Portable 12-lead electrocardiosignal monitoring system and method
Through improved U-Net network and gated attention mechanism, combined with lightweight hardware, the portable ECG monitoring equipment has been solved in the long-term monitoring and wear comfort, achieving a 12-lead core with high accuracy and portability Electric signal monitoring, supports 24-hour monitoring and remote diagnosis.
Patent Information
- Application Number
- CN202411575580.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-11-06
- Publication Date
- 2024-12-03
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The existing portable electrocardiogram monitoring equipment has shortcomings in long-term monitoring and wearing comfort. Some equipment is not convenient for long-term wear and is easily affected by motion artifacts, and the lead wire affects the wearing comfort.
Using the improved 12-lead ECK signal reconstruction method of U-Net network, residual module and gated attention mechanism, the II-lead ECK signal was obtained through 3 electrodes for multiple downsampling and upsampling, gradually reducing and Recover the spatial resolution of the signal, output 12-lead ECG signals, and combine lightweight hardware to realize a portable monitoring system.
It realizes 24-hour ECG monitoring with lightness, portability, high integration and high accuracy. It only requires 3 electrode tabs, which can run on embedded devices, and supports remote diagnosis and daily ECG monitoring of doctors.
Smart Images

Figure CN119055248A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of medical supplies, and in particular to a portable 12-lead electrocardiogram signal monitoring system and method. Background Art
[0002] As people pay more and more attention to their health, portable ECG monitoring devices have become popular smart wearable devices. These devices combine traditional ECG technology with the convenience of modern portable devices. They are small, light, easy to carry, and support real-time monitoring and data transmission, allowing users to understand their heart health at any time. There are some problems with the portable ECG monitoring devices currently on the market. For example, some handheld monitoring devices are not convenient enough for long-term monitoring, and bracelet devices are easily affected by motion artifacts. In addition, the lead wires of some devices may affect the wearing comfort, limiting the convenience of long-term wearing.
[0003] For example, CN118098565A discloses a 12-lead wearable ECG monitoring system and method based on an integrated algorithm for EASI lead reconstruction, wherein the EASI ECG is a derived 12-lead ECG, which requires five electrodes to obtain three bipolar lead ECG signals, and uses the EASI ECG signals as the input of the model; it divides the 12-lead ECG into two categories: high complexity and low complexity, and uses a deep learning method to reconstruct the high complexity, and a linear regression algorithm to reconstruct the low complexity, which increases the complexity of the system. It uses a Transformer-based model with a large number of parameters; For another example, CN113951892A discloses a data acquisition method based on a wearable 12-lead electrocardiograph, in which the 12-lead electrocardiograph acquisition system uses 10 electrodes to directly measure the 12-lead electrocardiograph. The greater the number of electrodes, the less comfortable and portable the wearer is. Summary of the invention
[0004] The present invention provides a portable 12-lead ECG signal monitoring system and method, which overcomes the shortcomings described in the background technology. The portable 12-lead ECG signal monitoring system and method in the present invention has a wide range of application scenarios, is flexible to wear, can be used for 24-hour ECG monitoring, and can obtain a complete 12-lead ECG signal with less hardware resources.
[0005] The technical solution adopted by the present invention to solve its technical problem is: A portable 12-lead electrocardiogram signal monitoring system, comprising: The monitoring system obtains the II lead ECG signal by measuring the subject, and the signal sampling frequency is 100Hz; The 12-lead ECG signal reconstruction method based on the improved U-Net network, residual module and gated attention mechanism takes the II-lead ECG signal obtained by the measuring subject as input, and gradually reduces the spatial resolution of the feature map by downsampling the II-lead ECG signal multiple times, leaving higher-level ECG signal features, and then restores the signal spatial resolution in the upsampling part, and outputs the 12-lead ECG signal. The monitoring system receives the II-lead ECG signal of any time length as input and outputs the 12-lead ECG signal of the corresponding time length; Among them, each downsampling layer in the improved U-Net network consists of a 3×3 convolution layer, a batch normalization layer, a Parametric ReLU activation function, and a residual module; The part connecting different downsampling levels is composed of Parametric ReLU activation function and maximum pooling layer; The bottom layer of the U-Net network consists of a 3×3 convolutional layer overlapped twice, a batch normalization layer, a Parametric ReLU activation function, and a transposed convolutional layer; Each upsampling layer consists of a connection layer, a 3×3 convolution layer, a batch normalization layer, a ParametricReLU activation function, and a residual module; The part connecting different upsampling levels is composed of Parametric ReLU activation function and transposed convolution layer; the corresponding layers of downsampling and upsampling are connected through gated attention mechanism, and the number of network layers is 5.
[0006] A better technical solution: The residual module includes a 3×3 convolution layer, a batch normalization layer, a Parametric ReLU activation function, a 3×3 convolution layer, and a skip connection that adds the input directly to the output.
[0007] A better technical solution: A gated attention mechanism is added between the layers corresponding to the downsampling and upsampling parts to eliminate unimportant features, thereby achieving adaptive modification of feature weights.
[0008] A better technical solution: The measuring body includes a shell, a measuring module, a power supply module, a data acquisition module, a computing platform, a data transmission module, and a PCB. The power supply module, the data acquisition module, the computing platform, and the data transmission module are welded on the PCB and fixed in the shell.
[0009] A better technical solution: the measuring module is an electrode sheet, and the measuring body is connected with the electrode sheet to measure the II lead electrocardiogram signal.
[0010] A preferred technical solution: the number of electrode sheets is three.
[0011] A preferred technical solution: a charging interface, a universal data acquisition interface, and a power supply module are arranged on the shell, and the charging interface and the universal data acquisition interface are located at the side end of the shell.
[0012] A preferred technical solution: The measuring body also includes a touch display screen.
[0013] A portable 12-lead electrocardiogram signal monitoring method comprises the following steps: S1: The user wears the measurement body, places the electrode sheet in the appropriate position and connects the measurement output end of the electrode sheet to the measurement body, wherein the measurement body includes the portable 12-lead ECG signal monitoring system mentioned above; S2: Start the measurement subject and select the measurement method; S3: The data acquisition module collects ECG signals and automatically filters them, and then resamples them through the computing platform with a sampling frequency of 100 Hz; S4: The collected II-lead ECG signal with a sampling frequency of 100Hz is used as the input of the 12-lead ECG signal reconstruction algorithm. The signal is processed through a series of convolution and maximum pooling operations in the downsampling module to gradually reduce the spatial resolution of the feature map, and finally leave the ECG signal features at a higher level. In the upsampling part, the transposed convolution operation is performed in conjunction with the gated attention mechanism of the jump connection part to restore the spatial resolution, and the features at different levels are fused, finally realizing the reconstruction of the 12-lead ECG signal from the II-lead ECG signal; S5: The data transmission module feeds back the 12-lead ECG signal to the user via wireless communication.
[0014] A better technical solution: The data acquisition module performs noise reduction processing through a preamplifier, a low-pass filter and a high-pass filter to remove noise such as baseline drift, electromyographic signal and power frequency interference in the electrocardiogram signal.
[0015] By adopting the above technical solution, the beneficial effects of the present invention are: 1. The measuring subject of the present invention is light to wear, highly portable, and has a certain decorative ability; only three electrodes are required, and a 12-lead electrocardiogram reconstruction algorithm can be directly run on the measuring subject.
[0016] 2. The monitoring system in the present invention has high integration and high monitoring accuracy. The algorithm is deployed on the computing platform without the need for terminal devices such as computers. The U-Net network improved by the residual module and the gated attention mechanism is used. The depth of the model is increased by adding the residual module, so that the model has better nonlinear ability. The feature weights are adjusted by the gated attention mechanism, so that the model reconstructs the 12-lead ECG signal with higher accuracy.
[0017] 3. The monitoring system of the present invention can enable doctors to observe and diagnose the patient's electrocardiogram remotely, and can also be used for daily electrocardiogram monitoring, which is highly practical. BRIEF DESCRIPTION OF THE DRAWINGS
[0018] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.
[0019] Figure 1 It is a schematic diagram of the overall framework of the present invention; Figure 2 It is a structural schematic diagram of a data acquisition module of the present invention; Figure 3 It is a schematic diagram of the framework of the 12-lead ECG signal reconstruction algorithm model of the present invention; Figure 4 A schematic diagram of the electrode placement position required for the II lead signal of the present invention; Figure 5 It is a schematic diagram of the housing of the present invention; Figure 6 This is a schematic diagram of the front side of the PCB of the present invention; Figure 7 This is a schematic diagram of the reverse side of the PCB of the present invention; Figure 8 It is a schematic diagram of the implementation process of the present invention; Fig. 9 Schematic diagram of input and output of the 12-lead ECG signal reconstruction algorithm model of the present invention.
[0020] Description of main reference numerals: 1. Measuring body; 2. Electrode sheet; 3. Touch screen display; 4. Charging interface; 5. Data acquisition interface; 6. Data transmission module; 7. Computing platform; 8. Data acquisition module; 9. Power supply module. DETAILED DESCRIPTION
[0021] Example 1 like Figure 1-Figure 7 As shown, this embodiment provides a portable 12-lead ECG signal monitoring system, including: The monitoring system obtains the II lead ECG signal by measuring subject 1, and the signal sampling frequency is 100 Hz; The 12-lead ECG signal reconstruction method based on the improved U-Net network, residual module and gated attention mechanism takes the II-lead ECG signal obtained by the measuring subject as input, and gradually reduces the spatial resolution of the feature map by downsampling the II-lead ECG signal multiple times, leaving higher-level ECG signal features, and then restores the signal spatial resolution in the upsampling part, and outputs the 12-lead ECG signal. The monitoring system receives the II-lead ECG signal of any time length as input and outputs the 12-lead ECG signal of the corresponding time length; Among them, each downsampling layer in the improved U-Net network consists of a 3×3 convolution layer, a batch normalization layer, a Parametric ReLU activation function, and a residual module; The part connecting different downsampling levels is composed of Parametric ReLU activation function and maximum pooling layer; The bottom layer of the U-Net network consists of a 3×3 convolutional layer overlapped twice, a batch normalization layer, a Parametric ReLU activation function, and a transposed convolutional layer; Each upsampling layer consists of a connection layer, a 3×3 convolution layer, a batch normalization layer, a ParametricReLU activation function, and a residual module; The part connecting different upsampling levels is composed of Parametric ReLU activation function and transposed convolution layer; the corresponding layers of downsampling and upsampling are connected through gated attention mechanism, and the number of network layers is 5.
[0022] Furthermore, the residual module includes a 3×3 convolutional layer, a batch normalization layer, a Parametric ReLU activation function, a 3×3 convolutional layer, and a skip connection to add the input directly to the output.
[0023] Furthermore, a gated attention mechanism is added between the layers corresponding to the downsampling and upsampling parts to remove unimportant features, thereby achieving adaptive modification of feature weights.
[0024] Furthermore, the measuring body includes a shell, a measuring module, a power supply module 9, a data acquisition module 8, a computing platform 7, a data transmission module 6, and a PCB. The power supply module 9, the data acquisition module 8, the computing platform 7, and the data transmission module 6 are welded on the PCB and fixed in the shell.
[0025] Furthermore, the measuring module is an electrode sheet 2, and the measuring body 1 is connected with the electrode sheet 2 to measure the II lead ECG signal, and the electrode sheet collects the weak electrical signal generated by the heart activity by being attached to the designated position. The computing platform executes the 12-lead ECG signal reconstruction algorithm and outputs the 12-lead ECG signal.
[0026] Furthermore, the number of the electrode sheets 2 is three.
[0027] Furthermore, a charging interface 4, a data acquisition interface 5, and a power supply module 9 are provided on the shell, and the charging interface 4 and the data acquisition interface 5 are located at the side end of the shell.
[0028] Furthermore, the measuring body 1 also includes a touch display screen.
[0029] like Figure 1As shown, the portable 12-lead ECG signal monitoring system of the present invention is composed of a data acquisition module, a computing platform, a data transmission module, and a 12-lead ECG signal reconstruction algorithm; the computing platform can run the 12-lead ECG signal reconstruction algorithm and switch the measurement mode according to the user's selection, including 24-hour monitoring and short-term measurement; the data transmission module can connect to the user's smart terminal device via Bluetooth or WiFi and transmit the final 12-lead ECG signal, and can also transmit the signal to the cloud via MQTT to realize remote monitoring and diagnosis by doctors; like Figure 2 As shown, the data acquisition module can perform noise reduction processing on the II lead ECG signal. The specific structure includes a preamplifier, a low-pass filter and a high-pass filter. The function of the high-pass filter is to filter out the baseline drift caused by breathing and other reasons during the measurement process, and the function of the low-pass filter is to filter out noise signals such as electromyographic signals and power frequency interference; like Figure 3 As shown, the 12-lead ECG signal reconstruction algorithm model can automatically output a 12-lead ECG signal with the same time length as the input signal according to different modes selected by the controller; like Figure 4 As shown, the measuring subject 1 is worn in a neck-hanging manner and the electrode sheet 2 is connected to the measuring subject 1 via a universal signal output interface; like Figure 5 As shown, the housing of the measuring body 1 is composed of a touch screen display 3, a charging interface 4 and a data acquisition interface 5. The charging interface 4 can specifically be a Type-c interface, a Micro USB interface, etc.
[0030] The input signal used by the monitoring system in this embodiment is: the II lead ECG in the standard 12-lead ECG, and the number of electrodes required is only 3. The monitoring system in this embodiment reconstructs the 12-lead ECG using a unified method based on deep learning. In this article, the model is only composed of a residual module and a gated attention mechanism, and the number of parameters is much smaller than the existing technology; the model in this embodiment can be deployed in embedded devices to ensure that the device can obtain a complete 12-lead ECG signal while being portable, which has obvious advantages. The measurement main body device is very small and can be worn without feeling. At the same time, the 12-lead ECG signal reconstruction algorithm can realize the conversion of II lead to 12-lead ECG signals with high precision.
[0031] Example 2 like Figure 8 As shown, this embodiment provides a portable 12-lead ECG signal monitoring method, comprising the following steps: S1: The user wears the measuring body 1, places the electrode sheet 2 in a proper position and connects the measuring output end of the electrode sheet 2 to the measuring body 1, wherein the measuring body 1 includes the portable 12-lead ECG signal monitoring system in Example 1; the measuring body 1 can be worn by hanging it around the neck through a rope; S2: Start the measuring subject 1 and select the measuring method; S3: The data acquisition module collects ECG signals and automatically filters them, and then resamples them through the computing platform at a sampling frequency of 100 Hz. The computing platform can provide basic communication functions, obtain analog signals output by the data acquisition module through the AD module, and run deep learning algorithms to reconstruct 12-lead ECG signals. S4: The collected II-lead ECG signal with a frequency of 100Hz is used as the input of the 12-lead ECG signal reconstruction algorithm. The signal is processed through a series of convolution and maximum pooling operations in the downsampling module to gradually reduce the spatial resolution of the feature map, leaving higher-level ECG signal features. In the upsampling part, the transposed convolution operation is performed in conjunction with the gated attention mechanism of the jump connection part to restore the spatial resolution, and the features of different levels are fused to finally reconstruct the 12-lead ECG signal from the II-lead ECG signal. The final result is as follows: Fig. 9 As shown; S5: The data transmission module feeds back the 12-lead ECG signal to the user via wireless communication.
[0032] Furthermore, the data acquisition module performs noise reduction processing through a preamplifier, a low-pass filter and a high-pass filter to remove noise such as baseline drift, electromyographic signal and power frequency interference in the electrocardiogram signal.
[0033] An intelligent terminal device can be introduced to receive information from the monitoring system through a data transmission module. The intelligent terminal device is only used to receive signals, and the 12-lead electrocardiogram reconstruction algorithm runs on the computing platform 7 of the measurement subject.
[0034] The above description is only a preferred embodiment of the present invention, and therefore cannot be used to limit the scope of the present invention. That is, equivalent changes and modifications made according to the patent scope of the present invention and the contents of the specification should still fall within the scope of the present invention.
Claims
1. A portable 12-lead ECG signal monitoring system, characterized in that: The monitoring system obtains the II lead ECG signal by measuring the subject, and the signal sampling frequency is 100 Hz; The 12-lead ECG signal reconstruction method based on the improved U-Net network, residual module and gated attention mechanism takes the II-lead ECG signal obtained by the measuring subject as input, and gradually reduces the spatial resolution of the feature map by downsampling the II-lead ECG signal multiple times, leaving higher-level ECG signal features, and then restores the signal spatial resolution in the upsampling part, and outputs a 12-lead ECG signal. The monitoring system receives the II-lead ECG signal of any time length as input and outputs the 12-lead ECG signal of the corresponding time length; Each downsampling layer in the improved U-Net network is composed of a 3×3 convolution layer, a batch normalization layer, a Parametric ReLU activation function, and a residual module; The part connecting different downsampling levels is composed of Parametric ReLU activation function and maximum pooling layer; The bottom layer of the U-Net network consists of a 3×3 convolutional layer overlapped twice, a batch normalization layer, a Parametric ReLU activation function, and a transposed convolutional layer; Each upsampling layer consists of a connection layer, a 3×3 convolution layer, a batch normalization layer, a Parametric ReLU activation function, and a residual module; The part connecting different upsampling levels is composed of Parametric ReLU activation function and transposed convolution layer; the corresponding layers of downsampling and upsampling are connected through gated attention mechanism, and the number of network layers is 5.
2. The portable 12-lead ECG signal monitoring system according to claim 1, characterized in that: The residual module consists of a 3×3 convolutional layer, a batch normalization layer, a Parametric ReLU activation function, a 3×3 convolutional layer, and a skip connection that adds the input directly to the output.
3. The portable 12-lead ECG signal monitoring system according to claim 1, characterized in that: The gated attention mechanism is added between the layers corresponding to the downsampling and upsampling parts to remove unimportant features, thereby achieving adaptive modification of feature weights.
4. The portable 12-lead ECG signal monitoring system according to claim 1, characterized in that: The measuring body comprises a housing, a measuring module, a power supply module, a data acquisition module, a computing platform, a data transmission module and a PCB. The power supply module, the data acquisition module, the computing platform and the data transmission module are welded on the PCB and fixed in the housing.
5. The portable 12-lead ECG signal monitoring system according to claim 4, characterized in that: The measuring module is an electrode sheet, and the measuring body is connected with the electrode sheet to measure the II lead electrocardiogram signal.
6. The portable 12-lead ECG signal monitoring system according to claim 5, characterized in that: The number of the electrode sheets is three.
7. The portable 12-lead ECG signal monitoring system according to claim 4, characterized in that: The shell is provided with a charging interface, a universal data acquisition interface, and a power supply module, and the charging interface and the universal data acquisition interface are located at the side end of the shell.
8. The portable 12-lead ECG signal monitoring system according to claim 4, characterized in that: The measuring body also includes a touch display screen.
9. A portable 12-lead ECG signal monitoring method, characterized in that: The following steps are involved: S1: The user wears the measuring body, places the electrode sheet in an appropriate position and connects the measuring output end of the electrode sheet to the measuring body, wherein the measuring body comprises the portable 12-lead ECG signal monitoring system according to any one of claims 1 to 8; S2: Start the measurement subject and select the measurement method; S3: The data acquisition module collects ECG signals and automatically filters them, and then resamples them through the computing platform, with a signal sampling frequency of 100 Hz; S4: The collected II-lead ECG signal with a sampling frequency of 100Hz is used as the input of the 12-lead ECG signal reconstruction algorithm. The signal is processed through a series of convolution and maximum pooling operations in the downsampling module to gradually reduce the spatial resolution of the feature map, and finally leave the ECG signal features at a higher level. In the upsampling part, the transposed convolution operation is performed in conjunction with the gated attention mechanism of the jump connection part to restore the spatial resolution, and the features at different levels are fused, finally realizing the reconstruction of the 12-lead ECG signal from the II-lead ECG signal; S5: The data transmission module feeds back the 12-lead ECG signal to the user via wireless communication.
10. The portable 12-lead ECG signal monitoring method according to claim 9, characterized in that: The data acquisition module performs noise reduction processing through a preamplifier, a low-pass filter and a high-pass filter to remove baseline drift, electromyographic signal and power frequency interference noise in the electrocardiogram signal.
Citation Information
Patent Citations
Fast magnetic resonance imaging method based on residual U-net convolutional neural network
CN109993809A
Rapid magnetic resonance imaging method based on AR2 U-Net neural network
CN111028306A
Detection of cardiac conditions from reduced lead set ECG
WO2023196651A1
