CN116649932A - Wearable continuous blood pressure measurement method and system based on removing age confusion factor - Google Patents
Wearable continuous blood pressure measurement method and system based on removing age confusion factor Download PDFInfo
- Publication number
- CN116649932A CN116649932A CN202310058024.1A CN202310058024A CN116649932A CN 116649932 A CN116649932 A CN 116649932A CN 202310058024 A CN202310058024 A CN 202310058024A CN 116649932 A CN116649932 A CN 116649932A
- Authority
- CN
- China
- Prior art keywords
- blood pressure
- layer
- age
- pulse wave
- size
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 31
- 238000009530 blood pressure measurement Methods 0.000 title claims abstract description 20
- 230000036772 blood pressure Effects 0.000 claims abstract description 74
- 238000007781 pre-processing Methods 0.000 claims abstract description 15
- 238000012549 training Methods 0.000 claims abstract description 9
- 238000013528 artificial neural network Methods 0.000 claims abstract description 8
- 230000006870 function Effects 0.000 claims description 16
- 230000004913 activation Effects 0.000 claims description 12
- 210000001367 artery Anatomy 0.000 claims description 10
- 230000000747 cardiac effect Effects 0.000 claims description 10
- 230000035488 systolic blood pressure Effects 0.000 claims description 8
- 230000004872 arterial blood pressure Effects 0.000 claims description 7
- 230000035487 diastolic blood pressure Effects 0.000 claims description 7
- 238000012952 Resampling Methods 0.000 claims description 6
- 238000001914 filtration Methods 0.000 claims description 6
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 claims description 4
- 238000003062 neural network model Methods 0.000 claims description 3
- 230000010349 pulsation Effects 0.000 claims description 3
- 238000005070 sampling Methods 0.000 claims description 3
- 230000003321 amplification Effects 0.000 claims description 2
- 230000005611 electricity Effects 0.000 claims description 2
- 238000003199 nucleic acid amplification method Methods 0.000 claims description 2
- 238000010606 normalization Methods 0.000 abstract description 2
- 238000005259 measurement Methods 0.000 description 7
- 238000013473 artificial intelligence Methods 0.000 description 3
- 238000011161 development Methods 0.000 description 3
- 238000000691 measurement method Methods 0.000 description 3
- 238000012544 monitoring process Methods 0.000 description 3
- 210000002321 radial artery Anatomy 0.000 description 3
- 208000024172 Cardiovascular disease Diseases 0.000 description 2
- 239000008280 blood Substances 0.000 description 2
- 230000017531 blood circulation Effects 0.000 description 2
- 238000013135 deep learning Methods 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 210000004369 blood Anatomy 0.000 description 1
- 230000008081 blood perfusion Effects 0.000 description 1
- 210000004204 blood vessel Anatomy 0.000 description 1
- 230000001364 causal effect Effects 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 239000012634 fragment Substances 0.000 description 1
- PCHJSUWPFVWCPO-UHFFFAOYSA-N gold Chemical compound [Au] PCHJSUWPFVWCPO-UHFFFAOYSA-N 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 210000000056 organ Anatomy 0.000 description 1
- 230000008288 physiological mechanism Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000001737 promoting effect Effects 0.000 description 1
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/02—Detecting, measuring or recording pulse, heart rate, blood pressure or blood flow; Combined pulse/heart-rate/blood pressure determination; Evaluating a cardiovascular condition not otherwise provided for, e.g. using combinations of techniques provided for in this group with electrocardiography or electroauscultation; Heart catheters for measuring blood pressure
- A61B5/021—Measuring pressure in heart or blood vessels
- A61B5/02108—Measuring pressure in heart or blood vessels from analysis of pulse wave characteristics
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
- A61B5/316—Modalities, i.e. specific diagnostic methods
- A61B5/318—Heart-related electrical modalities, e.g. electrocardiography [ECG]
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
- A61B5/316—Modalities, i.e. specific diagnostic methods
- A61B5/318—Heart-related electrical modalities, e.g. electrocardiography [ECG]
- A61B5/346—Analysis of electrocardiograms
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/68—Arrangements of detecting, measuring or recording means, e.g. sensors, in relation to patient
- A61B5/6801—Arrangements of detecting, measuring or recording means, e.g. sensors, in relation to patient specially adapted to be attached to or worn on the body surface
- A61B5/6802—Sensor mounted on worn items
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7203—Signal processing specially adapted for physiological signals or for diagnostic purposes for noise prevention, reduction or removal
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7235—Details of waveform analysis
- A61B5/7264—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Molecular Biology (AREA)
- General Health & Medical Sciences (AREA)
- Biophysics (AREA)
- Pathology (AREA)
- Veterinary Medicine (AREA)
- Biomedical Technology (AREA)
- Heart & Thoracic Surgery (AREA)
- Medical Informatics (AREA)
- Public Health (AREA)
- Surgery (AREA)
- Animal Behavior & Ethology (AREA)
- Cardiology (AREA)
- Physiology (AREA)
- Signal Processing (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Psychiatry (AREA)
- Vascular Medicine (AREA)
- Evolutionary Computation (AREA)
- Fuzzy Systems (AREA)
- Mathematical Physics (AREA)
- Measuring Pulse, Heart Rate, Blood Pressure Or Blood Flow (AREA)
- Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)
Abstract
The invention discloses a wearable continuous blood pressure measurement method and system based on removing age confusion factors, wherein the system specifically comprises a data acquisition module, a data preprocessing module, a neural network blood pressure estimation model and a blood pressure display module. Blood pressure may be estimated by a wearable device acquiring an Electrocardiogram (ECG) signal and a Photoplethysmogram (PPG) signal. Firstly, denoising, time synchronization, normalization and other series of operations are carried out on signals acquired by the wearable equipment, then the signals are put into a wearable continuous blood pressure estimation model for removing age confusion factors for training, and finally, the trained model is used for estimating blood pressure. The example data proves that the model accuracy and generalization are improved, and the robustness is better.
Description
Technical Field
The invention belongs to the field of blood pressure measurement of medical instruments, and mainly relates to a wearable continuous blood pressure measurement system based on removing age confusion factors.
Technical Field
Blood pressure is the lateral pressure caused to the inner wall of a blood vessel when blood continuously flows, is the power for promoting blood flow and maintaining blood perfusion and circulation of tissues and organs, and the numerical value of the blood pressure is an important basis for judging various cardiovascular diseases, so that the measurement of the blood pressure has important medical significance.
There are two current methods of measuring blood pressure, direct and indirect. The direct measurement method directly inserts the device into the artery of the human body to measure the arterial blood pressure, has the advantages of continuous blood pressure, high precision, gold standard for blood pressure measurement, invasive and troublesome implementation, and is mostly used for critical and open-cavity operations; indirect measurement methods mainly include Korotkoff sound method, oscillometric method, pulse wave velocity method, radial artery applanation method and arterial volume compensation method. The indirect measurement method is widely applied to clinic due to the characteristics of non-invasiveness, relatively simple implementation and the like, but still has some problems. For example, the Korotkoff method and the oscillometric method cannot acquire continuous blood pressure, and the cuff is required to be inflated and deflated during measurement, which causes discomfort to a patient; continuous blood pressure can be obtained by the radial artery applanation method and the arterial volume compensation method, but invasive measurement is needed, measurement equipment is complex, and professional staff is needed to operate the device, so that the device has a plurality of inconveniences. Although pulse wave velocity method can continuously measure blood pressure and is non-invasive, the accuracy is not high and the generalization is poor.
In recent years, the development direction of blood pressure measurement is mainly noninvasive, undisturbed and continuous measurement, which has very important significance for monitoring cardiovascular diseases of human bodies. The development of wearing sensing and artificial intelligence algorithms provides a new idea for sleeveless blood pressure. The method mainly adopts electrocardiosignals or pulse wave signals as input, inputs the electrocardiosignals or pulse wave signals into a neural network, predicts the blood pressure through the neural network, but has the problems of low measurement precision and poor generalization in the blood pressure measurement of the method because of more factors influencing the blood pressure, such as age, weight, height and gender of a user.
Age is an important reason for plagued blood pressure measurements. With age, the waveforms of the electrocardiosignal and the pulse wave signal are changed, and the blood pressure is increased with age. Based on the premise that the electrocardiographic and pulse wave signals can reflect the blood pressure changes, the artificial intelligence correlation algorithm can measure the blood pressure through an Electrocardiograph (ECG) signal or a photoplethysmoscope (PPG) signal. The reason-electrocardio signal and pulse wave signal and result-blood pressure are simultaneously influenced by age, so that the problem of attribution confusion of estimation occurs, and further the blood pressure measurement precision is reduced, and the generalization is reduced. To solve the above-mentioned problems, there is a need for research and improvement of cuff-free blood pressure measurement from the physiological mechanism.
Causal characterization learning has been a rapid development in recent years, and there has been a great deal of research in the field of deep learning. The method has the advantages that false relevance among data learned by the model can be reduced when complex application scenes are processed, so that the model is more robust and better in generalization. Its main feature is that problems due to confusion caused by potential variables are considered in modeling the problem, which is generally called a confusion factor.
Disclosure of Invention
The invention aims to solve the technical problem of providing a wearable cuff-free continuous blood pressure measurement system based on removing age confusion factors, which can realize continuous blood pressure measurement and has higher accuracy.
In order to solve the above problems, the present invention provides a wearable sleeveless continuous blood pressure measurement system based on removing age aliasing factors, the system comprising: the system comprises a data acquisition module, a data preprocessing module, a neural network blood pressure estimation model and a blood pressure display module;
the data acquisition module acquires a reflected light signal or a pressure signal of a certain part of a human body and simultaneously acquires an electrocardiosignal and blood pressure; transmitting all the acquired signals to a data preprocessing module;
the data preprocessing module performs noise filtering and amplification on all signals acquired by the data acquisition device, and then converts the obtained reflected light signals or pressure signals into pulse waves; normalizing the preprocessed electrocardiosignal and the preprocessed pulse wave signal, and resampling.
A wearable sleeveless continuous blood pressure measurement method based on removing age confusion factors mainly comprises the following steps:
step 1: acquiring pulse wave information and electrocardio information by using wearable sleeveless band measuring equipment, measuring heart and artery pulsation information related to a certain part of a human body by using wearable electricity, light and pressure sensing to obtain electrocardio and photoelectric volume pulse wave and body surface artery pressure signals, and measuring related blood pressure changes in the signals;
step 2: preprocessing the electrocardiosignal and the pulse wave signal as shown in fig. 2;
in the step 2, "electrocardiosignal and pulse wave signal preprocessing", specifically comprising the following steps:
step 2.1: filtering high-frequency noise of the electrocardiosignal and the pulse wave signal by using a butterworth filter;
step 2.2: aligning the electrocardiosignals and pulse wave signals, and dividing electrocardiosignals and pulse wave signals corresponding to 5 continuous cardiac cycles into a group of input signals;
step 2.3: normalizing the electrocardiosignal and the pulse wave signal;
step 2.4: interpolation resampling is carried out on the electrocardiosignal and the pulse wave signal, and 100 points are sampled in each cardiac cycle;
step 2.5: the electrocardiosignal and the pulse wave signal are respectively calculated with the reference electrocardiosignal and the pulse wave signal, and the signal of which the signal quality index is smaller than a set threshold value is removed;
step 3: building a blood pressure estimation network for removing age confusion factors, and outputting estimated systolic pressure or diastolic pressure through a full-connection layer by the network as shown in figure 1;
in the step 3, "building a blood pressure estimation network for removing age confusion factors", the method specifically comprises the following steps:
step 3.1: the input_size of the network is 2 x 500, specifically, the electrocardiosignals and pulse wave signals of 5 cardiac cycles, wherein the sampling points of each cycle are 100;
the neural network model comprises two layers of ANNs, an age confusion factor layer, a GRU layer and a full connection layer;
step 3.2: the first layer is an ANN layer with input_size of 2 x 500 and output_size of 2 x 1000, the activation function is ReLu, the second layer is an ANN layer with input_size of 2 x 1000 and output_size of 2 x 250, and the activation function is ReLu;
step 3.3: and the third layer carries out age confusion factor removal treatment on the output characteristics of the second layer, wherein the expression is as follows:
wherein the method comprises the steps ofQ for the age-out confusion feature t For the second layer output features, K is the average feature of each age group, Z is the average feature, p (Z) input features are probabilities of each age group, σ is the input feature q t Is a length of (2);
the average characteristic part of each age group is divided into three groups of 20-89 years old samples, namely 20-39 years old, 40-59 years old and 60-89 years old; pre-training a model in each group, and extracting the average characteristic of each age group; p (z) is connected with a softmax layer through a 5-layer perceptron to carry out probability prediction, wherein the input_size is 2 x 500, specifically, electrocardiosignals and pulse wave signals of 5 cardiac cycles; the output is the probability of ages 20-39 years, 40-59 years, 60-89 years, respectively;
step 3.4: putting the obtained age group probability into an ANN layer, outputting the size of 2 x 100, and finally constructing an age confusion factor removing layer, wherein the input size is 2 x 250, and the output size is 2 x 100;
step 3.5: the fourth layer splices the output of the third layer and the output of the second layer together, and the output size is 2 x 350;
step 3.6: the fifth layer is a GRU layer, the input size is 2 x 350, and the output size is 2 x 128;
step 3.7: the sixth layer is a full-connection layer, the input size is 256, the output size is 50, and the activation function is ReLu; the seventh layer is a full-connection layer, the input size is 50, the output size is 1, the predicted systolic pressure or diastolic pressure is obtained, and the activation function is ReLu;
step 4: the sample data are put into a blood pressure estimation network for training, parameters of the blood pressure estimation network are updated, and finally the blood pressure estimation network for removing the age confusion factors is obtained;
step 5: and inputting the data acquired by the intelligent wearable equipment into a trained blood pressure estimation network for removing the age confusion factors, and displaying the obtained blood pressure estimation value on a blood pressure display module.
In the step 1, the wearing of the cuff-free continuous blood pressure measuring device specifically comprises the following steps:
step 1.1: selecting a body surface artery of a subject as a subject;
step 1.2: continuously measuring the blood pressure at the body surface artery selected by the subject by adopting cuff type continuous blood pressure measuring equipment;
step 1.3: wearing the wearable measuring device to the skin surface of an artery on the other side of the same subject while measuring the cuff blood pressure;
step 1.4: simultaneously recording signals generated by arterial pulsation and heart;
in the step 4, "training in the blood pressure estimation network with age confusion factor removed" specifically includes:
calculating a loss value by using a root mean square error function, and updating a network weight parameter by using the back propagation of an Adam algorithm to gradually converge the model;
the learning rate was 0.0001, the weight decay was 0.0001, and the batch_size was 128.
The invention adopts the scheme and has the following beneficial effects:
1. compared with the traditional inflatable cuff blood pressure measurement, the invention can realize continuous and noninvasive blood pressure prediction and is more suitable for continuous blood pressure monitoring.
2. Compared with the traditional artificial intelligence blood pressure estimation algorithm, the method has better generalization.
3. The invention is based on a deep learning algorithm, and does not need to manually extract the characteristics.
4. Compared with an algorithm for estimating blood pressure by only using pulse wave signals, the method provided by the invention has better robustness by using the electrocardiosignals and the pulse wave signals.
5. The signal acquisition and algorithm processing of the invention can be completed by the micro-control chip, can realize the monitoring of miniaturized wearing equipment, and can be applied to daily life scenes.
Drawings
FIG. 1 is a diagram of a wearable continuous blood pressure estimation network structure model with age aliasing factors removed;
FIG. 2 is a signal preprocessing flow chart;
FIG. 3 is a set of network-input electrocardiographic signals;
fig. 4 is a set of pulse wave signals input by the network.
Detailed Description
In order to achieve the purpose, the invention samples physiological information of arteries and hearts of a human body by the wearable equipment to obtain electrocardiosignals, pulse wave signals and arterial blood pressure, and then preprocesses the signals. The electrocardiosignals and pulse wave signals of 5 cardiac cycles are taken as a group, and the age aliasing factor removing model is input for training, so that the average blood pressure of 5 cycles is estimated.
The wearable continuous blood pressure measurement system for removing the age confusion factor specifically comprises the following steps:
step one: measuring radial artery blood pressure of the subject by adopting continuous blood pressure measuring equipment;
step two: wearing the wearable device on the skin surface of a subject, recording electrocardiosignals and pulse wave signals of the subject, and synchronously measuring arterial blood pressure;
step three: preprocessing the signals acquired in the first step and the second step, wherein the preprocessing mainly comprises noise filtering, normalization, period alignment, resampling and the like;
step four: according to the de-aliasing factor concept, a blood pressure estimation model for removing the age aliasing factor is established, and arterial blood pressure is estimated in real time as shown in fig. 1.
And (3) removing the age confusion factor blood pressure estimation model in the fourth step, wherein the blood pressure estimation model is specifically as follows:
1) The input size of the network is 2×500, specifically, an electrocardiosignal and a pulse wave signal of 5 cardiac cycles, wherein the sampling point of each cycle is 100.
2) The first layer inputs the ANN layer with the size of 2 x 500, the output size of 2 x 1000, the activation function is ReLu, the second layer is the ANN layer with the input size of 2 x 1000, the output size of 2 x 250, and the activation function is ReLu;
3) And the third layer performs age confusion factor removal treatment on the output characteristics of the second layer, and then puts the features into the ANN layer, wherein the output size is 2 x 100.
4) A fourth layer, splicing the output of the third layer and the output of the second layer together, wherein the output size is 2 x 350;
5) The fifth layer is a GRU layer, the input size is 2 x 350, and the output size is 2 x 128;
6) The sixth layer is a full-connection layer, the input size is 256, the output size is 50, and the activation function is ReLu;
7) The seventh layer is the full connection layer, the input size is 50, the output size is 1, the predicted systolic pressure or diastolic pressure is obtained, and the activation function is ReLu.
Step five: the sample data is input to a blood pressure estimation model for removing the age confusion factor for training, and the method specifically comprises the following steps: the loss value is calculated using RMSE root mean square error as the loss function:
wherein Y is a reference blood pressure value,for predicting the blood pressure value, N is the number of samples of the input model, i is each predicted sample, σ is the input feature q t Is a length of (c).
The network weight is updated by back propagation of Adam algorithm, the network model gradually converges, the learning rate is 0.0001, the weight attenuation is 0.0001, and the batch_size is 128.
Step six: inputting the pulse wave signals acquired in real time into a trained blood pressure estimation model for removing age confusion factors to obtain a real-time predicted blood pressure value
In order to verify the method provided by the invention, the electrocardiosignals and pulse wave signals of 96 patients are acquired through experiments, and the ages are 20 to 89 years.
The specific experimental steps are as follows:
step one: collecting electrocardiosignals and pulse wave signals of a patient by using a wearable device, and synchronously collecting a systolic blood pressure value and a diastolic blood pressure value;
step two: the input electrocardiographic signals are shown in fig. 3, the input pulse wave signals are shown in fig. 4, and preprocessing is performed, as shown in fig. 2. Filtering high-frequency noise of the electrocardiosignal and the pulse wave signal by using a butterworth filter, resampling the electrocardiosignal and the pulse wave signal, aligning time synchronization periods of the electrocardiosignal and the pulse wave signal, eliminating time delay between the two signals, calculating a signal quality index with a reference signal, and removing a signal with the signal quality index smaller than 0.75;
step three: the preprocessed signals and reference blood pressure values (systolic pressure and diastolic pressure) are put into a blood pressure estimation model for removing age confusion factors for training, and the RMSE root mean square error is used as a loss function to calculate a loss value:
wherein Y is the reference blood pressure value,for predicting the blood pressure value, N is the number of samples of the input model, i is each predicted sample. Then, the Adam algorithm is used for back propagation to update the network weight, the network model gradually converges, the learning rate is 0.0001, the weight attenuation is 0.0001, and the batch_size is 128.
Step four: the trained model is used for carrying out blood pressure estimation on the electrocardiosignals and pulse wave signals of 20 persons with 2000 input fragments, and the accuracy of the model on systolic pressure estimation is shown in table 1.
Table 1 comparison of conventional neural network model and age-confounding factor removal model measurement Performance BHS criteria
Claims (4)
1. A wearable sleeveless continuous blood pressure measurement method based on removing age confusion factors mainly comprises the following steps:
step 1: acquiring pulse wave information and electrocardio information by using wearable sleeveless band measuring equipment, measuring heart and artery pulsation information related to a certain part of a human body by using wearable electricity, light and pressure sensing to obtain electrocardio and photoelectric volume pulse wave and body surface artery pressure signals, and measuring related blood pressure changes in the signals;
step 2: preprocessing electrocardiosignals and pulse wave signals;
in the step 2, "electrocardiosignal and pulse wave signal preprocessing", specifically comprising the following steps:
step 2.1: filtering high-frequency noise of the electrocardiosignal and the pulse wave signal by using a butterworth filter;
step 2.2: aligning the electrocardiosignals and pulse wave signals, and dividing electrocardiosignals and pulse wave signals corresponding to 5 continuous cardiac cycles into a group of input signals;
step 2.3: normalizing the electrocardiosignal and the pulse wave signal;
step 2.4: interpolation resampling is carried out on the electrocardiosignal and the pulse wave signal, and 100 points are sampled in each cardiac cycle;
step 2.5: the electrocardiosignal and the pulse wave signal are respectively calculated with the reference electrocardiosignal and the pulse wave signal, and the signal of which the signal quality index is smaller than a set threshold value is removed;
step 3: building a blood pressure estimation network for removing age confusion factors, and outputting estimated systolic pressure or diastolic pressure through a full-connection layer by the network;
in the step 3, "building a blood pressure estimation network for removing age confusion factors", the method specifically comprises the following steps:
step 3.1: the input_size of the network is 2 x 500, specifically, the electrocardiosignals and pulse wave signals of 5 cardiac cycles, wherein the sampling points of each cycle are 100;
the neural network model comprises two layers of ANNs, an age confusion factor layer, a GRU layer and a full connection layer;
step 3.2: the first layer is an ANN layer with input_size of 2 x 500 and output_size of 2 x 1000, the activation function is ReLu, the second layer is an ANN layer with input_size of 2 x 1000 and output_size of 2 x 250, and the activation function is ReLu;
step 3.3: and the third layer carries out age confusion factor removal treatment on the output characteristics of the second layer, wherein the expression is as follows:
wherein the method comprises the steps ofQ for the age-out confusion feature t For the second layer output features, K is the average feature of each age group, Z is the average feature, p (Z) input features are probabilities of each age group, σ is the input feature q t Is a length of (2);
the average characteristic part of each age group is divided into three groups of 20-89 years old samples, namely 20-39 years old, 40-59 years old and 60-89 years old; pre-training a model in each group, and extracting the average characteristic of each age group; p (z) is connected with a softmax layer through a 5-layer perceptron to carry out probability prediction, wherein the input_size is 2 x 500, specifically, electrocardiosignals and pulse wave signals of 5 cardiac cycles; the output is the probability of ages 20-39 years, 40-59 years, 60-89 years, respectively;
step 3.4: putting the obtained age group probability into an ANN layer, outputting the size of 2 x 100, and finally constructing an age confusion factor removing layer, wherein the input size is 2 x 250, and the output size is 2 x 100;
step 3.5: the fourth layer splices the output of the third layer and the output of the second layer together, and the output size is 2 x 350;
step 3.6: the fifth layer is a GRU layer, the input size is 2 x 350, and the output size is 2 x 128;
step 3.7: the sixth layer is a full-connection layer, the input size is 256, the output size is 50, and the activation function is ReLu; the seventh layer is a full-connection layer, the input size is 50, the output size is 1, the predicted systolic pressure or diastolic pressure is obtained, and the activation function is ReLu;
step 4: the sample data are put into a blood pressure estimation network for training, parameters of the blood pressure estimation network are updated, and finally the blood pressure estimation network for removing the age confusion factors is obtained;
step 5: and inputting the data acquired by the intelligent wearable equipment into a trained blood pressure estimation network for removing the age confusion factors, and displaying the obtained blood pressure estimation value on a blood pressure display module.
2. A wearable, sleeveless, continuous blood pressure measurement method based on age-aliasing factor removal according to claim 1, wherein the method in step 1 comprises the steps of:
step 1.1: selecting a body surface artery of a subject as a subject;
step 1.2: continuously measuring the blood pressure at the body surface artery selected by the subject by adopting cuff type continuous blood pressure measuring equipment;
step 1.3: wearing the wearable measuring device to the skin surface of an artery on the other side of the same subject while measuring the cuff blood pressure;
step 1.4: while recording the signals produced by the arterial pulses and the heart.
3. A wearable, sleeveless, continuous blood pressure measurement method based on age-aliasing factor removal according to claim 1, wherein the specific steps in step 4 include:
calculating a loss value by using a root mean square error function, and updating a network weight parameter by using the back propagation of an Adam algorithm to gradually converge the model;
the learning rate was 0.0001, the weight decay was 0.0001, and the batch_size was 128.
4. A system employing the method of claim 1, the system comprising: the system comprises a data acquisition module, a data preprocessing module, a neural network blood pressure estimation model and a blood pressure display module;
the data acquisition module acquires a reflected light signal or a pressure signal of a certain part of a human body and simultaneously acquires an electrocardiosignal and blood pressure; transmitting all the acquired signals to a data preprocessing module;
the data preprocessing module performs noise filtering and amplification on all signals acquired by the data acquisition device, and then converts the obtained reflected light signals or pressure signals into pulse waves; normalizing the preprocessed electrocardiosignal and pulse wave signal, and resampling;
the neural network blood pressure estimation model carries out blood pressure estimation according to the preprocessed data;
and finally, displaying by a blood pressure display module.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310058024.1A CN116649932A (en) | 2023-01-16 | 2023-01-16 | Wearable continuous blood pressure measurement method and system based on removing age confusion factor |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310058024.1A CN116649932A (en) | 2023-01-16 | 2023-01-16 | Wearable continuous blood pressure measurement method and system based on removing age confusion factor |
Publications (1)
Publication Number | Publication Date |
---|---|
CN116649932A true CN116649932A (en) | 2023-08-29 |
Family
ID=87724821
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202310058024.1A Pending CN116649932A (en) | 2023-01-16 | 2023-01-16 | Wearable continuous blood pressure measurement method and system based on removing age confusion factor |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN116649932A (en) |
-
2023
- 2023-01-16 CN CN202310058024.1A patent/CN116649932A/en active Pending
Similar Documents
Publication | Publication Date | Title |
---|---|---|
WO2021208490A1 (en) | Blood pressure measuring method and device based on deep neural network | |
CN108498089B (en) | Noninvasive continuous blood pressure measuring method based on deep neural network | |
CN106413534B (en) | Continuous blood pressure measuring device, measuring model establishing method and system | |
CN106821356A (en) | High in the clouds continuous BP measurement method and system based on Elman neutral nets | |
CN111248879B (en) | Hypertension old people activity analysis method based on multi-mode attention fusion | |
CN108143407A (en) | A kind of heart sound segmentation method for automatically extracting heart sound envelope characteristic | |
CN113143230B (en) | Peripheral arterial blood pressure waveform reconstruction system | |
CN112089405A (en) | Pulse wave characteristic parameter measuring and displaying device | |
CN112617850B (en) | Premature beat and heart beat detection system for electrocardiosignals | |
CN111839488B (en) | Non-invasive continuous blood pressure measuring device and method based on pulse wave | |
CN110558960A (en) | continuous blood pressure non-invasive monitoring method based on PTT and MIV-GA-SVR | |
Myint et al. | Blood pressure measurement from photo-plethysmography to pulse transit time | |
CN110840428B (en) | Noninvasive blood pressure estimation method based on one-dimensional U-Net network | |
CN113040738B (en) | Blood pressure detecting device | |
Pan et al. | Improved blood pressure estimation using photoplethysmography based on ensemble method | |
Ghosh et al. | Introduction of boosting algorithms in continuous non-invasive cuff-less blood pressure estimation using pulse arrival time | |
KR20220105092A (en) | Continuous blood pressure measurement method by inputting the difference between electrocardiogram and the photoplethysmography signal into artificial neural network | |
LU502768B1 (en) | Intelligent blood pressure prediction method based on multi-scale residual network and ppg signal | |
JP3721743B2 (en) | Cardiac function diagnostic device | |
CN114947782A (en) | System and method for central arterial pressure waveform reconstruction based on PPG signal | |
CN116649932A (en) | Wearable continuous blood pressure measurement method and system based on removing age confusion factor | |
CN113940646B (en) | Continuous blood pressure monitoring system of no sleeve area based on causality | |
Moghaddam Mansouri et al. | Systolic Blood Pressure Estimation from Electrocardiogram and Photoplethysmogram Signals Using Convoluional Neural Networks | |
Villata | Cuffless blood pressure measurement | |
CN116077034A (en) | Noninvasive blood pressure and blood glucose detection method and system based on overrun learning machine |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination |