Electrocardiogram classification method, training method, device, equipment and storage medium

Abstract

The application relates to a classification method and a training method of an electrocardiogram, a device, equipment and a storage medium. The method comprises the following steps: waveform form recognition is performed on an electrocardiogram signal in a to-be-classified electrocardiogram to obtain a waveform category of the electrocardiogram signal; a classification model corresponding to the waveform category of the electrocardiogram signal is obtained; the to-be-classified electrocardiogram is input into the classification model to perform heartbeat type classification, and a classification result is obtained. The classification model is obtained by pre-training based on sample data of an electrocardiogram of the same waveform category. In the above method, since the classification model is determined according to the waveform category of the electrocardiogram signal before the to-be-classified electrocardiogram is classified, the classification model is a classification model matched with the to-be-classified electrocardiogram, and then the to-be-classified electrocardiogram is classified by using the classification model, so that an accurate classification result can be obtained.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of medical images, in particular to a classification method and a training method for electrocardiograms, devices, equipment and storage media. BACKGROUND

[0002] With the improvement of people's living standards and the acceleration of the pace of life, the incidence of cardiovascular diseases is rapidly rising, which has become one of the main factors threatening human physical health. The electrocardiogram of the human body can objectively reflect the physiological conditions of each part of the heart, and provides an important basis for the diagnosis of heart disease and the evaluation of heart function. Therefore, the method of automatic detection, analysis and classification based on electrocardiogram has been widely studied.

[0003] At present, the classification method of electrocardiogram is generally based on a deep learning classification model, for example, a unified classification model is first trained based on a large amount of labeled data, and then in the actual electrocardiogram analysis process, the classification model is used to identify the electrocardiogram to be classified to achieve the classification purpose and obtain the classification result.

[0004] However, the above classification method has the problems of high complexity of the classification model and low accuracy of the classification result. SUMMARY

[0005] Therefore, it is necessary to provide a classification method, training method, device, equipment and storage medium for electrocardiogram which can improve the classification accuracy.

[0006] In a first aspect, the present application provides a classification method for electrocardiogram. The method comprises:

[0007] performing waveform morphology recognition on the electrocardiogram signal in the electrocardiogram to be classified to obtain the waveform category of the electrocardiogram signal;

[0008] obtaining a classification model corresponding to the waveform category of the electrocardiogram signal; the classification model is obtained by pre-training based on sample data of electrocardiogram of the same waveform category;

[0009] inputting the electrocardiogram to be classified into the classification model to classify the heart beat type and obtain the classification result.

[0010] In one embodiment, the method further comprises:

[0011] extracting a dominant signal from the electrocardiogram signal in the electrocardiogram to be classified;

[0012] performing waveform morphology recognition on the dominant signal to obtain the waveform category of the electrocardiogram signal.

[0013] In one of the embodiments, the extracting the dominant signal from the electrocardio signal in the electrocardiogram to be classified comprises:

[0014] extracting the electrocardio signal in a preset time period in the electrocardiogram to be classified as the dominant signal.

[0015] In one of the embodiments, the obtaining the classification model corresponding to the waveform category of the electrocardiogram to be classified comprises:

[0016] obtaining a mapping relationship between the preset waveform category and the classification model;

[0017] determining the classification model corresponding to the waveform category of the electrocardio signal according to the mapping relationship between the waveform category and the classification model.

[0018] In one of the embodiments, the waveform category comprises any one of a waveform from top to bottom, a waveform from bottom to top, a waveform from bottom to top first and from top to bottom second, and a waveform from top to bottom first and from bottom to top second.

[0019] In a second aspect, the application further provides a training method of a classification model, which is used to train a plurality of initial classification models to obtain the classification model of the first aspect, and the training method comprises:

[0020] obtaining a sample data set of electrocardiograms of a plurality of waveform categories;

[0021] inputting the sample data set of electrocardiograms of each of the waveform categories into a corresponding initial classification model for training to obtain a plurality of classification models corresponding to the electrocardiograms of the plurality of waveform categories.

[0022] In one of the embodiments, after obtaining the plurality of classification models corresponding to the electrocardiograms of the plurality of waveform categories, the method further comprises:

[0023] establishing a mapping relationship between the plurality of waveform categories and the plurality of classification models.

[0024] In a third aspect, the application further provides an electrocardiogram classification device, which comprises:

[0025] a recognition module, configured to perform waveform pattern recognition on an electrocardio signal in an electrocardiogram to be classified to obtain a waveform category of the electrocardio signal;

[0026] an obtaining module, configured to obtain a classification model corresponding to the waveform category of the electrocardio signal; the classification model is obtained by pre-training based on sample data of an electrocardiogram of the same waveform category;

[0027] a classification module, configured to input the electrocardiogram to be classified into the classification model for heart beat type classification to obtain a classification result.

[0028] In a fourth aspect, the present application provides a computer device. The computer device comprises a memory and a processor, the memory stores a computer program, and the processor implements the following steps when executing the computer program:

[0029] performing waveform pattern recognition on the electrocardiosignal in the electrocardiogram to be classified to obtain a waveform category of the electrocardiosignal;

[0030] obtaining a classification model corresponding to the waveform category of the electrocardiosignal; the classification model is obtained by pre-training based on sample data of electrocardiograms of the same waveform category;

[0031] inputting the electrocardiogram to be classified into the classification model to perform heartbeat type classification, and obtaining a classification result.

[0032] In a fifth aspect, the present application provides a computer readable storage medium. The computer readable storage medium stores a computer program, and the computer program is executed by a processor to implement the following steps:

[0033] performing waveform pattern recognition on the electrocardiosignal in the electrocardiogram to be classified to obtain a waveform category of the electrocardiosignal;

[0034] obtaining a classification model corresponding to the waveform category of the electrocardiosignal; the classification model is obtained by pre-training based on sample data of electrocardiograms of the same waveform category;

[0035] inputting the electrocardiogram to be classified into the classification model to perform heartbeat type classification, and obtaining a classification result.

[0036] The electrocardiogram classification method, the training method, the device, the equipment and the storage medium, by waveform morphology recognition of the electrocardiogram signal in the electrocardiogram to be classified, obtain the waveform category of the electrocardiogram signal, obtain the classification model corresponding to the waveform category of the electrocardiogram signal, input the electrocardiogram to be classified into the classification model to classify the heart beat type, and obtain the classification result. The classification model is obtained by pre-training based on the sample data of the electrocardiogram of the same waveform category. In the above method, since the classification model is determined according to the waveform category of the electrocardiogram signal before the electrocardiogram to be classified is classified, the classification model is the classification model matched with the electrocardiogram to be classified, and then the classification model is used to classify the electrocardiogram to be classified, so that an accurate classification result can be obtained. In addition, since the classification model is obtained by pre-training based on the sample data of the electrocardiogram of the same waveform category, that is, different classification models are trained for sample data of different waveform forms, and the sample data is classified by considering individual differences in advance, the complexity of the classification model trained based on the sample data of each category will be relatively low, and the dependence on the model training sample is reduced, and the practicability of the method is improved. BRIEF DESCRIPTION OF DRAWINGS

[0037] Figure 1 An application environment diagram of an electrocardiogram classification method is shown in one embodiment;

[0038] Figure 2 A flowchart of an electrocardiogram classification method is shown in one embodiment;

[0039] Figure 2A A waveform category diagram is shown in one embodiment;

[0040] Figure 2B A waveform category diagram is shown in another embodiment;

[0041] Figure 2C A waveform category diagram is shown in another embodiment;

[0042] Figure 2D A waveform category diagram is shown in another embodiment;

[0043] Figure 2E A waveform category diagram is shown in another embodiment;

[0044] Figure 3 A Figure 2 A flowchart of one implementation of S101 in an embodiment is shown;

[0045] Figure 4 A Figure 2 A flowchart of one implementation of S102 in an embodiment is shown;

[0046] Figure 4A a flowchart of a mapping table in one embodiment;

[0047] Figure 5 a flowchart of a classification method of an electrocardiogram in another embodiment;

[0048] Figure 5A a flowchart of a classification method of an electrocardiogram in another embodiment;

[0049] Figure 6 a flowchart of a training method of a classification model in one embodiment;

[0050] Figure 7 a flowchart of a training method of a classification model in one embodiment; Figure 6 a flowchart of one implementation of S501 in an embodiment;

[0051] Figure 8 a flowchart of a training method of a classification model in one embodiment;

[0052] Figure 9 a structural block diagram of a classification device of an electrocardiogram in one embodiment;

[0053] Figure 10 a structural block diagram of a classification device of an electrocardiogram in another embodiment;

[0054] Figure 11 a structural block diagram of a classification device of an electrocardiogram in another embodiment;

[0055] Figure 12 a structural block diagram of a training device of a classification model in one embodiment;

[0056] Figure 13 a structural block diagram of a training device of a classification model in another embodiment;

[0057] Figure 14 an internal structural diagram of a computer device in one embodiment. DETAILED DESCRIPTION

[0058] In order to make the purposes, technical solutions and advantages of the present application clearer, the present application is further described in detail below with reference to the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and not used to limit the present application.

[0059] The classification method of an electrocardiogram provided in the embodiments of the present application can be applied to, for example, Figure 1The illustrated electrocardio classification system includes an electrocardio tester 102 and a computer device 104. The computer device 104 can be a terminal, a server, or a workstation. The electrocardio tester 102 is used to monitor the electrocardio of a test subject and save the monitored electrocardiogram as a to-be-classified electrocardiogram, or upload the to-be-classified electrocardiogram to a cloud database; the computer device 104 can read the to-be-classified electrocardiogram from the electrocardio tester 102 and classify the to-be-classified electrocardiogram by using a corresponding classification algorithm or classification method to obtain a classification result. The computer device 104 can also classify the to-be-classified electrocardiogram obtained from the cloud database.

[0060] Those skilled in the art can understand that, Figure 1 The structure illustrated in the figure is only a block diagram of part of the structure related to the scheme of the present application, and does not constitute a limitation on the electrocardio classification system to which the scheme of the present application is applied. A specific electrocardio classification system can include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0061] In one embodiment, as Figure 2 illustrated, a classification method of electrocardiogram is provided, which is applied to the computer device in Figure 1 for example, and includes the following steps.

[0062] S101, waveform morphology recognition is performed on the electrocardio signal in the to-be-classified electrocardiogram to obtain a waveform category of the electrocardio signal.

[0063] The electrocardio signal can be referred to as a QRS complex, and the waveform morphology recognition of the electrocardio signal is also the recognition of the waveform morphology of the QRS complex. Generally, the waveform morphology recognition of the electrocardio signal is performed according to the waveform direction or offset of the QRS complex. The waveform direction or offset of the QRS complex determines the letters added to each part of the QRS complex (see the waveform category of the QRS complex illustrated in Figure 2A ). For example, in the QRS complex, if the first offset is negative, it is called Q wave. The first positive offset is called R wave (which can have Q wave or not). Any negative offset after the R wave is called S wave. If a second positive offset appears, it is called R' wave.

[0064] Based on the analysis of the QRS complex, the waveform category of the electrocardio signal can be obtained by performing waveform morphology recognition on the electrocardio signal, which includes waveform from top to bottom (see Figure 2B ), waveform from bottom to top (see Figure 2C ), waveform from bottom to top and then from top to bottom (see Figure 2D ), and waveform from top to bottom and then from bottom to top (see Figure 2EQRS wave, the wave from bottom to top is called R wave, the wave from bottom to top first and then from top to bottom is called Rs wave, and the wave from top to bottom first and then from bottom to top is called Qr wave. It can be understood that the specific wave category of the electrocardiosignal is not limited to the above four wave categories, and can also include other wave categories representing the direction and offset of the wave, such as the other wave categories shown in Figure 2A The specific wave category can be determined according to the actual wave.

[0065] In this embodiment, the computer device can be connected with the electrocardio tester and obtain the electrocardiogram from the electrocardio tester as the electrocardiogram to be classified, or directly obtain the electrocardiogram to be classified from the cloud database. When the electrocardiogram to be classified is obtained, the electrocardiosignal can be further extracted from the electrocardiogram to be classified, and the wave form of the electrocardiosignal is identified or classified by using the corresponding wave form recognition algorithm or other wave classification method, to obtain the wave category of the electrocardiosignal. For example, the computer device identifies which one of the wave from top to bottom, the wave from bottom to top, the wave from bottom to top first and then from top to bottom, the wave from top to bottom first and then from bottom to top, or other wave categories is the wave category of the electrocardiosignal. Optionally, when the computer device obtains the electrocardiogram to be classified, the computer device can first process the abnormal wave of the electrocardiogram to be classified to remove the abnormal wave, and then further identify the wave form of the processed electrocardiosignal by using the corresponding wave form recognition algorithm or other wave classification method, to obtain a more accurate wave category.

[0066] In S102, a classification model corresponding to the wave category of the electrocardiosignal is obtained.

[0067] The classification model is obtained by training the sample data of the electrocardiogram of the same wave category in advance. The classification model is used to identify the beat type of the electrocardiosignal in the electrocardiogram.

[0068] In this embodiment, when the computer device obtains the wave category of the electrocardiosignal based on the foregoing steps, the classification model corresponding to the wave category of the electrocardiosignal can be selected from the plurality of classification models trained in advance, so as to use the classification model to identify the beat type of the electrocardiosignal. It should be noted that different classification models correspond to different wave categories, for example, Figures 2B-2E The four wave categories shown in FIG. 4 correspond to four different classification models respectively, and each classification model can be used to identify the beat type of the electrocardiosignal.

[0069] In S103, the electrocardiogram to be classified is input into the classification model to classify the beat type, to obtain a classification result.

[0070] Generally, according to the American Heart Association (AAMI), electrocardiogram heartbeats are divided into five categories, namely, normal or bundle branch block rhythm (N), supraventricular abnormal rhythm (S), ventricular abnormal rhythm (V), fusion rhythm (F), and unclassified rhythm (Q). The five categories can further include subcategories, as shown in Table 1:

[0071] Table 1

[0072]

[0073] Based on the above analysis, the classification result in this embodiment can include the classification result of the above five categories, or the classification result of the fifteen subcategories in Table 1.

[0074] In this embodiment, the computer device can directly input the to-be-classified electrocardiogram into the classification model matched with the to-be-classified electrocardiogram to classify the heartbeat type, so as to obtain the classification result of the to-be-classified electrocardiogram. For example, through the above steps, it can be determined that the type of the heartbeat of the to-be-classified electrocardiogram is one of N, S, V, F, and Q, or one of the subcategories in Table 1.

[0075] The above-mentioned electrocardiogram classification method classifies the heartbeat type by performing waveform morphology recognition on the electrocardiogram signal in the to-be-classified electrocardiogram, obtaining the waveform category of the electrocardiogram signal, obtaining the classification model corresponding to the waveform category of the electrocardiogram signal, inputting the to-be-classified electrocardiogram into the classification model to classify the heartbeat type, and obtaining the classification result. The classification model is obtained by pre-training based on sample data of electrocardiogram of the same waveform category. In the above method, since the classification model is determined according to the waveform category of the electrocardiogram signal before the to-be-classified electrocardiogram is classified, the classification model is the classification model matched with the to-be-classified electrocardiogram, and then the to-be-classified electrocardiogram is classified using the classification model, so that an accurate classification result can be obtained. In addition, since the classification model is obtained by pre-training based on sample data of electrocardiogram of the same waveform category, that is, different classification models are obtained by training sample data of different waveform morphologies, and the sample data is classified by considering individual differences in advance, the complexity of the classification model obtained based on sample data of each category will be relatively low, and the dependence on model training samples is also reduced, thereby improving the practicability of the method.

[0076] In one embodiment, a specific implementation of waveform morphology recognition of an electrocardiogram signal is provided, as shown in Figure 3 The above S101 “performing waveform morphology recognition on the electrocardiogram signal in the to-be-classified electrocardiogram to obtain the waveform category of the electrocardiogram signal” includes:

[0077] S201, extract a dominant signal from the electrocardiosignal in the electrocardiogram to be classified.

[0078] The waveform category of the dominant signal can represent the waveform category of the electrocardiosignal in the electrocardiogram to be classified.

[0079] In this embodiment, when the computer device obtains the electrocardiogram to be classified, the computer device can further extract the electrocardiosignal from the electrocardiogram to be classified, and then extract the dominant signal from the electrocardiosignal according to the predefined dominant signal. Optionally, the computer device can extract the electrocardiosignal in a preset time period in the electrocardiogram to be classified as the dominant signal. The preset time period can be determined according to actual application requirements. For example, if the preset time period is 30 seconds, the computer device can extract the electrocardiosignal in the first 30 seconds of the electrocardiogram to be classified as the dominant signal. For another example, if the preset time period is 20 seconds, the computer device can extract the electrocardiosignal from 10 seconds to 30 seconds of the electrocardiogram to be classified as the dominant signal. It should be noted that the dominant signal can be the signal in any preset time period of the electrocardiosignal, such as the signal in the middle preset time period, the signal in the beginning preset time period, or the signal in the end preset time period.

[0080] S202, performing waveform morphology recognition on the dominant signal to obtain the waveform category of the electrocardiosignal.

[0081] In this embodiment, after the computer device determines the dominant signal of the electrocardiosignal, the computer device can use a corresponding waveform morphology recognition algorithm or other waveform classification method to recognize the waveform morphology of the dominant signal, and obtain the waveform category of the dominant signal. For example, the computer device can recognize which one of the following waveform categories the dominant signal belongs to: waveform from top to bottom, waveform from bottom to top, waveform from bottom to top first and from top to bottom second, waveform from top to bottom first and from bottom to top second, or other waveform categories. Optionally, when the computer device determines the dominant signal, the computer device can first perform abnormal waveform processing on the dominant signal to remove abnormal waveforms, and then further use a corresponding waveform morphology recognition algorithm or other waveform classification method to recognize the waveform morphology of the processed dominant signal, and obtain a more accurate waveform category. Finally, the computer device can determine the waveform category of the dominant waveform as the waveform category of the electrocardiosignal in the electrocardiogram to be classified. Since the dominant waveform is only the signal in the preset time period of the electrocardiosignal in the electrocardiogram to be classified, the data amount is small, and therefore, the recognition efficiency can be improved, and the efficiency of the classification method provided in this embodiment can be improved.

[0082] In one embodiment, a specific implementation of an acquisition classification model is provided, such as Figure 4As shown, the S102 "obtains a classification model corresponding to the waveform category of the electrocardiosignal" includes:

[0083] S301, obtaining a mapping relationship between a preset waveform category and a classification model.

[0084] The waveform category and the classification model correspond to each other. The mapping relationship between the waveform category and the classification model can be established by the computer device in advance and recorded in a mapping table for saving, for example, referring to the mapping table shown in Figure 4A .

[0085] In this embodiment, the computer device can pre-train different classification models based on sample data of different waveform categories, obtain a plurality of waveform categories and a plurality of corresponding classification models, then construct a mapping relationship between the plurality of waveform categories and the plurality of classification models, and save the mapping relationship in the database of the computer device or other cloud databases, so as to be directly called for use when needed later. Alternatively, the computer device writes the mapping relationship between the plurality of waveform categories and the plurality of classification models into a data structure, so that when the computer device needs to determine the classification model corresponding to the waveform category, the data structure can be called to achieve.

[0086] S302, determining the classification model corresponding to the waveform category of the electrocardiosignal according to the mapping relationship between the waveform category and the classification model.

[0087] In this embodiment, when the computer device determines the classification model to be used, a table or file recording the mapping relationship between the waveform category and the classification model can be found in the database or other places, then the classification model corresponding to the waveform category to be found can be determined from the table or file according to the waveform category to be found, so as to be used later.

[0088] In summary of all the above embodiments, a classification method of electrocardiogram is provided, as shown in Figure 5 , the method includes:

[0089] S401, obtaining an electrocardiogram to be classified.

[0090] S402, extracting an electrocardiosignal in a preset time period in the electrocardiogram to be classified as a dominant signal.

[0091] S403, performing waveform pattern recognition on the dominant signal to obtain a waveform category of the electrocardiosignal.

[0092] S404, obtaining a mapping relationship between a preset waveform category and a classification model.

[0093] S405, determining the classification model corresponding to the waveform category of the electrocardiosignal according to the mapping relationship between the waveform category and the classification model.

[0094] S406, inputting the electrocardiogram to be classified into the classification model for heartbeat type classification to obtain a classification result.

[0095] The classification method corresponding to the above steps can be seen from Figure 5A the block diagram, and the steps in the above steps or block diagram are described in the foregoing embodiments. For details, please refer to the foregoing description, which will not be described here.

[0096] In actual application, the above Figures 2-5 The classification model used in any embodiment is obtained by pre-training the computer device, so the present application provides a method for training a classification model, as shown in Figure 6 The training method comprises the following steps:

[0097] S501, obtaining a sample data set of electrocardiograms of multiple waveform categories.

[0098] When training the classification model, a large amount of sample data of electrocardiograms needs to be obtained first, and then the waveform form of the electrocardiogram signal of each electrocardiogram is analyzed to determine the waveform category of each electrocardiogram signal, and a large number of electrocardiograms of different waveform categories are selected as sample data sets. For example, when the waveform categories include eleven waveform categories as shown in Figure 2A , the sample data set of electrocardiograms of the eleven waveform categories can be obtained, and the number of samples in the sample data set of each category can be determined according to actual training requirements; when the waveform categories include four waveform categories as shown in Figures 2B-2E , the sample data set of electrocardiograms of the four waveform categories can be obtained, and the number of samples in the sample data set of each waveform category can be determined according to actual training requirements.

[0099] S502, inputting the sample data of electrocardiograms of each waveform category into the corresponding initial classification model for training to obtain multiple classification models corresponding to electrocardiograms of multiple waveform categories.

[0100] The initial classification model can be a neural network or a deep learning network. The specific category of the initial classification model can be determined according to actual application. For example, the initial classification model in the present embodiment can be a CNNC network model, which has high sensitivity in waveform category recognition. It should be noted that the initial classification models corresponding to different waveform categories can be designed based on different neural network model structures, for example, referring to Figures 2-2E different initial classification models corresponding to different architectures. Optionally, similar waveform categories can also correspond to the same initial classification model.

[0101] In this embodiment, when different waveform categories correspond to different initial classification models need to be set, the computer device can construct an initial classification model corresponding to the number of waveform categories contained in the sample data, and the architecture of each initial classification model is not the same; when the waveform category and the initial classification model are not in a one-to-one correspondence, the computer device can construct the initial classification model it needs according to the waveform category, and when similar waveform categories correspond to the same initial classification model, an initial classification model can be reused without repeated construction. Then during training, the computer device can input the sample data of electrocardiogram of different waveform categories into the corresponding initial classification model for training, and the parameters of each initial classification model are adjusted during the training process, and the initial evaluation results such as sensitivity, specificity, accuracy, f1, etc. are given by the initial classification model for the current waveform, and then the condition for determining whether the initial classification model is trained is completed can be the comprehensive index f1, if each index is improved or the comprehensive index f1 is improved, it is considered that the classification model of the current waveform can achieve the expected effect, the training is completed, and the obtained multiple classification models are the multiple classification models trained. It should be noted that the initial classification model corresponding to different waveform categories can be constructed in advance according to different methods, for example, the initial classification model corresponding to the waveform category 1 can be constructed in advance according to the neural network model, and the initial classification model corresponding to the waveform category 2 can be constructed in advance according to the deep learning network, and the method of constructing the initial model can be determined according to the corresponding waveform type, for example, a simple initial classification model can be constructed for a simple waveform category, and a more complex classification model can be constructed for a complex waveform category, so as to design the initial classification model specifically. For example, the direction or offset of the QRS complex wave determines the naming of the wave. In the QRS complex, if the first offset is negative, it is called Q wave. The first positive offset is called R wave (which can have Q wave or no Q wave). Any negative offset after the R wave is called S wave. If a second positive offset appears, it is called R' wave, etc. Therefore, according to the particularity of the waveform, the corresponding initial network model is trained, so that different waveform categories correspond to different initial network models. The classification model obtained after training can be more targeted for classification, which can improve the classification efficiency while ensuring the accuracy of classification. It can be understood that during the training process, different initial network models are trained for different waveform categories, and each initial network model is an independent model that can be trained with its own parameters. Even if different waveform categories correspond to the same initial network model, different training end conditions, training times, iteration times, training samples or other conditions, etc. can be set during the training process to achieve that the trained classification model can present good classification effect for different waveform categories.

[0102] In one embodiment, a specific implementation method of obtaining the sample data set of electrocardiograms of multiple waveform categories is provided, as shown in Figure 7 The method comprises the following steps:

[0103] S601, obtaining a preset number of electrocardiograms.

[0104] S602, analyzing the waveform form of each electrocardiogram to obtain the waveform category of each electrocardiogram.

[0105] S603, classifying the preset number of electrocardiograms according to the waveform category of each electrocardiogram to obtain a sample data set of electrocardiograms of multiple waveform categories.

[0106] Specifically, the computer device can obtain a large number of electrocardiograms from an electrocardiogram tester or a database, and analyze the waveform category of the electrocardiogram signal in each electrocardiogram, or analyze the waveform category of the dominant signal in each electrocardiogram, classify the electrocardiograms belonging to the same waveform category into a sample data set, and finally obtain the sample data set of electrocardiograms of different waveform categories, so as to input the sample data set of electrocardiograms of different waveform categories into different initial classification models for training, so as to obtain the classification model corresponding to different waveform categories. The above process can be referred to the block diagram shown in Figure 8

[0107] The training method provided in the above embodiment realizes training of different classification models for different waveform categories, reduces the complexity of each classification model, and improves the classification accuracy. In addition, in the above training method, the waveform category of the electrocardiogram is analyzed in advance, and then the sample data set is classified according to the analysis result, which realizes the pre-classification of the sample data set, and finally different classification models are trained based on the classified sample data set. Such training method can reduce the dependence of the classification model on the input data in the later use process.

[0108] When the computer device completes the training and obtains multiple classification models, the mapping relationship between the multiple waveform categories and the multiple classification models can be established and saved, so that when the classification model is used, the mapping relationship is directly called to match the corresponding classification model for classification of the electrocardiogram to be classified, thereby realizing the waveform category targeted classification of the electrocardiogram, and greatly improving the classification accuracy.

[0109] ​It should be understood that although each step in the flowchart involved in each embodiment as described above is shown in sequence according to the direction of the arrow, these steps are not necessarily executed in the order indicated by the arrow. Unless otherwise specified herein, there is no strict order limitation for the execution of these steps, and these steps can be executed in other orders. Moreover, at least part of the steps in the flowchart involved in each embodiment as described above can include multiple steps or multiple stages, which are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily sequential, but can be alternately or alternately executed with at least part of other steps or steps or stages in other steps.

[0110] Based on the same inventive concept, the embodiments of the present application also provide an electrocardiogram classification device for implementing the above-mentioned electrocardiogram classification method. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme described in the above method, so the specific limitations in one or more electrocardiogram classification device embodiments provided below can refer to the limitations of the electrocardiogram classification method in the above text, and will not be repeated here.

[0111] In one embodiment, as shown in Figure 9 An electrocardiogram classification device is provided, comprising:

[0112] The identification module 10 is configured to perform waveform morphology identification on the electrocardiogram signal in the electrocardiogram to be classified to obtain a waveform category of the electrocardiogram signal.

[0113] The acquisition module 11 is configured to acquire a classification model corresponding to the waveform category of the electrocardiogram signal; the classification model is obtained by pre-training based on sample data of an electrocardiogram of the same waveform category.

[0114] The classification module 12 is configured to input the electrocardiogram to be classified into the classification model to perform heartbeat type classification to obtain a classification result.

[0115] In one embodiment, as shown in Figure 10 The above-mentioned identification module 10 comprises:

[0116] The extraction unit 101 is configured to extract a dominant signal from the electrocardiogram signal in the electrocardiogram to be classified.

[0117] The identification unit 102 is configured to perform waveform morphology identification on the dominant signal to obtain a waveform category of the electrocardiogram signal.

[0118] In one embodiment, the above-mentioned extraction unit 101 is specifically configured to extract the electrocardiogram signal in a preset time period in the electrocardiogram to be classified as the dominant signal.

[0119] In one embodiment, as shown in Figure 11 The acquisition module 11 comprises:

[0120] The acquisition unit 111 is configured to acquire a mapping relationship between preset waveform categories and classification models.

[0121] The determination unit 112 is configured to determine a classification model corresponding to a waveform category of the electrocardiogram signal according to the mapping relationship between the waveform categories and the classification models.

[0122] In one embodiment, the waveform categories comprise any one of a waveform from top to bottom, a waveform from bottom to top, a waveform from bottom to top first and from top to bottom second, and a waveform from top to bottom first and from bottom to top second.

[0123] In one embodiment, as shown in Figure 12 The training device of the classification model comprises:

[0124] The acquisition module 20 is configured to acquire sample data sets of electrocardiograms of multiple waveform categories.

[0125] The training module 21 is configured to input the sample data sets of electrocardiograms of each waveform category into a corresponding initial classification model for training, to obtain multiple classification models corresponding to electrocardiograms of the multiple waveform categories.

[0126] In one embodiment, as shown in Figure 13 The training device of the classification model further comprises:

[0127] The construction module 21 is configured to establish a mapping relationship between the multiple waveform categories and the multiple classification models.

[0128] Each module in the electrocardiogram classification device can be realized by software, hardware, and a combination thereof, in whole or in part. Each module can be embedded in or independent of a processor in a computer device in a hardware form, or can be stored in a memory in a computer device in a software form, so as to be called and executed by a processor to perform operations corresponding to each module.

[0129] In one embodiment, a computer device is provided, which can be a terminal. An internal structure diagram of the computer device can be as shown in Figure 14As shown in the figure. The computer device includes a processor, a memory, a communication interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used for wired or wireless communication with external terminals. Wireless mode can be achieved through WIFI, mobile cellular network, NFC (near field communication) or other technologies. The computer program is executed by the processor to implement an electrocardiogram classification method. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen. The input device of the computer device can be a touch layer overlaid on the display screen, or a key, trackball or touchpad arranged on the shell of the computer device, or an external keyboard, touchpad or mouse, etc.

[0130] Those skilled in the art can understand that, Figure 14 The structure shown in the figure is only a block diagram of part of the structure related to the scheme of the present application, and does not constitute a limitation on the computer device to which the scheme of the present application is applied. The specific computer device can include more or fewer components than those shown in the figure, or combine certain components, or have a different component arrangement.

[0131] In one embodiment, a computer device is provided, comprising a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the following steps:

[0132] Waveform morphology recognition is performed on the electrocardio signal in the electrocardiogram to be classified to obtain the waveform category of the electrocardio signal;

[0133] A classification model corresponding to the waveform category of the electrocardio signal is obtained; the classification model is obtained by pre-training based on sample data of electrocardiogram of the same waveform category;

[0134] The electrocardiogram to be classified is input into the classification model for heart beat type classification to obtain a classification result.

[0135] In one embodiment, a computer readable storage medium is provided, which stores a computer program, and the computer program is executed by a processor to implement the following steps:

[0136] Waveform morphology recognition is performed on the electrocardio signal in the electrocardiogram to be classified to obtain the waveform category of the electrocardio signal;

[0137] obtaining a classification model corresponding to the waveform category of the electrocardiosignal; the classification model is obtained by training sample data of electrocardiograms of the same waveform category in advance;

[0138] inputting the electrocardiogram to be classified into the classification model to perform heart beat type classification, and obtaining a classification result.

[0139] In one embodiment, a computer program product is provided, comprising a computer program which, when executed by a processor, implements the following steps:

[0140] performing waveform pattern recognition on the electrocardiosignal in the electrocardiogram to be classified to obtain a waveform category of the electrocardiosignal;

[0141] obtaining a classification model corresponding to the waveform category of the electrocardiosignal; the classification model is obtained by training sample data of electrocardiograms of the same waveform category in advance;

[0142] inputting the electrocardiogram to be classified into the classification model to perform heart beat type classification, and obtaining a classification result.

[0143] It should be noted that the user information (including but not limited to user equipment information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data authorized by the user or authorized by all parties.

[0144] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer readable storage medium, and when the computer program is executed, the processes of the above-mentioned embodiments of the methods can be included. Any reference to memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (Read-Only Memory, ROM), magnetic tape, floppy disk, flash memory, optical storage, high-density embedded non-volatile memory, resistive memory (ReRAM), magnetoresistive random access memory (Magnetoresistive Random Access Memory, MRAM), ferroelectric memory (Ferroelectric Random Access Memory, FRAM), phase change memory (Phase Change Memory, PCM), graphene memory, etc. Volatile memory can include random access memory (Random Access Memory, RAM) or external cache memory, etc. As an illustration but not limitation, RAM can be in various forms, such as static random access memory (Static Random Access Memory, SRAM) or dynamic random access memory (Dynamic Random Access Memory, DRAM), etc. The database involved in the embodiments provided in the present application can include at least one of a relational database and a non-relational database. The non-relational database can include a distributed database based on a block chain, etc., without being limited thereto. The processor involved in the embodiments provided in the present application can be a general-purpose processor, a central processing unit, a graphics processing unit, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., without being limited thereto.

[0145] Any combination of the technical features of the above embodiments can be made. In order to make the description simple, all possible combinations of the technical features in the above embodiments are not described, however, as long as the combination of the technical features does not exist contradictory, it should be considered as the scope of the present application.

[0146] The above embodiments only express several implementation manners of the present application, and the description is more specific and detailed, but it should not be understood as a limitation on the scope of the patent of the present application. It should be pointed out that for ordinary skilled in the art, without departing from the concept of the present application, a number of modifications and improvements can be made, which are within the scope of protection of the present application. Therefore, the protection scope of the present application should be subject to the appended claims.

Claims

1. A method of classifying electrocardiograms, characterized by, The method comprises: waveform direction and offset of the electrocardiosignal in the electrocardiogram to be classified, to obtain a waveform category of the electrocardiosignal; the waveform category comprises any one of waveform from top to bottom, waveform from bottom to top, waveform from bottom to top first and then from top to bottom, and waveform from top to bottom first and then from bottom to top; obtaining a classification model corresponding to the waveform category of the electrocardiosignal; the classification model is obtained by pre-training sample data of electrocardiograms of the same waveform category; inputting the electrocardiogram to be classified into the classification model to classify the heart beat type, to obtain a classification result.

2. The method of claim 1, wherein, The waveform direction and offset of the electrocardiosignal in the electrocardiogram to be classified are used for waveform morphology recognition of the electrocardiosignal, to obtain a waveform category of the electrocardiosignal, comprising: extracting a dominant signal from the electrocardiosignal in the electrocardiogram to be classified; performing waveform morphology recognition on the dominant signal to obtain the waveform category of the electrocardiosignal.

3. The method of claim 2, wherein, The dominant signal is extracted from the electrocardiosignal in the electrocardiogram to be classified, comprising: extracting the electrocardiosignal in a preset time period in the electrocardiogram to be classified as the dominant signal.

4. The method according to claim 1 or 2, characterized in that, The classification model corresponding to the waveform category of the electrocardiosignal is obtained, comprising: obtaining a mapping relationship between preset waveform categories and classification models; determining the classification model corresponding to the waveform category of the electrocardiosignal according to the mapping relationship between the waveform categories and the classification models.

5. The method of claim 1, wherein, The method further comprises: performing abnormal waveform processing on the electrocardiogram to be classified; The waveform direction and offset of the electrocardiosignal in the electrocardiogram to be classified are used for waveform morphology recognition of the electrocardiosignal, to obtain a waveform category of the electrocardiosignal, comprising: performing waveform morphology recognition on the waveform direction and offset of the electrocardiosignal in the electrocardiogram after the abnormal waveform processing, to obtain the waveform category of the electrocardiosignal. 6.A method for training a classification model, the method comprising: The training method is used for training a plurality of initial classification models to obtain the classification model in claim 1, and the training method comprises: obtaining a sample data set of electrocardiograms of a plurality of waveform categories; the waveform categories comprise any one of waveform from top to bottom, waveform from bottom to top, waveform from bottom to top first and then from top to bottom, and waveform from top to bottom first and then from bottom to top; inputting the sample data set of electrocardiograms of each waveform category into a corresponding initial classification model for training, to obtain a plurality of classification models corresponding to the electrocardiograms of the plurality of waveform categories.

7. The method of claim 6, wherein, After obtaining the plurality of classification models corresponding to the electrocardiograms of the plurality of waveform categories, the method further comprises: establishing a mapping relationship between the plurality of waveform categories and the plurality of classification models.

8. An electrocardiogram classification apparatus characterized by comprising: The device comprises: an identification module configured to perform waveform morphology recognition on the electrocardiosignal in the electrocardiogram to be classified according to the waveform direction and offset of the electrocardiosignal, to obtain a waveform category of the electrocardiosignal; the waveform category comprises any one of waveform from top to bottom, waveform from bottom to top, waveform from bottom to top first and then from top to bottom, and waveform from top to bottom first and then from bottom to top. An acquisition module is configured to acquire a classification model corresponding to a waveform category of the electrocardiosignal, wherein the classification model is obtained by pre-training based on sample data of electrocardiograms of the same waveform category; A classification module is configured to input the electrocardiogram to be classified into the classification model to perform heartbeat type classification, and obtain a classification result. 9.A computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the computer device is configured to perform the method according to any one of claims 1-8 when the computer program is executed by the processor. The processor executes the computer program to implement the steps of the method in any one of claims 1 to 7.

10. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the steps of the method in any one of claims 1 to 7.

Images