Processing method, device and processing equipment of QRS wave group anomaly classification model

Abstract

The invention provides a QRS wave group anomaly classification model processing method, device and processing equipment, a QRS wave group anomaly classification mechanism based on 4-lead or 8-lead is introduced, the QRS wave group anomaly classification model constructed in this way is proved through experiments, the detection precision of different types of abnormal QRS wave groups in an 8-lead format is remarkably improved, and the QRS wave group anomaly classification model can be applied to the QRS wave group anomaly classification model. In addition, due to simplification of the data, especially simplification in a 4-lead format, the aspects of application cost such as data processing time, data application time and data storage time are also obviously improved, and meanwhile, the workload of personnel in the aspect of clinical work is also obviously reduced, so that the method has better application value.

Description

Technical Field

[0001] The present application relates to the field of medical technology, and in particular to a processing method, device and processing equipment for a QRS complex abnormal classification model. Background Art

[0002] The QRS complex in the electrocardiogram (ECG) signal represents the depolarization of the ventricles and is a key component of the heart's electrical activity. Specifically, the QRS complex reflects changes in the depolarization potential and time of the left and right ventricles. The first downward wave is the Q wave, the upward wave is the R wave, and the next downward wave is the S wave. It is closely related to the occurrence of heart failure, sudden death, and structural heart disease. Therefore, correct detection of ECG signals with abnormal QRS complex morphology is crucial for appropriate treatment, while incorrect interpretation may lead to adverse results.

[0003] Currently, ECG signal detection based on artificial intelligence (AI) has been widely studied, mainly focusing on electrocardiograms with abnormal QRS complex morphology. However, the inventors of the present application found that there is basically no research focusing on the refined detection of a wide range of QRS complex abnormalities.

[0004] In addition, the diagnostic guidelines for ECG abnormalities rely on 12-lead ECG, and the relevant algorithms are also developed in the form of 12-lead ECG. Although this provides comprehensive ECG change information, it also significantly increases the time for data processing, data application, and data storage, resulting in higher application costs.

[0005] That is to say, the existing detection schemes for ECG signals with abnormal QRS complex morphology have the problems of limited precision and high application cost. Summary of the invention

[0006] The present application provides a processing method, an apparatus and a processing device for a QRS complex abnormality classification model. By introducing a QRS complex abnormality classification mechanism based on 4 leads or 8 leads, the QRS complex abnormality classification model constructed in this way has been proved through experiments that the detection accuracy of different types of abnormal QRS complexes in the 8-lead format has been significantly improved. In addition, due to the simplification of data, especially the simplification in the 4-lead format, several aspects of application costs such as data processing time, data application time and data storage time have also been significantly improved. At the same time, the workload of clinical work personnel has been significantly reduced, so it has better application value.

[0007] In a first aspect, the present application provides a method for processing a QRS complex abnormal classification model, the method comprising:

[0008] Acquire sample ECG signals of different types of QRS complexes, wherein the sample ECG signals include ECG signals of normal QRS complexes and ECG signals of abnormal QRS complexes, and the sample ECG signals are specifically ECG signals of a preset lead format, and the preset lead format is specifically a 4-lead format or an 8-lead format;

[0009] Configure corresponding annotations for the sample ECG signal to mark the corresponding QRS complex type;

[0010] Based on the labeled sample ECG signal, a QRS complex abnormality classification model is trained, wherein the QRS complex abnormality classification model is a deep learning model architecture, and the QRS complex abnormality classification model is used to carry out corresponding QRS complex abnormality classification processing based on the ECG signal input by the model in a preset lead format.

[0011] In a second aspect, the present application provides a processing device for a QRS complex abnormal classification model, the device comprising:

[0012] An acquisition unit, used to acquire sample ECG signals of different types of QRS complexes, wherein the sample ECG signals include ECG signals of normal QRS complexes and ECG signals of abnormal QRS complexes, and the sample ECG signals are specifically ECG signals of a preset lead format, and the preset lead format is specifically a 4-lead format or an 8-lead format;

[0013] A labeling unit, used for configuring corresponding labels for the sample ECG signal to label the corresponding QRS wave group type;

[0014] A training unit is used to train a QRS complex abnormality classification model based on the labeled sample ECG signal, wherein the QRS complex abnormality classification model is a deep learning model architecture, and the QRS complex abnormality classification model is used to carry out corresponding QRS complex abnormality classification processing based on the ECG signal input by the model in a preset lead format.

[0015] In a third aspect, the present application provides a processing device, including a processor and a memory, wherein a computer program is stored in the memory, and when the processor calls the computer program in the memory, the method provided in the first aspect of the present application or any possible implementation method of the first aspect of the present application is executed.

[0016] In a fourth aspect, the present application provides a computer-readable storage medium, which stores multiple instructions, and the instructions are suitable for a processor to load to execute the method provided by the first aspect of the present application or any possible implementation of the first aspect of the present application.

[0017] From the above content, it can be concluded that the present application has the following beneficial effects:

[0018] Aiming at the detection target of ECG signals with abnormal QRS complex morphology, the present application introduces a QRS wave group abnormality classification mechanism based on 4 leads or 8 leads. The QRS wave group abnormality classification model constructed in this way has been proved through experiments that the detection accuracy of different types of abnormal QRS wave groups in the 8-lead format has been significantly improved. In addition, due to the simplification of data, especially the simplification in the 4-lead format, several aspects of application costs such as data processing time, data application time and data storage time have also been significantly improved. At the same time, it also significantly reduces the workload of clinical personnel, and therefore has a better application value. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For those skilled in the art, other drawings can be obtained based on these drawings without creative work.

[0020] Figure 1 A flowchart of a method for processing a QRS complex abnormal classification model of the present application;

[0021] Figure 2 A schematic diagram of a scenario of the model architecture involved in this application;

[0022] Figure 3 This is a schematic diagram of an example of the ROC curve evaluation indicator of this application;

[0023] Figure 4 This is a schematic diagram of an example of the confusion matrix evaluation indicator of this application;

[0024] Figure 5 A structural schematic diagram of a processing device for the abnormal QRS complex classification model of the present application;

[0025] Figure 6 A schematic diagram of the structure of the processing equipment of this application. DETAILED DESCRIPTION

[0026] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of this application.

[0027] The terms "first", "second", etc. in the specification and claims of this application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments described here can be implemented in an order other than that illustrated or described here. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or modules is not necessarily limited to those steps or modules clearly listed, but may include other steps or modules that are not clearly listed or inherent to these processes, methods, products or devices. The naming or numbering of steps in this application does not mean that the steps in the method flow must be executed in the time / logical sequence indicated by the naming or numbering. The process steps that have been named or numbered can change the execution order according to the technical purpose to be achieved, as long as the same or similar technical effects can be achieved.

[0028] The division of modules in this application is a logical division. There may be other division methods when it is implemented in actual applications. For example, multiple modules can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, and the indirect coupling or communication connection between modules can be electrical or other similar forms, which are not limited in this application. In addition, the modules or submodules described as separate components may or may not be physically separated, may or may not be physical modules, or may be distributed in multiple circuit modules, and some or all of the modules may be selected according to actual needs to achieve the purpose of the present application.

[0029] Before introducing the processing method of the QRS complex abnormal classification model provided by the present application, the background content involved in the present application is first introduced.

[0030] The processing method, device and computer-readable storage medium of the QRS complex abnormal classification model provided in the present application can be applied to processing equipment. By introducing a QRS complex abnormal classification mechanism based on 4 leads or 8 leads, the QRS complex abnormal classification model constructed in this way has been proved through experiments that the detection accuracy of different types of abnormal QRS complexes in the 8-lead format has been significantly improved. In addition, due to the simplification of data, especially the simplification in the 4-lead format, several aspects of application costs such as data processing time, data application time and data storage time have also been significantly improved. At the same time, the workload of clinical work personnel has been significantly reduced, so it has better application value.

[0031] The processing method of the QRS wave group abnormal classification model mentioned in the present application, its execution subject can be a processing device of the QRS wave group abnormal classification model, or a server, a physical host or a user equipment (User Equipment, UE) and other different types of processing devices integrated with the processing device of the QRS wave group abnormal classification model. Among them, the processing device of the QRS wave group abnormal classification model can be implemented in hardware or software, and the UE can be a terminal device such as a smart phone, a tablet computer, a laptop computer, a desktop computer or a personal digital assistant (Personal Digital Assistant, PDA), and the processing device can be set in the form of a device cluster.

[0032] It can be understood that the processing device that executes the processing method of the QRS wave group abnormal classification model of the present application, or the processing device equipped with the corresponding application service for processing the QRS wave group abnormal classification model of the present application, considers that the focus of the present application scheme is on the configuration or training of the model, and therefore only needs to have the corresponding data processing capabilities.

[0033] If the processing equipment needs to continue to involve the actual application of the model, it needs to be combined with clinical applications. It can be deployed on equipment that is specifically responsible for diagnostic work based on ECG signals or QRS complex abnormality analysis, or it can be deployed on equipment that is also responsible for other tasks (such as other clinical diagnoses or even routine office work).

[0034] Of course, the processing equipment can also be split into a first part responsible for model training and a second part responsible for model application according to the model training and application stages. For example, the backend server performs model training and then sends the model to the frontend (doctor side) through remote services for specific application. This is also possible.

[0035] It can be seen that the specific device type and device deployment form of the processing device can be flexibly adjusted according to actual needs, and this application does not make specific limitations.

[0036] Next, the processing method of the QRS complex abnormal classification model provided by the present application is introduced.

[0037] First, see Figure 1 , Figure 1 A flow chart of a processing method of the abnormal QRS wave group classification model of the present application is shown. The processing method of the abnormal QRS wave group classification model provided by the present application may specifically include the following steps S101 to S103:

[0038] Step S101, obtaining sample ECG signals of different types of QRS complexes, wherein the sample ECG signals include ECG signals of normal QRS complexes and ECG signals of abnormal QRS complexes, and the sample ECG signals are specifically ECG signals of a preset lead format, and the preset lead format is specifically a 4-lead format or an 8-lead format;

[0039] It is understandable that when the model needs to be configured / trained, the corresponding training samples need to be configured. Corresponding to the QRS complex abnormality classification model to be constructed in this application, the training samples are specifically sample ECG signals (or sample ECG signal data).

[0040] The sample ECG signal includes two types of ECG signals, namely, ECG signals of normal QRS complexes and ECG signals of abnormal QRS complexes, and the ECG signals of abnormal QRS complexes can be further divided into different types of abnormal QRS complexes for further classification.

[0041] The sample ECG signal can be a historical ECG signal collected clinically (), or an ECG signal obtained by secondary processing based on the historical ECG signal collected clinically, or an ECG signal generated manually or automatically under a preset generation strategy. These are all possible in actual situations to meet application requirements in different situations.

[0042] In addition, if the sample ECG signal involves data of real patients, it is understandable that, considering the privacy of patients, the setting of anonymized data can be adopted in the specific operation to retain the core data content of the signal content.

[0043] It is worth noting that in the existing solutions, the diagnostic guidelines for ECG abnormalities rely on 12-lead ECG, and the relevant algorithms are also developed in the form of 12-lead ECG, while this application focuses on 4-lead and 8-lead, corresponding to sample ECG signals in 4-lead format or 8-lead format.

[0044] Under the overall detection goal of different types of abnormal ECG signals including abnormal QRS complexes, although the 12 leads used in the existing scheme seem to be able to provide comprehensive ECG change information, the present application finds that the detection of ECG signals with abnormal QRS complexes does not bring obvious contribution, and even increases the operation time and workload of clinical staff. As a result, on the other hand, it increases the time for data processing, data application and data storage, resulting in higher application costs, which in actual situations affects the application value of the detection scheme for ECG signals with abnormal QRS complexes.

[0045] As an exemplary embodiment, the leads involved in the 4-lead format of the present application may specifically include:

[0046] Lead I, Lead II, Lead V1, and Lead V5.

[0047] On the other hand, as an exemplary embodiment, the leads involved in the 8-lead format of the present application may specifically include:

[0048] Lead I, Lead II, Lead V1, Lead V2, Lead V3, Lead V4, Lead V5 and Lead V6.

[0049] It can be understood that these leads themselves are included in the category of 12 leads, so the 4 leads or 8 leads involved in this application can be understood as being simplified from the existing 12 leads. Considering that each lead itself is not the focus of improvement of this application and belongs to the prior art, the embodiment here focuses on providing a specific application solution that can involve 4 lead types or 8 lead types, and no specific description will be given.

[0050] In addition, as for the signal characteristics of the ECG signal itself, as an exemplary embodiment, the sample ECG signal here and the ECG signal input to the subsequent model have a signal length of 10 seconds and a sampling frequency of 500 Hz.

[0051] It can be understood that the embodiment here is a specific signal format of the ECG signal involved in the solution of the present application, and provides a specific implementation solution.

[0052] Step S102, configuring corresponding annotations for the sample ECG signal to mark the corresponding QRS complex type;

[0053] It can be understood that after obtaining the initial training samples, labeling is also involved. The purpose of labeling the training samples is to provide the model with predicted true values, so as to provide a reference for the theoretical predicted values ​​in the subsequent specific model training process, and to guide the model to optimize the model parameters based on the predicted values ​​combined with the loss function. Considering that this belongs to the scope of conventional model training, no specific explanation will be given here.

[0054] In specific operations, the annotation process can be carried out manually, or based on the configuration of automatic annotation logic, the annotation process can be carried out using relevant automatic annotation tools. Both are acceptable.

[0055] Among them, for the annotation results or labels, the content involves the QRS complex type, including the normal QRS complex type and different abnormal QRS complex types.

[0056] As an exemplary embodiment, the QRS complex types involved in the present application solution may specifically include a normal type, a LBBB type, a RBBB type, an AVPR type, a WPW-A type, and a WPW-B type.

[0057] Among them, LBBB type refers to left bundle branch block (LBBB) type, RBBB type refers to right bundle branch block (RBBB) type, AVPR type refers to artificial ventricular pacing rhythm (APVR) type, WPW-A type refers to Wolff-Parkinson-White syndrome type-A (WPW-A) type, and WPW-B type refers to Wolff-Parkinson-White syndrome type-B (WPW-B) type.

[0058] These abnormal QRS complex types all show wide QRS complexes. For the WPW-A type, there is an abnormal atrioventricular connection, which leads to preexcitation of the left ventricle and an electric vector to the chest leads. Therefore, the WPW-A type will have positive delta waves and QRS waves in all chest leads. These features are often confused with the RBBB type. Similarly, for the WPW-B type, the abnormal atrioventricular connection between the right atrium and the right ventricle leads to preexcitation of the right ventricle. Therefore, WPW-B will have negative QRS waves in the V1 lead and positive delta waves and QRS waves in the V4-V6 leads, which is often misdiagnosed as the LBBB type. The AVPR type also shows a wide waveform similar to the LBBB type, which further increases the difficulty of diagnosing these ECG abnormalities, and the defects of the existing schemes mentioned above will be further expanded.

[0059] In this case, the output result of the subsequent QRS complex abnormality classification model may specifically be the probability value (which may be represented by P) of the ECG signal input to the model belonging to the above six QRS complex types.

[0060] Step S103, based on the labeled sample ECG signal, train a QRS complex abnormality classification model, wherein the QRS complex abnormality classification model is a deep learning model architecture, and the QRS complex abnormality classification model is used to carry out corresponding QRS complex abnormality classification processing based on the ECG signal input by the model in a preset lead format.

[0061] After obtaining the labeled sample ECG signal, it is obvious that the specific model training process can be carried out. In the model training process, the following processing contents are usually involved:

[0062] In each round of model training, a sample ECG signal is input into the model, so that the model can perform QRS wave group abnormality classification processing, determine the QRS wave group type to which the sample ECG signal input by the model belongs / is adapted, and realize forward propagation. Then, based on the QRS wave group abnormality classification result output by the model and the corresponding annotation result, the loss function is calculated, and the model parameters are optimized according to the loss function calculation result to realize reverse propagation. In this way, when the preset model training requirements such as training time, training times or detection accuracy are met, the model training can be completed, and a QRS wave group abnormality classification model that can be put into practical use can be obtained.

[0063] Among them, it can be understood that in the process of model training, the model architecture, model training architecture and loss function of the specific model itself can be flexibly configured according to actual needs. You can use the existing solution, optimize it based on the existing solution, or use self-developed and novel solutions. These are all possible in actual situations.

[0064] As an exemplary embodiment, the present application provides a set of specific implementation solutions for the specific model architecture of the QRS complex abnormality classification model.

[0065] Specifically, the QRS complex abnormality classification model of the present application can be constructed based on the ResNet model architecture.

[0066] ResNet is a network that combines the advantages of convolutional neural networks (CNN) and solves the problem of gradient vanishing when the number of network layers increases. It reduces the difficulty of optimization by introducing shortcut connections, can deepen the number of network layers, and also has better network learning effects.

[0067] refer to Figure 2A schematic diagram of a scenario of the model architecture involved in the present application is shown. The QRS complex abnormal classification model is based on the ResNet model architecture and includes 37 layers with 15 residual blocks in the middle. Each residual block is composed of two convolutional layers connected by shortcuts. The entire network can still be trained through end-to-end back propagation, where the residual block is the Residual block and the convolutional layer is Conv2D. In addition, each convolutional layer is connected to a BN layer and a ReLU layer. The BN layer normalizes the input to help improve the stability of the model during training. The ReLU layer introduces nonlinearity to enable the model to learn complex patterns. The BN layer is the batch normalization layer, Batch Normalization, and the ReLU layer is the nonlinear activation layer, Rectified Linear Unit.

[0068] In addition, the maximum pooling layer compresses the feature map processed by the previous layer to reduce the computational burden, where the maximum pooling layer is Global max polling.

[0069] In addition, the global average pooling layer generates the final features and reduces the risk of overfitting by significantly reducing the number of parameters. The global average pooling layer is Global average poling.

[0070] In addition, the last layer is a softmax layer, which provides the posterior probability of each category through the softmax function.

[0071] Furthermore, in specific operations, other aspects of model configuration may be involved. For example, the model configuration framework can be implemented based on TensorFlow and run on an NVIDIA Tesla V100 graphics card. In the CNN structure, the kernel size of the main convolution is fixed to (1,5), the pooling size of the maximum pooling layer is fixed to 2, and the pooling step is fixed to 2. An effective optimization method called mom can be used to achieve efficient computing. A weighted loss function called cross entropy loss is used during training. The model hyperparameters are set to {100, 0.00008, 500}, which represent the batch size, learning rate, and number of training rounds, respectively.

[0072] At the same time, in order to more vividly understand the improvement in detection performance brought about by the QRS complex abnormality classification model constructed by the above solution content of this application, the following examples can also be used to help understand.

[0073] As an example, this application compares the 4-lead (I, II, V1 and V5), 8-lead (I, II, V1-V6) format and the 12-lead format (I, II, III, aVR, aVL, aVF, V1-V6) when evaluating six abnormalities.

[0074] The 8-lead format showed the highest mean F1 score of 0.969 (0.943-0.997) and mean AUC score of 0.997 (0.975-1.00); in comparison, the 4-lead format had a mean F1 score of 0.952 (0.931-0.969) and an AUC score of 0.996 (0.976-1.00); in comparison, the 12-lead format had a mean F1 score of 0.95 (0.927-0.97) and an AUC score of 0.994 (0.971-1.00).

[0075] Meanwhile, the 8-lead format showed the highest mean sensitivity of 0.969 (0.94-0.995) and mean specificity score of 0.994 (0.988-1.00); in comparison, the 4-lead format had a mean sensitivity of 0.952 (0.932-0.964) and a mean specificity score of 0.99 (0.971-1.00); in comparison, the 12-lead format had a mean sensitivity of 0.95 (0.930-0.965) and a specificity of 0.99 (0.968-1.00).

[0076] From this, it can be seen that compared with the existing solutions that rely on the 12-lead ECG format, the QRS wave group abnormality classification model constructed based on 8 leads in this application has better detection effects in details, and the QRS wave group abnormality classification model constructed based on 4 leads, although the difference is not significant when the detection accuracy can be improved, it can greatly simplify the amount of data involved in the ECG signal input signal, thereby significantly reducing the application costs in terms of data processing time, data application time and data storage time, and at the same time, it can also significantly reduce the workload (task volume) of clinical work personnel, which is also very beneficial.

[0077] In addition, focusing on the 8-lead format, as another example, the QRS complex abnormality classification model in the 8-lead format was accurate in 456 of the 480 cases (95.0%) in the test data set, which was better than the accuracy of the physician assessment, which was 432 (89.9%). The average F1 score of the QRS complex abnormality classification model for the six abnormalities in the 8-lead format was 0.969 (0.943-0.997), which exceeded the average score of 0.898 (0.863-0.932) of the physicians.

[0078] Refer to Table 1 below for a set of examples of evaluation indicators for this application:

[0079] Table 1 - Examples of evaluation indicators

[0080]

[0081] The F1 scores of the QRS complex abnormality classification model in the six abnormalities were higher than the average scores of the six doctors. Specifically, the F1 scores of the QRS complex abnormality classification model in the LBBB type were 0.983 (0.958-0.999), the RBBB type was 0.982 (0.955-0.997), the AVPR type was 0.981 (0.958-0.997), the WPW-A type was 0.950 (0.91-0.989), and the WPW-B type was 0.951 (0.913-0.989), showing comparable or even superior performance to that of doctors.

[0082] The average AUC score of the QRS complex abnormality classification model for these six categories was 0.997 (0.975-1.00), which was higher than the doctors' AUC score of 0.941 (0.918-0.963), and the model's AUC score for each abnormality was higher than the average score of the six doctors.

[0083] The QRS complex abnormality classification model showed an average sensitivity of 0.969 (0.94-0.995) and specificity of 0.994 (0.988-1.00), which were comparable to the physician-assessed sensitivity of 0.902 (0.858-0.946) and specificity of 0.980 (0.97-0.99).

[0084] See also Figure 3 The figure shows an example of the ROC curve evaluation index of the present application (doctors' individual performance is represented by dots, and doctors' average performance is represented by asterisks). Regarding prediction sensitivity, when comparing the QRS complex abnormality classification model with the ROC curves of six doctors, the model performs better. Figure 4 An example schematic diagram of the confusion matrix evaluation index of the present application is shown. The confusion matrix of the QRS complex abnormality classification model shows a distribution trend similar to that of the doctor's performance. These results indicate that the QRS complex abnormality classification model is significantly better than the doctor's evaluation in the diagnosis of most of the six categories.

[0085] After completing the training / configuration of the model, it is easy to understand that the present application scheme can also involve the application of the QRS complex abnormality classification model.

[0086] In this regard, as an exemplary embodiment, after training the QRS complex abnormality classification model based on the labeled sample ECG signal, the method of the present application may further include:

[0087] Acquire target ECG signal;

[0088] Inputting the target ECG signal into the QRS complex abnormality classification model;

[0089] The QRS complex abnormality classification result corresponding to the target ECG signal output by the QRS complex abnormality classification model is extracted.

[0090] The target ECG signal corresponds to the sample ECG signal, which is an ECG signal in a 4-lead format or an 8-lead format. The QRS complex abnormal classification result can output a most likely QRS complex classification result and mark the corresponding probability value, or output the probability value of each QRS complex abnormal classification result. For example, the output number of the model can be configured as 6, and the output value is the probability value of the signal belonging to 6 categories, and is arranged in order from large to small to concisely and clearly reflect the determined category (corresponding to the maximum probability value).

[0091] In addition, after obtaining the abnormal QRS complex results, further data applications may be involved in accordance with clinical work needs, such as result display, result storage, result forwarding, output completion processing prompts or corresponding data analysis and processing. The specific data application can be configured according to actual needs, and this application does not make specific limitations.

[0092] Finally, for the above scheme content, in general, for the detection target of ECG signals with abnormal QRS complex morphology, this application introduces a QRS wave group abnormality classification mechanism based on 4 leads or 8 leads. The QRS wave group abnormality classification model constructed in this way has been proved through experiments that the detection accuracy of different types of abnormal QRS wave groups in the 8-lead format has been significantly improved. In addition, due to the simplification of data, especially the simplification in the 4-lead format, several aspects of application costs such as data processing time, data application time and data storage time have also been significantly improved. At the same time, it also significantly reduces the workload of clinical personnel, so it has better application value.

[0093] The above is an introduction to the processing method of the QRS complex abnormal classification model provided by the present application. In order to facilitate better implementation of the processing method of the QRS complex abnormal classification model provided by the present application, the present application also provides a processing device for the QRS complex abnormal classification model from the perspective of functional modules.

[0094] See also Figure 5 , Figure 5 : This is a structural diagram of a processing device for the abnormal QRS wave group classification model of the present application. In the present application, the processing device 500 for the abnormal QRS wave group classification model may specifically include the following structure:

[0095] An acquisition unit 501 is used to acquire sample ECG signals of different types of QRS complexes, wherein the sample ECG signals include ECG signals of normal QRS complexes and ECG signals of abnormal QRS complexes, and the sample ECG signals are specifically ECG signals of a preset lead format, and the preset lead format is specifically a 4-lead format or an 8-lead format;

[0096] A labeling unit 502, configured to configure corresponding labels for the sample ECG signal to label the corresponding QRS complex type;

[0097] The training unit 503 is used to train a QRS complex abnormality classification model based on the labeled sample ECG signal, wherein the QRS complex abnormality classification model is a deep learning model architecture, and the QRS complex abnormality classification model is used to carry out corresponding QRS complex abnormality classification processing based on the ECG signal input by the model in a preset lead format.

[0098] In an exemplary embodiment, the leads involved in the 4-lead format specifically include:

[0099] Lead I, Lead II, Lead V1, and Lead V5.

[0100] In another exemplary embodiment, the leads involved in the 8-lead format specifically include:

[0101] Lead I, Lead II, Lead V1, Lead V2, Lead V3, Lead V4, Lead V5 and Lead V6.

[0102] In yet another exemplary embodiment, the QRS complex type specifically includes a normal type, a LBBB type, a RBBB type, an AVPR type, a WPW-A type, and a WPW-B type;

[0103] The output result of the QRS complex abnormality classification model is the probability value of the ECG signal input to the model belonging to the six QRS complex types.

[0104] In another exemplary embodiment, the QRS complex abnormality classification model includes 37 layers based on the ResNet model architecture, with 15 residual blocks in the middle, and each residual block consists of two convolutional layers connected by shortcuts;

[0105] Each convolutional layer is connected to a BN layer and a ReLU layer. The BN layer normalizes the input, and the ReLU layer introduces nonlinearity.

[0106] The maximum pooling layer compresses the feature map processed by the previous layer;

[0107] The global average pooling layer generates the final features;

[0108] The last layer is the softmax layer, which provides the posterior probability of each category through the softmax function.

[0109] In yet another exemplary embodiment, the sample ECG signal and the ECG signal input to the model have a signal length of 10 seconds and a sampling frequency of 500 Hz.

[0110] In yet another exemplary embodiment, the apparatus further includes an application unit 504, configured to:

[0111] Acquire target ECG signal;

[0112] Inputting the target ECG signal into the QRS complex abnormality classification model;

[0113] The QRS complex abnormality classification result corresponding to the target ECG signal output by the QRS complex abnormality classification model is extracted.

[0114] This application also provides a processing device from the perspective of hardware structure, see Figure 6 , Figure 6 601, a memory 602, and an input / output device 603. The processor 601 is used to execute the computer program stored in the memory 602 to implement the following Figure 1 The steps of the processing method of the QRS complex abnormal classification model in the corresponding embodiment; or, the processor 601 is used to execute the computer program stored in the memory 602 to implement the following Figure 5 Corresponding to the functions of each unit in the embodiment, the memory 602 is used to store the processor 601 executing the above Figure 1 The computer program required for the processing method of the QRS complex abnormal classification model in the corresponding embodiment.

[0115] Exemplarily, the computer program may be divided into one or more modules / units, one or more modules / units are stored in the memory 602, and executed by the processor 601 to complete the present application. One or more modules / units may be a series of computer program instruction segments capable of completing specific functions, and the instruction segments are used to describe the execution process of the computer program in the computer device.

[0116] The processing device may include, but is not limited to, a processor 601, a memory 602, and an input / output device 603. Those skilled in the art will appreciate that the illustration is merely an example of a processing device and does not constitute a limitation on the processing device, and may include more or fewer components than shown in the illustration, or a combination of certain components, or different components. For example, the processing device may also include a network access device, a bus, etc., and the processor 601, the memory 602, the input / output device 603, etc. are connected via a bus.

[0117] The processor 601 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc. The processor is the control center of the processing device, and uses various interfaces and lines to connect various parts of the entire device.

[0118] The memory 602 can be used to store computer programs and / or modules. The processor 601 implements various functions of the computer device by running or executing the computer programs and / or modules stored in the memory 602 and calling the data stored in the memory 602. The memory 602 can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, at least one application required for a function, etc.; the data storage area can store data created according to the use of the processing device, etc. In addition, the memory can include a high-speed random access memory, and can also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (SecureDigital, SD) card, a flash card (Flash Card), at least one disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0119] When the processor 601 is used to execute the computer program stored in the memory 602, the following functions can be implemented:

[0120] Acquire sample ECG signals of different types of QRS complexes, wherein the sample ECG signals include ECG signals of normal QRS complexes and ECG signals of abnormal QRS complexes, and the sample ECG signals are specifically ECG signals of a preset lead format, and the preset lead format is specifically a 4-lead format or an 8-lead format;

[0121] Configure corresponding annotations for the sample ECG signal to mark the corresponding QRS complex type;

[0122] Based on the labeled sample ECG signal, a QRS complex abnormality classification model is trained, wherein the QRS complex abnormality classification model is a deep learning model architecture, and the QRS complex abnormality classification model is used to carry out corresponding QRS complex abnormality classification processing based on the ECG signal input by the model in a preset lead format.

[0123] Those skilled in the art can clearly understand that for the convenience and simplicity of description, the specific working process of the processing device, processing equipment and corresponding units of the QRS complex abnormal classification model described above can refer to the following. Figure 1 The description of the processing method of the QRS complex abnormal classification model in the corresponding embodiment will not be repeated here.

[0124] A person of ordinary skill in the art will appreciate that all or part of the steps in the various methods of the above embodiments may be completed by instructions, or by controlling related hardware through instructions. The instructions may be stored in a computer-readable storage medium and loaded and executed by a processor.

[0125] To this end, the present application provides a computer-readable storage medium, in which a plurality of instructions are stored, and the instructions can be loaded by a processor to execute the present application as follows: Figure 1 The steps of the processing method of the QRS complex abnormal classification model in the corresponding embodiment, the specific operation can refer to the following Figure 1 The description of the processing method of the QRS complex abnormal classification model in the corresponding embodiment will not be repeated here.

[0126] The computer-readable storage medium may include: a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0127] Due to the instructions stored in the computer-readable storage medium, the present application can be executed. Figure 1 The steps of the processing method of the QRS complex abnormal classification model in the corresponding embodiment, therefore, the present application can be implemented as follows Figure 1 The beneficial effects that can be achieved by the processing method of the abnormal QRS complex classification model in the corresponding embodiment are detailed in the previous description and will not be repeated here.

[0128] The above is a detailed introduction to the processing method, device, processing equipment and computer-readable storage medium of the QRS complex abnormal classification model provided by the present application. Specific examples are used in this article to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method of the present application and its core idea; at the same time, for technical personnel in this field, according to the idea of ​​the present application, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as a limitation on the present application.

Claims

1. A processing method for a QRS complex abnormal classification model, characterized in that: The method comprises: Acquire sample ECG signals of different types of QRS complexes, wherein the sample ECG signals include ECG signals of normal QRS complexes and ECG signals of abnormal QRS complexes, and the sample ECG signals are specifically ECG signals of a preset lead format, and the preset lead format is specifically a 4-lead format or an 8-lead format; Configuring corresponding annotations for the sample ECG signal to mark the corresponding QRS complex type; Based on the labeled sample ECG signal, a QRS complex abnormality classification model is trained, wherein the QRS complex abnormality classification model is a deep learning model architecture, and the QRS complex abnormality classification model is used to carry out corresponding QRS complex abnormality classification processing based on the ECG signal input by the model of the preset lead format.

2. The method according to claim 1, characterized in that The leads involved in the 4-lead format specifically include: Lead I, Lead II, Lead V1, and Lead V5.

3. The method according to claim 1, characterized in that: The leads involved in the 8-lead format specifically include: Lead I, Lead II, Lead V1, Lead V2, Lead V3, Lead V4, Lead V5 and Lead V6.

4. The method according to claim 1, characterized in that The QRS complex types specifically include normal type, LBBB type, RBBB type, AVPR type, WPW-A type and WPW-B type; The output result of the QRS complex abnormality classification model is the probability value of the ECG signal input to the model belonging to the six QRS complex types.

5. The method according to claim 1, characterized in that The QRS complex abnormality classification model is based on the ResNet model architecture and includes 37 layers with 15 residual blocks in the middle, and each residual block is composed of two convolutional layers connected by shortcuts; Each of the convolutional layers is connected to a BN layer and a ReLU layer, wherein the BN layer normalizes the input and the ReLU layer introduces nonlinearity; The maximum pooling layer compresses the feature map processed by the previous layer; The global average pooling layer generates the final features; The last layer is the softmax layer, which provides the posterior probability of each category through the softmax function.

6. The method according to claim 1, characterized in that The sample ECG signal and the ECG signal input to the model have a signal length of 10 seconds and a sampling frequency of 500 Hz.

7. The method according to claim 1, characterized in that After training the QRS complex abnormality classification model based on the labeled sample ECG signal, the method further includes: Acquire target ECG signal; Inputting the target ECG signal into the QRS complex abnormality classification model; The QRS complex abnormality classification result corresponding to the target ECG signal output by the QRS complex abnormality classification model is extracted.

8. A processing device for a QRS complex abnormal classification model, characterized in that: The device comprises: An acquisition unit, used for acquiring sample ECG signals of different types of QRS complexes, wherein the sample ECG signals include ECG signals of normal QRS complexes and ECG signals of abnormal QRS complexes, and the sample ECG signals are specifically ECG signals of a preset lead format, and the preset lead format is specifically a 4-lead format or an 8-lead format; A labeling unit, used for configuring a corresponding label for the sample ECG signal to label the corresponding QRS complex type; A training unit is used to train a QRS complex abnormality classification model based on the labeled sample ECG signal, wherein the QRS complex abnormality classification model is a deep learning model architecture, and the QRS complex abnormality classification model is used to carry out corresponding QRS complex abnormality classification processing based on the ECG signal input by the model of the preset lead format.

9. A processing device, characterized in that: The method comprises a processor and a memory, wherein the memory stores a computer program, and the processor executes the method according to any one of claims 1 to 7 when calling the computer program in the memory.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a plurality of instructions, and the instructions are suitable for being loaded by a processor to execute the method according to any one of claims 1 to 7.

Images