CN102955889A - Pulse wave reconstruction method for extracting time domain feature points - Google Patents
Pulse wave reconstruction method for extracting time domain feature points Download PDFInfo
- Publication number
- CN102955889A CN102955889A CN 201110250129 CN201110250129A CN102955889A CN 102955889 A CN102955889 A CN 102955889A CN 201110250129 CN201110250129 CN 201110250129 CN 201110250129 A CN201110250129 A CN 201110250129A CN 102955889 A CN102955889 A CN 102955889A
- Authority
- CN
- China
- Prior art keywords
- frequency
- pulse wave
- signal
- peak
- original
- 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.)
- Granted
Links
Images
Landscapes
- Measuring Pulse, Heart Rate, Blood Pressure Or Blood Flow (AREA)
Abstract
The pulse wave reconstruction method that the invention discloses a kind of for extracting temporal signatures point includes,Following steps: a) recording body surface pulse wave signal using pulse wave signal measuring device,And be converted to digital signal; B) digital signal is converted to by power spectrum chart by Fast Fourier Transform (FFT); C) fundamental frequency size is calculated,That is frequency value F 0 corresponding to peak-peak; D) using this frequency values as gap size,Frequency range be divided into 0.5--1.5F0,1.5--2.5F0...6.5--7.5F07 region,Maximizing ai and corresponding frequency values fi is distinguished in each region,I is the natural number from 1~7; E) in phase frequency figure,Calculate separately the corresponding phase value θ i of frequency fi; F) ai,Fi,θ i substitutes into formula
Obtain the reconstruction signal of original signal equal length; Wherein t=0:1/fs:4000/fs. Similarity is very high in the time domain with original waveform for the waveform that the present invention reconstructs.
Description
Technical field
The present invention relates to a kind of disposal route for human pulse ripple and other periodicity biomedicine signals.
Background technology
Diagnosis by feeling the pulse is the ancient medical diagnostic method of China, but it is strong to exist subjectivity, relies on the shortcoming of doctor's experience, modern medicine require the expression of pulse condition and analysis objectify, digitizing, scientific.Pulse wave measurement and analytical technology also become current important scientific research field.Pulse wave and other periodicity biomedicine signals from body surface measures owing to a variety of causes, often when having certain rule, have inevitably with a lot of noises, so that there is certain difficulty in the further analytical calculation of pulse wave.
Traditional way is to adopt the whole bag of tricks that pulse wave is carried out first smoothing processing in time domain, common are median method, the linear slide method of average, five-spot triple smoothing etc. (seeing Zhang Zhi's prosperous " engineering analysis of pulse wave and clinical practice ").But the effect to eliminating noise of these smoothing methods is limited.For example, median method and the linear slide method of average are relatively good for the rising edge smooth effect of pulse wave, but for negative edge and wave trough position, still have a lot of noises after level and smooth.And five-spot triple smoothing, smooth effect is bad at least for level and smooth number of times, and level and smooth number of times is too much, then can make peak reduction, build broadens, and the grown form of waveform produces larger change.And common filtering technique, for example band is logical and low-pass filter, still can not be with near the miniwatt noise removal the passband.
Summary of the invention
This patent has proposed a kind of pulse wave reconstructing method for extracting temporal signatures point, and the method can keep the chief component of original signal well, removes simultaneously the low-power noise.
For realizing above target, a kind of pulse wave reconstructing method for extracting temporal signatures point of the present invention comprises the steps:
A) adopt the pulse wave signal measurement mechanism to record the body surface pulse wave signal, and be converted to digital signal;
B) by Fast Fourier Transform (FFT) described digital signal is converted to power spectrum chart;
C) calculate fundamental frequency size, the i.e. corresponding frequency value F of peak-peak
0
D) take this frequency values as gap size, frequency range is divided into 0.5---1.5F
0, 1.5---2.5F
0... 6.5---7.5F
07 zones, difference maximizing a in each zone
iAnd corresponding frequency values f
i, i is the natural number from 1～7;
E) in phase frequency figure, difference calculated rate f
iCorresponding phase value θ i;
F) a
i, f
i, θ i substitution formula
Namely obtain the reconstruction signal of original signal equal length; T=0:1/fs:4000/fs wherein.
The present invention adopts several cosine functions formation that superposes successively, and the information such as the required frequency of cosine signal, amplitude, phase place obtain from the amplitude-frequency spectrum (or power spectrum) of signal and phase frequency spectrum respectively; And frequency spectrum is to adopt fast fourier transform algorithm to obtain, amplitude is utilized the characteristics of amplitude-frequency spectrum (or power spectrum) peak intervals equalization of pulse wave, adopting the by stages to get peaked method obtains, frequency values is the corresponding horizontal ordinate of peak value, phase place is the phase value on the corresponding horizontal ordinate in the phase frequency curve, therefore, the waveform of reconstruct and original waveform similarity on time domain is very high.
Description of drawings
Fig. 1 is the calculation process schematic diagram of a kind of pulse wave reconstructing method for extracting temporal signatures point of the present invention.
Fig. 2 is the nothing wound radial artery pulse wave original waveform figure based on pressure.
Fig. 3 is for to carry out getting the power spectrum chart that squared magnitude obtains after the FFT conversion to one section original pulse wave signal.
Fig. 4 is one of original pulse wave signal and reconstruct pulse wave signal comparison of wave shape schematic diagram.
Fig. 5 is two of original pulse wave signal and reconstruct pulse wave signal comparison of wave shape schematic diagram.
Fig. 6 is three of original pulse wave signal and reconstruct pulse wave signal comparison of wave shape schematic diagram.
Wherein, 1,2,3,4,5,6,7 are respectively the crest of the 1st, 2,3,4,5,6,7 harmonic wave in the power spectrum chart, and 5a, 6a, 7a are the waveform of original signal in the example, and 5b, 6b, 7b are respectively the corresponding reconfiguration waveform of pulse wave.
Embodiment
Human body surface pulse wave signal reconstructing method for the extraction of temporal signatures point of the present invention, concrete methods of realizing is as follows:
Record the body surface pulse wave signal with the pulse wave signal measurement mechanism, this kind device generally is divided into pressure resistance type according to the difference of sensor, piezoelectric type, and ultrasonic signal etc.; According to whether invade human body, be divided into have the wound signal and without the wound signal; Different according to measured position, be divided into the radial artery signal, the arteria carotis signal, the femoral artery signal, and the ankle arterial signal etc.
The above various pulse wave signal all can be used method of the present invention and be reconstructed, and does further computational analysis.Other other biosome cyclical signal, such as brain wave, electrocardio ripple etc., but also adopting said method is reconstructed and further computational analysis.
One section steady pulse wave signal recording, by analog to digital conversion, the pre-service such as amplification deposit computing machine in, just can utilize the calculation procedure in the computing machine to calculate.The present invention utilizes the FFT (Fast Fourier Transform (FFT)) in the matlab software that pulse wave signal is transformed into frequency domain, and its basic calculating formula is:
Wherein, N is that the calculating of carrying out Fast Fourier Transform (FFT) is counted, and by formula as can be known: after transforming to frequency domain, it is the sequence of complex numbers of N that data become a row length, so the frequency spectrum that is used for analyzing has real frequency spectrum according to different needs, and empty frequency spectrum, amplitude-frequency spectrum, phase frequency spectrum.The present invention is because at first the power of analytic signal distributes, therefore first amplitude frequency diagram being done a square processing obtains power spectrum chart, from power spectrum chart, can see, that the power spectrum of pulse wave generally can obviously be distinguished by naked eyes, interval uniform 6---8 spikes form, as shown in Figure 3, wherein general first amplitude is the highest, and lower successively the back, and 6---8 spikes substantially no longer include obvious power later and distribute.And since signal limit arranged, cause inevitable spectral leakage, so peak value broadens and secondary lobe, generally speaking, a spike only includes a cosine signal at the peak point place, so can regard as: original signal is comprised of a series of equally spaced cosine signals.Because pulse wave signal is by periodic aroused in interest causing, has certain periodicity, this also can match by the principle that Fourier series is decomposed with periodic signal.
The characteristics that equidistant arrangement is arranged according to the spike of power spectrum, obtain first the frequency of peak-peak (i.e. the first peak value), it is the frequency of original signal---fundamental frequency, be worth as the interval take this, 1 to 7 spike in-scope is divided into 7 zones, difference maximizing in each zone, it namely is respectively the peak value of 7 spikes, it namely is the amplitude of each cosine function that the peak value size is asked root, the frequency that peak value is corresponding namely is the frequency of each cosine function, on the phase frequency spectrum, ask the value of corresponding frequency, i.e. the initial phase of each cosine function.Because phase frequency curve changes violent, so frequency should obtain accurately as far as possible, under the certain prerequisite of signal length, should adopt the high frequency spectrum of fineness, nfft is large as far as possible even conversion is counted.The parameter substitution formula such as corresponding amplitude, frequency, phase place are then obtained the signal of reconstruct, the absolute value of reconstruction signal is specifically demarcated in conjunction with the absolute value of original signal.
Reconstructing method of the present invention is owing to the high-power composition in the reconstruction signal retention rate original signal, so the similarity very high with the waveform maintenance of original signal.And because after the stack of the limited cosine signal of number, the waveform smooth degree is very good, so when extracting temporal signatures point, very large advantage is arranged, be that the data after the original smooth treatment can not be compared.
In addition, because this method is extracted reconstruction parameter by frequency-region signal, with simple triangle cosine function reconfiguration waveform, technical easy realization, can be finished fast by signal processing software commonly used.Applied widely, because no matter how the shape of time domain waveform changes, because it is periodic that the periodic contractile motion of human heart, pulse always approach, so (except the serious ARR situation) that above method always is suitable for for general human body pulse wave.
As example of the present invention, that utilizes that this laboratory builds voluntarily gathers the radial artery waveform of different people, sample frequency 446HZ, sampling length 4000 points based on the pulse wave acquisition system of pressure, Fast Fourier Transform (FFT) is counted and is got 131096, and software for calculation utilizes matlab7.0 to write calculation procedure.Its basic step is as follows:
For reducing the impact of spectral leakage, original signal is the hamming window through windowing process herein.Through Fast Fourier Transform (FFT), obtain power spectrum chart.
Ask the fundamental frequency size, i.e. the corresponding frequency value F of peak-peak
0
Take this frequency values as gap size, frequency range is divided into 0.5---1.5F
0, 1.5---2.5F
0... .6.5---7 zones such as 7.5 grades, difference maximizing a in each zone
iAnd corresponding frequency values f
i, i is from 1---7 natural number.
In phase frequency figure, ask respectively frequency f
iCorresponding phase value θ i.
A
i, f
i, θ i substitution formula
The reconstruction signal that namely obtains the original signal equal length is t=0:1/fs:4000/fs wherein, namely obtains the reconstruction signal of original signal equal length.Oscillogram after the reconstruct and original waveform figure are to such as shown in Fig. 4,5,6, and as seen from the figure, the waveform after the reconstruct keeps the similarity very high with original waveform.
Claims (1)
1. a pulse wave reconstructing method that is used for extracting temporal signatures point comprises the steps:
A) adopt the pulse wave signal measurement mechanism to record the body surface pulse wave signal, and be converted to digital signal;
B) by Fast Fourier Transform (FFT) described digital signal is converted to power spectrum chart;
C) calculate fundamental frequency size, the i.e. corresponding frequency value F of peak-peak
0
D) take this frequency values as gap size, frequency range is divided into 0.5---1.5F
0, 1.5---2.5F
0... 6.5---7.5F
07 zones, difference maximizing a in each zone
iAnd corresponding frequency values f
i, i is the natural number from 1～7;
E) in phase frequency figure, difference calculated rate f
iCorresponding phase value θ i;
F) a
i, f
i, θ i substitution formula
Namely obtain the reconstruction signal of original signal equal length; T=0:1/fs:4000/fs wherein.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201110250129.4A CN102955889B (en) | 2011-08-29 | 2011-08-29 | A kind of pulse wave reconstruction method for extracting temporal signatures point |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201110250129.4A CN102955889B (en) | 2011-08-29 | 2011-08-29 | A kind of pulse wave reconstruction method for extracting temporal signatures point |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN102955889A true CN102955889A (en) | 2013-03-06 |
| CN102955889B CN102955889B (en) | 2015-12-16 |
Family
ID=47764689
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201110250129.4A Expired - Fee Related CN102955889B (en) | 2011-08-29 | 2011-08-29 | A kind of pulse wave reconstruction method for extracting temporal signatures point |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102955889B (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103767694A (en) * | 2014-01-06 | 2014-05-07 | 西安交通大学 | Method for accurately extracting cuff pressure shockwave |
| CN104021295B (en) * | 2014-06-12 | 2017-06-06 | 广州三星通信技术研究有限公司 | Cluster feature fusion method and device for moving identification |
| CN108289624A (en) * | 2015-11-27 | 2018-07-17 | 浜松光子学株式会社 | Blood pressure information computing device, blood pressure information computational methods, blood pressure information calculation procedure and the storage medium for storing the program |
| JPWO2017188099A1 (en) * | 2016-04-27 | 2019-02-14 | 旭化成株式会社 | Device, terminal and biometric information system |
| CN109645963A (en) * | 2019-02-25 | 2019-04-19 | 深圳奥思顿健康科技有限公司 | It is a kind of for detecting the intelligent analysis system of vital sign and diagnosis by feeling the pulse |
| CN111588345A (en) * | 2020-06-18 | 2020-08-28 | 歌尔科技有限公司 | Eye disease detection method, AR glasses and readable storage medium |
| CN113598755A (en) * | 2021-07-19 | 2021-11-05 | 燕山大学 | Human body rhythm motion parametric representation analysis method based on fast Fourier transform |
| CN114795168A (en) * | 2022-06-24 | 2022-07-29 | 昂科信息技术（上海）股份有限公司 | Method and system for calculating heart rate of vital sign parameter |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101156771A (en) * | 2007-09-28 | 2008-04-09 | 天津市先石光学技术有限公司 | Method and apparatus for improving vascellum hardness measurement precision base on pulse wave frequency spectrum analysis |
-
2011
- 2011-08-29 CN CN201110250129.4A patent/CN102955889B/en not_active Expired - Fee Related
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103767694A (en) * | 2014-01-06 | 2014-05-07 | 西安交通大学 | Method for accurately extracting cuff pressure shockwave |
| CN103767694B (en) * | 2014-01-06 | 2015-07-08 | 西安交通大学 | Method for accurately extracting cuff pressure shockwave |
| CN104021295B (en) * | 2014-06-12 | 2017-06-06 | 广州三星通信技术研究有限公司 | Cluster feature fusion method and device for moving identification |
| CN108289624A (en) * | 2015-11-27 | 2018-07-17 | 浜松光子学株式会社 | Blood pressure information computing device, blood pressure information computational methods, blood pressure information calculation procedure and the storage medium for storing the program |
| US11089965B2 (en) | 2015-11-27 | 2021-08-17 | Hamamatsu Photonics K.K. | Blood pressure information calculating device, blood pressure information calculating method, blood pressure information calculating program, and recording medium for recording said program |
| JPWO2017188099A1 (en) * | 2016-04-27 | 2019-02-14 | 旭化成株式会社 | Device, terminal and biometric information system |
| CN109645963A (en) * | 2019-02-25 | 2019-04-19 | 深圳奥思顿健康科技有限公司 | It is a kind of for detecting the intelligent analysis system of vital sign and diagnosis by feeling the pulse |
| CN111588345A (en) * | 2020-06-18 | 2020-08-28 | 歌尔科技有限公司 | Eye disease detection method, AR glasses and readable storage medium |
| CN113598755A (en) * | 2021-07-19 | 2021-11-05 | 燕山大学 | Human body rhythm motion parametric representation analysis method based on fast Fourier transform |
| CN113598755B (en) * | 2021-07-19 | 2023-02-21 | 燕山大学 | Human body rhythm motion parametric representation analysis method based on fast Fourier transform |
| CN114795168A (en) * | 2022-06-24 | 2022-07-29 | 昂科信息技术（上海）股份有限公司 | Method and system for calculating heart rate of vital sign parameter |
Also Published As
| Publication number | Publication date |
|---|---|
| CN102955889B (en) | 2015-12-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102955889B (en) | A kind of pulse wave reconstruction method for extracting temporal signatures point | |
| Islam et al. | Study and analysis of ecg signal using matlab &labview as effective tools | |
| CN110327036B (en) | Method for extracting respiratory signal and respiratory frequency from wearable electrocardiogram | |
| CN102697493B (en) | Method for rapidly and automatically identifying and removing ocular artifacts in electroencephalogram signal | |
| CN103610461B (en) | Based on the EEG Signal Denoising method of dual density small echo neighborhood dependent thresholds process | |
| Campolo et al. | ECG-derived respiratory signal using empirical mode decomposition | |
| CN104161509A (en) | Heart rate variability analyzing method based on amplitude spectrum and instruments | |
| CN103479349A (en) | Electrocardiosignal data acquisition and processing method and system | |
| WO2007021645A2 (en) | Method and system for limiting interference in electroencephalographic signals | |
| CN107693004A (en) | Fetal ECG extraction and fetal heart frequency recognition methods based on hilbert conversion | |
| CN105956388A (en) | Human body vital sign signal separation method based on VMD (Variational Mode Decomposition) | |
| CN105266800B (en) | One kind is based on fetus electrocardio blind separation under Low SNR | |
| Patro et al. | De-noising of ECG raw signal by cascaded window based digital filters configuration | |
| US20190142289A1 (en) | Robust real-time heart rate monitoring method based on heartbeat harmonics using small-scale radar | |
| CN110327029A (en) | A kind of heart rate and heart rate variability monitoring method based on microwave perception | |
| CN106539580B (en) | Continuous monitoring method for dynamic change of autonomic nervous system | |
| TW201717845A (en) | Devices, systems, and methods for determining heart rate of a subject from noisy electrocardiogram data | |
| Ranjan et al. | Cardiac artifact noise removal from sleep EEG signals using hybrid denoising model | |
| Ren et al. | Noise reduction based on ICA decomposition and wavelet transform for the extraction of motor unit action potentials | |
| CN114027804A (en) | Pulse condition diagnosis method, device and readable storage medium | |
| Wang et al. | Analysis of pulse waveforms preprocessing | |
| Arvind et al. | Automated detection of PD resting tremor using PSD with recurrent neural network classifier | |
| Sareen et al. | Nadi Yantra: a robust system design to capture the signals from the radial artery for assessment of the autonomic nervous system non-invasively | |
| Zheng et al. | Detection of QRS complexes based on biorthogonal spline wavelet | |
| Chen et al. | A fast ECG diagnosis using frequency-based compressive neural network |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20151216 Termination date: 20180829 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |