A blood pressure measurement method, device and readable storage medium based on calibration data

Abstract

The embodiment of the present application provides a blood pressure measurement method based on calibration data, including: obtaining the calibrated heart rate value and calibrated blood pressure value of the target user, determining the target group to which the target user belongs, collecting the target user's real-time PPG data and ECG data, and processing the PPG data and ECG data using the classification regression fitting model corresponding to the target group to obtain real-time blood pressure parameters, wherein the classification regression fitting model includes a feature extraction structure and a random forest regression fitting model. On the one hand, this solution divides the target groups by calibrating the heart rate value and blood pressure value, and applies the corresponding classification regression fitting model based on the characteristics of people with different blood pressure levels, thereby improving the targetedness and accuracy of the measurement and being able to more accurately identify the blood pressure condition; on the other hand, the comprehensive utilization of the peak area parameters of the PPG and ECG data, combined with the classification regression fitting model, effectively improves the feature extraction and processing capabilities of complex physiological data, thereby improving the accuracy and reliability of blood pressure measurement.

Description

Technical Field

[0001] The present application relates to the field of data processing, and in particular to a blood pressure measurement method, device, and readable storage medium based on calibration data. Background Art

[0002] In the healthcare sector, blood pressure measurement is an important tool for assessing cardiovascular health, playing a key role in disease diagnosis, treatment efficacy evaluation, and daily health management. By promptly understanding blood pressure fluctuations, people can better adjust their lifestyles and prevent cardiovascular disease. Therefore, the medical industry has long been committed to seeking more advanced blood pressure measurement technologies.

[0003] With the advancement of technology, combined measurement techniques based on PPG (photoplethysmography) and ECG (electrocardiogram) have gradually become an important research direction in the field of blood pressure measurement. This is because ECG signals can reflect cardiac electrophysiological activity, while PPG signals can reflect changes in peripheral vascular hemodynamics. Combining the two can comprehensively consider the interaction between the heart and blood vessels, theoretically providing more comprehensive information for blood pressure measurement.

[0004] Traditional blood pressure cuff-based blood pressure measurement methods have significant shortcomings. First, they cannot provide continuous measurement and can only capture blood pressure values ​​at a specific point in time, making it difficult to reflect dynamic changes in blood pressure. This makes them inadequate for patients who require real-time blood pressure monitoring, such as those in intensive care units and those with unstable hypertension. Second, the inflation and deflation of the cuff can cause discomfort, especially for those who require frequent blood pressure measurements. Furthermore, while PPG technology has made some progress in non-invasive blood pressure measurement, its accuracy is affected by various factors, such as the skin condition at the measurement site, ambient lighting, and individual movement. This results in unsatisfactory measurement stability, making it difficult to meet the requirements for accurate clinical diagnosis and treatment. Summary of the Invention

[0005] The present invention aims to provide a blood pressure measurement method based on calibration data, aiming to solve the problems of low data utilization and inaccurate detection in existing data. The blood pressure measurement method based on calibration data provided in this application includes:

[0006] Obtain the target user's calibrated heart rate and blood pressure values;

[0007] Determining the target group to which the target user belongs based on the relationship between the calibrated heart rate value and the calibrated blood pressure value, the target group including a low blood pressure group, a normal blood pressure group, and a high blood pressure group;

[0008] Collecting real-time PPG data and ECG data of the target user;

[0009] extracting peak area parameters of the PPG data and peak area parameters of the ECG data;

[0010] The peak area parameters of the PPG data and the peak area parameters of the ECG data are processed using a classification regression fitting model corresponding to the target group to obtain real-time blood pressure parameters, wherein the classification regression fitting model includes a feature extraction structure and a random forest regression fitting model.

[0011] Based on the blood pressure measurement method based on calibration data provided in the first aspect of the embodiment of the present application, optionally,

[0012] The feature extraction structure includes two LSTM layers and two fully connected layers.

[0013] Based on the blood pressure measurement method based on calibration data provided in the first aspect of the embodiment of the present application, optionally,

[0014] The peak area parameters of the PPG data and the peak area parameters of the ECG data are processed using the classification regression fitting model corresponding to the target group to obtain real-time blood pressure parameters including:

[0015] Inputting the peak area parameters of the PPG data and the peak area parameters of the ECG data into a feature extraction structure to obtain a plurality of potential features;

[0016] Acquire multiple representational features;

[0017] The multiple potential features and the multiple manifestation features are input into the random forest regression fitting model to obtain real-time blood pressure parameters.

[0018] Based on the blood pressure measurement method based on calibration data provided in the first aspect of the embodiment of the present application, optionally,

[0019] The classification regression fitting model includes: a classification regression fitting model corresponding to a lower blood pressure group, a classification regression fitting model corresponding to a normal blood pressure group, and a classification regression fitting model corresponding to a higher blood pressure group;

[0020] The classification regression fitting models are trained using historical data belonging to corresponding groups.

[0021] Based on the blood pressure measurement method based on calibration data provided in the first aspect of the embodiment of the present application, optionally, determining the target group to which the target user belongs based on the relationship between the calibrated heart rate value and the calibrated blood pressure value includes:

[0022] The target group to which the target user belongs is determined according to the following formula:

[0023]

[0024] To calibrate the heart rate, To calibrate diastolic blood pressure; is a constant value;

[0025] If the calibrated heart rate is within the above range, the target user is determined to belong to the normal blood pressure group;

[0026] If the calibrated heart rate meets , it is determined that the target user belongs to the lower blood pressure group;

[0027] If the calibrated heart rate meets , it is determined that the target user belongs to the high blood pressure group.

[0028] Based on the blood pressure measurement method based on calibration data provided in the first aspect of the embodiment of the present application, optionally, the appearance characteristics include: heart pressure difference value, calibrated pressure difference value, pulse wave transmission time, pulse wave transmission speed, calibrated heart rate, calibrated systolic pressure and calibrated diastolic pressure.

[0029] Based on the blood pressure measurement method based on calibration data provided in the first aspect of the embodiment of the present application, optionally, extracting the peak area parameters of the PPG data and the peak area parameters of the ECG data includes:

[0030] Before and after extracting the PPG peak point Parameters, before and after extracting the ECG peak point parameters, among which is a constant value, is the sampling rate.

[0031] A second aspect of an embodiment of the present application provides a blood pressure measurement device based on calibration data, comprising:

[0032] An acquisition unit, configured to acquire a calibrated heart rate value and a calibrated blood pressure value of a target user;

[0033] a determining unit, configured to determine a target group to which the target user belongs based on a relationship between the calibrated heart rate value and the calibrated blood pressure value, the target group comprising a low blood pressure group, a normal blood pressure group, and a high blood pressure group;

[0034] an acquisition unit, configured to acquire real-time PPG data and ECG data of the target user, and extract peak area parameters of the PPG data and peak area parameters of the ECG data;

[0035] A processing unit is used to process the peak area parameters of the PPG data and the peak area parameters of the ECG data using a classification regression fitting model corresponding to the target group to obtain real-time blood pressure parameters, wherein the classification regression fitting model includes a feature extraction structure and a random forest regression fitting model.

[0036] According to the blood pressure measurement device based on calibration data provided in the second aspect of the embodiment of the present application, optionally,

[0037] The feature extraction structure includes 2 layers of LSTM and 2 layers of fully connected layers.

[0038] According to the blood pressure measurement device based on calibration data provided in the second aspect of the embodiment of the present application, optionally,

[0039] The processing unit is specifically configured to:

[0040] Inputting the peak area parameters of the PPG data and the peak area parameters of the ECG data into a feature extraction structure to obtain a plurality of potential features;

[0041] Acquire multiple representational features;

[0042] The multiple potential features and the multiple manifestation features are input into the random forest regression fitting model to obtain real-time blood pressure parameters.

[0043] According to the blood pressure measurement device based on calibration data provided in the second aspect of the embodiment of the present application, optionally,

[0044] The classification regression fitting model includes: a classification regression fitting model corresponding to a lower blood pressure group, a classification regression fitting model corresponding to a normal blood pressure group, and a classification regression fitting model corresponding to a higher blood pressure group;

[0045] The classification regression fitting models are trained using historical data belonging to corresponding groups.

[0046] Based on the calibration data-based blood pressure measurement device provided in the second aspect of the embodiment of the present application, optionally, the determining unit is specifically configured to:

[0047] The target group to which the target user belongs is determined according to the following formula:

[0048]

[0049] To calibrate the heart rate, To calibrate diastolic blood pressure; is a constant value;

[0050] If the calibrated heart rate is within the above range, the target user is determined to belong to the normal blood pressure group;

[0051] If the calibrated heart rate meets , it is determined that the target user belongs to the lower blood pressure group;

[0052] If the calibrated heart rate meets , it is determined that the target user belongs to the high blood pressure group.

[0053] Based on the blood pressure measurement device based on calibration data provided in the second aspect of the embodiment of the present application, optionally, the appearance characteristics include: heart pressure difference value, calibrated pressure difference value, pulse wave transmission time, pulse wave transmission speed, calibrated heart rate, calibrated systolic pressure and calibrated diastolic pressure.

[0054] Based on the calibration data-based blood pressure measurement device provided in the second aspect of the embodiment of the present application, optionally, the step of the acquisition unit extracting the peak area parameters of the PPG data and the peak area parameters of the ECG data specifically includes:

[0055] Before and after extracting the PPG peak point Parameters, before and after extracting the ECG peak point parameters, among which is a constant value, is the sampling rate.

[0056] A third aspect of the embodiments of the present application provides a blood pressure measurement device based on calibration data, comprising:

[0057] CPU, memory, input and output interfaces, wired or wireless network interfaces, and power supply;

[0058] The memory is a transient storage memory or a persistent storage memory;

[0059] The central processing unit is configured to communicate with the memory and execute instruction operations in the memory on the device to perform the method described in any one of the first aspects of the embodiments of the present application.

[0060] A fourth aspect of the embodiments of the present application provides a computer-readable storage medium, comprising instructions, which, when executed on a computer, enable the computer to execute the method described in any one of the first aspects of the embodiments of the present application.

[0061] A fifth aspect of the embodiments of the present application provides a computer program product comprising instructions, which, when executed on a computer, enables the computer to execute any one of the methods described in the first aspect of the embodiments of the present application.

[0062] It can be seen from the above technical solution that the embodiment of the present application has the following advantages: the embodiment of the present application provides a blood pressure measurement method based on calibration data, including: obtaining a calibrated heart rate value and a calibrated blood pressure value of a target user; determining the target group to which the target user belongs based on the relationship between the calibrated heart rate value and the calibrated blood pressure value, the group including a low blood pressure group, a normal blood pressure group and a high blood pressure group; collecting real-time PPG data and ECG data of the target user, extracting the peak area parameters of the PPG data and the peak area parameters of the ECG data; using the classification regression fitting model corresponding to the target group to process the peak area parameters of the PPG data and the peak area parameters of the ECG data to obtain real-time blood pressure parameters, the classification regression fitting model including a feature extraction structure and a random forest regression fitting model. On the one hand, this solution divides the target groups by calibrating heart rate and blood pressure values, and uses corresponding classification and regression fitting models based on the characteristics of people with different blood pressure levels, which greatly improves the pertinence and accuracy of the measurement and can more accurately reflect the blood pressure conditions of different people. On the other hand, it comprehensively utilizes the peak area parameters of PPG and ECG data, combined with classification and regression fitting models, to fully tap the value of physiological data, effectively improve the feature extraction and processing capabilities of complex physiological data, and thus improve the accuracy and reliability of blood pressure measurement. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] To more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are merely embodiments of the present application. A person of ordinary skill in the art can also derive other drawings based on the provided drawings without inventive effort. It should be understood that the drawings provided in this section are only used to better understand the present solution and do not constitute a limitation of the present application.

[0064] Figure 1 A flow chart of an embodiment of a blood pressure measurement method based on calibration data provided in this application;

[0065] Figure 2 This is another flowchart of an embodiment of the blood pressure measurement method based on calibration data provided by this application;

[0066] Figure 3 This is an example graph showing a linear relationship between systolic and diastolic blood pressure;

[0067] Figure 4 Schematic diagram of ECG and PPG data;

[0068] Figure 5 A schematic diagram of the network architecture used in this application;

[0069] Figure 6 A structural diagram of an embodiment of a blood pressure measurement device based on calibration data provided by this application;

[0070] Figure 7 This is another structural schematic diagram of an embodiment of a blood pressure measurement device based on calibration data provided in this application. DETAILED DESCRIPTION

[0071] In order to enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application are clearly and completely described below. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work should fall within the scope of protection of this application. At the same time, for the sake of clarity and conciseness, the description of well-known functions and structures is omitted in the following description.

[0072] The terms "first," "second," "third," "fourth," and so forth (if any) in the specification and claims of this application and in the accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a particular order or precedence. It should be understood that the terms used in this manner are interchangeable where appropriate, so that the embodiments described herein can be implemented in an order other than that shown or described herein. In addition, the terms "including" and "having," and any variations thereof, are intended to cover non-exclusive inclusions. For example, a process, method, product, or apparatus comprising a series of steps or elements is not necessarily limited to those steps or elements expressly listed, but may include other steps or elements not expressly listed or inherent to such process, method, product, or apparatus.

[0073] In the healthcare sector, blood pressure measurement is an important tool for assessing cardiovascular health, playing a key role in disease diagnosis, treatment efficacy evaluation, and daily health management. By promptly understanding blood pressure fluctuations, people can better adjust their lifestyles and prevent cardiovascular disease. Therefore, the medical industry has long been committed to seeking more advanced blood pressure measurement technologies.

[0074] With the advancement of technology, combined measurement techniques based on PPG (photoplethysmography) and ECG (electrocardiogram) have gradually become an important research direction in the field of blood pressure measurement. This is because ECG signals can reflect cardiac electrophysiological activity, while PPG signals can reflect changes in peripheral vascular hemodynamics. Combining the two can comprehensively consider the interaction between the heart and blood vessels, theoretically providing more comprehensive information for blood pressure measurement.

[0075] The traditional cuff-type blood pressure measurement method has obvious shortcomings. On the one hand, it cannot achieve continuous measurement and can only obtain blood pressure values ​​at a certain point in time. It is difficult to reflect the dynamic changes in blood pressure. For patients who need real-time blood pressure monitoring, such as patients in intensive care units and patients with unstable hypertension, it cannot meet the needs; on the other hand, the inflation and deflation process of the cuff will cause discomfort to the patient, especially for those who need to measure blood pressure frequently, this discomfort will be more obvious. In addition, although PPG technology has made some progress in the field of non-invasive blood pressure measurement, its accuracy is affected by many factors, such as the skin condition of the measurement site, ambient light, individual movement, etc., resulting in unsatisfactory stability of the measurement results, making it difficult to meet the requirements of clinical accurate diagnosis and treatment.

[0076] To resolve the above issues, please refer to Figure 1 An embodiment of a blood pressure measurement method based on calibration data provided in this application includes: steps 101 to 105.

[0077] 101. Obtain the target user's calibrated heart rate value and calibrated blood pressure value.

[0078] Calibrated heart rate and blood pressure values ​​are measured in a calm state and serve as subsequent assessment criteria. These data can be obtained using a wrist-worn wearable device. The measurement environment is not limited to a quiet indoor environment; as long as the user is calm, measurements can be taken outdoors in a relatively quiet location free from significant interference. However, extreme environments such as high and low temperatures and strong electromagnetic interference should be avoided, as these may affect sensor performance and lead to inaccurate measurement results. The human body's physiological state may vary over time, and blood pressure and heart rate may also fluctuate. If multiple measurements are required to obtain more accurate calibration values, measurements can be taken at different times and the average value calculated as the final calibrated heart rate and blood pressure values.

[0079] 102. Determine the target group to which the target user belongs based on the relationship between the calibrated heart rate value and the calibrated blood pressure value.

[0080] The target group to which the target user belongs is determined based on the relationship between the calibrated heart rate value and the calibrated blood pressure value. These groups include a low blood pressure group, a normal blood pressure group, and a high blood pressure group. A certain relationship exists between the heart rate and blood pressure values ​​of the average person, and the group to which the target user belongs can be determined based on this numerical relationship. For example, by analyzing data from a large number of healthy individuals, a linear regression equation for systolic blood pressure and heart rate is established: systolic blood pressure = a × heart rate + b (where a and b are regression coefficients). For the target user, the residual of their calibrated data in this regression equation is calculated. If the residual is greater than a certain threshold, the user may be classified as belonging to the high or low blood pressure group; otherwise, the user belongs to the normal group. In practical applications, factors such as individual age, gender, and physical condition may also be considered to influence the relationship between heart rate and blood pressure. For example, the relationship between heart rate and blood pressure may differ between the elderly and the young, and the physiological characteristics of women may also lead to differences between their heart rate and blood pressure relationship and those of men. Therefore, personalized group classification models can be established based on different population characteristics to improve classification accuracy.

[0081] 103. Collect the real-time PPG data and ECG data of the target user.

[0082] Specifically, wrist-worn wearable devices can be used to collect data. Photoplethysmography (PPG) data can be obtained by illuminating the skin with light of a specific wavelength (such as green, red, or infrared). Hemoglobin in the blood absorbs some of the light, while the remaining light is reflected or transmitted to a detector. Changes in vascular volume cause changes in the amount of light absorbed, generating a waveform signal synchronized with the pulse. ECG data (electrocardiogram) data can be obtained using a two-electrode or multi-electrode system. It is understood that other devices or methods may also be used to collect the corresponding data in actual implementation, and the specifics are not limited here.

[0083] 104. Extract peak area parameters of the PPG data and peak area parameters of the ECG data.

[0084] Specifically, the peak values ​​of PPG data and ECG data are determined, and parameters near the peak values ​​are collected. In actual implementation, the amount of data collected may be determined according to the performance of the equipment. As much data as possible should be collected while meeting the real-time requirements, and the collected peak area parameters should contain all data within one or more heart rate cycles to meet the requirements of the subsequent analysis process. It is understandable that the data may also be filtered and processed during the collection process, which is not limited here.

[0085] 105. Use the classification regression fitting model corresponding to the target group to process the peak area parameters of the PPG data and the peak area parameters of the ECG data to obtain real-time blood pressure parameters.

[0086] Specifically, the peak area parameters of the PPG data and the peak area parameters of the ECG data are processed using a classification regression fitting model corresponding to the target group to obtain real-time blood pressure parameters. The classification regression fitting model includes a feature extraction structure and a random forest regression fitting model.

[0087] Specifically, the classification regression fitting model includes: a classification regression fitting model corresponding to the lower blood pressure group, a classification regression fitting model corresponding to the normal blood pressure group, and a classification regression fitting model corresponding to the higher blood pressure group; the applicable one is selected according to the specific situation of the target user, and the classification regression fitting models corresponding to different groups are trained by historical data that are suitable for their situation.

[0088] The feature extraction architecture can include two LSTM layers, which use their internal gating mechanism to learn dynamic features in time series. Following the LSTM layers, two fully connected (FC) layers perform feature fusion and compression. Each FC layer multiplies the output of the previous layer by a weight matrix and a bias, and then performs a nonlinear transformation using an activation function (such as ReLU). This allows the model to extract more potential features from the raw data.

[0089] After extracting the latent features, they are fed into the random forest regression model. Other features can also be added during this input process. The specific features added depend on the model training process and are not limited here. The random forest regression model is an ensemble learning method that combines the outputs of multiple decision trees for prediction. The random forest regressor has excellent generalization capabilities, can handle high-dimensional and nonlinear data, and is highly robust to inter-feature correlation and noise.

[0090] It can be seen from the above technical solutions that the embodiments of the present application have the following advantages: The embodiments of the present application provide a blood pressure measurement method based on calibration data, including obtaining the calibrated heart rate value and calibrated blood pressure value of the target user, determining the target group to which the target user belongs, collecting the target user's real-time PPG data and ECG data, and using the classification regression fitting model corresponding to the target group to process the PPG data and ECG data to obtain real-time blood pressure parameters. The classification regression fitting model includes a feature extraction structure and a random forest regression fitting model. On the one hand, this solution divides the target groups by calibrating the heart rate value and blood pressure value, and uses the corresponding classification regression fitting model according to the characteristics of people with different blood pressure levels, thereby improving the pertinence and accuracy of the measurement and being able to more accurately identify the blood pressure condition; on the other hand, the comprehensive use of the peak area parameters of the PPG and ECG data, combined with the classification regression fitting model, effectively improves the processing capability of the feature extraction capability of complex physiological data, thereby improving the accuracy and reliability of blood pressure measurement.

[0091] In order to facilitate the use of this method in actual implementation, this application also provides a more detailed embodiment that can be optionally implemented, please refer to Figure 2 An embodiment of a blood pressure measurement method based on calibration data provided in this application includes: steps 201 to 207.

[0092] 201. Obtain the target user's calibrated heart rate value and calibrated blood pressure value.

[0093] Specifically, the user's calibrated heart rate and blood pressure are obtained. Calibrated blood pressure values ​​include systolic and diastolic pressures. Systolic pressure (SBP) refers to the highest blood pressure value produced when blood is pumped into the arteries during heart contraction. Diastolic pressure (DBP) refers to the lowest pressure value in the arteries during heart relaxation. In the periodic waveform of the arterial blood pressure (ABP) signal, systolic pressure and diastolic pressure correspond to the maximum and minimum values ​​of the signal, respectively.

[0094] Ideally, the process of obtaining calibrated heart rate and blood pressure values ​​includes:

[0095] 1) The subject should not drink coffee or alcohol within 30 minutes before measuring blood pressure, should not engage in strenuous activities, should be calm, and should have an empty bladder. The subject should be in a quiet state and rest for at least 5 minutes, avoiding any activities that may affect the measurement.

[0096] 2) Record the subject's age, gender, height, weight, arrhythmia type, whether or not they are taking antihypertensive drugs, and medication information.

[0097] 3) Use a sphygmomanometer, ABP monitoring device, or mercury stethoscope to measure blood pressure and obtain the current heart rate, systolic blood pressure (SBP), and diastolic blood pressure (DBP) values.

[0098] 4) In order to reduce the interference of measurement errors and factors such as changes in body position, blood pressure usually needs to be measured multiple times. After removing outliers, the average of three or more measurements is calculated as the final calibrated blood pressure value, which is marked as and and resting calibrated heart rate respectively.

[0099] In actual implementation, the above steps may be adjusted and modified accordingly to obtain more accurate calibrated heart rate and blood pressure values. The specific methods are not limited at this time. For example, for users of wrist-worn devices, the user can be prompted to complete the test by measuring the calibration values, or the user can enter the calibrated heart rate and blood pressure values ​​independently.

[0100] 202. Determine the target group to which the target user belongs based on the relationship between the calibrated heart rate value and the calibrated blood pressure value.

[0101] Generally, there is a physiological connection between blood pressure and heart rate. A low resting heart rate usually means that the heart is pumping blood more efficiently, so the diastolic blood pressure may be relatively low; on the other hand, a high resting heart rate may be accompanied by a higher diastolic blood pressure. For patients with hypertension, regulating heart rate is not only a good way to lower blood pressure, but also can relieve cardiac stress. Because patients with hypertension generally have a relatively high resting heart rate, a fast heart rate can trigger a series of adverse reactions, such as palpitations, chest tightness, and dizziness.

[0102] For users with different heart rate and blood pressure relationships, their blood pressure parameters have different characteristics. Therefore, using different methods to measure and calculate them according to different situations can further improve the accuracy of the obtained blood pressure. Specifically, the division method can compare the expected systolic blood pressure value calculated by the resting calibrated heart rate with the individual's actual measured calibrated systolic blood pressure value, and divide each person into different blood pressure groups, including people whose calibrated systolic blood pressure is within the calibrated heart rate calculation range, people whose calibrated systolic blood pressure is higher than the resting calibrated heart rate calculation range, and people whose calibrated systolic blood pressure is lower than the resting calibrated heart rate calculation range. In this way, different blood pressure data sets can be divided.

[0103] Divided into different blood pressure groups, assuming the expected diastolic blood pressure The range can be determined by adjusting the resting heart rate to a certain value. The specific formula is:

[0104]

[0105] If the calibrated heart rate is within the above range, the target user is determined to belong to the normal blood pressure group;

[0106] If the calibrated heart rate meets , it is determined that the target user belongs to the lower blood pressure group;

[0107] If the calibrated heart rate meets , it is determined that the target user belongs to the high blood pressure group.

[0108] General The value can be set to 20 mmHg.

[0109] Figure 3 This is an example diagram showing a linear relationship between systolic and diastolic pressures. Assuming that there is a certain linear relationship between the expected systolic pressure and the expected diastolic pressure, the expected systolic pressure is defined as a linear function of the expected diastolic pressure. The formula is as follows:

[0110]

[0111] in, is the proportionality coefficient, which indicates the strength of the relationship between the desired systolic and diastolic pressures, such as 1.1. It is also a constant that represents the offset between the desired systolic pressure and the desired diastolic pressure, such as 32.

[0112] From the above formula, we can know that With upper and lower limits respectively and ,in , .

[0113] Based on the above formula, the data can also be divided into three data sets:

[0114] (1) The resting calibrated systolic blood pressure is within the resting calibrated heart rate calculation range. This group is characterized by a calibrated systolic blood pressure that meets the expected systolic blood pressure range calculated from the calibrated heart rate. The specific formula is as follows:

[0115]

[0116] The dataset can be represented as:

[0117]

[0118] (2) The calibrated systolic blood pressure is higher than the resting calibrated heart rate calculation range. This group of people indicates that the expected maximum systolic blood pressure calculated by the resting calibrated heart rate is still lower than the calibrated systolic blood pressure. This group is characterized by a higher calibrated systolic blood pressure, which results in the expected maximum systolic blood pressure calculated based on the calibrated heart rate being lower than the actual resting calibrated systolic blood pressure value. The specific formula is as follows:

[0119]

[0120] The dataset can be represented as:

[0121]

[0122] (3) The calibrated systolic blood pressure is lower than the calibrated blood pressure calculated by the resting calibrated heart rate. This group of people indicates that the expected minimum systolic blood pressure calculated by the resting calibrated heart rate is still greater than the calibrated systolic blood pressure. This group is characterized by a low calibrated systolic blood pressure, which causes the expected minimum systolic blood pressure calculated based on the calibrated heart rate to be greater than the actual resting calibrated systolic blood pressure value. The specific formula is as follows:

[0123] The dataset can be represented as:

[0124]

[0125] The target user's group is determined based on the actual collected values ​​of their calibrated blood pressure and heart rate. It is understood that the range of expected diastolic blood pressure, which fluctuates based on the resting calibrated heart rate, is not limited to a fixed value and can also be adjusted based on actual conditions and the allowable error in blood pressure calculations. This is not specifically limited here.

[0126] 203. Collect the real-time PPG data and ECG data of the target user.

[0127] 204. Extract peak area parameters of the PPG data and peak area parameters of the ECG data.

[0128] Specifically, the real-time PPG data and ECG data of the target user are collected. In actual implementation, this can be achieved through wrist wearable devices. The heart rate detection range is usually 30bpm to 180bpm. Therefore, 0.5 to 3 peaks can be detected in a one-second data window, that is, at least 0.5 heart rate peaks can be generated in 1s. In order to ensure the effectiveness of the solution implementation, it is necessary to extract the forward and backward data. data, is the sampling rate, is a constant, such as .

[0129] ECG peak point Before and After Data: , The ECG peak point, which means that Extract forward and backward Data, such as Figure 4 ECG and PPG data are shown in the diagram. PPG peak point Before and After Data: , is the PPG peak point. In actual implementation, filtering and peak point determination using a specific algorithm may also be performed, which are not limited here.

[0130] 205. Input the peak area parameters of the PPG data and the peak area parameters of the ECG data into a feature extraction structure to obtain multiple potential features.

[0131] Specifically, this step enters the model processing process. The classification regression fitting model used in this solution is a random forest regression fitting model with a nested feature extraction structure. The classification regression fitting model includes: a classification regression fitting model corresponding to the lower blood pressure group, a classification regression fitting model corresponding to the normal blood pressure group, and a classification regression fitting model corresponding to the higher blood pressure group. The classification regression fitting models are trained using historical data belonging to the corresponding groups. The training process of the model can refer to the existing technology and will not be described in detail here. By dividing the data into three independent data sets, each data set is processed by LSTM, fully connected layer (FC) and random forest regression algorithm respectively, it can better adapt to the differences between data sets, avoid overfitting, and improve the generalization ability and prediction accuracy of the model. Figure 5 The following is a diagram of the network architecture used. Classification and fitting are performed separately for each dataset to ensure that the characteristics of each dataset are fully learned and the advantages of the segmentation provided by the calibrated systolic and diastolic blood pressures are utilized, ultimately effectively predicting the labeled systolic blood pressure (SBP) and diastolic blood pressure (DBP) values.

[0132] The feature extraction architecture consists of two LSTM layers and two fully connected layers. LSTM is a recurrent neural network (RNN) suitable for sequential data, particularly adept at capturing long-term dependencies in time series data. Because ECG and PPG signals are time series, LSTM is able to learn the patterns of these signals over time.

[0133] Here, the input data is ECG and PPG including the peak point before and after For each data point, the LSTM layer uses its internal gating mechanism to learn the dynamic features of the time series. Following the LSTM layer, two fully connected layers (FC) perform feature fusion and compression. Each FC layer multiplies the output of the previous layer by a weight matrix and a bias, and then performs a nonlinear transformation using an activation function (such as ReLU). This allows the model to extract more latent features from the raw data. The final output is 10 latent features processed by the deep learning network. These features combine the deep fusion results of ECG and PPG signals, and include the signal's temporal information and potential blood pressure-related patterns. The number of latent features can be adjusted based on model training and actual needs, and is not limited here.

[0134] It is worth noting that the purpose of selecting two LSTM layers and two fully connected layers in this solution is to adapt to the computing performance of the device as much as possible. In actual implementation, the feature extraction structure can also be adjusted, such as increasing the number of network layers, but the specific adjustments are not limited here.

[0135] 206. Obtain multiple representational features.

[0136] The appearance characteristics include: cardiac pressure difference value, calibrated pressure difference value, pulse wave transmission time, pulse wave transmission speed, calibrated heart rate, calibrated systolic pressure and calibrated diastolic pressure.

[0137] (1) Calibration of cardiac pressure difference :

[0138]

[0139] (2) Calibration pressure difference :

[0140]

[0141] (3) ECG and PPG before and after peak points data, namely the data collected in step 204.

[0142] (4) PWTT (Pulse Wave Transit Time) is a physiological parameter that measures the time it takes for the pulse wave to travel from the heart to the peripheral blood vessels. It is often used to assess blood vessel stiffness and blood pressure. Specifically, the peak point of the electrocardiogram is used. The corresponding time (ms) minus the peak point detected by PPG The time required (ms). The specific formula is:

[0143]

[0144] in, 、 、 like Figure 4 ECG and PPG data are shown in the schematic diagram.

[0145] (5) Pulse wave velocity (PWV).

[0146] PWV (Pulse Wave Velocity) is a measure of the speed at which the pulse wave travels from the heart to the peripheral blood vessels. A shorter PWTT means a faster pulse wave velocity, which means a higher PWV. In the human body, the speed of the pulse wave is related to factors such as height, age, and gender. Pulse wave propagation path There is usually a linear relationship between them, so sometimes the calculation can be simplified by making a simple assumption:

[0147]

[0148] This assumption applies if the distance from the heart to peripheral areas (such as ankles, fingers, earlobes) is approximately half the height. This simplifies the formula:

[0149]

[0150] It is understandable that in actual implementation, the type of representational features may be determined according to actual conditions, and other relevant features may also be extracted, such as PPG heart beat amplitude ratio, area ratio, etc., which are not limited here.

[0151] 207. Input the multiple potential features and the multiple manifestation features into the random forest regression fitting model to obtain real-time blood pressure parameters.

[0152] Random forest is an ensemble learning method that combines the outputs of multiple decision trees to make predictions. The random forest regressor has good generalization capabilities, can handle high-dimensional, nonlinear data, and is robust to feature correlation and noise.

[0153] The input data of the random forest regression fitting model includes: 10 comprehensive features extracted from the LSTM and FC layers, calibrated heart rate, calibrated systolic and diastolic blood pressure, PWTT, PWV, and calibrated heart pressure difference , calibrated pressure difference Through these features, the random forest regression algorithm can establish a mapping relationship between blood pressure and these features. Based on the random forest regression fitting model processing, the current systolic and diastolic blood pressure of the target user can be obtained.

[0154] It can be seen from the above technical solutions that the embodiments of the present application have the following advantages: The embodiments of the present application provide a blood pressure measurement method based on calibration data, including obtaining the calibrated heart rate value and calibrated blood pressure value of the target user, determining the target group to which the target user belongs, collecting the target user's real-time PPG data and ECG data, and using the classification regression fitting model corresponding to the target group to process the PPG data and ECG data to obtain real-time blood pressure parameters. The classification regression fitting model includes a feature extraction structure and a random forest regression fitting model. On the one hand, this solution divides the target groups by calibrating the heart rate value and blood pressure value, and uses the corresponding classification regression fitting model according to the characteristics of people with different blood pressure levels, thereby improving the pertinence and accuracy of the measurement and being able to more accurately identify the blood pressure condition; on the other hand, the comprehensive use of the peak area parameters of the PPG and ECG data, combined with the classification regression fitting model, effectively improves the processing capability of the feature extraction capability of complex physiological data, thereby improving the accuracy and reliability of blood pressure measurement.

[0155] The above content describes the blood pressure measurement method based on calibration data provided by this application. To support the implementation of the above embodiment, this application also provides a blood pressure measurement device based on calibration data. Figure 6 An embodiment of a blood pressure measurement device based on calibration data provided by the present application includes:

[0156] An acquisition unit 601 is configured to acquire a calibrated heart rate value and a calibrated blood pressure value of a target user;

[0157] a determining unit 602, configured to determine a target group to which the target user belongs based on a relationship between the calibrated heart rate value and the calibrated blood pressure value, the target group including a low blood pressure group, a normal blood pressure group, and a high blood pressure group;

[0158] The collecting unit 603 is used to collect the real-time PPG data and ECG data of the target user, and extract the peak area parameters of the PPG data and the peak area parameters of the ECG data;

[0159] The processing unit 604 is used to process the peak area parameters of the PPG data and the peak area parameters of the ECG data using a classification regression fitting model corresponding to the target group to obtain real-time blood pressure parameters, wherein the classification regression fitting model includes a feature extraction structure and a random forest regression fitting model.

[0160] According to the blood pressure measurement device based on calibration data provided in the second aspect of the embodiment of the present application, optionally,

[0161] The feature extraction structure includes 2 layers of LSTM and 2 layers of fully connected layers.

[0162] According to the blood pressure measurement device based on calibration data provided in the second aspect of the embodiment of the present application, optionally,

[0163] The processing unit is specifically configured to:

[0164] Inputting the peak area parameters of the PPG data and the peak area parameters of the ECG data into a feature extraction structure to obtain a plurality of potential features;

[0165] Acquire multiple representational features;

[0166] The multiple potential features and the multiple manifestation features are input into the random forest regression fitting model to obtain real-time blood pressure parameters.

[0167] According to the blood pressure measurement device based on calibration data provided in the second aspect of the embodiment of the present application, optionally,

[0168] The classification regression fitting model includes: a classification regression fitting model corresponding to a lower blood pressure group, a classification regression fitting model corresponding to a normal blood pressure group, and a classification regression fitting model corresponding to a higher blood pressure group;

[0169] The classification regression fitting models are trained using historical data belonging to corresponding groups.

[0170] Based on the calibration data-based blood pressure measurement device provided in the second aspect of the embodiment of the present application, optionally, the determining unit is specifically configured to:

[0171] The target group to which the target user belongs is determined according to the following formula:

[0172]

[0173] To calibrate the heart rate, To calibrate diastolic blood pressure; is a constant value;

[0174] If the calibrated heart rate is within the above range, the target user is determined to belong to the normal blood pressure group;

[0175] If the calibrated heart rate meets , it is determined that the target user belongs to the lower blood pressure group;

[0176] If the calibrated heart rate meets , it is determined that the target user belongs to the high blood pressure group.

[0177] Based on the blood pressure measurement device based on calibration data provided in the second aspect of the embodiment of the present application, optionally, the appearance characteristics include: heart pressure difference value, calibrated pressure difference value, pulse wave transmission time, pulse wave transmission speed, calibrated heart rate, calibrated systolic pressure and calibrated diastolic pressure.

[0178] Based on the calibration data-based blood pressure measurement device provided in the second aspect of the embodiment of the present application, optionally, the step of the acquisition unit extracting the peak area parameters of the PPG data and the peak area parameters of the ECG data specifically includes:

[0179] Before and after extracting the PPG peak point Parameters, before and after extracting the ECG peak point parameters, among which is a constant value, is the sampling rate.

[0180] In this embodiment, the processes performed by each unit in the device are the same as those described above. Figure 1 The method flow described in the corresponding embodiment is similar and will not be repeated here.

[0181] Figure 71 is a structural diagram of a blood pressure measurement device based on calibration data provided in an embodiment of the present application. The blood pressure measurement device based on calibration data 700 may include one or more central processing units (CPU) 701 and a memory 705, wherein the memory 705 stores one or more applications or data.

[0182] In this embodiment, the specific functional module division in the central processing unit 701 can be the same as the above Figure 6 The functional module division method of each unit described in is similar and will not be repeated here.

[0183] Memory 705 can be volatile or persistent storage. The program stored in memory 705 can include one or more modules, each of which can include a series of instruction operations on the server. Furthermore, the central processing unit 701 can be configured to communicate with memory 705 and execute the series of instruction operations in memory 705 on the blood pressure measurement device 700 based on the calibration data.

[0184] The blood pressure measurement device 700 based on calibration data may further include one or more power supplies 702 , one or more wired or wireless network interfaces 703 , one or more input / output interfaces 704 , and / or one or more operating systems.

[0185] The CPU 701 can execute the aforementioned Figure 1 The operations performed by the blood pressure measurement method based on calibration data in the illustrated embodiment will not be described in detail here.

[0186] An embodiment of the present application also provides a computer storage medium for storing computer software instructions used for the above-mentioned blood pressure measurement method based on calibration data, which includes a program designed for executing the blood pressure measurement method based on calibration data.

[0187] The blood pressure measurement method based on calibration data can be as described above. Figure 1 or Figure 2 The blood pressure measurement method based on calibration data described in .

[0188] The present application also provides a computer program product, which includes computer software instructions that can be loaded by a processor to implement the above Figure 1 Figure 2 The process of any one of the blood pressure measurement methods based on calibration data.

[0189] In the several embodiments provided in this application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are merely schematic. For example, the equivalent transformation of circuits and the division of units are only a kind of logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, 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 an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0190] 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 units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0191] In addition, the functional units in the various embodiments of the present application may be integrated into a single processing unit, or each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0192] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions or improvements made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A blood pressure measurement method based on calibration data, characterized in that: include: Obtain the target user's calibrated heart rate and blood pressure values; Determining the target group to which the target user belongs based on the relationship between the calibrated heart rate value and the calibrated blood pressure value, the target group including a low blood pressure group, a normal blood pressure group, and a high blood pressure group; Collecting real-time PPG data and ECG data of the target user; extracting peak area parameters of the PPG data and peak area parameters of the ECG data; Processing the peak area parameters of the PPG data and the peak area parameters of the ECG data using a classification regression fitting model corresponding to the target group to obtain the current systolic and diastolic blood pressures of the target user, wherein the classification regression fitting model includes a feature extraction structure and a random forest regression fitting model; The feature extraction structure includes 2 layers of LSTM and 2 layers of fully connected layers; The peak area parameters of the PPG data and the peak area parameters of the ECG data are processed using the classification regression fitting model corresponding to the target group to obtain the current systolic and diastolic blood pressures of the target user, including: Inputting the peak area parameters of the PPG data and the peak area parameters of the ECG data into a feature extraction structure to obtain a plurality of potential features; Acquire multiple representational features; Inputting the multiple potential features and the multiple appearance features into the random forest regression fitting model to obtain the current systolic and diastolic blood pressure of the target user; The appearance characteristics include: cardiac pressure difference value, calibrated pressure difference value, pulse wave transmission time, pulse wave transmission speed, calibrated heart rate, calibrated systolic pressure and calibrated diastolic pressure.

2. The blood pressure measurement method based on calibration data according to claim 1, characterized in that: The classification regression fitting model includes: a classification regression fitting model corresponding to a lower blood pressure group, a classification regression fitting model corresponding to a normal blood pressure group, and a classification regression fitting model corresponding to a higher blood pressure group; Each of the classification regression fitting models is trained using historical data belonging to a corresponding group.

3. The blood pressure measurement method based on calibration data according to claim 1, characterized in that: Determining the target group to which the target user belongs based on the relationship between the calibrated heart rate value and the calibrated blood pressure value includes: The target group to which the target user belongs is determined according to the following formula: To calibrate the heart rate, To calibrate diastolic blood pressure; is a constant value; If the calibrated heart rate is within the above range, the target user is deemed to belong to the normal blood pressure group; If the calibrated heart rate meets , it is determined that the target user belongs to the lower blood pressure group; If the calibrated heart rate meets , it is determined that the target user belongs to the high blood pressure group.

4. The blood pressure measurement method based on calibration data according to claim 1, characterized in that: Extracting the peak area parameters of the PPG data and the peak area parameters of the ECG data includes: Before and after extracting PPG peak points Parameters, before and after extracting ECG peak points parameters, among which is a constant value, is the sampling rate.

5. A blood pressure measurement device based on calibration data, characterized in that: include: An acquisition unit, configured to acquire a calibrated heart rate value and a calibrated blood pressure value of a target user; a determining unit, configured to determine a target group to which the target user belongs based on a relationship between the calibrated heart rate value and the calibrated blood pressure value, the target group comprising a low blood pressure group, a normal blood pressure group, and a high blood pressure group; an acquisition unit, configured to acquire real-time PPG data and ECG data of the target user, and extract peak area parameters of the PPG data and peak area parameters of the ECG data; a processing unit, configured to process the peak area parameters of the PPG data and the peak area parameters of the ECG data using a classification regression fitting model corresponding to the target group to obtain the current systolic and diastolic blood pressures of the target user, the classification regression fitting model including a feature extraction structure and a random forest regression fitting model; The feature extraction structure includes 2 layers of LSTM and 2 layers of fully connected layers; The processing unit is specifically configured to: input the peak area parameters of the PPG data and the peak area parameters of the ECG data into a feature extraction structure to obtain a plurality of potential features; Acquire multiple representational features; Inputting the multiple potential features and the multiple appearance features into the random forest regression fitting model to obtain the current systolic and diastolic blood pressure of the target user; The appearance characteristics include: cardiac pressure difference value, calibrated pressure difference value, pulse wave transmission time, pulse wave transmission speed, calibrated heart rate, calibrated systolic pressure and calibrated diastolic pressure.

6. A blood pressure measurement device based on calibration data, characterized in that: include: CPU, memory, input and output interfaces, wired or wireless network interfaces, and power supply; The memory is a transient storage memory or a persistent storage memory; The central processing unit is configured to communicate with the memory and execute instructions in the memory on the device to perform the blood pressure measurement method based on calibration data according to any one of claims 1 to 4.

7. A computer-readable storage medium, characterized in that The method comprises instructions, which, when executed on a computer, cause the computer to perform the method according to any one of claims 1 to 4.

Images