CN107928654A - A kind of pulse wave signal blood pressure detecting method based on neutral net - Google Patents
A kind of pulse wave signal blood pressure detecting method based on neutral net Download PDFInfo
- Publication number
- CN107928654A CN107928654A CN201711309884.9A CN201711309884A CN107928654A CN 107928654 A CN107928654 A CN 107928654A CN 201711309884 A CN201711309884 A CN 201711309884A CN 107928654 A CN107928654 A CN 107928654A
- Authority
- CN
- China
- Prior art keywords
- mrow
- pulse wave
- msub
- point
- msup
- 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
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
- A61B5/72—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
-
- 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
- A61B5/725—Details of waveform analysis using specific filters therefor, e.g. Kalman or adaptive filters
Landscapes
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Engineering & Computer Science (AREA)
- Heart & Thoracic Surgery (AREA)
- Public Health (AREA)
- Physiology (AREA)
- Biophysics (AREA)
- Pathology (AREA)
- Signal Processing (AREA)
- Biomedical Technology (AREA)
- Veterinary Medicine (AREA)
- Medical Informatics (AREA)
- Molecular Biology (AREA)
- Surgery (AREA)
- Animal Behavior & Ethology (AREA)
- General Health & Medical Sciences (AREA)
- Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Psychiatry (AREA)
- Cardiology (AREA)
- Vascular Medicine (AREA)
- Measuring Pulse, Heart Rate, Blood Pressure Or Blood Flow (AREA)
Abstract
The present invention proposes a kind of pulse wave signal blood pressure detecting method based on neutral net, belongs to vital sign parameter signals process field and artificial intelligence field.The method comprising the steps of：Using a data acquisition module, pulse wave signal is gathered；It is main to remove the high frequency catastrophe point in signal and the baseline drift in Wavelet Algorithm removal signal using low-pass filtering for handling original pulse wave signal using a data processing module；The characteristic point in pulse wave is extracted, characteristic point includes pulse wave cycle, maximum and minimum value in the cycle, waveform time domain area parameters, the series of features point such as point of maximum curvature and curvature smallest point in the cycle；Relevant feature parameters are calculated, the present invention is extracted 15 characteristic parameters；Using 15 characteristic parameters as input layer, the BP algorithm of appropriate artificial intelligence is chosen, systolic pressure and diastolic pressure are gone out based on neural computing.
Description
Technical field
Related algorithm the present invention relates to processing and the machine learning of pulse wave signal belongs to vital sign parameter signals processing
Field and artificial intelligence field.
Background technology
Blood pressure is one of important parameter for weighing cardiovascular system of human body, in order to break away from the beam of traditional blood inflation cuff
Tie up, realize long-continued blood pressure detecting, be always one of the hot spot of biomedical sector research.When blood flows through peripheral blood vessel
In arteriole, capillary and venule when capilary, the volumetric blood of the part capilary is same under heartbeat
Pulsating nature change can be presented.The pulsating nature of this volumetric blood can be recorded by photoplethysmographic graphical method.Arteries and veins
The wave-shape amplitude of ripple of fighting and form contain many important physiologic informations of heart and cardiovascular system, it is detected and is divided
Analysis, the prevention to angiocardiopathy is with having great significance and acting in terms of clinical conditions.
Study is a kind of important intelligent behavior that the mankind have, and computer also tentatively possesses this ability at present.
Machine learning just refers to " behavior that computer improves self performance using experience automatically ".In short, machine learning refers to pass through
Inherent law information in computer learning data, obtains new experience and knowledge, to improve the intelligent of computer, makes meter
Calculation machine can go decision-making as people.Machine learning using deep learning as representative is currently closest to the intelligence of human brain
Habit and cognitive process, have fully used for reference the multi-segment structure of human brain, the connection interaction of neuron, distributed sparse storage and table
Sign, the bed-by-bed analysis of information and treatment mechanism, achieve breakthrough, in many applications in voice, image recognition etc.
Field obtains huge business success.
Existing non-invasive blood pressure measuring method has：Stethoscopy, oscillographic method, arterial tonometry, volume-compensation method.Wherein
Stethoscopy is oppressed to the discomfort caused by human body due to blood vessel, is not suitable for using in the observation of long-time continuous blood pressure；Show
Ripple method weakens the radio-frequency component of reaction blood pressure, and intelligence is used for intermittent blood pressure measurement；Arterial tonometry is to keeping sensing
Device measurement position is more difficult；Volume-compensation method keeps certain pressure due to needing at tested position, and comfort is poor in addition
In place of the deficiencies of measuring device is complicated.
In order to overcome the shortcomings of above method, patent of the present invention proposes a kind of noninvasive continuous based on artificial neural network
The new method of blood pressure measurement.
The content of the invention
The purpose of the present invention is to overcome the shortcoming of prior art, propose a kind of pulse wave letter based on neutral net
Number blood pressure detecting method.This method calculates the characteristic parameter of pulse wave, is finally based on machine by extracting the characteristic point of pulse wave
Study calculates systolic pressure and diastolic pressure, and the general frame of the present invention is shown in attached drawing 1.
Pulse wave signal blood pressure detecting method proposed by the present invention based on neutral net, it is characterised in that：
For existing non-invasive blood pressure measuring method there are various deficiencies, machine learning is the research hotspot of artificial intelligence, it is managed
By the duplication for being widely used in solving the problems, such as engineer application and scientific domain with method, obtained in many application fields
Huge business success.The present invention is applied to vital sign parameter field, is made using the machine learning algorithm of comparative maturity
The physical sign parameters that must be measured are more accurate.
A kind of pulse wave signal blood pressure detecting method based on neutral net, method proposed by the present invention mainly include following
Several steps：
Step 1: using a data acquisition module, for gathering pulse wave signal；
Step 2: using a data processing module, for handling the pulse wave signal collected；
Step 3: handling the pulse wave signal after denoising, the characteristic point of pulse wave is extracted；
Step 4: the feature point extraction pulse wave characteristic parameters based on pulse wave；
Step 5: being inputted characteristic parameter as the parameter of neutral net, systolic pressure and diastolic pressure are calculated.
Beneficial effects of the present invention essentially consist in the accuracy for improving blood pressure measurement.It is specific as follows：
1. improve the accuracy of measurement
The method of the present invention is used as neutral net (Artificial Neural Network, ANN) using the characteristic parameter extracted
Vector input, using the vector output of the corresponding systolic pressure of PPG signals (SBP) and diastolic pressure (DBP) as ANN of often fighting, progress
Training.Experimental data comes from THE MIMIC DATABASE, the data-base recording a variety of physiological signals and vital sign.Root
The absolute error for testing SBP and DBP according to the experiment has reached the blood manometer accuracy that Used In The Regulation of Medical Device In Usa promotes association to issue
International standard.This method realizes that the noninvasive continuous measurement of blood pressure provides certain reference value for intelligent wearable device.
Brief description of the drawings
Attached drawing 1：Flow entire block diagram of the present invention.
Attached drawing 2：The recognition result of pulse wave characteristic point.
Attached drawing 3：Characteristic parameter list.
Attached drawing 4：Based on neutral net blood pressure detecting model.
Embodiment
The purpose of the present invention is to overcome the shortcoming of prior art, propose a kind of pulse wave letter based on neutral net
Number blood pressure detecting method.This method calculates the characteristic parameter of pulse wave, is finally based on machine by extracting the characteristic point of pulse wave
Study calculates systolic pressure and diastolic pressure.
Pulse wave signal blood pressure detecting method proposed by the present invention based on neutral net, it is characterised in that：
For existing non-invasive blood pressure measuring method there are various deficiencies, machine learning is the research hotspot of artificial intelligence, it is managed
By the duplication for being widely used in solving the problems, such as engineer application and scientific domain with method, obtained in many application fields
Huge business success.The present invention is applied to vital sign parameter field, is made using the machine learning algorithm of comparative maturity
The physical sign parameters that must be measured are more accurate.Embodiment of the present invention is described in detail below in conjunction with the accompanying drawings.
A kind of pulse wave signal blood pressure detecting method based on neutral net, it is comprised the following steps that：
Step 1: using a data acquisition module, for gathering pulse wave signal.
Step 2: using a data processing module, for handling the pulse wave collected letter.
Comprise the following steps that：
1) low-pass filter is set, removes the high frequency catastrophe point included in signal；
2) filtered signal is subjected to Wavelet Denoising Method processing；
3) selection of threshold function table.Hard threshold function continuity is bad, can cause Pseudo-Gibbs artifacts, and soft-threshold function
Although overall continuity is good, distortion easily always occurs there are constant deviation between estimate and actual value, has necessarily
Limitation.Therefore the selection for threshold function table, should ensure its continuity, prevent its distortion again.Choosing to threshold function table
It is selected as：
The function by two state modulators of k and m, the value range of k for (0,1], the asymptote of parameter k determining functions,
As k=1, which levels off to hard threshold function, and as k → 0, which levels off to soft-threshold function.M is fine setting function,
We can obtain optimal function by adjusting the size of m；
4) selection of wavelet basis.The selection of wavelet basis will take into account orthogonality, symmetry, smoothness and the regularity of small echo.
The symmetry of small echo retains more preferable phase information, its regularity and smoothness are to obtain more preferable reconstruction signal.Wherein,
Except db1 (Haar) small echo has symmetry in the small wave systems of Daubechies, others all do not have；Symlets (symN) small echo
It is that the small echo near symmetry is proposed on the basis of Daubechies small echos.Therefore selection for wavelet basis can be
Selected in Haar small echos and symN small echos；
5) after utilizing Noise Elimination from Wavelet Transform, then reconstruct to obtain the pulse wave signal after denoising to wavelet coefficient.
Step 3: handling the pulse wave signal after denoising, the characteristic point of pulse wave is extracted.
Comprise the following steps that：
1) period divisions are carried out to pulse wave using threshold method, threshold value P is determined by following formula
P=h (Amax-Amin)
Wherein η is the coefficient that a large amount of pulse wave statistical analyses obtain, AmaxAnd AminThe maximum of respectively each group pulse wave
Value and minimum value, then think that minimum point is the terminal of previous cycle if more than threshold value, the starting point in latter cycle；
2) minimum point and maximum of points are found in the range of each pulse wave cycle, corresponds to B points and C points respectively；
3) minimum point and maximum of points are found in the range of each pulse wave cycle, corresponds to F points and G points respectively；
4) second differnce maximum of points, character pair point D are found in the range of each pulse wave cycle；
5) point of slope minimum, character pair point E are found between D points and F points；
Step 4: the feature point extraction pulse wave characteristic parameters based on pulse wave.Defined feature parameter, includes following characteristics
Parameter：Main wave-amplitude H1, wave-amplitude H before dicrotic wave2, dicrotic notch amplitude H3, dicrotic pulse wave-amplitude H4, main ripple rate of rise V, pulse
Ripple shrinks period T1, pulse wave diastole period T2, pulse cycle time T and pulse wave systole phase area Sa, pulse wave diastole face
Product Sb, pulse wave time domain waveform gross area S.
Step 5: using characteristic parameter as the parameter input layer of neutral net, systolic pressure and diastolic pressure are calculated.Such as attached drawing
Shown in 4, input layer is that 15 characteristic parameters choosing use signmoid functions as input layer, in the middle hidden layer in attached drawing 3,
Training uses Levenberg-Marquardt algorithms, i.e. trainlm functions.Hidden layer includes 30 neurons, with the PPG that often fights
The corresponding systolic pressure of signal (SBP) and diastolic pressure (DBP) are used as output layer.Experimental data comes from THE MIMIC DATABASE,
Extract pulse wave data part and corresponding SBP and DBP in the database.These data are divided into ANN model training and are surveyed
Try two parts.Wherein 70% data are used as model measurement data as model training data, 30%.Final output systolic pressure and
Unfold the result of pressure.
Claims (4)
- A kind of 1. pulse wave signal blood pressure detecting method based on neutral net, it is characterised in that：The detection algorithm by with Lower step is realized：Step 1: using a data acquisition module, for gathering pulse wave signal；Step 2: using a data processing module, for handling the pulse wave signal collected；Step 3: handling the pulse wave signal after denoising, the characteristic point of pulse wave is extracted；Step 4: the feature point extraction pulse wave characteristic parameters based on pulse wave；Step 5: using the characteristic parameter of selection as the parameter input layer of neutral net, systolic pressure and diastolic pressure are calculated.
- 2. a kind of pulse wave signal blood pressure detecting method based on neutral net is it is characterized in that, step according to claim 1 Rapid two detailed process is：Step 2 one, set low-pass filter, removes the high frequency catastrophe point included in signal；Step 2 two, by filtered signal carry out Wavelet Denoising Method processing；The selection of step 221, threshold function table, hard threshold function continuity is bad, can cause Pseudo-Gibbs artifacts, and soft threshold Although value function entirety continuity is good, always distortion easily occurs, has there are constant deviation between estimate and actual value Certain limitation, therefore for the selection of threshold function table, should ensure its continuity, prevent its distortion again, to threshold value letter Several selected as：<mrow> <msub> <mover> <mi>w</mi> <mo>~</mo> </mover> <mrow> <mi>j</mi> <mo>,</mo> <mi>k</mi> </mrow> </msub> <mo>=</mo> <mfenced open = "{" close = ""> <mtable> <mtr> <mtd> <mrow> <msub> <mi>w</mi> <mrow> <mi>j</mi> <mo>,</mo> <mi>k</mi> </mrow> </msub> <mo>+</mo> <mrow> <mo>(</mo> <mi>k</mi> <mo>-</mo> <mn>1</mn> <mo>)</mo> </mrow> <mi>&lambda;</mi> <mo>-</mo> <mn>0.5</mn> <mfrac> <mrow> <mi>k</mi> <mo>&times;</mo> <msup> <mi>&lambda;</mi> <mi>m</mi> </msup> </mrow> <mrow> <msup> <msub> <mi>w</mi> <mrow> <mi>j</mi> <mo>,</mo> <mi>k</mi> </mrow> </msub> <mrow> <mi>m</mi> <mo>-</mo> <mn>1</mn> </mrow> </msup> </mrow> </mfrac> </mrow> </mtd> <mtd> <mrow> <msub> <mi>w</mi> <mrow> <mi>j</mi> <mo>,</mo> <mi>k</mi> </mrow> </msub> <mo>></mo> <mi>&lambda;</mi> </mrow> </mtd> </mtr> <mtr> <mtd> <mrow> <mn>0.5</mn> <mfrac> <mrow> <mo>|</mo> <msub> <mi>w</mi> <mrow> <mi>j</mi> <mo>,</mo> <mi>k</mi> </mrow> </msub> <msup> <mo>|</mo> <mrow> <mi>m</mi> <mo>+</mo> <mo>&lsqb;</mo> <mn>2</mn> <mo>-</mo> <mi>k</mi> <mo>/</mo> <mi>k</mi> <mo>&rsqb;</mo> </mrow> </msup> </mrow> <msup> <mi>&lambda;</mi> <mrow> <mi>m</mi> <mo>+</mo> <mo>&lsqb;</mo> <mn>2</mn> <mo>-</mo> <mi>k</mi> <mo>/</mo> <mi>k</mi> <mo>&rsqb;</mo> </mrow> </msup> </mfrac> </mrow> </mtd> <mtd> <mrow> <mo>|</mo> <msub> <mi>w</mi> <mrow> <mi>j</mi> <mo>,</mo> <mi>k</mi> </mrow> </msub> <mo>|</mo> <mo><</mo> <mi>&lambda;</mi> </mrow> </mtd> </mtr> <mtr> <mtd> <mrow> <msub> <mi>w</mi> <mrow> <mi>j</mi> <mo>,</mo> <mi>k</mi> </mrow> </msub> <mo>-</mo> <mrow> <mo>(</mo> <mi>k</mi> <mo>-</mo> <mn>1</mn> <mo>)</mo> </mrow> <mo>-</mo> <mn>0.5</mn> <mfrac> <mrow> <mi>k</mi> <mo>&times;</mo> <msup> <mi>&lambda;</mi> <mi>m</mi> </msup> </mrow> <mrow> <msup> <msub> <mi>w</mi> <mrow> <mi>j</mi> <mo>,</mo> <mi>k</mi> </mrow> </msub> <mrow> <mi>m</mi> <mo>-</mo> <mn>1</mn> </mrow> </msup> </mrow> </mfrac> </mrow> </mtd> <mtd> <mrow> <msub> <mi>w</mi> <mrow> <mi>j</mi> <mo>,</mo> <mi>k</mi> </mrow> </msub> <mo><</mo> <mo>-</mo> <mi>&lambda;</mi> </mrow> </mtd> </mtr> </mtable> </mfenced> </mrow>The function by two state modulators of k and m, the value range of k for (0,1], the asymptote of parameter k determining functions, works as k= When 1, which levels off to hard threshold function, and as k → 0, which levels off to soft-threshold function, and m is fine setting function, Wo Menke With the size by adjusting m, optimal function is obtained；The selection of step 2 two or two, wavelet basis, the selection of wavelet basis will take into account the orthogonality of small echo, symmetry, smoothness and just Then property, the symmetry of small echo retain more preferable phase information, its regularity and smoothness be in order to obtain more preferable reconstruction signal, Wherein, except db1 (Haar) small echo has symmetry in the small wave systems of Daubechies, others all do not have；Symlets(symN) Small echo is that the small echo near symmetry is proposed on the basis of Daubechies small echos, therefore the selection for wavelet basis can be with Selected in Haar small echos and symN small echos；Step 2 two or three, using Noise Elimination from Wavelet Transform after, then reconstruct to obtain the pulse wave signal after denoising to wavelet coefficient.
- 3. a kind of pulse wave signal blood pressure detecting method based on neutral net is it is characterized in that, step according to claim 1 Rapid three detailed process is：Step 3 one, using threshold method carry out period divisions to pulse wave, and threshold value P determines by following formulaP=h (Amax-Amin)Wherein η is the coefficient that a large amount of pulse wave statistical analyses obtain, AmaxAnd AminThe maximum of respectively each group pulse wave and Minimum value, then thinks that minimum point is the terminal of previous cycle if more than threshold value, the starting point in latter cycle；Step 3 two, find minimum point and maximum of points in the range of each pulse wave cycle, corresponds to B points and C points respectively；Step 3 three, find minimum point and maximum of points in the range of each pulse wave cycle, corresponds to F points and G points respectively；Step 3 four, find second differnce maximum of points, character pair point D in the range of each pulse wave cycle；Step 3 five, between D points and F points find slope minimum point, character pair point E.
- 4. a kind of pulse wave signal blood pressure detecting method based on neutral net is it is characterized in that, step according to claim 1 Rapid four detailed process is：The value of step 4 one, extraction feature ginseng point；Step 4 two, defined feature parameter, include following characteristics parameter：Main wave-amplitude H1, wave-amplitude H before dicrotic wave2, dicrotic notch Amplitude H3, dicrotic pulse wave-amplitude H4, main ripple rate of rise V, pulse wave shrink period T1, pulse wave diastole period T2, the pulsation period when Between T and pulse wave systole phase area Sa, pulse wave diastole area Sb, pulse wave time domain waveform gross area S.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711309884.9A CN107928654A (en) | 2017-12-11 | 2017-12-11 | A kind of pulse wave signal blood pressure detecting method based on neutral net |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711309884.9A CN107928654A (en) | 2017-12-11 | 2017-12-11 | A kind of pulse wave signal blood pressure detecting method based on neutral net |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107928654A true CN107928654A (en) | 2018-04-20 |
Family
ID=61946527
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711309884.9A Pending CN107928654A (en) | 2017-12-11 | 2017-12-11 | A kind of pulse wave signal blood pressure detecting method based on neutral net |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107928654A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109259745A (en) * | 2018-10-25 | 2019-01-25 | 贵州医科大学附属医院 | A kind of wearable cardiovascular and cerebrovascular disease intelligent monitor system and method |
| CN111214218A (en) * | 2020-01-13 | 2020-06-02 | 京东方科技集团股份有限公司 | Detection equipment for multiple physiological parameters |
| CN112089405A (en) * | 2020-08-26 | 2020-12-18 | 上海大学 | Pulse wave characteristic parameter measuring and displaying device |
| CN112220449A (en) * | 2019-07-15 | 2021-01-15 | 爱科维申科技（天津）有限公司 | Photoelectric chicken embryo activity detection device and method |
| CN113100720A (en) * | 2021-05-25 | 2021-07-13 | 北京测态培元科技有限公司 | Pulse condition discrimination method based on deep learning |
| CN113116321A (en) * | 2021-03-03 | 2021-07-16 | 吉林大学珠海学院 | Non-invasive continuous blood pressure measuring system based on PSO-GRNN neural network |
| WO2021164345A1 (en) * | 2020-02-21 | 2021-08-26 | 乐普（北京）医疗器械股份有限公司 | Blood pressure prediction method and device |
| WO2023236658A1 (en) * | 2022-06-10 | 2023-12-14 | 荣耀终端有限公司 | Blood pressure measuring device and system |
-
2017
- 2017-12-11 CN CN201711309884.9A patent/CN107928654A/en active Pending
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109259745A (en) * | 2018-10-25 | 2019-01-25 | 贵州医科大学附属医院 | A kind of wearable cardiovascular and cerebrovascular disease intelligent monitor system and method |
| CN112220449A (en) * | 2019-07-15 | 2021-01-15 | 爱科维申科技（天津）有限公司 | Photoelectric chicken embryo activity detection device and method |
| CN112220449B (en) * | 2019-07-15 | 2023-09-15 | 爱科维申科技（天津）有限公司 | Photoelectric chicken embryo survival detection device and method |
| CN111214218A (en) * | 2020-01-13 | 2020-06-02 | 京东方科技集团股份有限公司 | Detection equipment for multiple physiological parameters |
| CN111214218B (en) * | 2020-01-13 | 2024-02-09 | 京东方科技集团股份有限公司 | Multi-physiological parameter detection equipment |
| WO2021164345A1 (en) * | 2020-02-21 | 2021-08-26 | 乐普（北京）医疗器械股份有限公司 | Blood pressure prediction method and device |
| CN112089405A (en) * | 2020-08-26 | 2020-12-18 | 上海大学 | Pulse wave characteristic parameter measuring and displaying device |
| CN112089405B (en) * | 2020-08-26 | 2023-03-14 | 上海大学 | Pulse wave characteristic parameter measuring and displaying device |
| CN113116321A (en) * | 2021-03-03 | 2021-07-16 | 吉林大学珠海学院 | Non-invasive continuous blood pressure measuring system based on PSO-GRNN neural network |
| CN113100720A (en) * | 2021-05-25 | 2021-07-13 | 北京测态培元科技有限公司 | Pulse condition discrimination method based on deep learning |
| WO2023236658A1 (en) * | 2022-06-10 | 2023-12-14 | 荣耀终端有限公司 | Blood pressure measuring device and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107928654A (en) | A kind of pulse wave signal blood pressure detecting method based on neutral net | |
| Tanveer et al. | Cuffless blood pressure estimation from electrocardiogram and photoplethysmogram using waveform based ANN-LSTM network | |
| El-Hajj et al. | A review of machine learning techniques in photoplethysmography for the non-invasive cuff-less measurement of blood pressure | |
| Kachuee et al. | Cuff-less high-accuracy calibration-free blood pressure estimation using pulse transit time | |
| Forouzanfar et al. | Oscillometric blood pressure estimation: past, present, and future | |
| WO2021208490A1 (en) | Blood pressure measuring method and device based on deep neural network | |
| CN102429649B (en) | Continuous blood pressure measuring device | |
| Yan et al. | Novel deep convolutional neural network for cuff-less blood pressure measurement using ECG and PPG signals | |
| CN112089405B (en) | Pulse wave characteristic parameter measuring and displaying device | |
| CN108053708B (en) | Dynamic pulse wave simulation module | |
| WO2023185873A1 (en) | Non-cuff blood pressure measuring apparatus based on multi-stage multi-modal learning and method | |
| Li et al. | A new signal decomposition to estimate breathing rate and heart rate from photoplethysmography signal | |
| CN111839488B (en) | Non-invasive continuous blood pressure measuring device and method based on pulse wave | |
| Argha et al. | A novel automated blood pressure estimation algorithm using sequences of Korotkoff sounds | |
| CN108403101A (en) | A kind of model training method and device, the method and device for measuring blood pressure | |
| Esmaelpoor et al. | Cuffless blood pressure estimation methods: Physiological model parameters versus machine-learned features | |
| CN110840428B (en) | Noninvasive blood pressure estimation method based on one-dimensional U-Net network | |
| Dash et al. | Estimation of arterial blood pressure waveform from photoplethysmogram signal using linear transfer function approach | |
| Wang et al. | Photoplethysmography-based blood pressure estimation using deep learning | |
| Ertuğrul et al. | A noninvasive time-frequency-based approach to estimate cuffless arterial blood pressure | |
| Pan et al. | Improved blood pressure estimation using photoplethysmography based on ensemble method | |
| CN103767694B (en) | Method for accurately extracting cuff pressure shockwave | |
| Yen et al. | Blood Pressure and Heart Rate Measurements Using Photoplethysmography with Modified LRCN. | |
| Ghosh et al. | Introduction of boosting algorithms in continuous non-invasive cuff-less blood pressure estimation using pulse arrival time | |
| Pu et al. | Cuff-less blood pressure estimation from electrocardiogram and photoplethysmography based on VGG19-LSTM network |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20180420 |
|
| WD01 | Invention patent application deemed withdrawn after publication |