A blood pressure measurement model training method and a pulse blood pressure measurement system

Abstract

This invention provides a blood pressure measurement model training method and a pulse blood pressure measurement system. A pulse wave measurement module obtains pulse wave signals. The pulse wave from the user's wrist acts on the pressure of a pressurization module to obtain pressure change signals. A signal denoising module denoises the pulse wave signals and pressure change signals before transmitting them to the blood pressure measurement module. The blood pressure measurement module executes a blood pressure measurement model training method, training an initial neural network including a time-series prediction network, a feature extraction network, and a decoding network using pulse wave signal datasets and pressure change signal datasets. The time-series prediction network and feature extraction network output personal vascular features and high-dimensional pulse features, respectively. The decoding network outputs blood pressure data prediction results. The initial neural network model is iteratively updated to minimize the deviation between the predicted and actual blood pressure data, thus obtaining a blood pressure measurement model. Blood pressure data is obtained through the blood pressure measurement model, enabling continuous blood pressure measurement.

Description

Technical Field

[0001] This invention relates to the field of blood pressure and pulse measurement technology, and in particular to a blood pressure measurement model training method and a pulse blood pressure measurement system. Background Technology

[0002] In today's society, people pay insufficient attention to blood pressure management, the measurement process is cumbersome, and the quality management of blood pressure is inadequate. The general public usually does not measure or rarely measures their blood pressure, resulting in low awareness, treatment, and control rates of hypertension. The overall prevalence of hypertension remains high and is increasing year by year. With the continuous advancement of medical technology, the application of blood pressure in clinical monitoring is receiving increasing attention. Traditional intermittent blood pressure measurement can no longer meet the needs of accurately monitoring the dynamic changes in patients' blood pressure during treatment. Ambulatory blood pressure measurement has gradually become an important part of clinical practice because it can provide more detailed and accurate data, blood pressure change trends, and help identify potential blood pressure problems.

[0003] In current medical practice, ambulatory blood pressure monitoring uses a pressurized measurement method. This involves inflating a cuff to create pressure in the artery and then releasing the pressure to measure blood pressure. This method suffers from excessive pressure and noise during inflation, which can interfere with users' daily lives and work, making comfortable monitoring difficult and impacting user experience. Furthermore, measurements are sometimes inaccurate in hospital settings. To address the challenge of comfortable, continuous blood pressure monitoring, researchers have proposed a non-contact method. While this improves user comfort, it still has drawbacks. First, it cannot provide continuous measurement; patients must maintain a specific posture within a designated area, making monitoring impossible while the patient is moving. Second, the measurement range is limited and accuracy is low. Wireless measurement is susceptible to signal attenuation, with accuracy dropping significantly beyond a certain distance. Even within the device's operating range, its accuracy is significantly lower than that of medical blood pressure monitors, with a low accuracy rate meeting international standard errors, resulting in relatively unreliable results. Therefore, it cannot currently replace medical devices. Finding a portable, comfortable, and continuous method to monitor blood pressure changes remains a significant challenge for public health. Summary of the Invention

[0004] In view of this, embodiments of the present invention provide a blood pressure measurement model training method and a pulse blood pressure measurement system to eliminate or improve one or more defects existing in the prior art, and solve the problem that the prior art cannot easily and comfortably measure dynamic blood pressure continuously.

[0005] One aspect of the present invention provides a method for training a blood pressure measurement model, the method comprising the following steps:

[0006] A training sample set is obtained, which includes a pulse wave signal dataset and a pressure change signal dataset obtained by applying pressure to the user's wrist using the pulse wave of the user's wrist. The corresponding blood pressure measurement results obtained by applying pressure throughout the entire blood pressure measurement are added as labels to the samples in the training sample set. Applying pressure throughout the entire blood pressure measurement means applying low-pressure and high-pressure signals to the user's wrist using a blood pressure measuring device throughout the blood pressure measurement process.

[0007] An initial neural network model is obtained, comprising a temporal prediction network, a feature extraction network, and a decoding network. The temporal prediction network takes the pulse wave signal dataset and the pressure change signal dataset as input, and outputs a predicted pulse wave signal and a predicted pressure change signal throughout the entire blood pressure measurement process, and obtains personal vascular features according to a preset calculation model. The feature extraction network takes the pulse wave signal dataset as input and outputs high-dimensional pulse features. The personal vascular features and the high-dimensional pulse features are fused and then input into the decoding network. The decoding network learns the relationship between the personal vascular features, the high-dimensional pulse features, and the blood pressure data, and outputs the user's blood pressure data prediction result.

[0008] The initial neural network model is updated by iteratively updating its parameters to minimize the deviation between the predicted and actual blood pressure data, thereby obtaining a blood pressure measurement model.

[0009] In some embodiments, the time-series prediction network includes a sequence block layer, a first attention and normalization layer, a first feedforward fully connected and normalization layer, and a fully connected mapping layer arranged sequentially; the inputs of the feedforward fully connected and normalization layer and the inputs of the fully connected mapping layer establish residual short connections.

[0010] In some embodiments, the feature extraction network includes a convolutional layer, a second attention and normalization layer, a second feedforward fully connected and normalization layer, a third attention and normalization layer, a third feedforward fully connected and normalization layer, a fourth attention and normalization layer, a fourth feedforward fully connected and normalization layer, and a category label extraction layer, all arranged sequentially. The input and output of the second attention and normalization layer establish residual short connections. The input and output of the second feedforward fully connected and normalization layer establish residual short connections. The input and output of the third attention and normalization layer establish residual short connections. The input and output of the third feedforward fully connected and normalization layer establish residual short connections. The input and output of the fourth attention and normalization layer establish residual short connections. The input and output of the fourth feedforward fully connected and normalization layer establish residual short connections.

[0011] In some embodiments, the decoding network includes a first fully connected layer, a first nonlinear activation function layer, a second fully connected layer, a second nonlinear activation function layer, and a third fully connected layer arranged sequentially.

[0012] In some embodiments, the decoding network employs a convolutional neural network, a multilayer perceptron, or a deep learning model.

[0013] In some embodiments, fusing the individual vascular features and the high-dimensional pulse features includes:

[0014] The personal vascular features and the high-dimensional pulse features are mapped to the same feature space through a preset mapping function; the personal vascular features include vascular diameter, vascular wall thickness, blood flow velocity, and vascular compliance parameters; the high-dimensional pulse features include pulse wave waveform features, pulse wave conduction features, pulse wave frequency features, pulse wave dynamic features, and pulse wave pulse features.

[0015] The individual vascular features and the high-dimensional pulse features, which are mapped to the same feature space, are weighted according to feature importance and directly concatenated to obtain the fused features.

[0016] In some embodiments, the method further includes: inputting the pulse wave signal dataset and the pressure change signal dataset into a time-series prediction network to obtain the predicted pulse wave signal and the full-range pressure change signal; the predicted pulse wave signal responds to the full-range pressure change signal to obtain the pulse wave change signal; and obtaining personal vascular characteristics according to a preset calculation model.

[0017] On the other hand, the present invention also provides a pulse blood pressure measurement system, the system being deployed on a wrap-around device worn on the wrist, the system comprising:

[0018] The pulse wave measurement module includes a PPG sensor assembly and a filter assembly, which are used to acquire the PPG signal of the user's wrist and obtain the pulse wave signal after filtering by the filter assembly;

[0019] The pressurization module includes a miniature air pump assembly and a high-precision air pressure sensor, which is used to control the pressure applied to the user's wrist and to detect the pressure change signal obtained by the pulse wave signal acting on the pressure through the high-precision air pressure sensor.

[0020] The signal denoising module receives the pulse wave signal and the pressure change signal and performs denoising processing through a microcontroller.

[0021] The signal transmission module includes a Bluetooth component and a wireless transmission component, used to transmit the noise-reduced pulse wave signal and the pressure change signal to the blood pressure measurement module;

[0022] A blood pressure measurement module, wherein the module is used to execute any of the above-described blood pressure measurement model training methods to obtain a blood pressure measurement model, and to obtain blood pressure data through the pulse wave signal and the pressure change signal;

[0023] The storage module is used to store the blood pressure data, the denoised pulse wave signal, and the pressure change signal.

[0024] In some embodiments, collecting the user's PPG signal includes:

[0025] The light-emitting diodes deployed in the PPG sensor emit light of a specific wavelength toward the skin of the user's wrist; the specific wavelength of light is able to penetrate the skin;

[0026] After the light penetrates the skin, some of the light is absorbed by arteries, veins and other tissues, while the unabsorbed light is reflected to the photodiode deployed in the PPG sensor.

[0027] When the blood volume in the artery increases, the light intensity detected by the photodiode decreases; when the blood volume in the artery decreases, the light intensity detected by the photodiode increases.

[0028] The photodiode converts the change in light intensity into a PPG signal.

[0029] In some embodiments, denoising via a microcontroller includes reducing noise in the pulse wave signal and the pressure change signal through beamforming, wavelet transform, and variational mode decomposition filtering.

[0030] The beneficial effects of the present invention are at least as follows:

[0031] In the blood pressure measurement model training method and pulse blood pressure measurement system described in this invention, the pulse wave measurement module obtains the pulse wave signal. The pulse wave from the user's wrist acts on the pressure of the pressurization module to obtain the pressure change signal. The pressurization module precisely controls the pressure applied to the user's wrist so that the pressure level does not cause discomfort to the user. The blood pressure measurement module executes the blood pressure measurement model training method, training an initial neural network including a time-series prediction network, a feature extraction network, and a decoding network through the pulse wave signal dataset and the pressure change signal dataset. The time-series prediction network and the feature extraction network output personal vascular features and high-dimensional pulse features, respectively. The time-series prediction network obtains the predicted pulse wave signal and the pressure change signal throughout the measurement process, reducing the discomfort and damage caused by prolonged pressure on the wrist during blood pressure measurement, making it suitable for all-weather blood pressure measurement. After fusion, the blood pressure data prediction result is output through the decoding network. The deviation between the blood pressure data prediction result and the actual result is minimized to iteratively update the parameters of the initial neural network model to obtain the blood pressure measurement model. The fusion of personal vascular features and high-dimensional pulse features enables accurate and robust blood pressure measurement and the acquisition of blood pressure data.

[0032] Furthermore, this invention denoises the pulse wave signal and pressure change signal by using beamforming, wavelet transform, and variational mode decomposition filtering to reduce noise in the pulse wave signal and pressure change signal, thereby improving the clarity and accuracy of the signal and providing a high-quality input signal for blood pressure measurement.

[0033] Additional advantages, objects, and features of the invention will be set forth in part in the description which follows, and will also become apparent in part to those skilled in the art upon studying the description, or may be learned by practice of the invention. The objects and other advantages of the invention can be realized and obtained by means of the structures specifically pointed out in the description and drawings.

[0034] Those skilled in the art will understand that the objectives and advantages achievable with the present invention are not limited to those specifically described above, and that the above and other objectives achievable with the present invention will become clearer from the following detailed description. Attached Figure Description

[0035] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this application, are not intended to limit the scope of the invention. In the drawings:

[0036] Figure 1 This is a schematic flowchart of a blood pressure measurement model training method according to an embodiment of the present invention.

[0037] Figure 2 This is a schematic diagram of the blood pressure measurement model according to an embodiment of the present invention.

[0038] Figure 3 This is a schematic diagram of the pulse blood pressure measurement system according to an embodiment of the present invention. Detailed Implementation

[0039] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the embodiments and accompanying drawings. Here, the illustrative embodiments and descriptions of this invention are used to explain the invention, but are not intended to limit the invention.

[0040] It should also be noted that, in order to avoid obscuring the invention with unnecessary details, only the structures and / or processing steps closely related to the solution according to the invention are shown in the accompanying drawings, while other details that are not closely related to the invention are omitted.

[0041] It should be emphasized that the term "including / comprises" as used herein refers to the presence of a feature, element, step, or component, but does not exclude the presence or addition of one or more other features, elements, steps, or components.

[0042] It should also be noted that, unless otherwise specified, the term "connection" in this article can refer not only to a direct connection, but also to an indirect connection involving an intermediary.

[0043] In the following description, embodiments of the invention will be illustrated with reference to the accompanying drawings. In the drawings, the same reference numerals represent the same or similar parts, or the same or similar steps.

[0044] In existing technologies, ambulatory blood pressure monitoring uses a pressurized measurement method, which inflates the cuff to create pressure in the artery and then releases the pressure to measure blood pressure. This method suffers from excessive pressure and noise during inflation, interfering with the user's normal work and life, making comfortable monitoring and impacting the user experience. While non-contact measurement methods improve user comfort, they still suffer from drawbacks such as inability to perform continuous measurements, a small measurement range, and low accuracy. They cannot monitor while the patient is moving, and accuracy decreases with distance. Even within the device's operating range, its measurement accuracy is significantly lower than that of medical blood pressure measuring devices, with a low accuracy rate meeting international standard errors, resulting in relatively unreliable results. Therefore, it cannot currently replace medical equipment. This invention proposes a blood pressure measurement model. The training method and pulse blood pressure measurement system involve a pulse wave measurement module that obtains pulse wave signals. The pulse wave from the user's wrist acts on the pressure of the pressurization module to obtain pressure change signals. The pulse wave signal and pressure change signal are denoised by a signal denoising module and then transmitted to the blood pressure measurement module. The blood pressure measurement module executes a blood pressure measurement model training method, training an initial neural network including a time-series prediction network, a feature extraction network, and a decoding network using the pulse wave signal dataset and the pressure change signal dataset. The time-series prediction network and the feature extraction network output personal vascular features and high-dimensional pulse features, respectively. After fusion, these features are passed through the decoding network to output blood pressure data prediction results. The initial neural network model is iteratively updated to minimize the deviation between the blood pressure data prediction results and the actual results, thereby obtaining a blood pressure measurement model. Blood pressure data is obtained through the blood pressure measurement model.

[0045] Figure 1 This is a schematic flowchart of a blood pressure measurement model training method according to an embodiment of the present invention. Figure 2 This is a schematic diagram of the structure of a pressurized blood pressure measurement model according to an embodiment of the present invention. Specifically, one aspect of the present invention provides a blood pressure measurement model training method, which includes the following steps S101 to S103:

[0046] Step S101: Obtain a training sample set, which includes a pulse wave signal dataset and a pressure change signal dataset obtained by applying pressure to the user's wrist using the pulse wave of the user's wrist; add the corresponding blood pressure measurement results obtained after full-range pressure as labels to the samples in the training sample set; full-range pressure means applying low-pressure and high-pressure signals to the user's wrist using the blood pressure measuring device throughout the blood pressure measurement process.

[0047] Step S102: Obtain the initial neural network model, which includes a temporal prediction network, a feature extraction network, and a decoding network. The temporal prediction network takes the pulse wave signal dataset and the pressure change signal dataset as input, and outputs the predicted pulse wave signal and the predicted pressure change signal throughout the entire blood pressure measurement process, and obtains personal vascular features according to the preset calculation model. The feature extraction network takes the pulse wave signal dataset as input and outputs high-dimensional pulse features. The personal vascular features and high-dimensional pulse features are fused and then input into the decoding network. The decoding network learns the relationship between the personal vascular features, high-dimensional pulse features, and blood pressure data, and outputs the user's blood pressure data prediction result.

[0048] Step S103: The parameters of the initial neural network model are iteratively updated by minimizing the deviation between the predicted and actual blood pressure data, thereby obtaining the blood pressure measurement model.

[0049] In step S101, when the pulse wave of the user's wrist acts on the pressure applied to the user's wrist to obtain the pressure change signal dataset, the pressure applied to the user's wrist is low pressure. This avoids affecting the user's experience due to excessive pressure and long application time during long-term blood pressure measurement, making it difficult to perform measurements comfortably for extended periods. The labels of the training samples are the measurement results obtained by the blood pressure measuring device through the entire pressurization process. The blood pressure measuring device of this invention adopts the gold standard blood pressure device, and the entire pressurization process is a process of gradually increasing the pressure applied to the user's wrist from low pressure to high pressure.

[0050] In step S102, an initial neural network model for measuring blood pressure is trained. The time-series prediction network is input with the applied pulse wave signal and pressure change signal. The pulse wave of the user's wrist affects the pressure applied to the wrist to obtain the pressure change signal. The pressure is applied and controlled by the pressurization module. The pulse wave signal is obtained by the pulse wave measurement module and processed by the signal denoising module. The time-series prediction network obtains the pulse wave change signal by the response of the pulse wave signal to the pressure change signal. The pulse wave signal response to the pressure change signal throughout the entire pressurization process is analyzed to understand the interaction between blood vessels and pressure, and to calculate the user's personal vascular characteristics.

[0051] In some embodiments, the time-series prediction network includes a sequence block layer, a first attention and normalization layer, a first feedforward fully connected and normalization layer, and a fully connected mapping layer arranged sequentially; residual short connections are established between the inputs of the feedforward fully connected and normalization layer and the inputs of the fully connected mapping layer.

[0052] Specifically, the sequence segmentation layer divides the input pulse wave signal and pressure change signal into segments, improving the model's ability to understand and express data features, while reducing computational complexity to improve training efficiency. Furthermore, the sequence is segmented at different levels to enhance the model's predictive ability and accuracy. Finally, the input pulse wave signal and pressure change signal are passed through a fully connected mapping layer, and the predicted pulse wave signal and the entire pressure change signal are obtained through an autoregressive prediction mechanism.

[0053] In some embodiments, the method further includes: inputting a pulse wave signal dataset and a pressure change signal dataset into a time-series prediction network to obtain a predicted pulse wave signal and a full-range pressure change signal; the predicted pulse wave signal responds to the full-range pressure change signal to obtain a pulse wave change signal; and personal vascular characteristics are obtained according to a preset calculation model.

[0054] Furthermore, in some embodiments, fusing individual vascular features and high-dimensional pulse features includes steps S1021-S1022:

[0055] Step S1021: Map personal vascular features and high-dimensional pulse features to the same feature space through a preset mapping function; personal vascular features include vascular diameter, vascular wall thickness, blood flow velocity and vascular compliance parameters; high-dimensional pulse features include pulse wave waveform features, pulse wave conduction features, pulse wave frequency features, pulse wave dynamic features and pulse wave pulse features.

[0056] Step S1022: The personal vascular features and high-dimensional pulse features mapped to the same feature space are weighted according to feature importance and directly spliced ​​to obtain the fused features.

[0057] Specifically, pulse wave variation signals are obtained by using the predicted pulse wave signal output by the time-series prediction network and the full-process pressure change signal, and the influence of different pressurization conditions on the pulse wave signal is analyzed. By measuring the speed at which the pulse wave signal propagates in the blood vessels, the vessel diameter, vessel wall thickness, blood flow velocity, and vessel compliance parameters are calculated. The changes in vessel diameter are analyzed by the periodic changes of the pulse wave signal. The pulse wave waveform characteristics are analyzed to determine the peak time, peak width, and peak area of ​​the pulse wave. The pulse wave conduction characteristics are analyzed to determine the conduction time and speed of the pulse wave signal in the blood vessels. The pulse wave frequency characteristics are analyzed to determine the heart rate, pulse wave count, and frequency distribution. The pulse wave dynamic characteristics are analyzed to determine the trend and fluctuation of the pulse wave over time. The pulse wave pulse characteristics are analyzed to determine the amplitude of the pulse peak and the time interval between peaks, so as to better understand the user's pulse status.

[0058] Furthermore, the preset calculation model is a personalized vascular feature model of the user, used to accurately analyze and judge the user's personal vascular features based on the predicted pulse wave signal and the predicted pressure change signal throughout the blood pressure measurement. Mapping personal vascular features and high-dimensional pulse features to the same feature space through a preset mapping function solves the problem of differences in dimensions and distributions that occur when different features are fused. Weighting personal vascular features and high-dimensional pulse wave features is used to distinguish the importance and influence of the two features. Using a direct concatenation method to obtain fused features simplifies the processing flow and obtains new feature vectors simply and directly.

[0059] In some embodiments, the feature extraction network includes a convolutional layer, a second attention and normalization layer, a second feedforward fully connected and normalization layer, a third attention and normalization layer, a third feedforward fully connected and normalization layer, a fourth attention and normalization layer, a fourth feedforward fully connected and normalization layer, and a category label extraction layer arranged sequentially; residual short connections are established between the input and output of the second attention and normalization layer; residual short connections are established between the input and output of the second feedforward fully connected and normalization layer; residual short connections are established between the input and output of the third attention and normalization layer; residual short connections are established between the input and output of the third feedforward fully connected and normalization layer; residual short connections are established between the input and output of the fourth attention and normalization layer; residual short connections are established between the input and output of the fourth feedforward fully connected and normalization layer.

[0060] Specifically, the feature extraction network converts the pulse wave signal into a data representation that reflects the essential characteristics of the pulse wave signal. The high-dimensional pulse features contain detailed information about the pulse wave signal for analysis and processing. Furthermore, after the input pulse wave signal is processed through a convolutional layer, learnable category labels are added to identify and distinguish features of different categories, thereby improving the model's recognition accuracy and classification ability. By learning different category labels, richer and more accurate information support is provided for subsequent decision-making and prediction. When a softmax function layer is set after the category label extraction layer, the probability distribution of each category can be output to perform multi-class classification tasks and obtain classification results.

[0061] In some embodiments, the decoding network includes a first fully connected layer, a first nonlinear activation function layer, a second fully connected layer, a second nonlinear activation function layer, and a third fully connected layer arranged sequentially. In some embodiments, the decoding network employs a convolutional neural network, a multilayer perceptron, or a deep learning model. The decoding network structure perceptually fuses personal vascular features and high-dimensional pulse features to calculate the user's blood pressure data, generating accurate blood pressure measurement results.

[0062] In step S103, by minimizing the predicted result and the actual result and continuously updating the parameters of the initial neural network model, the blood pressure measurement model obtained by continuously adjusting the parameters can more accurately predict blood pressure data and improve the efficiency of blood pressure data measurement.

[0063] Figure 3 This is a schematic diagram of the structure of a pulse blood pressure measurement system according to an embodiment of the present invention. On the other hand, the present invention also provides a pulse blood pressure measurement system, which is deployed on a wrap-around device worn on the user's wrist, and the system includes:

[0064] The pulse wave measurement module includes a PPG sensor assembly and a filtering assembly. It collects PPG signals from the user's wrist and filters them to obtain the pulse wave signal. Specifically, the PPG sensor assembly collects PPG signals, which are the portion of light emitted by a light-emitting diode (LED) that is not absorbed by the tissue and returns to a photodiode. The blood volume in the artery affects the intensity of the light returning to the photodiode; after detecting changes in light intensity, the photodetector converts this into a PPG signal. The PPG signal changes periodically with the contraction and relaxation of the heartbeat. The pulse wave signal is obtained by sensing and filtering the PPG signal. The photodiode converts the reflected light into a digital PPG signal.

[0065] In some embodiments, collecting the user's PPG signal includes steps S201 to S204:

[0066] Step S201: The light-emitting diodes deployed in the PPG sensor emit light of a specific wavelength toward the skin of the user's wrist; the specific wavelength of light is able to penetrate the skin.

[0067] Step S202: After the light penetrates the skin, some of the light is absorbed by arteries, veins and other tissues, while the unabsorbed light is reflected to the photodiode deployed in the PPG sensor.

[0068] Step S203: When the blood volume in the artery increases, the light intensity detected by the photodiode decreases; when the blood volume in the artery decreases, the light intensity detected by the photodiode increases.

[0069] Step S204: The photodiode converts the change in light intensity into a PPG signal.

[0070] The pressurization module, comprising a miniature air pump assembly and a high-precision pressure sensor, controls the pressure applied to the user's wrist and detects pressure changes by detecting the pulse wave's influence on the pressure. Specifically, the pressurization module is deployed in a wrap-around pressurization device worn on the user's wrist. This device applies wrap-around pressure to the user's wrist, controlling the pressure applied and ensuring the user's measurement experience is not affected by the pressurization process. The miniature air pump assembly provides the applied pressure to the wrist, and the high-precision pressure sensor measures the pressure changes by detecting the pulse wave's influence on the pressure.

[0071] The signal denoising module receives pulse wave signals and pressure change signals and performs denoising processing via a microcontroller. In some embodiments, denoising processing via the microcontroller includes reducing noise in the pulse wave signals and pressure change signals through beamforming, wavelet transform, and variational mode decomposition filtering. Specifically, the denoising process yields high-definition and high-precision pulse wave signals and pressure change signals, providing high-quality input signals for the blood pressure measurement module during blood pressure measurement.

[0072] The signal transmission module, including a Bluetooth component and a wireless transmission component, is used to transmit the noise-reduced pulse wave signal and pressure change signal to the blood pressure measurement module.

[0073] The blood pressure measurement module executes any of the above-mentioned blood pressure measurement model training methods to obtain blood pressure data through pulse wave signals and pressure change signals.

[0074] The storage module stores blood pressure data, denoised pulse wave signals, and pressure change signals. Specifically, it stores these data as a vascular feature database to analyze the relationship between blood pressure, pulse, and externally applied pressure. Data from different scenarios is integrated and processed for later review, analysis, and decision-making support, thereby optimizing the measurement process.

[0075] The present invention will now be described with reference to a specific embodiment:

[0076] This invention proposes a method for accurately predicting a user's blood pressure using a pulse blood pressure measurement system to calculate individual vascular characteristics. The first challenge is designing the pulse blood pressure measurement system, which employs a PPG sensor to collect pulse wave signals and a compression module to apply compression pressure to the user's wrist using a wrap-around compression device. The core technology then focuses on how to use the pulse blood pressure measurement system to calculate individual vascular characteristics and accurately measure blood pressure. A time-series prediction network is designed to predict the pulse wave signal throughout the compression period and the pulse wave changes in response to pressure, thereby obtaining individual vascular characteristics. The pulse wave signal is then processed through a feature extraction network to obtain high-dimensional pulse features. Finally, the individual vascular characteristics and the high-dimensional pulse features are input into a decoding network to achieve accurate and robust blood pressure measurement.

[0077] 1. Overall procedure for pulse and blood pressure measurement.

[0078] (1) Data was collected by wearing the wrap-around device on the subject's wrist.

[0079] (2) The collected signals are processed and the pulse wave signal is obtained by filtering.

[0080] (3) Input the pulse wave signal during the pressurization process into the time-series prediction network to obtain individual vascular characteristics.

[0081] (4) Input the pulse wave signal and personal vascular characteristics into the decoding network to measure blood pressure.

[0082] 2. The pulse and blood pressure measurement system includes a pulse measurement module, a pressurization module, a signal denoising module, a signal transmission module, a blood pressure measurement module, and a storage module.

[0083] The pulse measurement module uses a PPG sensor to detect changes in blood volume within the blood vessels, collecting rich and detailed data from the subject's wrist. This data will then be further fed into a signal denoising module, providing a high-quality input source for subsequent neural network processing.

[0084] The pressurization module includes a miniature air pump and a high-precision air pressure sensor to precisely control the pressure applied to the user. The pressure change signal obtained by the wrist artery in response to the pressure is ultimately detected by the high-precision air pressure sensor and further sent to the signal denoising module. The microcontroller performs data synchronization and processing to prepare for the next step of blood pressure measurement.

[0085] The signal denoising module uses a microcontroller integrated on the wristband and employs signal enhancement technology. It reduces noise in the sensed signal through advanced signal processing methods, including but not limited to beamforming, wavelet transform, and variational mode decomposition filtering, thereby improving the clarity and accuracy of the signal.

[0086] The signal transmission module enables communication between the signal and related subsequent tools, including a Bluetooth module and a wireless transmission module, to further transmit the measured signal to the subsequent blood pressure measurement module.

[0087] The blood pressure measurement module executes a blood pressure measurement model training method to obtain a blood pressure measurement model, and obtains blood pressure data through pulse wave signals and pressure change signals.

[0088] The storage module is used to store blood pressure data, noise-reduced pulse wave signals, and pressure change signals.

[0089] 3. Blood pressure measurement incorporates pulse wave signals during the inflation process. Through the collaborative work of a temporal prediction network, a feature extraction network, and a decoding network, it achieves efficient representation and utilization of physiological signals. First, a temporal prediction network is defined to obtain the pulse wave change signal after the entire inflation process using the pulse wave signal under inflation. By inputting the pulse wave signal's response to the inflation and depressurization processes, the network obtains the predicted pulse wave signal's response to the complete inflation process, thereby calculating the user's individual vascular characteristics.

[0090] Furthermore, a feature extraction network encodes the user's pulse wave signal, extracting high-dimensional features related to hemodynamics. These hemodynamic features are then concatenated with individual vascular features to obtain a reliable and robust high-dimensional representation. Finally, the decoding network consists of three fully connected layers separated by nonlinear activation functions. The decoding network fully perceives the extracted fused features and calculates the user's blood pressure value. This end-to-end processing flow generates more accurate measurement results for each user, organically combining personal information with physiological signals, thus improving the overall performance and applicability of the blood pressure measurement model. During the training of the blood pressure measurement model, measurement results from the gold standard blood pressure device are used as sample labels. The blood pressure measurement process of the blood pressure measurement model includes:

[0091] (1) Input pulse wave signal and pressure change signal into time series prediction network to obtain personal vascular characteristics.

[0092] (2) Input the pulse wave signal into the feature extraction network to obtain high-dimensional pulse features.

[0093] (3) Combine individual vascular characteristics with high-dimensional pulse characteristics.

[0094] (4) Predict user blood pressure data by decoding network.

[0095] In summary, this invention provides a blood pressure measurement model training method and a pulse blood pressure measurement system. A pulse wave measurement module obtains a pulse wave signal. The pulse wave from the user's wrist acts on the pressure of a pressurization module to obtain a pressure change signal. A signal denoising module denoises the pulse wave signal and the pressure change signal and transmits them to the blood pressure measurement module. The blood pressure measurement module executes a blood pressure measurement model training method, training an initial neural network using a pulse wave signal dataset and a pressure change signal dataset. The initial neural network model includes a temporal prediction network, a feature extraction network, and a decoding network. The temporal prediction network and the feature extraction network output personal vascular features and high-dimensional pulse features, respectively. After fusion, these features are passed through the decoding network to output a blood pressure data prediction result. The deviation between the predicted and actual blood pressure data results is minimized by iteratively updating the parameters of the initial neural network model to obtain a blood pressure measurement model. Blood pressure data is obtained through this blood pressure measurement model.

[0096] Furthermore, this invention denoises the pulse wave signal and pressure change signal by using beamforming, wavelet transform, and variational mode decomposition filtering to reduce noise in the pulse wave signal and pressure change signal, thereby improving the clarity and accuracy of the signal and providing a high-quality input signal for blood pressure measurement.

[0097] This invention also provides a computer device, which may include a processor and a memory, wherein the processor and the memory may be connected via a bus or other means.

[0098] The processor can be a central processing unit (CPU). The processor can also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, or combinations of the above types of chips.

[0099] Memory, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs, non-transitory computer-executable programs, and modules, such as the program instructions / modules corresponding to the button blocking method of the vehicle display device in this embodiment of the invention. The processor executes various functional applications and data processing by running the non-transitory software programs, instructions, and modules stored in the memory.

[0100] The memory may include a program storage area and a data storage area. The program storage area may store the operating system and applications required for at least one function; the data storage area may store data created by the processor, etc. Furthermore, the memory may include high-speed random access memory and non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, the memory may optionally include memory remotely located relative to the processor, which can be connected to the processor via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0101] The one or more modules are stored in the memory, and when executed by the processor, they perform the method described in this embodiment.

[0102] This invention also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the aforementioned edge computing server deployment method. The computer-readable storage medium can be a tangible storage medium, such as random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, floppy disks, hard disks, removable storage disks, CD-ROMs, or any other form of storage medium known in the art.

[0103] Those skilled in the art will understand that the exemplary components, systems, and methods described in conjunction with the embodiments disclosed herein can be implemented in hardware, software, or a combination of both. Whether implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this invention. When implemented in hardware, it can be, for example, electronic circuits, application-specific integrated circuits (ASICs), appropriate firmware, plug-ins, function cards, etc. When implemented in software, the elements of this invention are programs or code segments used to perform the desired tasks. The programs or code segments can be stored in a machine-readable medium or transmitted over a transmission medium or communication link via data signals carried in a carrier wave.

[0104] It should be clarified that the present invention is not limited to the specific configurations and processes described above and shown in the figures. For the sake of brevity, detailed descriptions of known methods are omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of the present invention is not limited to the specific steps described and shown. Those skilled in the art can make various changes, modifications, and additions, or change the order of steps, after understanding the spirit of the present invention.

[0105] In this invention, features described and / or illustrated for one embodiment may be used in the same or similar manner in one or more other embodiments, and / or combined with or in place of features of other embodiments.

[0106] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, various modifications and variations of the embodiments of the present invention are possible. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for training a blood pressure measurement model, characterized in that, The method includes the following steps: A training sample set is obtained, which includes a pulse wave signal dataset and a pressure change signal dataset obtained by applying pressure to the user's wrist using the pulse wave of the user's wrist. The corresponding blood pressure measurement results obtained by applying pressure throughout the entire blood pressure measurement are added as labels to the samples in the training sample set. Applying pressure throughout the entire blood pressure measurement means applying low-pressure and high-pressure signals to the user's wrist using a blood pressure measuring device throughout the blood pressure measurement process. An initial neural network model is obtained, comprising a temporal prediction network, a feature extraction network, and a decoding network. The temporal prediction network takes the pulse wave signal dataset and the pressure change signal dataset as input, and outputs a predicted pulse wave signal and a predicted pressure change signal throughout the entire blood pressure measurement process, and obtains personal vascular features according to a preset calculation model. The feature extraction network takes the pulse wave signal dataset as input and outputs high-dimensional pulse features. The personal vascular features and the high-dimensional pulse features are fused and then input into the decoding network. The decoding network learns the relationship between the personal vascular features, the high-dimensional pulse features, and the blood pressure data, and outputs the user's blood pressure data prediction result. The initial neural network model is iteratively updated by minimizing the deviation between the predicted and actual blood pressure data, thereby obtaining a blood pressure measurement model. The time-series prediction network includes a sequence block layer, a first attention and normalization layer, a first feedforward fully connected and normalization layer, and a fully connected mapping layer, all set in succession. The inputs of the feedforward fully connected and normalization layer and the inputs of the fully connected mapping layer establish residual short connections. The feature extraction network comprises a series of convolutional layers, a second attention and normalization layer, a second feedforward fully connected and normalization layer, a third attention and normalization layer, a third feedforward fully connected and normalization layer, a fourth attention and normalization layer, a fourth feedforward fully connected and normalization layer, and a category label extraction layer. Residual short connections are established between the input and output of the second attention and normalization layer; residual short connections are established between the input and output of the second feedforward fully connected and normalization layer; residual short connections are established between the input and output of the third attention and normalization layer; residual short connections are established between the input and output of the third feedforward fully connected and normalization layer; residual short connections are established between the input and output of the fourth attention and normalization layer; residual short connections are established between the input and output of the fourth feedforward fully connected and normalization layer. The decoding network includes a first fully connected layer, a first nonlinear activation function layer, a second fully connected layer, a second nonlinear activation function layer, and a third fully connected layer, which are arranged sequentially. The fusion of the individual vascular features and the high-dimensional pulse features includes: The personal vascular features and the high-dimensional pulse features are mapped to the same feature space through a preset mapping function; the personal vascular features include vascular diameter, vascular wall thickness, blood flow velocity, and vascular compliance parameters; the high-dimensional pulse features include pulse wave waveform features, pulse wave conduction features, pulse wave frequency features, pulse wave dynamic features, and pulse wave pulse features. The individual vascular features and the high-dimensional pulse features mapped to the same feature space are weighted according to feature importance and directly concatenated to obtain the fused features; The method further includes: inputting the pulse wave signal dataset and the pressure change signal dataset into a time-series prediction network to obtain the predicted pulse wave signal and the full-range pressure change signal; the predicted pulse wave signal responds to the full-range pressure change signal and obtains the pulse wave change signal; and obtaining individual vascular characteristics according to a preset calculation model.

2. The blood pressure measurement model training method according to claim 1, characterized in that, The decoding network employs a convolutional neural network, a multilayer perceptron, or a deep learning model.

3. A pulse blood pressure measurement system, characterized in that, The system is deployed on a wraparound device worn on the user's wrist, and the system includes: The pulse wave measurement module includes a PPG sensor assembly and a filter assembly, which are used to acquire the PPG signal of the user's wrist and obtain the pulse wave signal after filtering by the filter assembly; The pressurization module includes a miniature air pump assembly and a high-precision air pressure sensor, which is used to control the pressure applied to the user's wrist and detect pressure change signals obtained by the pulse wave acting on the pressure through the high-precision air pressure sensor. The signal denoising module receives the pulse wave signal and the pressure change signal and performs denoising processing through a microcontroller. The signal transmission module includes a Bluetooth component and a wireless transmission component, used to transmit the noise-reduced pulse wave signal and the pressure change signal to the blood pressure measurement module; A blood pressure measurement module, wherein the module is used to execute the blood pressure measurement model training method as described in any one of claims 1 to 2 to obtain a blood pressure measurement model, and to obtain blood pressure data through the pulse wave signal and the pressure change signal; The storage module is used to store the blood pressure data, the denoised pulse wave signal, and the pressure change signal.

4. The pulse blood pressure measurement system according to claim 3, characterized in that, The acquisition of the user's PPG signal includes: The light-emitting diodes deployed in the PPG sensor emit light of a specific wavelength toward the skin of the user's wrist; the specific wavelength of light is able to penetrate the skin; After the light penetrates the skin, some of the light is absorbed by arteries, veins and other tissues, while the unabsorbed light is reflected to the photodiode deployed in the PPG sensor. When the blood volume in the artery increases, the light intensity detected by the photodiode decreases; when the blood volume in the artery decreases, the light intensity detected by the photodiode increases. The photodiode converts the change in light intensity into a PPG signal.

5. The pulse blood pressure measurement system according to claim 3, characterized in that, Denoising processing using a microcontroller includes reducing noise in the pulse wave signal and the pressure change signal through beamforming, wavelet transform, and variational mode decomposition filtering.

Images