Deep learning-based illness state early warning method and system

Abstract

The invention provides an illness state early warning method and system based on deep learning, and relates to the technical field of medical data analysis. The method comprises the following steps: acquiring an electrocardiogram signal of a patient; performing denoising processing on the electrocardiogram signal to obtain a denoised electrocardiogram signal; based on the de-noised electrocardiogram signal, performing data enhancement processing through an SMOTE algorithm to obtain an enhanced electrocardiogram signal; constructing an illness state prediction classification model based on deep learning; taking the enhanced electrocardiogram signal as input, and outputting a category prediction result through an illness state prediction classification model based on deep learning; and performing myocardial infarction condition early warning according to the category prediction result. According to the method, the data quality can be effectively improved, complex features can be automatically learned, and key time sequence association can be captured by performing denoising processing on the electrocardiogram signals and constructing the disease condition prediction classification model based on deep learning, so that the accuracy of a prediction result is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of medical data analysis, and in particular to a disease early warning method and system based on deep learning. Background Art

[0002] With the rapid development of deep learning technology, disease early warning methods based on deep learning have become a key research direction in the medical field. By deeply analyzing large amounts of medical data, deep learning models can identify potential health risks and provide timely disease warnings, thereby effectively improving the accuracy of early diagnosis.

[0003] Currently, deep learning-based disease early warning technologies primarily rely on models such as deep neural networks (DNNs), convolutional neural networks (CNNs), and recurrent neural networks (RNNs). These models are capable of processing and analyzing large amounts of medical data, including electronic health records (EHRs), medical images, genomic data, and laboratory test data. In medical image analysis, CNNs have been widely used in areas such as cancer detection and fracture identification.

[0004] However, existing methods perform poorly in data-scarce environments due to the limited availability of case data and the difficulty in obtaining labeled data. Furthermore, due to the strong heterogeneity, high dimensionality, and noise of medical data, models may not fully capture key temporal correlations during feature extraction, resulting in inaccurate predictions. Summary of the Invention

[0005] In view of the above shortcomings of the existing technology, the purpose of the embodiments of the present invention is to provide a deep learning-based disease early warning method that can address the technical problem that existing methods, due to the limited number of case data and the difficulty in obtaining labeled data, lead to poor model performance in data-scarce situations. Furthermore, due to the strong heterogeneity, high dimensionality, and noise of medical data, the model may not fully capture key temporal correlations during feature extraction, resulting in inaccurate prediction results.

[0006] A first aspect of an embodiment of the present invention provides a disease early warning method based on deep learning, comprising:

[0007] S1: Acquire the patient's electrocardiogram signal;

[0008] S2: performing denoising processing on the electrocardiogram signal to obtain a denoised electrocardiogram signal;

[0009] S3: Based on the denoised electrocardiogram signal, performing data enhancement processing using a SMOTE algorithm to obtain an enhanced electrocardiogram signal;

[0010] S4: Build a disease prediction classification model based on deep learning;

[0011] S5: using the enhanced electrocardiogram signal as input, outputting a category prediction result through the deep learning-based disease prediction classification model;

[0012] S6: Perform a myocardial infarction condition warning based on the category prediction results.

[0013] A second aspect of an embodiment of the present invention provides a disease early warning system based on deep learning, comprising: a processor and a memory;

[0014] The memory stores programs or instructions that can be run on the processor, and when the programs or instructions are executed by the processor, the steps of the deep learning-based disease warning method as described in the first aspect are implemented.

[0015] According to a third aspect of an embodiment of the present invention, a readable storage medium is proposed, on which a program or instruction is stored. When the program or instruction is executed by a processor, the steps of the deep learning-based disease warning method as described in the first aspect are implemented.

[0016] The beneficial effects brought about by the technical solution provided by the embodiment of the present invention include at least:

[0017] In this embodiment of the present invention, the SMOTE algorithm is used for data enhancement to alleviate the problem of insufficient data and thus improve model performance. Furthermore, by denoising the electrocardiogram signal and building a deep learning-based disease prediction classification model, data quality can be effectively improved, complex features can be automatically learned, and key temporal correlations can be captured, thereby improving the accuracy of myocardial infarction prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The accompanying drawings are only for the purpose of illustrating specific embodiments and are not to be considered as limiting the present invention. Throughout the drawings, the same reference symbols represent the same components. Obviously, the drawings described below are only some embodiments of the present invention. It is clear that those skilled in the art can derive other drawings based on these drawings without inventive effort.

[0019] Figure 1 This is a flowchart of a disease early warning method based on deep learning provided by an embodiment of the present invention;

[0020] Figure 2 This is a structural diagram of a deep learning-based disease warning system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0021] In order to enable those skilled in the art to better understand the technical solutions in the embodiments of the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the embodiments described are part of the embodiments of the present invention, rather than all of the embodiments. It should be understood that these descriptions are merely exemplary and are not intended to limit the scope of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative work should fall within the scope of protection of the present invention.

[0022] The following describes in detail the deep learning-based disease warning method provided by the embodiment of the present invention through specific embodiments and their application scenarios in conjunction with the accompanying drawings.

[0023] Reference Manual Figure 1 , which shows a flow chart of a disease warning method based on deep learning provided by an embodiment of the present invention.

[0024] An embodiment of the present invention provides a disease early warning method based on deep learning, which may include the following steps:

[0025] S1: Obtain the patient's electrocardiogram signal.

[0026] An electrocardiogram (ECG) is a technique used to record the heart's electrical activity. Electrodes are placed on the skin to capture the electrical signals generated by each heartbeat. ECG signals are commonly used to diagnose heart disease and monitor heart health.

[0027] It should be noted that electrocardiogram features such as ST segment elevation, Q waves, and T wave inversion are important indicators for early warning of myocardial infarction. Therefore, during signal acquisition, it is important to ensure that the electrodes are properly placed to ensure that the collected electrocardiogram signal contains all key information about the heart.

[0028] Specifically, by using an electrocardiogram (ECG) monitoring device, electrodes are placed on the patient's skin to record the heart's electrical activity signals in real time to obtain the patient's ECG signal.

[0029] In embodiments of the present invention, an electrocardiogram (ECG) monitoring device can record electrical signals in real time during a patient's cardiac activity. This is crucial for detecting key features associated with myocardial infarction, such as ST-segment elevation, Q wave formation, and T-wave inversion. Furthermore, properly placed electrodes can capture comprehensive and accurate cardiac electrical activity, ensuring that the monitored signals reflect the heart's true condition.

[0030] S2: De-noising the electrocardiogram signal to obtain a de-noised electrocardiogram signal.

[0031] In a possible implementation, S2 specifically includes:

[0032] S201: Decompose the electrocardiogram signal into multiple sub-signals in different frequency ranges using discrete wavelet transform:

[0033]

[0034] Among them, X(t) represents the electrocardiogram signal at time t, ψ j (t) represents the value of the wavelet basis function at time t under the j-th scale, W j Represents the wavelet coefficients at the j-th scale.

[0035] The discrete wavelet transform (DWT) is a mathematical tool used to analyze and process signals. It decomposes signals into distinct frequency components, enabling analysis at multiple scales. DWT, a form of wavelet transform, recursively decomposes signals while preserving both time and frequency domain information. DWT has a wide range of applications in fields such as signal denoising, data compression, and feature extraction, and is particularly important in electrocardiogram (ECG) signal processing and image processing.

[0036] In the embodiments of the present invention, DWT helps preserve signal characteristics during denoising. Important waveforms in electrocardiogram signals are typically located in the low-frequency portion, while noise (such as myoelectric noise and baseline wander) is typically located in the high-frequency portion. Therefore, by decomposing the wavelet coefficients, the signal and noise can be processed separately, achieving better denoising effects.

[0037] S202: Use soft threshold denoising to denoise the wavelet coefficients of each layer after decomposition:

[0038]

[0039] in, It represents the wavelet coefficient at the j-th scale after denoising, sign represents the sign function, λ represents the soft threshold, and max represents the maximum value.

[0040] The basic idea of ​​soft threshold denoising is to remove noise by applying a threshold to the wavelet transform coefficients (or other transform coefficients). The denoising process modifies the transform coefficients so that coefficients smaller than the threshold are suppressed to zero, while coefficients larger than the threshold are reduced to a certain extent but not completely eliminated.

[0041] In this embodiment of the present invention, the coefficients after wavelet transformation can be divided into low-frequency and high-frequency components. Noise is generally located in the high-frequency component. Soft threshold denoising can effectively remove high-frequency noise while retaining the key components of the electrocardiogram signal, especially the useful information in the low-frequency component.

[0042] S203: Reconstruct the electrocardiogram signal according to the denoised wavelet coefficients to obtain a denoised electrocardiogram signal:

[0043]

[0044] in, represents the denoised ECG signal at time t, Represents the wavelet coefficients after denoising.

[0045] In the embodiment of the present invention, the accuracy of the electrocardiogram signal is improved through reconstruction, and the error introduced by noise is reduced. For clinical applications, accurate electrocardiogram signals can improve the reliability and accuracy of diagnosis.

[0046] S3: Based on the denoised ECG signal, data enhancement processing is performed using the SMOTE algorithm to obtain an enhanced ECG signal.

[0047] SMOTE (Synthetic Minority Over-sampling Technique) is a data augmentation method commonly used to address class imbalance. It balances the class distribution by synthesizing samples of the minority class, thereby improving the machine learning model's prediction ability for the minority class.

[0048] In one possible implementation, based on the denoised electrocardiogram signal, data enhancement processing is performed using the following formula to obtain an enhanced electrocardiogram signal:

[0049] X new =X i +rand(0,1)×(X i,j -X i )

[0050] Among them, X new represents the generated minority class samples, X i represents the i-th heartbeat sample in the minority class sample, X i,j represents the neighboring samples of the i-th heartbeat sample, and rand(0,1) represents a random number between 0 and 1. It should be noted that data augmentation using the SMOTE algorithm helps generate more samples when minority class samples (such as myocardial infarction) are insufficient. This is particularly useful for rare events such as myocardial infarction. The SMOTE algorithm can synthesize new samples, thereby balancing the dataset and helping deep learning models better learn the characteristics of minority class samples.

[0051] In this embodiment of the present invention, applying the SMOTE algorithm to denoised ECG signals effectively balances class imbalance, enhances minority class samples, and improves the model's classification capabilities, particularly in medical signal analysis applications such as abnormal heartbeat detection. By generating synthetic minority class samples, SMOTE enhances the model's generalization, robustness, and ability to identify minority classes, providing more reliable results for ECG signal classification tasks.

[0052] S4: Build a disease prediction classification model based on deep learning.

[0053] Among them, deep learning-based disease prediction and classification models are systems that use deep neural networks to predict and classify diseases. By automatically learning features from large amounts of data, deep learning technology can extract useful information from complex biomedical signals (such as electrocardiograms, electroencephalograms, and medical imaging), helping doctors diagnose diseases and select treatment options.

[0054] S5: Using the enhanced ECG signal as input, the deep learning-based disease prediction classification model outputs the category prediction results.

[0055] In a possible implementation, S5 specifically includes:

[0056] S501: Using dilated causal convolution to capture the spatial features of enhanced ECG signals:

[0057]

[0058] in, represents the spatial features of the t-th time step, ReLU represents the activation function, and W dil Represents the weight matrix of the dilated convolution, * represents the symbol of the convolution operation, x t-k:t represents the signal segment from the tkth time step to the tth time step, b dil Represents the bias term of the dilated convolution.

[0059] It should be noted that dilated causal convolution is used to extract spatial features related to myocardial infarction, such as ST segment elevation, Q wave changes, and T wave inversion.

[0060] In the embodiment of the present invention, dilated convolution can effectively capture important local spatial features in the electrocardiogram signal.

[0061] S502: Introduce a gated recurrent unit to capture and enhance the timing characteristics of the ECG signal:

[0062]

[0063] in, represents the time series feature of the tth time step, xt represents the electrocardiogram signal at the t-th time step, represents the temporal hidden features of the t-1th time step, W gru Represents the learnable weight matrix in the GRU network.

[0064] It's important to note that the GRU is used to capture temporal features in ECG signals. Here, the GRU specifically focuses on temporal changes associated with myocardial infarction, such as persistent ST segment elevation and T wave inversion. By leveraging long-term temporal dependencies, the model can learn how these features change over time, thereby identifying the onset of myocardial infarction.

[0065] In the embodiment of the present invention, the electrocardiogram signal is a typical time series signal. The GRU can effectively capture the time series information and learn the long-term time dependency in the signal, thereby identifying the important time series features in the electrocardiogram signal.

[0066] S503: Perform a preliminary fusion of spatial features and temporal features through the feature cross-attention mechanism to obtain preliminary fusion features:

[0067]

[0068] in, represents the preliminary fusion feature of the t-th time step, A represents the cross attention weight, Softmax represents the activation function, Q represents the query vector obtained after the spatial feature is linearly transformed, K represents the key vector obtained after the temporal feature is linearly transformed, T represents the transpose operation, and d represents the scaling factor.

[0069] In this embodiment of the present invention, the cross-attention mechanism helps the model focus on changes in ST-segment elevation and Q waves, thereby more accurately integrating their information and improving prediction accuracy. Furthermore, the cross-attention mechanism helps capture the nonlinear relationship between spatial and temporal features, improving the model's understanding of complex patterns in the electrocardiogram.

[0070] S504: Use the dynamic gating mechanism to perform fine fusion on the preliminary fusion features to obtain fine fusion features:

[0071]

[0072] in, represents the fine fusion feature of the t-th time step, g t represents the weighting factor, ⊙ represents element-wise multiplication, σ represents the Sigmoid activation function, W g represents the weight matrix, b g Represents the bias term used to adjust the gate value calculation.

[0073] In the embodiment of the present invention, the dynamic gating mechanism automatically adjusts the fusion of spatial and temporal features, so that the model can prioritize the characteristics of myocardial infarction and improve prediction accuracy.

[0074] S505: Based on the refined fusion features, the final fusion features are obtained through multi-layer convolution and skip connection processing:

[0075]

[0076] in, represents the final fusion feature of the t-th time step, α l Represents the weight coefficient used to weight the convolution results of each layer, l = 1, 2, ..., L, L represents the number of jump connection layers, Conv1D represents a one-dimensional convolution operation, It represents the refined fusion features after the l-th layer convolution operation at the t-th time step.

[0077] In the embodiment of the present invention, through multi-layer convolution and skip connections, the model can fuse features from different levels, thereby enhancing the model's ability to learn features of different scales.

[0078] S506: Input the final fusion features into the fully connected layer for classification and output the category prediction result.

[0079] In a possible implementation, S506 specifically includes:

[0080] S5061: Input the final fusion features into the fully connected layer for classification to obtain the score value of each category:

[0081]

[0082] Among them, y m Represents the score value of the mth category, W cls represents the weight matrix of the fully connected layer, b cls represents the bias term of the fully connected layer.

[0083] S5062: Convert the score values ​​of each category into category prediction results through the Softmax function:

[0084]

[0085] Among them, P m represents the predicted probability of the mth category, e represents the exponential function, n represents the index of the category, y n Represents the raw score value of the nth category.

[0086] In summary, through dilated convolution, GRU, attention mechanism and gating mechanism, the model can effectively extract spatiotemporal features from ECG signals, enhance the understanding of the signals, and significantly improve the accuracy of the classification model, especially for the recognition of minority classes.

[0087] In a possible implementation, the category prediction result specifically includes: a predicted probability of a normal heartbeat and a predicted probability of myocardial infarction.

[0088] Optionally, the disease prediction classification model based on deep learning is optimized by an improved Adam optimizer.

[0089] Specifically, the loss function of the disease prediction classification model is set to the mean square error loss function:

[0090]

[0091] Among them, MSE represents the mean square error loss function, k represents the total number of samples, Indicates the predicted value of the b-th sample, y b Represents the true value of the b-th sample.

[0092] Initialization parameters, including model parameters θ0, learning rate η, batch size N B , decay rates β1, β2 and constant ε.

[0093] Initialize the first-order moment estimate m0=0, the second-order moment estimate v0=0, and

[0094] A training data set is obtained, which includes normal samples and myocardial infarction heartbeat samples, and each sample is labeled.

[0095] Extract a batch of training samples from the training data set and calculate the current parameter θ t-1 Gradient of the objective function:

[0096]

[0097] Among them, G t represents the gradient at time t, x ik Represents the input data of the kth sample, y ik represents the label of the kth sample, θ t-1 represents the model parameters in the previous iteration, I represents the cross entropy loss function, L((x ik ,y ik ),θ t-1 ) represents the output of the model, N B Indicates the total number of samples.

[0098] Update the first-order moment estimate using the current gradient:

[0099] m t =β1·m t-1 +(1-β1)G t

[0100] Among them, m t represents the first-order moment estimate at time t, m t-1 represents the first-order moment estimate at time t-1, β1 represents the decay rate of the first-order moment, G t represents the gradient at time t.

[0101] Find the maximum value of the first-order moment at the previous moment and the current gradient.

[0102] Calculate the target ratio based on the updated first-order moment estimate and the maximum value of the first-order moment and the current gradient at the previous moment:

[0103]

[0104] Where Λ(t) represents the target ratio.

[0105] In an embodiment of the present invention, by dynamically judging the target ratio, abnormal situations that occur during the training process (such as a sudden increase in gradient) can be identified, and the optimization strategy can be automatically switched, thereby improving the robustness and fault tolerance of the model training.

[0106] Compute the second-order moment estimate by applying the target ratio to control the magnitude of the update:

[0107] v t =β2·v t-1 +(1-β2)·|G t | Λ(t)

[0108] Among them, v t represents the second-order moment estimate at time t, v t-1 It represents the second-order moment estimate at time t-1, and β2 represents the decay rate of the second-order moment.

[0109] When the target ratio is less than 2, switch to the AMSGrad optimizer for optimization, set amsgrad = True, and update the parameter θ t :

[0110]

[0111] Among them, θ t represents the model parameters at time t, θ t-1 represents the model parameters at time t-1, η represents the learning rate, m t represents the first-order moment estimate at time t, ε represents a constant, represents the modified second-order moment estimate in the AMSGrad optimizer, represents the revised second moment estimate at the previous moment.

[0112] When the target ratio is greater than or equal to 2, continue to use the Adam optimizer to update the parameter θ t :

[0113]

[0114] When the iteration reaches the preset number of iterations, the iteration is stopped and the final optimization parameters are output.

[0115] In an embodiment of the present invention, by introducing a target ratio and dynamically switching between the two optimization strategies of Adam and AMSGrad, it is possible to balance rapid convergence and optimization stability during the model training process. When the gradient fluctuates greatly, Adam is used to maintain efficient updates; when the training tends to be stable or faces a local optimum, AMSGrad is switched to suppress overshoot, effectively improving the robustness and accuracy of the training. In addition, this method has adaptive adjustment capabilities, which can automatically control the parameter update amplitude, avoid oscillation and divergence, and reduce the burden of manual parameter adjustment. This allows the trained model to meet medical scenarios with high stability requirements.

[0116] S6: Provide early warning of myocardial infarction based on category prediction results.

[0117] In a possible implementation, S6 specifically includes:

[0118] S601: Sort the category probabilities of the categories in the category prediction results in descending order, and select the target category corresponding to the highest prediction probability.

[0119] S602: If the target category is normal heartbeat, continue monitoring. If the target category is myocardial infarction, issue an alarm.

[0120] In an embodiment of the present invention, by setting different response mechanisms for different categories, the system can precisely control when to trigger an alert. For example, if a normal heartbeat is predicted, the system continues monitoring without interfering with normal operations. However, if a myocardial infarction is predicted, the system will immediately issue an alert. This approach avoids excessive alerts and improves system efficiency. At the same time, the system will not easily trigger an alert if there is no myocardial infarction. An alert is only issued when the model predicts a myocardial infarction with a high probability, which can effectively reduce the interference or unnecessary emergency response caused by excessive alerts.

[0121] In a possible implementation, issuing an alarm specifically includes:

[0122] Sound an alarm via warning lights.

[0123] and / or sound an alarm via a buzzer.

[0124] In the embodiment of the present invention, the warning light and buzzer alarm each provide different sensing methods: the warning light provides a visual signal, while the buzzer alarm provides an auditory signal. The combination of the two ensures that the alarm information is not missed due to ambient noise, equipment failure, or personnel failing to notice a particular signal source.

[0125] The beneficial effects brought about by the technical solution provided by the embodiment of the present invention include at least:

[0126] In this embodiment of the present invention, the SMOTE algorithm is used for data enhancement to alleviate the problem of insufficient data and thus improve model performance. Furthermore, by denoising the electrocardiogram signal and building a deep learning-based disease prediction classification model, data quality can be effectively improved, complex features can be automatically learned, and key temporal correlations can be captured, thereby improving the accuracy of myocardial infarction prediction.

[0127] Reference Manual Figure 2 , which shows a structural diagram of a disease warning system based on deep learning provided by an embodiment of the present invention.

[0128] The embodiment of the present invention provides a disease early warning system 20 based on deep learning, comprising: a processor 201 and a memory 202;

[0129] The memory 202 stores programs or instructions that can be run on the processor 201. When the programs or instructions are executed by the processor 201, the steps of the above-mentioned deep learning-based disease warning method are implemented, and the same technical effects can be achieved. To avoid repetition, the present invention will not be described in detail.

[0130] It should be understood that the processor 201 in the embodiment of the present invention may be a central processing unit (CPU), and the processor may also be 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. The general-purpose processor may be a microprocessor or any conventional processor, etc.

[0131] It should also be understood that the memory 202 in the embodiment of the present invention can be a volatile memory or a non-volatile memory, or can include both volatile and non-volatile memories. Among them, the non-volatile memory can be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. The volatile memory can be a random access memory (RAM), which is used as an external cache. By way of example and not limitation, many forms of random access memory (RAM) are available, such as static RAM (SRAM), dynamic random access memory (DRAM), synchronous DRAM (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link DRAM (SLDRAM), and direct rambus RAM (DR RAM).

[0132] The above embodiments can be implemented in whole or in part through software, hardware (such as circuits), firmware, or any other combination. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer program are loaded or executed on a computer, the processes or functions described in accordance with the embodiments of the present invention are generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via a wired method (such as infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that contains a collection of one or more available media. The available medium can be a magnetic medium (such as a floppy disk, hard disk, or magnetic tape), an optical medium (such as a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state drive.

[0133] It should be understood that in various embodiments of the present invention, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0134] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present invention.

[0135] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described equipment, devices and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0136] In the several embodiments provided by the present invention, it should be understood that the disclosed devices, apparatuses and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the units is merely a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another device, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interface, indirect coupling or communication connection of the device or unit, which can be electrical, mechanical or other forms.

[0137] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0138] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0139] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0140] An embodiment of the present invention provides a readable storage medium including: a program or instruction is stored on the readable storage medium, and when the program or instruction is executed by the processor, the steps of the above-mentioned deep learning-based disease warning method are implemented, and the same technical effect can be achieved. To avoid repetition, the present invention will not be described in detail.

[0141] Finally, it should be noted that the above embodiments are merely illustrative of the technical solutions of the embodiments of the present invention, and are not intended to limit them. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they may still modify the technical solutions described in the aforementioned embodiments, or replace some of the technical features therein with equivalents; and such modifications or replacements do not deviate from the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention. Any changes or replacements that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be covered by the scope of protection of the present invention.

Claims

1. A disease warning method based on deep learning, characterized in that: include: S1: Acquire the patient's electrocardiogram signal; S2: performing denoising processing on the electrocardiogram signal to obtain a denoised electrocardiogram signal; S3: Based on the denoised electrocardiogram signal, performing data enhancement processing using a SMOTE algorithm to obtain an enhanced electrocardiogram signal; S4: Build a disease prediction classification model based on deep learning; S5: using the enhanced electrocardiogram signal as input, outputting a category prediction result through the deep learning-based disease prediction classification model; S6: Perform a myocardial infarction condition warning based on the category prediction results.

2. The disease early warning method based on deep learning according to claim 1, characterized in that: The S2 specifically includes: S201: Decomposing the electrocardiogram signal into a plurality of sub-signals in different frequency ranges using discrete wavelet transform; S202: Using soft threshold denoising to denoise the wavelet coefficients of each layer after decomposition; S203: Reconstructing the electrocardiogram signal according to the denoised wavelet coefficients to obtain the denoised electrocardiogram signal.

3. The disease early warning method based on deep learning according to claim 1, characterized in that: The S3 is specifically: Based on the denoised electrocardiogram signal, data enhancement processing is performed using the following formula to obtain the enhanced electrocardiogram signal: X new =X i +rand(0,1)×(X i,j -X i ); Among them, X new represents the generated minority class samples, X i represents the i-th heartbeat sample in the minority class sample, X i,j represents the neighboring samples of the i-th heartbeat sample, and rand(0,1) represents a random number between 0 and 1.

4. The disease early warning method based on deep learning according to claim 1, characterized in that: The S5 specifically includes: S501: Using dilated causal convolution to capture the spatial features of the enhanced electrocardiogram signal; S502: introducing a gated recurrent unit to capture the temporal characteristics of the enhanced electrocardiogram signal; S503: Preliminarily fusing the spatial features and the temporal features through a feature cross-attention mechanism to obtain a preliminary fused feature; S504: Use a dynamic gating mechanism to perform fine fusion on the preliminary fusion features to obtain fine fusion features: in, represents the fine fusion feature of the t-th time step, g t represents the weighting factor, ⊙ represents element-wise multiplication, represents the spatial features of the t-th time step, represents the time series feature of the t-th time step, σ represents the Sigmoid activation function, W g represents the weight matrix, represents the initial fusion feature of the t-th time step, b g Represents the bias term used to adjust the gate value calculation; S505: Based on the refined fusion features, the final fusion features are obtained through multi-layer convolution and skip connection processing: in, represents the final fusion feature of the t-th time step, α l Represents the weight coefficient used to weight the convolution results of each layer, l = 1, 2, ..., L, L represents the number of jump connection layers, Conv1D represents a one-dimensional convolution operation, Represents the refined fusion features after the l-th layer convolution operation at the t-th time step; S506: Input the final fusion feature into the fully connected layer for classification, and output the category prediction result.

5. The disease warning method based on deep learning according to claim 4 is characterized in that: The S506 specifically includes: S5061: Inputting the final fusion feature into the fully connected layer for classification to obtain a score value for each category; S5062: Convert the score value of each category into the category prediction result through the Softmax function.

6. The disease early warning method based on deep learning according to claim 1, characterized in that: The category prediction results specifically include: the predicted probability of normal heartbeat and the predicted probability of myocardial infarction.

7. The disease warning method based on deep learning according to claim 6, characterized in that: The S6 specifically includes: S601: sorting the category probabilities of the categories in the category prediction results in descending order, and selecting the target category corresponding to the highest prediction probability; S602: When the target category is normal heartbeat, continue monitoring; when the target category is myocardial infarction, issue an alarm.

8. The disease warning method based on deep learning according to claim 7, characterized in that: The issuing of an alarm specifically includes: Sound an alarm through warning lights; and / or sound an alarm via a buzzer.

9. A disease early warning system based on deep learning, characterized in that: include: processor and memory; The memory stores programs or instructions that can be run on the processor, and when the programs or instructions are executed by the processor, the steps of the deep learning-based disease warning method as described in any one of claims 1 to 8 are implemented.

10. A readable storage medium, characterized in that: The readable storage medium stores a program or instruction, and when the program or instruction is executed by the processor, the steps of the deep learning-based disease warning method as described in any one of claims 1 to 8 are implemented.

Images