintain the atrial fibrillation identification model according to the verification result;
acquiring an atrial fibrillation detection result of the to-be-detected electrocardiosignal, and analyzing the to-be-detected electrocardiosignal based on the atrial fibrillation identification model to obtain the to-be-detected electrocardiosignal.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711334437.9A CN108186011B (en) | 2017-12-13 | 2017-12-13 | Atrial fibrillation detection method, atrial fibrillation detection device and readable storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711334437.9A CN108186011B (en) | 2017-12-13 | 2017-12-13 | Atrial fibrillation detection method, atrial fibrillation detection device and readable storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108186011A CN108186011A (en) | 2018-06-22 |
| CN108186011B true CN108186011B (en) | 2021-06-01 |
Family
ID=62574247
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711334437.9A Active CN108186011B (en) | 2017-12-13 | 2017-12-13 | Atrial fibrillation detection method, atrial fibrillation detection device and readable storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108186011B (en) |
Families Citing this family (22)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110717597A (en) * | 2018-06-26 | 2020-01-21 | 第四范式（北京）技术有限公司 | Method and device for acquiring time sequence characteristics by using machine learning model |
| CN109033990B (en) * | 2018-07-02 | 2021-08-17 | 河南大学 | CNN model heartbeat classification method based on intra-class and inter-class distances |
| CN109077714B (en) * | 2018-07-05 | 2021-03-23 | 广州视源电子科技股份有限公司 | Signal identification method, device, equipment and storage medium |
| CN109077719A (en) * | 2018-07-05 | 2018-12-25 | 广州视源电子科技股份有限公司 | Signal recognition method, device, equipment and storage medium |
| CN110786847B (en) * | 2018-08-02 | 2022-11-04 | 深圳市理邦精密仪器股份有限公司 | Electrocardiogram signal library building method and analysis method |
| CN109325402B (en) * | 2018-08-06 | 2022-04-15 | 高维度（深圳）生物信息智能应用有限公司 | Signal processing method, system and computer storage medium |
| CN110881966A (en) * | 2018-09-10 | 2020-03-17 | 深圳市游弋科技有限公司 | Algorithm for processing electrocardiogram data by using LSTM network |
| CN109171712B (en) * | 2018-09-28 | 2022-03-08 | 东软集团股份有限公司 | Atrial fibrillation identification method, atrial fibrillation identification device, atrial fibrillation identification equipment and computer readable storage medium |
| EP3870039A4 (en) * | 2018-10-26 | 2022-08-10 | Mayo Foundation for Medical Education and Research | Neural networks for atrial fibrillation screening |
| CN109567789B (en) * | 2018-12-03 | 2021-08-31 | 东软集团股份有限公司 | Electrocardiogram data segmentation processing method and device and readable storage medium |
| CN109740481B (en) * | 2018-12-26 | 2019-12-10 | 山东科技大学 | CNN and LSTM combined atrial fibrillation signal classification system based on jump connection |
| CN109871808B (en) * | 2019-02-21 | 2021-03-02 | 天津惊帆科技有限公司 | Atrial fibrillation model training and detecting method and device |
| CN109770893B (en) * | 2019-03-08 | 2022-11-18 | 东南大学 | Method and device for rapidly positioning atrial fibrillation position in Holter analysis system |
| TWI701681B (en) * | 2019-05-16 | 2020-08-11 | 中國醫藥大學附設醫院 | Prediction model of atrial fibrillation and prediction system thereof |
| CN110179456B (en) * | 2019-05-23 | 2021-11-02 | 中国航天员科研训练中心 | Electrocardio noise recognition model training and electrocardio noise detection method and device |
| CN110179465A (en) * | 2019-05-27 | 2019-08-30 | 东南大学附属中大医院 | Mechanical ventilation off line quantitative estimation method, device, equipment and storage medium |
| CN110547792B (en) * | 2019-07-31 | 2022-02-15 | 平安科技（深圳）有限公司 | Atrial fibrillation detection method and device, computer equipment and storage medium |
| WO2021042372A1 (en) * | 2019-09-06 | 2021-03-11 | 中国医药大学附设医院 | Atrial fibrillation prediction model and prediction system thereof |
| CN110812699B (en) * | 2019-12-03 | 2021-09-28 | 四川大学 | Remote defibrillation protection system and control method thereof |
| CN112971790A (en) * | 2019-12-18 | 2021-06-18 | 华为技术有限公司 | Electrocardiosignal detection method, device, terminal and storage medium |
| CN111920405A (en) * | 2020-09-15 | 2020-11-13 | 齐鲁工业大学 | Atrial fibrillation signal identification system and method |
| CN112120691A (en) * | 2020-09-22 | 2020-12-25 | 浙江智柔科技有限公司 | Signal identification method and device based on deep learning and computer equipment |
Family Cites Families (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104783782B (en) * | 2015-04-13 | 2017-09-01 | 深圳市飞马与星月科技研究有限公司 | Electrocardiosignal automatic testing method and device |
| CN104873186B (en) * | 2015-04-17 | 2018-05-18 | 中国科学院苏州生物医学工程技术研究所 | A kind of wearable artery detection device and its data processing method |
| CN105726018A (en) * | 2016-02-06 | 2016-07-06 | 河北大学 | Automatic atrial fibrillation detection method irrelevant to RR interphase |
| CN106073755A (en) * | 2016-05-27 | 2016-11-09 | 成都信汇聚源科技有限公司 | The implementation method that in a kind of miniature holter devices, atrial fibrillation identifies automatically |
| CN106901723A (en) * | 2017-04-20 | 2017-06-30 | 济南浪潮高新科技投资发展有限公司 | A kind of electrocardiographic abnormality automatic diagnosis method |
| CN107203692B (en) * | 2017-05-09 | 2020-05-05 | 哈尔滨工业大学（威海） | Electrocardio data digital signal processing method based on deep convolutional neural network |
-
2017
- 2017-12-13 CN CN201711334437.9A patent/CN108186011B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN108186011A (en) | 2018-06-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108186011B (en) | Atrial fibrillation detection method, atrial fibrillation detection device and readable storage medium | |
| CN107981858B (en) | Automatic electrocardiogram heart beat identification and classification method based on artificial intelligence | |
| US11350868B2 (en) | Electrocardiogram information processing method and electrocardiogram workstation system | |
| CN107358196B (en) | Heart beat type classification method and device and electrocardiograph | |
| US8401624B2 (en) | ECG signal analysis tool | |
| WO2019100565A1 (en) | Method and device for self-learning dynamic electrocardiography analysis employing artificial intelligence | |
| WO2019100560A1 (en) | Artificial intelligence self-learning-based automatic electrocardiography analysis method and apparatus | |
| CN109411042B (en) | Electrocardio information processing method and electrocardio workstation | |
| US9042973B2 (en) | Apparatus and method for measuring physiological signal quality | |
| US8825148B2 (en) | System for monitoring and diagnosis of cardiac electrogram signals using multi-dimensional analysis | |
| CN106815570B (en) | Electrocardiosignal ST-T segment identification method based on dynamic pattern identification | |
| KR101366101B1 (en) | System and method for classificating normal signal of personalized ecg | |
| Klug et al. | The BeMoBIL Pipeline for automated analyses of multimodal mobile brain and body imaging data | |
| CN112426160A (en) | Electrocardiosignal type identification method and device | |
| CN115486849A (en) | Electrocardiosignal quality evaluation method, device and equipment | |
| JP5207172B2 (en) | Waveform analysis apparatus and waveform analysis program | |
| Vest et al. | Benchmarking heart rate variability toolboxes | |
| CN108836312B (en) | Clutter rejection method and system based on artificial intelligence | |
| Naseri et al. | An expert electrocardiogram quality evaluation algorithm based on signal mobility factors | |
| CN111685754B (en) | Heart rate calculation method and system for wearable ECG acquisition equipment | |
| CN116229521B (en) | Method, device and equipment for detecting heart information based on multi-scale features | |
| EP4371487A1 (en) | Method for generating multiple pieces of standard electrocardiogram data on basis of deep learning algorithm | |
| CN111723622B (en) | Heart beat classification method, heart beat classification device, wearable equipment and storage medium | |
| CN210354670U (en) | Physiological monitoring and analysis system based on hybrid sensing | |
| CN116831589A (en) | Optimal lead signal selection method for long-time wearable electrocardiosignal |
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 |