CN111643073A - Electrocardio data recognition device and method, equipment and computer readable storage medium - Google Patents
Electrocardio data recognition device and method, equipment and computer readable storage medium Download PDFInfo
- Publication number
- CN111643073A CN111643073A CN202010364989.XA CN202010364989A CN111643073A CN 111643073 A CN111643073 A CN 111643073A CN 202010364989 A CN202010364989 A CN 202010364989A CN 111643073 A CN111643073 A CN 111643073A
- Authority
- CN
- China
- Prior art keywords
- data
- feature
- electrocardiographic
- neural network
- electrocardiogram
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Abstract
The invention provides an electrocardiogram data recognition device, an electrocardiogram data recognition method, electrocardiogram data recognition equipment and a computer readable storage medium, and belongs to the technical field of data recognition. An electrocardiographic data recognition apparatus according to the present invention is a device for recognizing a category of electrocardiographic data, comprising: the device comprises a data feature extractor and a feature recognition classifier, wherein the data feature extractor is configured to perform feature extraction on electrocardiogram data to be recognized through a first neural network to obtain feature data of the electrocardiogram data to be recognized; the feature recognition classifier is configured to classify the feature data through a machine learning algorithm to realize class recognition of the electrocardiographic data.
Description
Technical Field
The invention belongs to the technical field of data identification, and particularly relates to an electrocardiogram data identification device, an electrocardiogram data identification method, electrocardiogram data identification equipment and a computer readable storage medium.
Background
The electrocardio monitoring is the most effective means for monitoring the heart rhythm, and whether the heart rhythm is normal or not can be found by monitoring the heart rhythm, so that the electrocardio monitoring is used for checking various symptoms such as arrhythmia and the like.
In the prior art, the electrocardio data of the user can be acquired through the medical instrument, but the judgment on whether the cardiac rhythm is normal or not still needs a doctor with professional experience, so that the workload is large.
Disclosure of Invention
The invention aims to solve at least one technical problem in the prior art and provides an electrocardiogram data recognition device which can accurately recognize the category of electrocardiogram data.
The technical scheme adopted for solving the technical problem of the invention is an electrocardiogram data identification device which is used for identifying the category of electrocardiogram data and is characterized by comprising the following components: a data characteristic extractor and a characteristic identification classifier, wherein,
the data feature extractor is configured to perform feature extraction on the electrocardiogram data to be identified through a first neural network to obtain feature data of the electrocardiogram data to be identified;
the feature recognition classifier is configured to classify the feature data through a machine learning algorithm to realize class recognition of the electrocardiographic data.
Optionally, the data feature extractor is configured to perform feature extraction on the electrocardiographic data to be identified through a convolutional neural network, so as to obtain feature data of the electrocardiographic data to be identified.
Further optionally, the feature recognition classifier is configured to perform classification processing on the feature data output by the data feature extractor through a support vector machine algorithm to realize class recognition on the electrocardiographic data.
Optionally, before identifying the category of the electrocardiographic data to be identified, the data feature extractor is further configured to train a first neural network by using sample electrocardiographic data; wherein the content of the first and second substances,
the sample electrocardiographic data comprises electrocardiographic data of different categories;
the data feature extractor is configured to obtain a parameter group F1 of a feature extraction function of a first neural network by training the first neural network using the electrocardiographic data of different classes as an input and feature data of the electrocardiographic data of different classes as an output, and to form the first neural network based on the parameter group F1.
Further optionally, before the category of the electrocardiographic data to be recognized is recognized, the feature recognition classifier is further configured to train a machine learning algorithm by using feature data output by the data feature extractor; wherein the content of the first and second substances,
the characteristic data output by the data characteristic extractor comprises characteristic data output after the first neural network performs characteristic extraction on the electrocardio data of different types;
the feature recognition classifier is configured to obtain a parameter set F2 of a machine learning algorithm by training the machine learning algorithm using the feature data as input and the electrocardiographic data category as output, and the machine learning algorithm based on the parameter set F2.
Another technical solution adopted to solve the technical problem of the present invention is an electrocardiographic data recognition method for recognizing the category of electrocardiographic data, including:
performing feature extraction on the electrocardiogram data to be identified through a neural network to obtain feature data of the electrocardiogram data to be identified;
and classifying the characteristic data through a machine learning algorithm to realize the category identification of the electrocardiogram data.
Optionally, the step of performing feature extraction on the electrocardiographic data to be identified to obtain feature data of the electrocardiographic data to be identified includes:
and performing feature extraction on the electrocardiogram data to be identified through a convolutional neural network to obtain feature data of the electrocardiogram data to be identified.
Further optionally, before identifying the category of the electrocardiographic data to be identified, training a first neural network by using sample electrocardiographic data is further included; wherein the content of the first and second substances,
the sample electrocardiographic data comprises electrocardiographic data of different categories;
the step of training the first neural network by adopting the sample electrocardiogram data comprises the following steps: the electrocardiogram data of different types are used as input, the feature data of the electrocardiogram data of different types are used as output, a first neural network is trained, and a parameter group F1 of a feature extraction function of the first neural network and the first neural network formed based on the parameter group F1 are obtained.
Further optionally, before the category of the electrocardiographic data to be recognized is recognized, training a machine learning algorithm by using feature data output by the data feature extractor; wherein the content of the first and second substances,
the characteristic data output by the data characteristic extractor comprises characteristic data output after the first neural network performs characteristic extraction on the electrocardio data of different types;
the step of training a machine learning algorithm by using the feature data output by the data feature extractor comprises: and training a machine learning algorithm by using the feature data as input and the electrocardiogram data type as output, and obtaining a parameter group F2 of the machine learning algorithm and a machine learning algorithm based on the parameter group F2.
Another technical solution adopted to solve the technical problem of the present invention is an electrocardiographic data recognition apparatus, comprising: at least one processor, at least one memory, and computer instructions stored in the memory, the processor being configured to implement one or more steps of any one of the methods described above when executing the computer instructions.
Another technical solution to solve the technical problem of the present invention is a computer-readable storage medium having computer instructions stored thereon, wherein the computer instructions, when executed by a processor, implement one or more steps of any one of the above methods.
Drawings
Fig. 1 and 2 are schematic views of an electrocardiographic data recognition apparatus according to an embodiment of the present invention;
fig. 3 is a schematic diagram of an electrocardiographic data recognition structure according to an embodiment of the present invention.
Detailed Description
In order to make the technical solutions of the present invention better understood, the present invention will be described in further detail with reference to the accompanying drawings and specific embodiments.
Example 1:
the present embodiment provides an electrocardiographic data recognition apparatus for recognizing the category of electrocardiographic data. This electrocardio data recognition device includes: the device comprises a data feature extractor and a feature recognition classifier, wherein the data feature extractor is configured to perform feature extraction on the electrocardiogram data to be recognized to obtain feature data of the electrocardiogram data to be recognized; the feature recognition classifier is configured to classify the feature data through a machine learning algorithm to realize class recognition on the electrocardio data.
In an embodiment of the invention, the category of the electrocardiographic data includes normal heart rate, atrial fibrillation, bradycardia, tachycardia, noise and the like. The electrocardiographic data of different categories have different characteristic data.
The data feature extractor in the electrocardiographic data recognition apparatus according to this embodiment can perform feature extraction on electrocardiographic data without a category identifier to obtain feature data, and then use the obtained feature data as an input of a feature recognition classifier, and perform classification and recognition on the feature data by a machine learning algorithm in the feature recognition classifier, thereby obtaining a category of the feature data, that is, obtaining a category of the electrocardiographic data. It can be seen that, in this embodiment, the electrocardiographic data to be recognized is not directly classified by the feature recognition classifier, but the electrocardiographic data is subjected to feature extraction by the data feature extractor to obtain feature data, and then is subjected to accurate classification, so that a more accurate classification result can be obtained.
In some embodiments, as shown in fig. 2, the data feature extractor is configured to perform feature extraction on the electrocardiographic data to be identified through a convolutional neural network, so as to obtain feature data of the electrocardiographic data to be identified.
Based on the features of common electrocardiographic data (electrocardiographic data in the prior art is usually one-dimensional signal data), the one-dimensional convolutional neural network adopted by the data feature extractor is used as a learning model for feature extraction. The one-dimensional convolutional neural network is used as a deep learning model, can effectively perform learning analysis on sequence data, and realizes extraction of characteristic data.
In this embodiment, the data feature extractor acquires feature data of the electrocardiographic data, and the data recognition classifier classifies the feature data. It should be noted here that the convolutional neural network can actually extract feature data and also can classify the feature data, and a common output result of the convolutional neural network learning model is a classification result. In this embodiment, as shown in fig. 2, only the feature extraction function of the convolutional neural network is applied, and the intermediate result (extracted feature data) of the convolutional neural network is output to the feature recognition classifier as an output result.
In this embodiment, the feature recognition classifier is configured to perform classification processing on the feature data through a machine learning algorithm to realize class recognition on the electrocardiographic data. That is, in this embodiment, the classifier in the convolutional neural network is replaced with a machine learning algorithm, and the feature data output by the data feature extractor is classified. In some embodiments, the feature recognition classifier may be specifically configured to classify the feature data output by the data feature extractor by a support vector machine algorithm (SVM algorithm).
Specifically, in the machine learning method, the SVM algorithm has a relatively accurate classification effect on data with a relatively small scale. The characteristic data of the electrocardio data extracted by the convolutional neural network is more abstract and lower in dimensionality, and the SVM algorithm has a more accurate classification effect on the data with smaller scale. The SVM algorithm is used for replacing the original classifier in the convolutional neural network, so that better classification performance can be obtained compared with the method that the convolutional neural network is directly used for classification.
In the electrocardiographic data recognition device provided by the embodiment, the one-dimensional convolutional neural network is combined with the traditional machine learning method, the classification algorithm in the machine learning is used for replacing the original classifier in the convolutional neural network, and the obtained improved algorithm model integrates the advantages of the two algorithms, can automatically recognize the type of electrocardiographic data, and has higher classification performance. Specifically, experiments prove that the value F1 of the electrocardiographic data recognition device provided by the embodiment can reach 0.7982, and the accuracy can reach 0.8450.
In some embodiments, prior to identifying the category of the electrocardiographic data to be identified, the data feature extractor is further configured to perform training of a first neural network using the sample electrocardiographic data; wherein, the sample electrocardio data comprises different types of electrocardio data. That is, the data feature extractor is further configured to derive the first neural network by training
Specifically, the data feature extractor is configured to obtain the parameter group F1 of the feature extraction function of the first neural network by training the first neural network using the electrocardiographic data of different classes as an input and the feature data of the electrocardiographic data of different classes as an output, and to form the first neural network based on the parameter group F1.
Wherein the first neural network may comprise a one-dimensional convolutional neural network. Specifically, the one-dimensional convolutional neural network can be constructed by a convolutional layer, a pooling layer, a nonlinear activation layer (BN layer), a global average pooling layer, and a fully-connected layer.
Specifically, in this embodiment, the electrocardiographic data of different categories are input to the data feature extractor, and specific category identifiers are labeled on the electrocardiographic data. The data feature extractor can output an intermediate result from the global average pooling layer after training is finished, namely feature data obtained by the convolutional neural network is output to be used as feature data for classifying by the subsequent feature recognition classifier.
In some embodiments, the feature extraction function of the convolutional neural network may comprise a softmax function (loss function). The definition of the softmax function is as follows:
wherein j represents a category, i represents a certain category in j, aiA value representing the classification. It is understood that in the present embodiment, the parameter group F1 may include the value a of the category ii。
In some embodiments, before identifying the category of the electrocardiographic data to be identified, the feature identification classifier is further configured to train a machine learning algorithm with the feature data output by the data feature extractor; the characteristic data output by the data characteristic extractor comprises characteristic data output after the first neural network performs characteristic extraction on the electrocardio data of different types; the feature recognition classifier is configured to use the feature data as input and the electrocardiographic data category as output, and to train the machine learning algorithm, obtain the parameter group F2 of the machine learning algorithm, and the machine learning algorithm based on the parameter group F2. In other words, the feature recognition classifier is also configured to derive a specific machine learning algorithm through training.
Specifically, the feature data of different categories extracted by the data feature extractor are input into the constructed SVM classifier, and the parameter set F2 of the kernel function is trained based on the feature data to determine the optimal parameter set of the SVM classifier, so as to obtain the SVM classifier based on the optimal parameter set through training.
In some embodiments, the objective function of the SVM classifier is defined as follows:
where w and b represent the weight and bias of the classification function, respectively, C represents a penalty parameter, ξiRepresenting the relaxation variable, and m represents the number of samples.
In some embodiments, the feature recognition classifier may be configured as an SVM classifier constructed using a Gaussian kernel as a kernel function. Gaussian nucleus KThe definition is as follows:
wherein x isi,xjRepresenting two data points and sigma representing the bandwidth of the kernel function. The feature recognition classifier can optimize the penalty parameter C of the Gaussian kernel and the sigma parameter of the kernel function of the parameter group by a GridSearch method and/or a 5-time cross verification method, and determine the optimal parameter penalty parameter C of the SVM classifier and the sigma parameter of the kernel function. And then inputting the characteristic data output by the data characteristic extractor into an SVM classifier for training to obtain an optimal parameter group for determining the SVM classifier. It is easy to understand that the parameter set of the SVM classifier may include a weight w and an offset b of the classification function, in addition to the parameter penalty parameter C and the σ parameter of the kernel function.
It is easy to understand that, in the electrocardiographic data recognition apparatus provided in this embodiment, a classifier with a better classification effect is used to replace a classifier in a convolutional neural network, and optionally, in the embodiment of the present invention, a feature recognition classifier may also be constructed and formed by using a decision tree and a random forest and other machine learning methods, so that, in combination with a data feature extractor constructed by using the convolutional neural network as a model, the electrocardiographic data is recognized more accurately and efficiently, and the classification performance of the electrocardiographic data recognition apparatus is improved.
Example 2:
this embodiment provides an electrocardiographic data recognition method that can recognize the type of electrocardiographic data by using the electrocardiographic data recognition apparatus provided in embodiment 1.
The method for recognizing the electrocardiographic data provided by the embodiment comprises two stages of training and practical operation of the electrocardiographic data recognition device, and specifically comprises the following steps:
a training stage:
s01, training a first neural network by adopting sample electrocardio data; wherein, the sample electrocardio data comprises different types of electrocardio data.
In this step, by training the first neural network using the electrocardiographic data of different types as input and the feature data of the electrocardiographic data of different types as output, the parameter group F1 of the feature extraction function of the first neural network and the first neural network formed based on the parameter group F1 are obtained.
It will be appreciated that different categories of electrocardiographic data have different characteristic data. Wherein the characteristic data may comprise waveform characteristic data.
In some embodiments, step S01 is preceded by a step S001 of constructing the first neural network. In some embodiments, the first neural network comprises a convolutional neural network, which may specifically comprise a one-dimensional convolutional neural network. Specifically, step S00 may include: and constructing a one-dimensional convolutional neural network comprising a convolution layer, a pooling layer, a BN layer, a global average pooling layer and a full-connection layer.
In some embodiments, step S01 includes:
and S011, inputting the electrocardiogram data into a convolutional neural network for training, and calculating the classification probability by using a softmax function. Specifically, in the convolutional neural network, feature data acquisition and electrocardiographic data classification are performed on electrocardiographic data of different types, a parameter group F1 of a classification function is adjusted according to a classification result, and then feature data acquisition is performed until the parameter group F1 of the classification function is stable, so that the convolutional neural network with the parameter group F1 is obtained.
And S012, outputting an intermediate result from the trained global average pooling layer of the convolutional neural network, that is, obtaining the feature data which is obtained by distributing the parameter group F1 and is used as the feature data for the subsequent feature recognition classifier to perform recognition classification.
And S02, training the machine learning algorithm by using the characteristic data output by the data characteristic extractor.
In this step, the first neural network (e.g., convolutional neural network) may be used to extract features of electrocardiographic data of different types and output feature data, and the electrocardiographic data type of each feature data may be used as output, so as to train the machine learning algorithm, thereby obtaining the parameter group F2 of the machine learning algorithm and the machine learning algorithm based on the parameter group F2.
Specifically, the feature data of different categories extracted by the data feature extractor are input into the SVM classifier, and the parameter group F2 of the kernel function is optimized based on the feature data to determine the optimal parameters of the SVM classifier, so as to train and obtain the SVM classifier based on the most parameter group.
In some embodiments, step S02 is preceded by a machine learning algorithm construction step S002. In some embodiments, the machine learning algorithm comprises an SVM classifier. Specifically, in some embodiments, the objective function of the SVM classifier is defined as follows:
where w and b represent the weight and bias of the classification function, respectively, C represents a penalty parameter, ξiRepresenting the relaxation variable, and m represents the number of samples.
In some embodiments, the feature recognition classifier may be configured as an SVM classifier constructed using a Gaussian kernel as a kernel function. The gaussian kernel K is defined as follows:
wherein x isi,xjRepresenting two data points and sigma representing the bandwidth of the kernel function.
Wherein, in some embodiments, step S002 further comprises: and optimizing the punishment parameter C of the Gaussian kernel and the sigma parameter of the kernel function by the parameter group through a GridSearch (grid search) method and/or a 5-time cross verification method, and determining the optimal parameter punishment parameter C of the SVM classifier and the sigma parameter of the kernel function. The specific optimizing steps may include: randomly dividing the data set into 5 parts by using 5 times of cross validation, wherein 4 parts are used as a training set, and 1 part is used as a validation set; determining the optimizing ranges of all candidate parameters (C and sigma); and traversing all candidate parameters, and taking the parameter with the best performance of the classifier as the optimal parameter.
Specifically, step S02 includes: and inputting the feature data output by the data feature extractor into an SVM classifier for training to obtain an optimal parameter group for determining the SVM classifier. It is easy to understand that the parameter set of the SVM classifier may include a weight w and an offset b of the classification function, in addition to the parameter penalty parameter C and the σ parameter of the kernel function.
The implementation stage is as follows:
and S11, performing feature extraction on the electrocardiogram data to be recognized to obtain feature data of the electrocardiogram data to be recognized.
And S12, classifying the characteristic data through a machine learning algorithm to realize the category identification of the electrocardio data.
A specific implementation procedure is introduced below:
and S21, performing feature extraction on the electrocardiogram data to be recognized through the convolutional neural network formed by training to obtain feature data of the electrocardiogram data to be recognized.
And S22, classifying the feature data through a support vector machine algorithm to realize the category identification of the electrocardio data.
In the method for recognizing the electrocardiographic data, firstly, a neural network and a machine learning algorithm are trained to enable the neural network to obtain the characteristic data of the electrocardiographic data, and the machine learning algorithm is trained by utilizing the characteristic data obtained by the neural network to enable the machine learning algorithm to recognize the category of the electrocardiographic data. And then, carrying out class identification on the input electrocardiogram data by using the trained neural network and the machine learning algorithm after training.
Example 3:
the present embodiment provides an electrocardiographic data recognition apparatus, wherein the electrocardiographic data recognition method according to embodiment 2 of the present invention can be implemented by the electrocardiographic data recognition apparatus. Fig. 3 shows a schematic diagram of a hardware structure of the electrocardiographic data recognition device according to the embodiment of the present invention.
The apparatus may include a processor and a memory storing computer instructions.
In particular, the processor may include a Central Processing Unit (CPU), or an Application Specific Integrated Circuit (ASIC), or a field programmable logic array FPGA, or an image processor GPU, or one or more integrated circuits that may be configured to implement the methods of embodiments of the present invention.
The memory may include mass storage for data or instructions. By way of example, and not limitation, memory may include a Hard Disk Drive (HDD), floppy Disk Drive, flash memory, optical Disk, magneto-optical Disk, magnetic tape, or Universal Serial Bus (USB) Drive or a combination of two or more of these. The memory may include removable or non-removable (or fixed) media, where appropriate. The memory may be internal or external to the data processing apparatus, where appropriate. In a particular embodiment, the memory is non-volatile solid-state memory. In a particular embodiment, the memory includes Read Only Memory (ROM). Where appropriate, the ROM may be mask-programmed ROM, Programmable ROM (PROM), Erasable PROM (EPROM), Electrically Erasable PROM (EEPROM), electrically rewritable ROM (EAROM), or flash memory or a combination of two or more of these.
The processor may read and execute the computer program instructions stored in the memory to implement one or more steps of any of the image reconstruction methods in the above embodiments.
In one example, the device may also include a communication interface and a bus. As shown in fig. 3, the processor, the memory, and the communication interface are connected via a bus to complete communication therebetween.
The communication interface is mainly used for realizing communication among modules, devices, units and/or equipment in the embodiment of the invention.
A bus comprises hardware, software, or both that couple the various components of the device to one another. By way of example, and not limitation, a bus may include an Accelerated Graphics Port (AGP) or other graphics bus, an Enhanced Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), a Hypertransport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an infiniband interconnect, a Low Pin Count (LPC) bus, a memory bus, a Micro Channel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a video electronics standards association local (VLB) bus, or other suitable bus or a combination of two or more of these. A bus may include one or more buses, where appropriate. Although specific buses have been described and shown in the embodiments of the invention, any suitable buses or interconnects are contemplated by the invention.
Example 4:
embodiments of the present invention may be implemented by providing a computer-readable storage medium. The computer readable storage medium having stored thereon computer instructions; the computer program instructions, when executed by a processor, implement one or more steps of any one of the electrocardiographic data recognition methods in the above embodiments.
It is to be understood that the invention is not limited to the specific arrangements and instrumentality described above and shown in the drawings. A detailed description of known methods is omitted herein for the sake of brevity. In the above embodiments, several specific steps are described and shown as examples. However, the method processes of the present invention are not limited to the specific steps described and illustrated, and those skilled in the art can make various changes, modifications and additions or change the order between the steps after comprehending the spirit of the present invention.
The functional blocks shown in the above structural block diagrams may be implemented as hardware, software, firmware, or a combination thereof. When implemented in hardware, it may be, for example, an electronic circuit, an Application Specific Integrated Circuit (ASIC), suitable firmware, plug-in, function card, or the like. When implemented in software, the elements of the invention are the programs or code segments used to perform the required tasks. The program or code segments may be stored in a machine-readable medium or transmitted by a data signal carried in a carrier wave over a transmission medium or a communication link. A "machine-readable medium" may include any medium that can store or transfer information. Examples of a machine-readable medium include electronic circuits, semiconductor memory devices, ROM, flash memory, Erasable ROM (EROM), floppy disks, CD-ROMs, optical disks, hard disks, fiber optic media, Radio Frequency (RF) links, and so forth. The code segments may be downloaded via a computer network such as the internet, a local area network, etc.
It should also be noted that the exemplary embodiments mentioned in this patent describe some methods or systems based on a series of steps or devices. However, the present invention is not limited to the order of the above-described steps, that is, the steps may be performed in the order mentioned in the embodiments, may be performed in an order different from the order in the embodiments, or may be performed simultaneously.
As will be apparent to those skilled in the art, for convenience and brevity of description, the specific working processes of the systems, modules and units described above may refer to the corresponding processes in the foregoing method embodiments, and are not described herein again. It should be understood that the scope of the present invention is not limited thereto, and any person skilled in the art can easily conceive various equivalent modifications or substitutions within the technical scope of the present invention, and these modifications or substitutions should be covered within the scope of the present invention.
It will be understood that the above embodiments are merely exemplary embodiments taken to illustrate the principles of the present invention, which is not limited thereto. It will be apparent to those skilled in the art that various modifications and improvements can be made without departing from the spirit and substance of the invention, and these modifications and improvements are also considered to be within the scope of the invention.
Claims (11)
1. An electrocardiographic data recognition apparatus for recognizing a category of electrocardiographic data, comprising: a data characteristic extractor and a characteristic identification classifier, wherein,
the data feature extractor is configured to perform feature extraction on the electrocardiogram data to be identified through a first neural network to obtain feature data of the electrocardiogram data to be identified;
the feature recognition classifier is configured to classify the feature data through a machine learning algorithm to realize class recognition of the electrocardiographic data.
2. The electrocardiogram data recognition apparatus according to claim 1, wherein the data feature extractor is configured to perform feature extraction on the electrocardiogram data to be recognized through a convolutional neural network, so as to obtain feature data of the electrocardiogram data to be recognized.
3. The electrocardiogram data recognition apparatus according to claim 2, wherein the feature recognition classifier is configured to classify the feature data outputted by the data feature extractor by a support vector machine algorithm, so as to realize class recognition of the electrocardiogram data.
4. The ecg data recognition device of claim 2, wherein prior to recognizing the ecg data class to be recognized, the data feature extractor is further configured to perform training of a first neural network using sample ecg data; wherein the content of the first and second substances,
the sample electrocardiographic data comprises electrocardiographic data of different categories;
the data feature extractor is configured to obtain a parameter group F1 of a feature extraction function of a first neural network by training the first neural network using the electrocardiographic data of different classes as an input and feature data of the electrocardiographic data of different classes as an output, and to form the first neural network based on the parameter group F1.
5. The ecg data recognition device of claim 4, wherein prior to recognizing the ecg data class to be recognized, the feature recognition classifier is further configured to train a machine learning algorithm using the feature data output by the data feature extractor; wherein the content of the first and second substances,
the characteristic data output by the data characteristic extractor comprises characteristic data output after the first neural network performs characteristic extraction on the electrocardio data of different types;
the feature recognition classifier is configured to obtain a parameter set F2 of a machine learning algorithm by training the machine learning algorithm using the feature data as input and the electrocardiographic data category as output, and the machine learning algorithm based on the parameter set F2.
6. An electrocardiogram data identification method is used for identifying the category of electrocardiogram data, and is characterized by comprising the following steps:
performing feature extraction on the electrocardiogram data to be identified through a neural network to obtain feature data of the electrocardiogram data to be identified;
and classifying the characteristic data through a machine learning algorithm to realize the category identification of the electrocardiogram data.
7. The method according to claim 6, wherein the step of extracting the characteristic of the electrocardiographic data to be recognized to obtain the characteristic data of the electrocardiographic data to be recognized comprises:
and performing feature extraction on the electrocardiogram data to be identified through a convolutional neural network to obtain feature data of the electrocardiogram data to be identified.
8. The method for recognizing the electrocardiographic data according to claim 7, further comprising training a first neural network by using sample electrocardiographic data before recognizing the category of the electrocardiographic data to be recognized; wherein the content of the first and second substances,
the sample electrocardiographic data comprises electrocardiographic data of different categories;
the step of training the first neural network by adopting the sample electrocardiogram data comprises the following steps: the electrocardiogram data of different types are used as input, the feature data of the electrocardiogram data of different types are used as output, a first neural network is trained, and a parameter group F1 of a feature extraction function of the first neural network and the first neural network formed based on the parameter group F1 are obtained.
9. The method for recognizing the electrocardiographic data according to claim 8, further comprising training a machine learning algorithm by using the feature data outputted from the data feature extractor before recognizing the category of the electrocardiographic data to be recognized; wherein the content of the first and second substances,
the characteristic data output by the data characteristic extractor comprises characteristic data output after the first neural network performs characteristic extraction on the electrocardio data of different types;
the step of training a machine learning algorithm by using the feature data output by the data feature extractor comprises: and training a machine learning algorithm by using the feature data as input and the electrocardiogram data type as output, and obtaining a parameter group F2 of the machine learning algorithm and a machine learning algorithm based on the parameter group F2.
10. An electrocardiographic data reading apparatus, comprising: at least one processor, at least one memory, and computer instructions stored in the memory, the processor being configured to implement one or more steps of the method of any one of claims 6-9 when executing the computer instructions.
11. A computer-readable storage medium having stored thereon computer instructions, which when executed by a processor, perform one or more steps of the method of any one of claims 6-9.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010364989.XA CN111643073A (en) | 2020-04-30 | 2020-04-30 | Electrocardio data recognition device and method, equipment and computer readable storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010364989.XA CN111643073A (en) | 2020-04-30 | 2020-04-30 | Electrocardio data recognition device and method, equipment and computer readable storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111643073A true CN111643073A (en) | 2020-09-11 |
Family
ID=72342203
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010364989.XA Pending CN111643073A (en) | 2020-04-30 | 2020-04-30 | Electrocardio data recognition device and method, equipment and computer readable storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111643073A (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103886184A (en) * | 2014-03-03 | 2014-06-25 | 浙江大学 | Construction method for heart pathology recognition model |
| CN104783782A (en) * | 2015-04-13 | 2015-07-22 | 深圳市飞马与星月科技研究有限公司 | Automatic detection method and device for electrocardiosignals |
| CN109887595A (en) * | 2019-01-16 | 2019-06-14 | 成都蓝景信息技术有限公司 | Heartbeat anomalous identification algorithm based on depth learning technology |
| CN110148466A (en) * | 2019-05-15 | 2019-08-20 | 东北大学 | A kind of heart impact signal atrial fibrillation computer aided diagnosing method based on transfer learning |
| CN110840402A (en) * | 2019-11-19 | 2020-02-28 | 山东大学 | Atrial fibrillation signal identification method and system based on machine learning |
| CN111053549A (en) * | 2019-12-23 | 2020-04-24 | 威海北洋电气集团股份有限公司 | Intelligent biological signal abnormality detection method and system |
-
2020
- 2020-04-30 CN CN202010364989.XA patent/CN111643073A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103886184A (en) * | 2014-03-03 | 2014-06-25 | 浙江大学 | Construction method for heart pathology recognition model |
| CN104783782A (en) * | 2015-04-13 | 2015-07-22 | 深圳市飞马与星月科技研究有限公司 | Automatic detection method and device for electrocardiosignals |
| CN109887595A (en) * | 2019-01-16 | 2019-06-14 | 成都蓝景信息技术有限公司 | Heartbeat anomalous identification algorithm based on depth learning technology |
| CN110148466A (en) * | 2019-05-15 | 2019-08-20 | 东北大学 | A kind of heart impact signal atrial fibrillation computer aided diagnosing method based on transfer learning |
| CN110840402A (en) * | 2019-11-19 | 2020-02-28 | 山东大学 | Atrial fibrillation signal identification method and system based on machine learning |
| CN111053549A (en) * | 2019-12-23 | 2020-04-24 | 威海北洋电气集团股份有限公司 | Intelligent biological signal abnormality detection method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109785976B (en) | Gout disease stage prediction system based on Soft-Voting | |
| CN112022141B (en) | Electrocardiosignal class detection method, electrocardiosignal class detection device and storage medium | |
| CN111460250A (en) | Image data cleaning method, image data cleaning device, image data cleaning medium, and electronic apparatus | |
| CN112426160A (en) | Electrocardiosignal type identification method and device | |
| CN111666865B (en) | Multi-lead electrocardiosignal convolutional neural network classification method and application method thereof | |
| CN116259415A (en) | Patient medicine taking compliance prediction method based on machine learning | |
| CN112690802B (en) | Method, device, terminal and storage medium for detecting electrocardiosignals | |
| CN109558827A (en) | A kind of finger vein identification method and system based on personalized convolutional neural networks | |
| US20200279148A1 (en) | Material structure analysis method and material structure analyzer | |
| CN111643073A (en) | Electrocardio data recognition device and method, equipment and computer readable storage medium | |
| CN111062345B (en) | Training method and device for vein recognition model and vein image recognition device | |
| CN112244863A (en) | Signal identification method, signal identification device, electronic device and readable storage medium | |
| CN115374882B (en) | Sleep staging method and device, electronic equipment and storage medium | |
| CN114898802B (en) | Terminal sequence frequency distribution characteristic determination method, evaluation method and device based on plasma free DNA methylation sequencing data | |
| Mohanty et al. | Tomato Plant Leaves Disease Detection using Machine Learning | |
| CN111611848B (en) | Cadaver iris recognition method and device | |
| CN113397563A (en) | Training method, device, terminal and medium for depression classification model | |
| Kaur et al. | Finger print Recognition Using Genetic Algorithm and Neural Network | |
| CN113724779A (en) | SNAREs protein identification method, system, storage medium and equipment based on machine learning technology | |
| CN108304746B (en) | Method and equipment for updating authentication reference information for electrocardio identity authentication | |
| CN111476282A (en) | Data classification method and device, storage medium and electronic equipment | |
| CN117893528B (en) | Method and device for constructing cardiovascular and cerebrovascular disease classification model | |
| CN115269939B (en) | Regular expression generation method and device, intelligent terminal and computer storage medium | |
| Mou et al. | Prediction and Rule Generation for Leukemia using Decision Tree and Association Rule Mining | |
| CN113229798B (en) | Model migration training method, device, computer equipment and readable storage medium |
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 |