CN104168821B - The system and method monitored for ECG - Google Patents
The system and method monitored for ECG Download PDFInfo
- Publication number
- CN104168821B CN104168821B CN201380015976.XA CN201380015976A CN104168821B CN 104168821 B CN104168821 B CN 104168821B CN 201380015976 A CN201380015976 A CN 201380015976A CN 104168821 B CN104168821 B CN 104168821B
- Authority
- CN
- China
- Prior art keywords
- group
- detail coefficients
- approximation coefficient
- coefficient
- ripple
- 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.)
- Expired - Fee Related
Links
- 238000000354 decomposition reaction Methods 0.000 claims abstract description 39
- 238000000034 method Methods 0.000 claims description 58
- 238000004422 calculation algorithm Methods 0.000 abstract description 13
- 235000013601 eggs Nutrition 0.000 abstract 1
- 238000010586 diagram Methods 0.000 description 11
- 238000003860 storage Methods 0.000 description 11
- 230000000694 effects Effects 0.000 description 8
- 238000011112 process operation Methods 0.000 description 5
- 238000005070 sampling Methods 0.000 description 5
- 210000003491 Skin Anatomy 0.000 description 4
- 238000004590 computer program Methods 0.000 description 4
- 230000005540 biological transmission Effects 0.000 description 3
- 230000000875 corresponding Effects 0.000 description 3
- 238000002592 echocardiography Methods 0.000 description 3
- 230000005611 electricity Effects 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 230000003287 optical Effects 0.000 description 3
- 210000003205 Muscles Anatomy 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 238000005304 joining Methods 0.000 description 2
- 238000005096 rolling process Methods 0.000 description 2
- WHWDWIHXSPCOKZ-UHFFFAOYSA-N 6,10,14-trimethylpentadecan-2-one Chemical compound CC(C)CCCC(C)CCCC(C)CCCC(C)=O WHWDWIHXSPCOKZ-UHFFFAOYSA-N 0.000 description 1
- 210000004165 Myocardium Anatomy 0.000 description 1
- 230000002457 bidirectional Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000000747 cardiac effect Effects 0.000 description 1
- 238000007906 compression Methods 0.000 description 1
- 230000001276 controlling effect Effects 0.000 description 1
- 230000002596 correlated Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 230000002349 favourable Effects 0.000 description 1
- 238000009499 grossing Methods 0.000 description 1
- 101710031899 moon Proteins 0.000 description 1
- 230000000737 periodic Effects 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 230000001052 transient Effects 0.000 description 1
- 230000002861 ventricular Effects 0.000 description 1
Abstract
The present invention provides a kind of and is attributable to, to remove or to reduce, the method for noise, system, equipment and the device that EMG and/or motion artifacts produce in described ECG signal for processing ECG signal.The novel algorithm that all aspects of described device share can comprise and ECG signal carries out wavelet decomposition to produce one group of approximation coefficient and to organize detail coefficients more, it is second order polynomial by the subset local fit of described group of approximation coefficient, described group of approximation coefficient is adjusted by the described second order polynomial through local fit, some coefficients in described detail coefficients are set to zero, and detail coefficients based on the described group of approximation coefficient revised or described many group amendments rebuilds the EGG signal of the noise with minimizing.
Description
The cross reference of related application
Subject application advocates U.S. Provisional Application case No. 61/616,890 and 2013 3 filed in 28 days March in 2012
The priority of U.S. Patent Application No. 13/797,889 filed in the moon 12, said two application case is at this clearly
It is incorporated herein in the way of it quotes in full.
Technical field
The present invention relates to ECG monitor and, systems for processing ECG signal to reduce making an uproar in signal
The system of sound, method and apparatus.
Background technology
Electrocardiogram (ECG) is the signal of telecommunication representing the pulse produced by the heart.ECG signal and waveform be by P, QRS and
T ripple characterizes, and as seen in Figure 1, it is at myocardium atrium and ventricular systole and to loosen period generation.Fig. 1 says
The example of bright typical case's ECG signal 10, it is comparatively speaking without any destruction or noise.ECG signal 10 demonstrates characteristic P
Ripple 12, Q ripple 14, R ripple 16, S ripple 18 and T ripple 20 feature.Can use on the skin adhering to experimenter
The electrode in region outside the heart obtains ECG signal.The sticker on a surface of electrode can be used these electricity
Pole attaches on skin.Electrode then can obtain signal from the electrical activity in the heart and around the heart.Term " ECG signal " and
" signal " can refer to that the simulation of ECG electrode exports, and treated or undressed sampled data point, such as,
Use the sampled data point that A/D converter produces.Analogue signal and/or sampled data point can be filtered.Can
Then process described signal further, store described signal and/or transmit described signal or be shipped to be configured to display letter
Number device.
The artifact comprising power line interference, motion artifacts and electromyogram (EMG) signal destroys ECG signal because of " noise "
And the performance of effect characteristics detection algorithm.It has an effect on the Accurate Diagnosis of clinicist.Movement owing to test experimenter
With the electrical activity of other muscle near the heart, motion and EMG artifact the most usually occur.Destroyed ECG signal (example
As, there is the ECG signal of motion and EMG artifact) it is found in Fig. 2 and 3.The destroyed length in ECG signal
Periodic noise (the most notable) be typically caused by subject motion and be referred to as motion artifacts.Fig. 2 and 3 liang
Upper frequency distortion in the destroyed ECG signal shown in person may be by around the heart or the electricity of the muscle near the heart
Activity causes.The distortion of this type is referred to as EMG artifact.
Summary of the invention
The system of the present invention, method and apparatus each have some aspects, and any single one in described aspect is the most alone
It is responsible for attribute needed for it.In the case of being not intended to the scope of the present invention as expressed by claims below, existing
Some features will discussed compactly.After considering this discussion, and especially at the chapters and sections reading entitled " embodiment "
Afterwards, it should be understood that how inventive feature provides advantage, comprise the experimenter that is attributable to reducing in ECG signal to transport
The noise that dynamic and non-cardiac activity produces.
In an aspect, it is provided that a kind of for processing ECG signal with the method reducing the noise in described ECG signal.
Described method comprises the ECG signal wavelet decomposition containing artifact, produces one group of approximation coefficient and organizes detail coefficients more.
Described approximation coefficient can be revised, remove motion artifacts.Can be by first by the neighbouring subset local fit of described approximation coefficient
Described approximation coefficient is revised for second order polynomial.After described approximation coefficient local fit is second order polynomial, can
By the described second order polynomial through local fit, by described second order polynomial is deducted from described corresponding approximation coefficient
Adjust described group of original approximation coefficient, thus produce the approximation coefficient of one group of amendment.Described detail coefficients can be carried out
Threshold process is to produce the detail coefficients of many group amendments, and the detail coefficients of described many group amendments may be used to remove described ECG
EMG artifact in signal.Described threshold process can be position threshold process, and can by be at least partially based on described carefully
Some detail coefficients in described detail coefficients are set to zero relative to the position of the R ripple of described ECG signal by joint coefficient
Perform.In another embodiment, described method comprises further from the approximation coefficient of described amendment with through threshold process
Detail coefficients rebuild described ECG signal, thus produce compared with described original ECG signal for there is minimizing
The ECG signal of noise.
In another aspect, it is provided that a kind of for processing ECG signal to reduce the system of the noise in described ECG signal
Or equipment.Described system or equipment comprises processor, and described processor is configured to little containing noisy ECG signal
Wave Decomposition, thus produce one group of approximation coefficient and organize detail coefficients more.Described processor is further configured with amendment described
Approximation coefficient and/or detail coefficients.In another embodiment, described processor can be further configured with from described amendment
Approximation coefficient and through threshold process (such as, amendment) detail coefficients rebuild described ECG signal, thus produce and institute
State the ECG signal of the noise for original ECG signal compares with minimizing.In yet another embodiment, described
System or equipment comprises ECG electrode, A/D converter further, and is configured for use in the health being adhered to experimenter
On antenna.
In another aspect, it is provided that a kind of for processing ECG signal to reduce the device of the noise in described ECG signal.
Described device comprise for by ECG signal wavelet decomposition with produce one group of approximation coefficient and organize more detail coefficients, amendment institute
State group approximation coefficient and described detail coefficients is carried out the device of threshold process.In one embodiment, described device can
Comprise further for rebuilding institute from the approximation coefficient of described amendment with through (such as, amendment) detail coefficients of threshold process
State the device of ECG signal.
Mthods, systems and devices described in the present invention can be implemented with hardware, software, firmware or its any combination.
If implemented with hardware, then can equipment be come real as integrated circuit, processor, discrete logic or its any combination
Existing.If implemented with software, then can be at such as microprocessor, special IC (ASIC), field programmable gate
One or more processors such as array (FPGA) or digital signal processor (DSP) perform software.Initially can be described by performing
The software of mthods, systems and devices is stored in computer-readable media and is loaded within a processor and performs described software.
Therefore, present invention also contemplates that a kind of computer-readable storage medium including instruction, described instruction is in processor
Make during execution described processor when receiving ECG signal, described ECG signal wavelet decomposition is near to produce one group
Like coefficient with organize detail coefficients more, revise described group of approximation coefficient, and described detail coefficients is carried out threshold process, thus
Reduce the noise in described ECG signal.
The details of the one or more aspects of the present invention is illustrated by the accompanying drawings and the following description.From describing and graphic and from power
Profit claim will be apparent to the further feature of technology described in the present invention, target and advantage.
Accompanying drawing explanation
Fig. 1 shows have characteristic P, the typical ECG signal of QRS and T ripple or the example of waveform.
Fig. 2 shows the example of the ECG signal destroyed by EMG and motion artifacts.
Fig. 3 shows the example of the ECG signal destroyed by EMG and motion artifacts.
Fig. 4 shows for processing ECG signal to remove or to reduce the side of the noise caused by EMG and/or motion artifacts
The process flow diagram flow chart of one embodiment of method.
Fig. 5 shows the process flow diagram flow chart of an embodiment of the method for amendment approximation coefficient.
Fig. 6 displaying revises the method for detail coefficients by least some detail coefficients in detail coefficients is set to zero
The process flow diagram flow chart of one embodiment.
Fig. 7 shows for processing ECG signal with removal or the process flow of an embodiment of the method reducing noise
Figure.
Fig. 8 shows the ECG signal that is corrupted by noise and treated thus between ECG signal after reducing noise ratio
Example relatively.
Fig. 9 shows the ECG signal that is corrupted by noise and treated thus between ECG signal after reducing noise ratio
Example relatively.
Figure 10 show the ECG signal that is corrupted by noise and treated thus between ECG signal after reducing noise
Example relatively.
Figure 11 shows for processing ECG signal to reduce or to remove the system of noise, equipment or the dress in ECG signal
The functional block diagram of the embodiment put.
Figure 12 A shows for processing ECG signal to remove or to reduce the system of noise, equipment or the dress in ECG signal
The functional block diagram of the embodiment put.
Figure 12 B show is for processing ECG signal to remove or to reduce the system of noise, equipment or the dress in ECG signal
The functional block diagram of the embodiment put.
Detailed description of the invention
Novel method described herein and system relate to processing ECG signal to reduce the noise in ECG signal.Root
According to an embodiment, it is provided that a kind of method for processing ECG signal.Described method comprises ECG signal small echo
Decompose to produce one group of approximation coefficient and to organize detail coefficients more, and revise described approximation coefficient and/or detail coefficients.At some
In embodiment, described method comprise further from amendment approximation coefficient and/or amendment (such as, at position threshold
Reason) detail coefficients reconstruction ECG signal.As described further below, the embodiment of these amendments is reducing ECG letter
In common noise source in number highly effective.
Wavelet decomposition relates to the resolution approximate signal being gradually lowered, the wherein difference of successive approximation each self-defined details letter
Number.Wavelet decomposition produces one group of approximation coefficient and organizes detail coefficients more.Decompose n-th order and produce n rank approximation coefficient An,
With many group detail coefficients D1To Dn.Produced approximation coefficient represents that the approximate function at discrete point is (such as, in decomposition
After approximate signal) value.Detail coefficients represents the difference of the value of the successive approximation function at discrete point.Wavelet decomposition describes
" the theory of multiresolution signal decomposition: Wavelet representation for transient (A Theory for Multiresolution Signal in Mallat, S.
Decomposition:The Wavelet Representation) " (IEEE Trans.Pattern Anal Mach.Intell, the
Volume 11, page 674 to 693, in July, 1989) in, the entire content of described file is hereby incorporated herein by
In.One method of wavelet decomposition is well known, but the method being previously not carried out coefficient processing described below.
Fig. 4 shows the embodiment of method processing ECG signal to remove or to reduce noise in ECG signal
Process flow diagram flow chart.At frame 110, receive the ECG signal being likely to be of at least one EMG and/or motion artifacts.?
At frame 112, the ECG signal received at frame 110 is performed wavelet decomposition.The wavelet decomposition of frame 112 produce one group near
Like coefficient (represented by frame 114) with organize detail coefficients (represented by frame 116) more.Frame 114 and 116 is not single
Operation, but the result of the wavelet decomposition of frame 112.The approximation coefficient of frame 114 represents the discrete point determined at frame 112
The value of the last approximate function (such as, approximate signal after decomposing) at place.The detail coefficients of frame 116 represents discrete
The difference between successive approximation function at Dian.If not revising approximation coefficient and the detail coefficients of frame 114 and 116, that
Substantially can from these coefficient reconstruction at frame 112 at receive original ECG signal.
At frame 118, amendment is from least some approximation coefficient in the approximation coefficient of frame 114.In some embodiments
In, the amendment at frame 118 reduces the approximation coefficient mean deviation away from baseline.Preferably, the amendment at frame 118 can comprise
Reduce the mean deviation away from baseline, relate generally to that there is the time scale longer than the time scale of R ripple present in data set
Skew.In some embodiments, the approximation coefficient revised in this way can be used at frame 122 to rebuild ECG
Signal, described ECG signal compared with the original ECG signal received at frame 110 for there is the motion of minimizing
Artifact.Below in reference to Fig. 5, a kind of appropriate algorithm is described.
At frame 120, revise from least some detail coefficients in the detail coefficients of frame 116, repair from place's generation is multiple
The detail coefficients changed.The well known threshold process form of a kind of noise reduced in signal is for " at value threshold value
Reason ", wherein the detail coefficients of the wavelet decomposition of signal is that the value being at least partially based on detail coefficients is (such as, based on details
Whether coefficient has the value more than or less than threshold magnitude) carry out threshold process (e.g. setting it to zero).Although
In the context of method described herein, this technology can also be used with to revise some detail coefficients, it has been found that have
Use sharply herein with the method for " position threshold process " form reference to the noise decrease completing in ECG signal,
Described " position threshold process " relates to the time location being at least partially based on some detail coefficients in one group of detail coefficients
Revise some detail coefficients.See below Fig. 6 more detailed description, for determining that the position of amendment can be relative in signal
The position of R ripple.In some embodiments, the distance away from R ripple is set to zero more than the detail coefficients of preset distance,
And make the distance away from R ripple keep constant less than or equal to the detail coefficients of preset distance.It addition, in some embodiments,
Can be by all detail coefficients (such as, D of array lower-order detail coefficients1、D2And/or D3) it is set to zero.The most detailed
The thin particular discussing this amendment operation.Revising some detail coefficients (such as, by thin by all organize
Joint coefficient be set to zero, or process by position threshold) after many groups detail coefficients constitute revise detail coefficients.One
In a little embodiments, can use the detail coefficients of amendment to rebuild ECG signal at frame 122, described ECG signal and
The original ECG signal received at frame 110 has the EMG artifact of minimizing for comparing.
At frame 122, use approximation coefficient and the detail coefficients (at least some in described approximation coefficient and/or detail coefficients
Coefficient is revised) rebuild the ECG signal of artifact with minimizing.In some embodiments, can use without repairing
The approximation coefficient (such as, substantially such as the approximation coefficient from frame 114) changed and detail coefficients (as described in detail coefficients extremely
Fewer detail coefficients are set to zero according to frame 120) rebuild ECG signal.In some embodiments, can use
The detail coefficients (such as, substantially such as the detail coefficients from frame 116) of unmodified and approximation coefficient (as described in approximation coefficient
In at least some approximation coefficient revise according to frame 118) rebuild ECG signal.The ECG letter rebuild at frame 122
Number can have minimizing or the motion through removing and/or EMG artifact.
Fig. 5 illustrates to revise the process flow of an embodiment of the method for at least some approximation coefficient in approximation coefficient
Figure.At frame 210, calculate one group of approximation coefficient A according to ECG signalK, wherein K can represent to calculate approximation system
The order of the decomposition of number.At frame 212, calculate and described in local fit, organize approximation coefficient AKSecond order polynomial, thus
Produce and reside at one group on the multinomial of the institute's matching point through local fitDescribed local fit can use local to add
Power method is carried out so that through the point of local fitOne stock-traders' know-how fits to the feature (as shown in FIG. 2) of motion artifacts,
There is the contribution (as demonstrated in Figure 1) of little primary waves feature (such as, PQRST ripple) from ECG signal.At frame
At 214, each is calculatedFrom the original A of its correspondenceKIn deduct to produce the approximation coefficient of one group of amendmentIn some embodiments, the approximation coefficient of the described group of amendment calculated at frame 214 it is usable inRebuild tool
The ECG signal of that be reduced or through removing motion artifacts.It will be appreciated that method as described above and the operation of method
Any combination can perform in substantially real time, such as, when obtaining ECG signal from experimenter.
In some embodiments, the local fit of frame 212 can use local weighted scatterplot to smooth (LOWESS) algorithm
Realizing, described arthmetic statement is in " sane local weighted recurrence and the smooth scatterplot (Robust of Cleveland, W.
Locally Weighted Regression and Smoothing Scatterplots)”(Journal of the American
Statistical Association, volume 74, page 829 to 836, in December, 1979) in, described file whole
Content is incorporated herein by reference.In some embodiments, by the subset of multinomial local fit approximation coefficient,
Wherein said subset is by from organizing AKApproximation coefficient and predetermined number the approximation coefficient of next-door neighbour (such as, described
Predetermined number coefficient before approximation coefficient and afterwards) define.For carrying out the described son of the approximation coefficient of local fit
The time cycle span of collection definition local fit.(such as, it is used as local to intend for carrying out the time cycle span of local fit
The number of approximation coefficient of the fragment closed) should compare with ECG waveform for relatively long but relatively come compared with motion artifacts
Say relatively short, in order to make ECG waveform feature that the effect of matching to be minimized.In certain embodiments, such as, greatly
The time cycle span of about 100 to 500 milliseconds (such as, represents the approximation coefficient of the signal data of 100 to 500 milliseconds
Subset) can be used for the subset of local fit approximation coefficient, the most in some embodiments, it is suitable for finding about 200 milliseconds
's.
The relative weighting (" proximity weight ") of each coefficient in fitting to polynomial subset can at least part of base
Proximity in the center of the subset of coefficient to coefficient (such as, is at least partially based on the center of the coefficient subset away from approximation coefficient
The distance of point), wherein in matching, give greater weight to the coefficient being relatively close to center.In some embodiments,
After by the subset local fit of approximation coefficient being second order polynomial (" the first local polynomial fitting "), based on approximation
Each approximation coefficient that the distance antithetical phrase between corresponding point on coefficient and the first local polynomial fitting is concentrated gives new
Weighted value (" remaining weight "), wherein gives greater weight to the coefficient being relatively close to the first local polynomial fitting.Extremely
This remaining weight of each coefficient being at least partly based in subset, is second order polynomial by the subset local fit of coefficient again
(" the second local polynomial fitting ").This remaining weighting is used to inhibit matching to wanted ECG signal for matching
Contribution, thus force matching closely to follow longer cycle motion artifacts.In some embodiments, can be by anti-for this algorithm
Multiple once above to calculate the last local fit multinomial of each subset.Preferably, by this algorithm repeatedly twice by terms of
Calculate the last local fit multinomial of each subset.In some embodiments, to described group of approximation coefficient AKIn
Each approximation coefficient performs this algorithm and (such as, defines subset and use this algorithm to AKInterior each approximation coefficient carries out office
Portion's matching).Can be by with the approximation coefficient A original from described groupKCorresponding point centered by the last local of subset
The value of polynomial fitting determines groupIn each second order polynomial point.In some embodiments, described group of amendment
Approximation coefficientIt is by adjusting described group of original approximation coefficient AK, by described group of second order polynomial point
By by described group of second order polynomial pointFrom described group of original approximation coefficient AKIn deduct and produce.Described group of amendment
Approximation coefficient may be used to rebuild the ECG signal of the motion artifacts with minimizing.
Fig. 6 explanation revises the method for detail coefficients by least some detail coefficients in detail coefficients is set to zero
The process flow diagram flow chart of one embodiment.At frame 310, calculate D according to ECG signal1To DKGroup detail coefficients,
Wherein K can represent the order calculating the decomposition of detail coefficients.At frame 312, can be by D1To DNGroup detail coefficients
It is set to zero.Depend on the sampling frequency obtaining ECG signal, detail coefficients (such as, the D of relatively low order group1、
D2And D3) may not store the bulk information about ECG waveform.Therefore, in some embodiments, depend on
Sample frequency and/or other factors, can be by D1To DNGroup detail coefficients is set to zero, without loss about ECG waveform
Any bulk information.At frame 314, group D can be at least partially based onN+1To DKIn detail coefficients in some are thin
More described detail coefficients are set to zero relative to the position of the position of R ripple (such as, R " peak value ") by joint coefficient.Extremely
Fewer detail coefficients be set in frame 312 and/or 314 the described group of detail coefficients of zero include described group revise thin
Joint coefficient.The detail coefficients of described group of amendment may be used to rebuild the ECG signal without EMG artifact.
In some embodiments, use position threshold to process in block 314 and some detail coefficients are set to zero, wherein
Distance away from R crest value is set to zero more than the detail coefficients of preset distance, and distance is less than or equal to preset distance
Detail coefficients keeps constant.For example, for one group of detail coefficients DKWith R rolling land point RK[i] (wherein i is for identifying letter
Number treated period across time cycle in the index of each independent R ripple), away from RK[i] reaches 4 DKData point
Or less than 4 DKThe D of data pointKAll detail coefficients unmodified, and away from RK[i] reaches 4 data above points
All detail coefficients are set to zero.In some embodiments, detail coefficients is carried out position threshold process place away from
Threshold process to be carried out can be at least partially based on from (number of the data point away from R such as, used in amendment frame 314)
Described group of detail coefficients in the maximum cycle of ECG waveform and the width of ECG derive.This distance also can at least portion
Point based on calculate detail coefficients institute according to the sampling frequency of ECG signal and change.In some embodiments, permissible
It is experimentally determined this distance.Distance (such as, the number of point) as position threshold can be for the details of different order groups
Coefficient and difference are (such as, for D5Coefficient, threshold value is 4 points, and for D4Coefficient, threshold value is 3 points).?
Some detail coefficients are set to zero (by being set to zero by all organizing detail coefficients in frame 312, or by frame 314
In carry out position threshold process) after many groups detail coefficients constitute described group amendment detail coefficients.It will be appreciated that frame 312
Can perform independently from each other with 314, and need not both execution to produce the detail coefficients of amendment.Implement at some
In scheme, the described group of detail coefficients revised can be used to rebuild have a minimizing or the ECG of EMG artifact through removing
Signal.It will be appreciated that any combination of the operation of method as described above and method can perform in substantially real time.
Fig. 7 explanation is for processing ECG signal with removal or the process flow of an embodiment of the method reducing noise
Figure.At frame 410, obtain and may contain EMG and/or the ECG signal of motion artifacts.At frame 412, use
ECG signal from frame 410 is decomposed by wavelet decomposition.In some embodiments, such as, by ECG signal small echo
Decompose up to 5 rank.But, in some embodiments, ECG signal can be decomposed into less rank or bigger rank (such as,
Up to n rank), as described above.(such as) biorthogonal 5.5 small echo can be used to realize this decomposition.The up to ECG on 5 rank
The wavelet decomposition of signal will produce 5 rank approximation coefficient A5, and many group detail coefficients D1To D5.As described above, little
Wave Decomposition relates to gradually reducing resolution approximate signal, wherein difference definition detail signal (such as, the details of successive approximation
Coefficient).Therefore, group A5Interior approximation coefficient (such as, is decomposing up to 5 corresponding to 5 rank approximate functions at discrete point
Approximate signal after rank) value, and detail coefficients D1To D5Represent gradually the difference between the approximate function of order.
At frame 414, with organizing approximation coefficient A described in second order polynomial local fit5To produce one group through local fit
Multinomial pointIn some embodiments, with organizing approximation coefficient described in second order polynomial local fit, so that through office
The point of portion's matchingOne stock-traders' know-how fits to feature (as shown in FIG. 2) rather than the spy of ECG signal of motion artifacts
Levy (such as, QRS wave is combined) (as demonstrated in Figure 1).This situation can (such as) be had more than data set by local fit
In the presence of the subset of time cycle span of time cycle of ECG waveform complete, as discussed above.One
In a little embodiments, use group approximation described in the algorithm second order polynomial fit described in detail in the discussion of Fig. 5 above
Coefficient.At frame 416, by the second order polynomial point through local fitFrom described group of approximation coefficient A5In deduct, with
Obtain the approximation coefficient of one group of amendment
At frame 418, it is at least partially based on derivativeDetermine the place of R ripple.In some embodiments, can pass through
To derivativeCarry out amplitude threshold and process the place determining R ripple.Because R rolling land point should have any of ECG waveform
The maximum derivative of feature/ripple, so the method can be used.In order to determine that the specific amplitude threshold value in the place of R ripple can be at least
It is based partially on derivativeMeansigma methods and standard deviation.When threshold value is that meansigma methods adds 1.5 times of standard deviation intervals, original
The remaining set of each R ripple of signalIn there typically will be a point.In some embodiments, available experiment method
Or known any other method determines the place of R ripple in use art.
At frame 420, detail coefficients can be at least partially based on relative to the R ripple such as determined in frame 418 place time
Between position revise D4And D5At least some detail coefficients in detail coefficients.In some embodiments, such as, as
Really D4And D5The detail coefficients distance away from R ripple in group detail coefficients is more than preset distance, then be at least partially based on as
The place of the R ripple determined in frame 418 is by D4And D5Detail coefficients in group detail coefficients is set to zero.Real at some
Execute in scheme, be at least partially based on the detail coefficients position relative to the place of R ripple, to D4And D5In detail coefficients
At least some detail coefficients carries out position threshold process, e.g. setting it to zero, substantially as above in the opinion of Fig. 6
Described in stating.At frame 422, by D1To D3Detail coefficients is set to zero.
At frame 424, be based, at least in part, in frame 416 amendment obtained approximation coefficient and/or at frame 420 and/or
The detail coefficients of the amendment obtained in 422 rebuilds ECG waveform.In some embodiments, with acquisition in frame 410
For original ECG signal compares, reconstructed ECG signal or waveform can have minimizing or through remove motion
And/or EMG artifact (such as, ECG signal can have minimizing or through remove noise).It will be appreciated that it is described above
Method and any combination of operation of method can perform in substantially real time.
Fig. 8 shows the primary signal of Fig. 2 and at the signal of treated Fig. 2 to reduce after noise.Original is (destroyed
) ECG signal has bigger motion artifacts, and run through the high-frequency noise of signal, the characteristic of EMG artifact.As by this
Figure demonstration, after carrying out processing to remove or reducing artifact, is easier in ECG signal identify PQRST ripple.Can
See, by LOWESS fit procedure, also effectively removes the baseline drift in signal, it has therefore proved that described LOWESS
Fit procedure is difficult to complete with curve fitting technique, and described curve fitting technique attempts with the whole original letter of fitting of a polynomial
Number span is to remove motion artifacts.
Fig. 9 shows the primary signal of the Fig. 3 with much noise and carrying out the letter that processes to reduce the Fig. 3 after noise
Number.In Fig. 8, before carrying out processing to remove artifact, it is difficult to identify ECG waveform 580 in ECG signal
P, Q, S and T ripple, and treated to remove or the ECG signal of minimizing artifact is easier to identify.
Figure 10 shows the frame 580 of Fig. 9 in more detail, described figure show the ECG signal that is corrupted by noise with treated with
Reduce the comparison between the ECG signal after noise.In Fig. 8 and 9, before carrying out processing to remove artifact,
It is difficult in ECG signal identify P, Q, S and T ripple of ECG waveform 580, and treated to remove or to reduce puppet
The ECG signal of shadow is easier to identify.Figure 10 also illustrates various time slices (such as, the QRS of ECG waveform 580
Width), it may be correlated with reading or analyzing in ECG signal.
Figure 11 explanation for process ECG signal 510 with reduce or remove noise in described ECG signal 510 be
The functional block diagram of one embodiment of system, equipment or device.In some embodiments, there is EMG and/or motion
The ECG signal 510 of artifact can be obtained by real-time data acquisition 512, such as, be adhered on experimenter one or
Multiple electrodes.The signal obtained by real-time data acquisition can be transferred to A/D converter 514, described A/D converter
514 can process signal and produce the time domain samples of (such as) signal.Also can be from data base 516, from the Internet 518 or from institute
In genus field, known other source any or storage media obtain ECG signal 510.
Then ECG signal being transmitted or be shipped to equipment 520, wherein ECG signal is by the input/output of device 522
(I/O) module receives.I/O device 522 can be known other dress any in (such as) antenna, FPDP or art
Put.Can be transmitted by signal or be shipped to processor 526, described processor is configured to process ECG signal to remove or to subtract
Few EMG and/or motion artifacts.As described below, processor 526 can be known any multiple place in art
Reason device or the combination of processor, including (for example) general processor, digital signal processor or special IC.Process
Device 526 can be configured to perform any one in process operation as detailed above or any combination, comprises ECG signal
510 wavelet decomposition are to produce one group of approximation coefficient and to organize detail coefficients more.Processor 526 can be further configured with amendment
Approximation coefficient and/or detail coefficients.In some embodiments, processor 526 can be configured to use place as detailed above
Manage any one in operation or coefficient is revised in any combination, comprise: with second order polynomial local fit approximation coefficient to produce
Raw second order polynomial point, deducts to obtain the approximation system of amendment by the second order polynomial point through local fit from approximation coefficient
Number, the derivative of the approximation coefficient being at least partially based on amendment determines the place of R ripple, and the detail coefficients some organized is set to
Zero, and/or be at least partially based on detail coefficients, relative to the position of the R ripple of ECG signal, some detail coefficients carried out position
Put threshold process to obtain the detail coefficients of amendment.Processor 526 also can be configured with from amendment approximation coefficient and/or repair
The detail coefficients changed rebuilds motion that is that have minimizing or that remove and/or the ECG signal 530 of EMG artifact.If will system
System is embodied as software, then can be programmed processor can being stored in the instruction in memorizer 524 or code (example to use
As, the form in module) perform these functions.Memorizer 524 can be can be by any storage media of computer access.?
Before process, period and/or afterwards, can by ECG signal and any part thereof (including (for example) described group of approximation coefficient or
The approximation coefficient of amendment) transmit or be shipped to memorizer 524 and be stored in memorizer 524, and can be from memorizer
Retrieve in 524 and use for processor 526.After being processed ECG signal 510 by processor 524, can be by treated
ECG signal transmission or be shipped to I/O module or device 528.
After the treatment, depending on operation or the method performed by processor 524, ECG signal 510 can be to have minimizing
Or through remove EMG and/or the ECG signal 530 of motion artifacts.In some embodiments, can then will have
Display 532, described display are transmitted or be shipped to the EMG reduced and/or the ECG signal 530 of the motion artifacts of minimizing
Device is configured to show ECG signal 530, so that can read and/or analyze described ECG signal.Also can be by ECG
Signal 530 transmits or is shipped to the Internet 534 or data base 536 for storage and/or subsequent transmission.It will be appreciated that figure
Any part of system, equipment or device illustrated in 11 can be configured to perform as detailed above appointing in substantially real time
What function or task, such as, when receiving signal from data base 516, the Internet 518 or is being adopted by real time data
During collection 512 acquisition signal, process signal.
Figure 12 A shows for processing ECG signal to remove or to reduce the system of noise, equipment or the dress in ECG signal
Put the functional block diagram of an embodiment of 610 (below is " device 610 ").Device 610 can comprise ECG electricity
Pole 612, it may be used to obtain ECG signal from experimenter.(such as) adhesion on a surface of electrode 612 can be used
ECG electrode 612 is adhered on experimenter by agent, in order to obtain signal.Electrode 612 can be adhered to (such as) experimenter
Skin near experimenter the region outside the heart at.Can be then by the signal transmission obtained by electrode 612 or delivery
To A/D converter 614, described A/D converter 614 can be configured to generate the signal that obtained by electrode 612 time
Territory sample.The time domain samples produced by A/D converter 614 be as obtained by electrode 612 have EMG and/
Or the expression of the ECG signal 616 of motion artifacts.The ECG signal 616 with EMG and/or motion artifacts can be passed
It is passed to processor 618 for carrying out processing to remove or reducing EMG and/or motion artifacts.
As described below, processor 618 can be any various processor known in art or the group of processor
Close, including (for example) general processor, digital signal processor or special IC.Processor 618 can be configured with
Perform any one in process operation as detailed above or any combination, comprise ECG signal 616 wavelet decomposition to produce
Give birth to one group of approximation coefficient and organize detail coefficients more.Processor 618 can be further configured to revise approximation coefficient and/or details
Coefficient.In some embodiments, processor 618 can be configured to use any one in process operation as detailed above
Or any combination amendment coefficient, comprise: with second order polynomial local fit approximation coefficient to produce second order polynomial point, general
From approximation coefficient, deduct to obtain the approximation coefficient of amendment through the second order polynomial point of local fit, be at least partially based on and repair
The derivative of the approximation coefficient changed determines the place of R ripple, and the detail coefficients some organized is set to zero, and/or at least partly
Relative to the position of the R ripple of ECG signal, some detail coefficients are carried out position threshold based on detail coefficients to process to obtain
The detail coefficients of amendment.Processor 618 also can be configured to rebuild from the approximation coefficient of amendment and/or the detail coefficients of amendment
Have minimizing or the motion through removing and/or the ECG signal 620 of EMG artifact.If system is embodied as software,
So processor programming can be stored in the instruction in the memorizer that can be accessed or code (example by processor 618 to use
As, the form in module) perform these functions.Then can will have the EMG of minimizing and/or the treated of motion artifacts
ECG signal 620 is transferred to I/O module or device 622.ECG signal 620 can be then transferred to through joining by I/O 622
Put to receive ECG signal 620 for the mesh realizing showing, storing, transmit or process further ECG signal 620
Any device, equipment or memorizer.It will be appreciated that device 610 is configured to perform in substantially real time as detailed above
Function and task.
Figure 12 B show is for processing ECG signal to remove or to reduce the system of noise, equipment or the dress in ECG signal
The functional block diagram of the embodiment put.In some embodiments, the system of Figure 12 B, equipment or device can be non-
Bed ECG monitoring system.Described system can comprise acquisition system, equipment or device 710 and (below is " device
710 "), such as, paster ECG monitor.Paster ECG monitor can comprise ECG electrode 712, A/D converter
714, processor or signal processing circuit 718, and antenna 720.The sticker on a side of electrode can be used ECG
Electrode 712 adheres to experimenter upper (such as, adhere on the skin of experimenter), and ECG electrode 712 can be configured with
ECG signal (such as, original sample data points) is obtained from experimenter.Then the signal obtained by electrode can be delivered
To A/D converter, described A/D converter is configured to produce the ECG signal 716 with EMG and/or motion artifacts
Time domain samples.ECG signal 716 can be sent to processor or signal processing circuit 718, described processor or signal
Process circuit 718 to be configured to process the signal obtained by electrode.
As described below, processor 718 can be any various processor known in art or the group of processor
Close, including (for example) general processor, digital signal processor or special IC.Processor 718 can be configured with
Perform any one in process operation as detailed above or any combination, comprise ECG signal 716 wavelet decomposition to produce
Give birth to one group of approximation coefficient and organize detail coefficients more.Processor 718 can be further configured to revise approximation coefficient and/or details
Coefficient.In some embodiments, processor 718 can be configured to use any one in process operation as detailed above
Or any combination amendment coefficient, comprise: with second order polynomial local fit approximation coefficient to produce second order polynomial point, general
From approximation coefficient, deduct to obtain the approximation coefficient of amendment through the second order polynomial point of local fit, be at least partially based on and repair
The derivative of the approximation coefficient changed determines the place of R ripple, and the detail coefficients some organized is set to zero, and/or at least partly
Relative to the position of the R ripple of ECG signal, some detail coefficients are carried out position threshold based on detail coefficients to process to obtain
The detail coefficients of amendment.Processor 718 also can be configured to rebuild from the approximation coefficient of amendment and/or the detail coefficients of amendment
Have minimizing or the motion through removing and/or the ECG signal 726 of EMG artifact.Or, processor 718 can be through joining
Put to process or compression ECG signal 716 is to be transferred to another system, equipment or dress by antenna 720 by described signal
Put 730 to perform the further process to signal 716.
Can wirelessly by antenna 720 treated ECG signal 716 or its compressed version be transferred to system,
Equipment or device 730 (below is " device 730 ").Device 730 can be mobile device, such as cell phone, flat
Plate computer or by antenna 722 receive data and by data delivery to processor or signal processing circuit 724 other just
Take formula electronic system.Processor 724 can be configured to perform any combination processing operation of novel algorithm as detailed above
All with reduce or remove the EMG in ECG signal 716 and/or motion artifacts with produce have minimizing or through going
The EMG removed and/or the ECG signal 726 of motion artifacts.Device 730 can comprise be configured to show ECG signal 726
Display 728.Display 728 can be configured to handle ECG signal by the keypad in mobile device/touch screen
726.Device 730 also can be configured with waveform or its compressed version are transferred to external network (such as, the Internet) with
In storage, for physical examination examination & verification etc..
It will be appreciated that the assembly of system, equipment or device illustrated in Figure 12 B need not be adhered to same physical lining together
, but previously described operation and the function performing device 710 and 730 can be divided in many ways at the end.Also should
Understanding, device 710 and 730 is configured to perform function as detailed above and task in substantially real time.
Can be real in conjunction with various illustrative logical, logical block, module and the algorithm operating described by embodiments disclosed herein
Execute as electronic hardware, computer software, or a combination of both.The interchangeability of hardware and software is the most substantially at functional aspect
It is been described by, and is illustrated in various Illustrative components as described above, block, module, circuit and operation.Institute
State and functional be implemented as hardware or software depends on application-specific and puts on the design constraint of whole system.
In conjunction with aspect disclosed herein describe in order to implement various illustrative logical, logical block, module and circuit
Hardware and data handling equipment can be practiced or carried out by the following: general purpose single-chip or multi-chip processor, numeral
Signal processor (DSP), special IC (ASIC), field programmable gate array (FPGA) or other FPGA
Device, discrete gate or transistor logic, discrete hardware components, or it is designed to perform function described herein
Any combination.General processor can be microprocessor or any conventional processors, controller, microcontroller or state machine.
Processor also can be embodied as calculating the combination of device, such as, DSP and the combination of microprocessor, the group of multi-microprocessor
Close, one or more microprocessor combine with DSP core, or any other this configure.In some embodiments, may be used
Specific operation and method is performed by the specific circuit of given function institute.
In in one or more aspects, (can comprise in this specification with hardware, Fundamental Digital Circuit, computer software, firmware
Disclosed structure and structural equivalents thereof) or implement described function with its any combination.Described in this specification
The embodiment of subject matter also can be embodied as one or more computer program, such as, be encoded in computer storage media with
For data handling equipment execution or one or more module of the computer program instructions of the operation controlling data handling equipment.
If implemented with software, then function can be stored in computer-readable media as one or more instruction or code
Above or transmitted via computer-readable media.The operation of methods disclosed herein or algorithm can may reside in meter
Processor on calculation machine readable media can perform to implement in software module.Computer-readable media comprises computer storage media
With communication medium, communication medium comprises any matchmaker that can make it possible to be sent at by computer program at another
Body.Storage media can be can be by any useable medium of computer access.Illustrative not limiting with example, this type of computer
Readable media can comprise RAM, ROM, EEPROM, CD-ROM or other optical disk storage apparatus, disk storage dress
Put or other magnetic storage device, or can be used for storing wanted program code and can be by calculating with instruction or data structure form
Other media any of machine access.Further, any connection can be properly termed as computer-readable media.Such as institute herein
Using, disk and CD comprise compact disc (CD), laser-optical disk, optical compact disks, digital image and sound optical disk (DVD), soft
Disk and Blu-ray Disc, wherein disk the most magnetically replicates data, and usage of CD-ROM laser optics ground replicates data.On
The combination of literary composition also should be included in the range of computer-readable media.It addition, the operation of method or algorithm can as code and
Any one or any combination or set in instruction reside at the machine-readable medium being incorporated in computer program
Or on computer-readable media.
Those skilled in the art can the easily apparent various amendments to embodiment described in the present invention,
And without departing from the spirit or scope of the present invention, one principle as defined herein can be applicable to other and implements
Scheme.Therefore, the present invention is not intended to be limited to embodiment shown herein, and should be endowed and be taken off herein
The principle shown and the consistent widest range of novel feature.Word " exemplary " is in this article for exclusively meaning
" serving as example, example or explanation ".Any embodiment here depicted as " exemplary " may not be explained
For more preferred than other embodiment or favourable.
In this manual also can be in single embodiment in some feature described in the context of independent embodiment
Implement in combination.On the contrary, the various features described in the context of single embodiment also can be respectively in multiple enforcements
Scheme is implemented or with any incompatible enforcement of suitable subgroup.Although additionally, may describe feature as above with some group
Close effect and even initially therefore advocate, but in some cases, can be by from one or more of claim combinations
Feature is deleted from combination, and the combination advocated can be about sub-portfolio or the change of sub-portfolio.
Similarly, although describe operation by certain order in the drawings, but this situation is understood not to require by being shown
Certain order or in order order perform this generic operation, or perform the operation being had been described, to realize desired result.
One or more example procedure may be the most schematically described it addition, graphic.But, can by do not describe its
Its operation is incorporated in the example procedure through schematically illustrating.For example, can before illustrated operation, afterwards,
Simultaneously or between perform one or more operation bidirectional.In some cases, multitask processes and parallel processing can be to have
Profit.Additionally, the separation of the various system components in embodiment as described above is understood not in all enforcements
Scheme requiring, this separates, and should be understood that described program assembly and system one can together be integrated in single software produce
In product or be encapsulated in multiple software product.It addition, other embodiment is in the range of following claims.One
In the case of Xie, the action described in claims can perform in different order and still realize desired result.
Claims (30)
1., for reducing a computer implemented method for the noise in ECG signal, described method includes:
With processor by ECG waveform wavelet decomposition to produce one group of approximation coefficient and to organize detail coefficients more；And
Revise described group of approximation coefficient with processor and be attributable to being subject to during the acquisition of described ECG waveform to reduce
Examination person move produce noise,
Described group of approximation coefficient of wherein said amendment includes:
It is second order polynomial by the neighbouring subset local fit of described approximation coefficient；And
Described group of approximation coefficient is adjusted to produce one group of amendment by the respective value deducting described second order polynomial
Approximation coefficient.
Method the most according to claim 1, it farther includes D1、D2And D3The value of group detail coefficients is set to
Zero.
3., according to the method described in any claim in claim 1 to 2, it farther includes by following step institute
The described detail coefficients stating decomposition carries out the EMG that position threshold process is attributable in described ECG waveform with minimizing
The noise that signal produces: the distance away from R ripple is set to zero more than the detail coefficients of preset distance, and makes away from R
The distance of ripple keeps constant less than or equal to the detail coefficients of preset distance.
Method the most according to claim 3, the wherein said detail coefficients processed through position threshold includes D4And D5
Group detail coefficients.
Method the most according to claim 3, it farther includes:
The derivative of the approximation coefficient being at least partially based on described amendment determines the place of the R ripple of described ECG waveform；With
And
By following step, the detail coefficients of described decomposition is carried out position threshold process, be attributable to minimizing described
The noise that EMG signal in ECG waveform produces: the distance away from R ripple is set more than the detail coefficients of preset distance
It is set to zero, and makes the distance away from R ripple keep constant less than or equal to the detail coefficients of preset distance.
6., according to the method described in any claim in claim 1,2,4 and 5, wherein said group of approximation coefficient includes
5 rank approximation coefficients.
7. an ECG signal processing equipment, described equipment includes:
Processor, it is configured to:
By ECG waveform wavelet decomposition to produce one group of approximation coefficient and to organize detail coefficients more；And
Revise described group of approximation coefficient to reduce the experimenter's fortune being attributable to during the acquisition of described ECG waveform
The noise that movable property is raw,
Wherein said processor is further configured to revise described group of approximation coefficient by following step:
It is second order polynomial by the neighbouring subset local fit of described approximation coefficient；And
Described group of approximation coefficient is adjusted to produce one group of amendment by the respective value deducting described second order polynomial
Approximation coefficient.
Equipment the most according to claim 7, wherein said processor be further configured with by following step to described
The detail coefficients decomposed carries out position threshold and processes the EMG signal being attributable in described ECG waveform with minimizing
The noise produced: the distance away from R ripple is set to zero more than the detail coefficients of preset distance, and makes away from R ripple
Distance keeps constant less than or equal to the detail coefficients of preset distance.
9., according to the equipment described in claim 7 or 8, wherein said equipment includes ECG electrode, A/D converter and warp
Configuration is for the antenna on the health adhering to experimenter.
10., for reducing a computer implemented method for the noise in ECG signal, described method includes:
With processor by ECG waveform wavelet decomposition to produce one group of approximation coefficient and to organize detail coefficients more；And
By following step, the detail coefficients of described decomposition is carried out position threshold process, be attributable to minimizing described
The noise that EMG signal in ECG waveform produces: the distance away from R ripple is set more than the detail coefficients of preset distance
It is set to zero, and makes the distance away from R ripple keep constant less than or equal to the detail coefficients of preset distance.
11. methods according to claim 10, it farther includes to revise described group of approximation coefficient to reduce with processor
It is attributable to the noise that the subject motion during the acquisition of described ECG waveform produces.
12. methods according to claim 11, wherein revise described group of approximation coefficient and include the neighbour of described approximation coefficient
Nearly subset local fit is second order polynomial, and is adjusted described by the respective value deducting described second order polynomial
Group approximation coefficient is to produce the approximation coefficient of one group of amendment.
13. 1 kinds of ECG signal processing equipments, described equipment includes:
Processor, it is configured to:
By ECG waveform wavelet decomposition to produce one group of approximation coefficient and to organize detail coefficients more；And
By following steps, the detail coefficients of described decomposition is carried out position threshold process, be attributable to minimizing described
The noise that EMG signal in ECG waveform produces: the distance away from R ripple is more than the detail coefficients of preset distance
It is set to zero, and makes the distance away from R ripple keep constant less than or equal to the detail coefficients of preset distance.
14. equipment according to claim 13, wherein said processor is further configured to revise described approximation coefficient
To reduce the noise that the subject motion being attributable to during the acquisition of described ECG waveform produces.
15. equipment according to claim 14, wherein said processor is further configured with by by described approximation system
The neighbouring subset local fit of number is second order polynomial and is adjusted by the respective value deducting described second order polynomial
Whole described group of approximation coefficient is to produce the approximation coefficient of one group of amendment to revise described approximation coefficient.
16. equipment according to claim 15, wherein said processor is further configured with by deducting described second order
Polynomial respective value adjusts described group of approximation coefficient and produces the approximation coefficient of one group of amendment.
17. equipment according to claim 16, wherein said processor is further configured to be at least partially based on described
The derivative of the approximation coefficient of group amendment determines the place of the R ripple of described ECG signal.
18. according to the equipment described in any claim in claim 13 to 17, wherein said equipment include ECG electrode,
A/D converter and the antenna being configured for use on the health adhering to experimenter.
19. 1 kinds of computer implemented methods being used for processing ECG signal, described method includes:
With processor by ECG waveform wavelet decomposition to produce one group of approximation coefficient and to organize detail coefficients more；
It is second order polynomial with processor by the neighbouring subset local fit of described approximation coefficient；
Described approximation coefficient is adjusted to produce the approximation of one group of amendment by deducting the analog value of described second order polynomial
Coefficient；
The derivative of the approximation coefficient being at least partially based on described group of amendment determines the place of the R ripple of described ECG signal；
And
By following step, the detail coefficients of described decomposition is carried out position threshold process with processor, to produce many groups
The detail coefficients of amendment: the distance away from R ripple is set to zero more than the detail coefficients of preset distance, and makes away from R
The distance of ripple keeps constant less than or equal to the detail coefficients of preset distance.
20. methods according to claim 19, its farther include with processor from the described group of approximation coefficient revised and
At least some of of detail coefficients of described many group amendments rebuilds described ECG waveform.
21. include 5 rank approximation coefficients according to the method described in claim 19 or 20, wherein said group of approximation coefficient.
22. methods according to claim 19, it farther includes D1、D2And D3The value of group detail coefficients is arranged
It is zero.
23. include D according to the method described in claim 19 or 22, the wherein said detail coefficients processed through position threshold4
And D5Group detail coefficients.
24. 1 kinds of equipment being used for reducing the noise in ECG signal, described equipment includes:
Processor, it is configured to:
By ECG waveform wavelet decomposition to produce one group of approximation coefficient and to organize detail coefficients more；
It is second order polynomial by the neighbouring subset local fit of described approximation coefficient；
Described approximation coefficient is adjusted to produce the near of one group of amendment by deducting the analog value of described second order polynomial
Like coefficient；
The derivative of the approximation coefficient being at least partially based on described group of amendment determines the ground of the R ripple of described ECG signal
Point；And
By following step, the detail coefficients of described decomposition is carried out position threshold to process to produce the thin of many group amendments
Joint coefficient: the distance away from R ripple is set to zero more than the detail coefficients of preset distance, and make away from R ripple away from
Keep constant from less than or equal to the detail coefficients of preset distance.
25. equipment according to claim 24, wherein said processor be further configured with from described group revise near
Like coefficient and the described ECG waveform of at least some of reconstruction of the detail coefficients of described many group amendments.
26. equipment according to claim 24, wherein said equipment includes ECG electrode, A/D converter and is configured
For the antenna on the health adhering to experimenter.
27. 1 kinds of devices being used for processing ECG signal, comprising:
For by ECG waveform wavelet decomposition to produce one group of approximation coefficient and the devices organizing detail coefficients more；
For being the device of second order polynomial by the neighbouring subset local fit of described approximation coefficient；
For by deduct the analog value of described second order polynomial adjust described approximation coefficient with produce one group amendment
The device of approximation coefficient；
For being at least partially based on the ground that the derivative of the approximation coefficient of described group of amendment determines the R ripple of described ECG signal
The device of point；And
Organize amendment for the detail coefficients of described decomposition being carried out position threshold process by following step to produce more
The device of detail coefficients: the distance away from R ripple is set to zero more than the detail coefficients of preset distance, and makes away from R
The distance of ripple keeps constant less than or equal to the detail coefficients of preset distance.
28. devices according to claim 27, it farther includes for from the described group of approximation coefficient revised and described
At least some of device rebuilding described ECG waveform of the detail coefficients of many group amendments.
29. according to the device described in claim 27 or 28, and wherein said approximation coefficient includes 5 rank approximation coefficients.
30. include D according to the device described in claim 27 or 28, the wherein said detail coefficients processed through position threshold4
And D5Group detail coefficients.
Applications Claiming Priority (5)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| US201261616890P | 2012-03-28 | 2012-03-28 | |
| US61/616,890 | 2012-03-28 | ||
| US13/797,889 | 2013-03-12 | ||
| US13/797,889 US9364160B2 (en) | 2012-03-28 | 2013-03-12 | Systems and methods for ECG monitoring |
| PCT/US2013/032109 WO2013148318A1 (en) | 2012-03-28 | 2013-03-15 | Systems and methods for ecg monitoring |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104168821A CN104168821A (en) | 2014-11-26 |
| CN104168821B true CN104168821B (en) | 2016-11-30 |
Family
ID=
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102274020A (en) * | 2011-06-30 | 2011-12-14 | 东北大学 | Low-power consumption portable electrocardiograph monitor and control method thereof |
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102274020A (en) * | 2011-06-30 | 2011-12-14 | 东北大学 | Low-power consumption portable electrocardiograph monitor and control method thereof |
Non-Patent Citations (2)
| Title |
|---|
| 基于小波变换系数的心电信号基线漂移噪声去除方法;任杰 等;《医疗卫生装备》;20101130;第31卷(第11期);第24-26页 * |
| 基于平稳小波变换的自适应心电信号去噪方法;董红生 等;《中国医疗器械杂志》;20091231;第33卷(第3期);第163-166页 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US9364160B2 (en) | Systems and methods for ECG monitoring | |
| Lee et al. | Hardware implementation of EMD using DSP and FPGA for online signal processing | |
| Kumar et al. | Efficient QRS complex detection algorithm based on Fast Fourier Transform | |
| CN111046824A (en) | Time series signal efficient denoising and high-precision reconstruction modeling method and system | |
| Noorbasha et al. | Removal of EOG artifacts and separation of different cerebral activity components from single channel EEG—an efficient approach combining SSA–ICA with wavelet thresholding for BCI applications | |
| CN111814656A (en) | Electrocardiosignal noise reduction method based on countermeasure generation network | |
| Annavarapu et al. | ECG-based atrial fibrillation detection using different orderings of Conjugate Symmetric–Complex Hadamard Transform | |
| Phinyomark et al. | Wavelet-based denoising algorithm for robust EMG pattern recognition | |
| CN114129171B (en) | Electrocardiosignal noise reduction method based on improved residual error dense network | |
| CN107361762A (en) | ECG baseline drift bearing calibration based on variation mode decomposition | |
| US9259167B2 (en) | Method and system for extracting ventricular fibrillation signals in electrocardiogram | |
| Alhelal et al. | FPGA-based denoising and beat detection of the ECG signal | |
| Ghorbanian et al. | Stochastic non-linear oscillator models of EEG: the Alzheimer's disease case | |
| El B’charri et al. | The ECG signal compression using an efficient algorithm based on the DWT | |
| Farahabadi et al. | Detection of QRS complex in electrocardiogram signal based on a combination of hilbert transform, wavelet transform and adaptive thresholding | |
| Bui et al. | Design of a nearly linear-phase IIR filter and JPEG compression ECG signal in real-time system | |
| Miller et al. | Physiological waveform imputation of missing data using convolutional autoencoders | |
| CN104168821B (en) | The system and method monitored for ECG | |
| Liu et al. | Event-driven processing for hardware-efficient neural spike sorting | |
| Karataş et al. | Design and implementation of FPGA-based arrhythmic ECG signals using VHDL for biomedical calibration applications | |
| Mansurjonovich et al. | Creation of Algorithms for Constructions of Wavelet Models Suitable for Ecg-Signals | |
| Nguyen et al. | A deep sparse autoencoder method for automatic EOG artifact removal | |
| Zhang et al. | An efficient ECG denoising method by fusing ECA-Net and CycleGAN | |
| Mourad | Automatic correction of short‐duration artefacts in single‐channel EEG recording: a group‐sparse signal denoising algorithm | |
| Palaniza et al. | Phonocardiosignal as a periodically correlated stochastic process preprocessing algorithm structure grounding |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20161130 Termination date: 20190315 |