CN104203091B - Use the real-time QRS detection of adaptive threshold - Google Patents
Use the real-time QRS detection of adaptive threshold Download PDFInfo
- Publication number
- CN104203091B CN104203091B CN201380014120.0A CN201380014120A CN104203091B CN 104203091 B CN104203091 B CN 104203091B CN 201380014120 A CN201380014120 A CN 201380014120A CN 104203091 B CN104203091 B CN 104203091B
- Authority
- CN
- China
- Prior art keywords
- signal
- filtered
- threshold signal
- ecg data
- filter
- 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.)
- Active
Links
- 238000001514 detection method Methods 0.000 title abstract description 20
- 230000003044 adaptive Effects 0.000 title description 4
- 238000005070 sampling Methods 0.000 claims abstract description 31
- 230000010247 heart contraction Effects 0.000 claims abstract description 19
- 238000001914 filtration Methods 0.000 claims description 25
- 230000000007 visual effect Effects 0.000 claims description 9
- 230000004044 response Effects 0.000 claims description 7
- 230000003111 delayed Effects 0.000 claims description 6
- 230000001629 suppression Effects 0.000 claims description 4
- 229920005549 butyl rubber Polymers 0.000 claims 4
- 230000003321 amplification Effects 0.000 abstract 1
- 238000003199 nucleic acid amplification method Methods 0.000 abstract 1
- 238000004088 simulation Methods 0.000 abstract 1
- 210000002216 Heart Anatomy 0.000 description 24
- 238000010586 diagram Methods 0.000 description 10
- 238000000034 method Methods 0.000 description 10
- 238000004422 calculation algorithm Methods 0.000 description 9
- 210000004027 cells Anatomy 0.000 description 9
- 230000005540 biological transmission Effects 0.000 description 8
- 239000000203 mixture Substances 0.000 description 7
- 238000004458 analytical method Methods 0.000 description 6
- 230000002999 depolarising Effects 0.000 description 6
- 210000003141 Lower Extremity Anatomy 0.000 description 5
- 208000009729 Ventricular Premature Complexes Diseases 0.000 description 5
- 210000001013 Sinoatrial Node Anatomy 0.000 description 4
- 210000003491 Skin Anatomy 0.000 description 4
- 230000015654 memory Effects 0.000 description 4
- 230000033764 rhythmic process Effects 0.000 description 4
- 201000002930 sinoatrial node disease Diseases 0.000 description 4
- 210000002837 Heart Atria Anatomy 0.000 description 3
- 210000004165 Myocardium Anatomy 0.000 description 3
- 238000010009 beating Methods 0.000 description 3
- 230000005611 electricity Effects 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 238000003860 storage Methods 0.000 description 3
- 206010029470 Nodal rhythm Diseases 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 238000004140 cleaning Methods 0.000 description 2
- 230000001808 coupling Effects 0.000 description 2
- 238000010168 coupling process Methods 0.000 description 2
- 238000005859 coupling reaction Methods 0.000 description 2
- 230000003412 degenerative Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000002107 myocardial Effects 0.000 description 2
- 230000000630 rising Effects 0.000 description 2
- 239000000243 solution Substances 0.000 description 2
- 230000001960 triggered Effects 0.000 description 2
- 210000001367 Arteries Anatomy 0.000 description 1
- 206010048623 Atrial hypertrophy Diseases 0.000 description 1
- 210000001992 Atrioventricular Node Anatomy 0.000 description 1
- 210000000481 Breast Anatomy 0.000 description 1
- 208000008787 Cardiovascular Disease Diseases 0.000 description 1
- 210000000170 Cell Membrane Anatomy 0.000 description 1
- 244000145845 Chattering Species 0.000 description 1
- 208000002330 Congenital Heart Defects Diseases 0.000 description 1
- 201000006306 Cor pulmonale Diseases 0.000 description 1
- 208000002907 Heart Valve Disease Diseases 0.000 description 1
- 206010020772 Hypertension Diseases 0.000 description 1
- 206010020880 Hypertrophy Diseases 0.000 description 1
- 210000003127 Knee Anatomy 0.000 description 1
- 210000002414 Leg Anatomy 0.000 description 1
- 210000003205 Muscles Anatomy 0.000 description 1
- 208000004186 Pulmonary Heart Disease Diseases 0.000 description 1
- 210000000115 Thoracic Cavity Anatomy 0.000 description 1
- 210000003462 Veins Anatomy 0.000 description 1
- 235000014121 butter Nutrition 0.000 description 1
- 230000000747 cardiac effect Effects 0.000 description 1
- 150000001768 cations Chemical class 0.000 description 1
- 230000001413 cellular Effects 0.000 description 1
- 210000000038 chest Anatomy 0.000 description 1
- 238000004590 computer program Methods 0.000 description 1
- 201000003137 congenital heart disease Diseases 0.000 description 1
- 230000000875 corresponding Effects 0.000 description 1
- 238000005520 cutting process Methods 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 239000003814 drug Substances 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000002347 injection Methods 0.000 description 1
- 239000007924 injection Substances 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 239000002184 metal Substances 0.000 description 1
- 230000004899 motility Effects 0.000 description 1
- 239000005022 packaging material Substances 0.000 description 1
- 238000004806 packaging method and process Methods 0.000 description 1
- 231100000915 pathological change Toxicity 0.000 description 1
- 230000036285 pathological change Effects 0.000 description 1
- 230000002028 premature Effects 0.000 description 1
- 230000002685 pulmonary Effects 0.000 description 1
- 238000004080 punching Methods 0.000 description 1
- 238000000718 qrs complex Methods 0.000 description 1
- 230000004213 regulation of atrial cardiomyocyte membrane depolarization Effects 0.000 description 1
- 230000034225 regulation of ventricular cardiomyocyte membrane depolarization Effects 0.000 description 1
- 230000001105 regulatory Effects 0.000 description 1
- 230000002336 repolarization Effects 0.000 description 1
- 230000000717 retained Effects 0.000 description 1
- 238000005096 rolling process Methods 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 230000001131 transforming Effects 0.000 description 1
- 230000001702 transmitter Effects 0.000 description 1
- 230000002861 ventricular Effects 0.000 description 1
Abstract
A kind of mobile system for analyzing ECG data, it includes the analog front-end module being coupled in mobile consumption device.Described analog front-end module is configured to collect from one or more ECG data led and operable so that simulation ECG data to be converted into digital ECG data.Described mobile consumption device such as smart mobile phone (400) coupled reception described numeral ECG data (150), and being configured to use wave filter (436) to perform QRS detection (451), the cut-off frequency of described wave filter is adapted to noise level in real time.ECG signal is by non-linear amplification (431) and obtains three Windowing threshold signals (D, E, J).Cut-off frequency for QRS detection is dynamically selected the function that (439) are threshold signal.Only just the sampling in amplified signal is identified as heart beating point when sampled value is equal to first threshold signal and exceedes filtered threshold signal.
Description
Technical field
Present invention relates in general to the analysis of ECG signal, and be specifically related to low cost consumption dress for this purpose
The use put.
Background technology
Electrocardiogram (ECG or EKG) is that the electrode by being attached on the outer surface of skin such as detects and by body
The explaining through breast (cross over thoracic cavity or chest) of cardiac electrical activity within a period that the device of external body is recorded.By this nothing
The record that invasive procedure produces is referred to as electrocardiogram (also referred to as ECG or EKG).Electrocardiogram (ECG) is that the electricity of record heart is lived
Dynamic test.
ECG is for measuring the speed of heart beating and the regular and size and location of ventricle, to any infringement of heart
Exist, and for regulating medicine or the effect of device (such as pacemaker) of heart.
ECG device detects and amplifies the small electric on the skin caused when myocardial depolarization during each heart beating
Change.When rest, each myocardial cell has the negative charge (transmembrane potential) crossing over its outer wall (or cell membrane).By this negative charge
Increase (by the injection of cation Na+ and Ca++) towards zero and be referred to as depolarization, causing in described depolarization active cell
The mechanism of cellular contraction.During each heart beating, healthy heart will have (the described depolarization wave that advances in order of depolarization wave
Triggered by cell in sinuatrial node), diffuse through atrium, through " inherent conducting path ", and expand at whole ventricle subsequently
Dissipate.This is detected as the small rising on the voltage that is placed between two electrodes of heart both sides and decline, described rising
And decline on a display screen or on paper, be shown as wave-like line.The overall rhythm and pace of moving things of this display instruction heart and cardiac muscle are not
Weakness with part.
Generally, use two or more electrode and they can be combined into multipair, such as: left arm (LA), right arm (RA) with
And left lower limb (LL) electrode to form three right: LA+RA, LA+LL, and RA+LL.It is referred to as leading from the output of every pair.Each lead
Connection it is said observes heart from different angles.The number that leads that different types of EKG can pass through to be recorded is named, and such as 3 lead
Connection, 5 lead or 12 lead EKG (sometimes simply referred to as " 12 lead ").12 EKG that lead are that wherein 12 different signals of telecommunication exist
The roughly the same moment is recorded and generally will act as the EKG of the disposable recording of EKG, and described disposable recording is made traditionally
Print for papery duplicate.3 lead and 5 EKG that lead often are continually monitored and only in the display of suitable supervising device
On screen observed, such as, intra-operative or when in ambulance transport time.According to the equipment used, it is understood that there may be or can
Can not have 3 to lead or the 5 any HC hard copy leading EKG.
Generally there is two kinds of existing heart monitor: 1) there is the Portable health-care formula watch-dog of limited capability,
The most only there is heart rate, and there is no ECG waveform, there is no early warning, and QRS detector performance is poor;2) medical grade watch-dog, institute
State watch-dog actually also non-portable, but it has good performance and provides early warning.ECG, EMG (electromyogram), EEG
Consumer is still prohibitively expensive by the cost of medical grade equipment such as (electroencephalograms).Some existing solutions are likely to be of well
Performance, but they are not real-time.Some other solutions are portable and real-time, but demonstrate bad performance, such as
Body-building heart rate monitor.Medical grade equipment has rational performance and operates substantially real time, but it is for as portable
For formula equipment the compactest and sufficiently expensive.
Summary of the invention
The present invention provides a kind of method and system for processing ECG data.The method includes: receives and includes that PQRST schemes
The stream of the raw ECG data sampling of case;By using nonlinear filter to be filtered described raw ECG data being formed
Filtered ECG data, in order to minimize baseline drift, and thus suppress the T ripple part of each PQRST pattern;To described
Filtered ECG data performs nonlinear operation to form the amplified signal at the R peak expanding each PQRST pattern;To institute
Stating amplified signal uses mobile maximum filter to obtain first threshold signal and the 3rd threshold signal;To described warp
The signal amplified uses mobile mean filter to obtain Second Threshold signal;Use wave filter that described 3rd threshold signal is entered
Row filtering is to form filtered threshold signal, and in described wave filter, cut-off frequency is dynamically selected for described first threshold
Value signal and the function of described Second Threshold signal;And only and exceed equal to described first threshold signal when the value of sampling
During described filtered threshold signal, the sampling in described amplified signal is identified as heart beating point, wherein to described original
ECG data is filtered including: use the cascade of median filter and low pass filter to filter described raw ECG data
Ripple;Postpone described raw ECG data;And deduct the output signal of described low pass filter from the raw ECG data postponed, with
Just the data of high-pass filtering are transmitted.A kind of computer implemented system for processing ECG data in real time, this system includes:
For receiving the stream of the raw ECG data sampling including PQRST pattern from the AFE (analog front end) being coupled on described computer
Device;For by using nonlinear filter to be filtered forming filtered ECG number to described raw ECG data
The most just minimize baseline drift and thus suppress the device of the T ripple part of each PQRST pattern;For to described filtered
ECG data perform nonlinear operation to form the device of amplified signal at the R peak expanding each PQRST pattern;For
Described amplified signal use mobile maximum filter obtain first threshold signal and the dress of the 3rd threshold signal
Put;For using mobile mean filter to obtain the device of Second Threshold signal described amplified signal;For using
Wave filter is filtered being formed the device of filtered threshold signal to described 3rd threshold signal, in described wave filter,
Cut-off frequency is dynamically chosen as described first threshold signal and the function of described Second Threshold signal;And for only when adopting
By adopting in described amplified signal when the value of sample is equal to described first threshold signal and exceedes described filtered threshold signal
Sample is identified as the device of heart beating point, and wherein the device for being filtered described raw ECG data is configured to: make
With the cascade of median filter and low pass filter, described raw ECG data is filtered;Postpone described raw ECG data;
And the output signal of described low pass filter is deducted from the raw ECG data postponed, in order to transmit the data of high-pass filtering.
Accompanying drawing explanation
Fig. 1 be a diagram that the block diagram of the one embodiment of the present of invention in mobile ECG system;
Fig. 2-3 illustrates the ECG curve analyzed by the mobile ECG system of Fig. 1;
Fig. 4 is the indicative icon of the real-time ECG wave filter implemented by the ECG system of Fig. 1;
Fig. 5 is the curve that the almost linear used by the wave filter of Fig. 4 returns;
Fig. 6,7A-7D illustrate generation during the analysis carried out ECG signal by the real-time ECG wave filter of Fig. 4
Waveform;
Fig. 8-9 be a diagram that by the ECG system of Fig. 1 sequential chart to the detection that alert if is carried out;
Figure 10 be a diagram that the ECG carried out by ECG system analyzes the flow chart of operation;
Figure 11 illustrates another embodiment of the mobile ECG system utilizing cell phone;And
Figure 12 A-12B illustrates the other embodiments of mobile ECG system.
Detailed description of the invention
Due to DC biasing and the existence of various interference signal, therefore the measurement of ECG signal is challenging.For
Typical electrode, this current potential can be up to 300mV.Interference signal includes disturbing from the 50/60-Hz of power supply, due to patient's shifting
Dynamic cause motion artifacts, from electrosurgery equipment, defibrillation pulse, pacemaker pulse, the radio frequency of other monitoring device etc.
Interference.Embodiments of the invention achieve processing in real time and analyzing heart abnormality with high detection performance, so that energy
Enough producing real-time early warning under degenerative conditions, described degenerative conditions may be by body kinematics, the friction of electrode or deformation, survey
The installation of the individual of examination causes.Embodiments of the invention can use the mobile consumption device through being suitably equipped with to implement, such as
Smart mobile phone, tablet PC, personal digital assistant, personal computer etc..Other embodiments can be fixed in such as medical grade
Or mobile platform is implemented.QRS detection algorithm be will be described in greater detail hereinafter, and described algorithm uses by IIR (unlimited pulse
Response) wave filter coupled to the wave filter of the some cascades in the adaptive threshold value filtered, and utilizes logarithmic scale to zoom to
The noise level of primary signal adjusts the cut-off frequency of described iir filter in real time.
QRS detection algorithm uses the threshold value through filtering relatively, described by using feedback adaptive iir filter
In iir filter, cut-off frequency changes with signal to noise ratio.Sliding window is used to calculate letter according to the ratiometer of peak value Yu meansigma methods
Make an uproar ratio.Therefore, described algorithm can adjust self in real time, even and if exist the vigorous exercise of user, noise or its
In the case of its artifact, it is also possible to provide QRS detection the most accurately and PQRST shape to retain.Therefore, it can accurately detect
To user heart early warning.
Fig. 1 be a diagram that the block diagram of the embodiments of the invention in mobile ECG system 100.AFE (analog front end) part 110 coupling
Close lead on (lead) 120 to one group, described in lead in the main body (subject) 130 being coupled to just be monitored.The master monitored
Body 130 is typically people;But, embodiments of the invention may be used for monitoring other type of main body, such as domestic animal, bird or climb
Action thing etc..What monitoring was led on 120 each aspects including being attached in main body 130 multiple leads.Generally, three, five, seven
Individual or 12 lead be used to ECG monitoring;But, embodiments of the invention are not limited to any certain number of lead.
AFE (analog front end) part 110 can include for selecting the multiplexer of various signal in 120 groups from described leading
111, the EMI in each input (electromagnetic interference) and low-pass filtering, and include the delta sigma (delta-of high-pass filtering
Sigma) analog-digital converter 112.One or more selected combination from input signal can obtain right leg drive (RLD)
Signal 113.
Standard supervision needs the frequency between 0.05Hz to 30Hz.Diagnosis monitoring needs the frequency in 0.05Hz to 1000Hz
Rate.High input impedance measuring amplifier (INA) can be utilized to eliminate some in 50Hz/60Hz common mode disturbances, and described height is defeated
Enter impedance measuring amplifiers and remove the AC circuit noise that the two input is common.For further limiting circuitry power supply noise, logical
Cross amplifier make signal inversion and be driven back in patient's body by this signal by right lower limb.Only need the electricity of several microamperes or less
Stream realizes significant CMR and improves and be maintained in UL544 restriction.It addition, 50/60Hz digital notch filter is used for into one
Step reduces this interference.
Analog portion 110 may be implemented within single integrated circuit such as ADS1294/6/8/4R/6R/8R, described
ADS1294/6/8/4R/6R/8R is family's multichannel, simultaneously sampling, the delta sigma analog-digital converter (ADC) of 24, described Δ-
∑ analog-digital converter has built-in programmable gain amplifier (PGA), internal reference and airborne agitator.ADS1294/
6/8/4R/6R/8R is incorporated in medical treatment electrocardiogram (ECG) and electroencephalogram (EEG) application all features typically required.Can
Available from Texas Instruments (Texas Instruments) in January, 2012 revise tables of data SBAS459I in further detail
Description ADS129x device, tables of data SBAS459I is incorporated herein by reference.
When using low resolution (16) ADC, signal demand is significantly enlarged (usual 100x to 200x) and obtains institute
The resolution needed.When using high-resolution (24) delta sigma ADC, the gain that signal demand is moderate, such as, 0 arrives 5x.Therefore,
The second gain stage and circuit needed for eliminating DC biasing can be removed.This causes the overall minimizing of area and cost.Δ-
The whole frequency content (content) of ∑ method stick signal and numeral post processing is given to the motility of abundance.
Signal processing 140 is coupled to receive the stream 150 of the digital ECG data from analog portion 110.Numeral letter
Number processor (DSP) 141 is coupled to store the nonvolatile memory 144 of the instruction defining various signal processing algorithm
On.Random access memory (RAM) 143 can carry out data while being used for performing various signal processing algorithm by DSP 141
Storage.Display driver 142 can be the microcontroller of management display graphic user interface (GUI) in display device 145
Unit (MCU), display device 145 can be that such as LCDs (LCD) is currently known or later exploitation any other
Display device.DSP 141 and MCU 142 may be implemented within single IC, such as, available from the OMAP of Texas Instruments
(open multimedia application platform) device etc..Other embodiments can use processor or the signal of currently known or later exploitation
Other embodiment of processor.Such as, different types of SOC(system on a chip) (SoC) IC can containing such as two to 16,
Or more DSP core.
Can provide one or more wireline interface 146, described wireline interface support ECG signal after processing passes
It is delivered in another system USB (USB (universal serial bus)), the RS232 for assessment further, or other type of wireline interface.
Support ZigBee, bluetooth or one or more wave points 147 of other type of wave point can be provided, this or
Multiple wave points ECG signal after processing is delivered in another system for assessment further.Further, it is also possible to
The ECG signal after processing via such as cell phone or data network is provided to be sent to the wave point of remote system
147。
Power management logic 102, clock, temperature sensing logic and fan logic control may also be included in that ECG system
In system 100, to provide power supply and temperature to control for analog systems 110 and dsp system 140.
Can obtain the major part being included in signal processing 140 or institute in cell phone such as smart mobile phone now
There are assembly, described cell phone can be programmed to provide required signal processing.Previously, the one-tenth of ECG, EMG and EEG equipment
This is prohibitively expensive for consumption type device.When smart mobile phone is programmed to perform ECG process, by AFE (analog front end) part
110 are coupled on smart mobile phone as hereafter can be provided with will be described in further derail cost-effective medical monitoring device.
Fig. 2 be a diagram that the curve 200 of typical heart beat cycle.As other biomedicine signals, ECG is considered as
Nonstatic process.Its meansigma methods and its standard deviation elapse in time and change.But, ECG contains useful information, described in have
Can be construed to be referred to as the pseudo-definitiveness pattern of PQRST pattern by information.During each heart beating, healthy heart will have
The advancing in order of the depolarization wave triggered by cell in sinuatrial node, carried out by atrium, through " inherence conducting path ", and
And spread at whole ventricle subsequently.Normal ECG (EKG) is by P ripple, QRS complex (complex) and T wave component.P wave table shows
Atrial depolarization and QRS represent ventricular depolarization.The phase place of the rapid repolarization of T ripple reflection ventricle.
There are the some exceptions that may notice that when analyzing PQRST ripple.Described exception can include following in one
Or multiple:
-lead reverse, wherein P is typically upright, or there is upright P ripple in aVr.This change is generally found at it
Under conditions of middle pulse is advanced through atrium via off path (such as dystopy atrium or A-V nodal rhythm);
The amplitude of-increase is indicated generally at atrial hypertrophy and especially in A-V valvular heart disease, hypertension, pulmonary heart disease, and
Congenital heart disease is found;
The width of-increase is indicated generally at Left atrium enlargement or pathological changes heart muscle;
-Bipolar the second half portions at P ripple are notable for, time negative, being the notable of Left atrium enlargement in III and V1 that lead
Mark;
Recess in-atrial hypertrophy (P-mitrale): P is typically wider and jagged, and ratio exists in the I that leads
Lead higher in III.When distance between crest was more than 0.04 second, recess is considered as significant;
-peak value instruction right atrium tension force.These high point P ripples are generally higher than in the I that leads in the III that leads.This is claimed
For pulmonary P wave (P-Pulmonale);
The disappearance of-P ripple occurs in A-V nodal rhythm and S-A block.
Fig. 3 be a diagram that referring back to Fig. 1 within the period of about 50 seconds in be received from the original of analog portion 110
The curve 300 of the sequence of ECG data 150.In this embodiment, modulus (ADC) code value is based on 24 bit pads.Other is implemented
Example can use different value based on different ADC transducer precision.Noise, artifact, from the EMG (myoelectricity of other muscle in health
Figure) disturb, breathe, the simple analysis of ECG may be made to become difficulty from the 60Hz interference etc. of power line.In order to retain PQRST
Information, it may be necessary to the bandwidth of substantially 3Hz to 30Hz is provided.In certain embodiments, the bandwidth being up to substantially 1000Hz is permissible
Increase accuracy.
Fig. 4 is the indicative icon of the real-time ECG wave filter 400 implemented by the ECG system of Fig. 1.Each QRS wave steady
Strong detection is important for the further digital processing of ECG sequence.Even if detector also should when signal to noise ratio (SNR) is degenerated
Work.In this embodiment, ECG wave filter is divided into three parts;Part 1 410 band filter, the logical filter of part 2 420 band
Ripple device, and part 3 430 QRS detector.Referring back to Fig. 1, the stream 150 of raw ECG is received from AFE (analog front end) part
110.Generally, as discussed about Fig. 3, the stream of raw ECG will include interference and distortion.Real-time ECG described herein
The embodiment of wave filter is tested for the MIT-BIH data base including based on file every 30 minutes 48 records.This is
It is widely used in the basis reference of test QRS detection algorithm.Embodiment described herein obtains more complicated with use non-
The medical grade equipment of real time algorithm similar 99.7% sensitivity and the forward predictability of 99.9%.
Part 2 420 be intended to provide prepare for cardiologist explain through bandpass filtering ECG.In this embodiment,
The sample frequency using 250Hz records initial data ECG by AFE (analog front end) 110.In another embodiment, it is possible to use
Higher or lower sample frequency.Use median filter (P point) 422 filtering ECG initial data in a non-linear manner so that go
Except all radio-frequency components and transport through low pass filter LP1 423 subsequently, in order to remove remaining median filter noise.
Same ECG initial data is made to postpone 424P sampling and deduct the output of 425 LP1 wave filter 423 from which so that suppression
The baseline drift (wander) of ECG.Median filter 422, LP1 wave filter 423 and subtracting from original delayed signal
The combination of method 425 is equivalent to high-pass filtering.Low pass filter LP2 426 is 5 rank wave filter, and described wave filter has cutting of 33Hz
Only frequency and generation are delayed by filtered ECG signal H of 427, in order to align with the result from part 3 430.Output
Signal H prepares to explain for cardiologist.The example of filtered ECG signal H is illustrated at 450.By using P's
Appropriate value, whole PQRST ripple is retained and non-distortion.Generally, the value of P can be chosen as such as P=Fs/2.
Part 1 410 is intended to prepare the ECG signal for beat detector 430.Except median filter 412 altered with
Having outside the point value of (N), part 1 is equal to part 2.By selecting the appropriate value of N, baseline drift, P ripple and T ripple are also
Can be inhibited in this part;The QRS part making only signal is kept as signal B 444, for processing further.
Generally, the value of N can be chosen as such as N=Fs/10.
Part 3 430 represents QRS detector self.In order to amplify the high frequency content of QRS wave, signal B is sought twice square
431 to produce signal C1.Three threshold signals are obtained as follows from signal C1.Signal C1 non-linearly bi-directional scaling is to be formed
Signal C2.Each sampling k of signal C2 is produced as postponing the denary logarithm of 432Q sampling, as by equation (1) institute
Instruction.
C2k=log10 (C1k) (1)
First threshold signal is signal D.As indicated by by equation (2), mobile maximization 433 is utilized to produce signal D
Each sampling k, the mobile M point sliding window maximized on 433 use signal C1, described sampling acquisition denary logarithm
And postpone Q-M sampling.Therefore, signal D represents the peak value of signal C1.
Dk=log10 (max (C1(k)...C1(k+M))) (2)
Second Threshold signal is signal E.As indicated by by equation (3), use on signal C1, use T point sliding window
Rolling averageization 434 produce each sampling k of signal E, described sampling obtains as denary logarithm and postpones Q-T
Individual sampling.Therefore, signal E represents the meansigma methods of the signal C1 in window ranges.
3rd threshold signal is signal G.As indicated by by equation (4), use on signal C1, use Q point sliding window
Mobile maximization obtain each k sample of signal G.
Gk=(max (C1(k)...C1(k+Q))) (4)
Signal G passes the wave filter LP3 436 with transmission function according to equation (5).
Wherein, coefficient a1, b0, b1 is according to the function that equation (6) dynamic calculation is the difference between signal D and signal E.
Fk=s (Dk-Ek) (6)
Wherein s is proportionality constant.Coefficient a0, a1, b0, b1 obtain, wherein subsequently from almost linear as shown in Figure 5 returns
Curve 502 represents coefficient b0 and b1, and curve 504 is with function FkRepresent coefficient a1.
Obtainable known method in the environment such as Matlab, Octave is used to calculate (to enter for Fc=[1-10Hz]
Row iteration) a1, b0, b1.For example, function [B, A]=BUTTER (N, Wn) is by using input N=exponent number and Wn=2*
Pi*Fc/ sample frequency makes b0 and b1 and a0 return in [B, A] vector.Algorithm after this is:
-select cut-off frequency Fc and filter order (Fc=[1-10Hz] and exponent number are 1);
-for every a pair Fc and exponent number, obtain the analog transfer function of correspondence;
-use the bilinear transformation applied to obtain the numerical coefficient of correspondence.
In this embodiment, b0, b1 and a1 are almost linear to the change of cut-off frequency Fc and can be write as: b0=
B1=0.034*Fc+0.064;A1=0.022*Fc+0.99.Cut-off frequency value Fc=[1-10Hz] can by suitable by than
Example scaling substitutes each sampling with the Fk from equation (6) in the equations.Such as, the value of Fk=0 changes into Fc=
1Hz, and the value that Fk is 100dB is corresponding to the Fc of 10Hz.
3rd threshold signal J is formed by the output of wave filter H (z) 436 represented with dB having moved down 437 threshold values 438, institute
State threshold value to be chosen as in this example: threshold value=6dB.As explained above, the difference between signal Fk to Dk and Ek is proportional,
Therefore proportional to the ratio (difference in dB) between the peak value of signal C1 and meansigma methods.When ECG signal has noise, peak
Being worth relatively low with the ratio of meansigma methods, therefore the cut-off frequency of H (z) is relatively low and J signal (described J signal is main threshold) is through strong
Strong low-pass filtering is to avoid error detection.When signal C1 is clean, the cut-off frequency of LP3 increases and J signal is through more weak
Low-pass filtering, and therefore can obtain the lower value of C1 signal.This is avoided loss to beat (false negative).All use signal
C2, D and J perform each judgement of beat detector 439.If C2==J and C2 > D, then QRS peak is considered as true
Beat in fact (wherein "==" meaning is that C2 is equal to J).It should be noted that and beat for each, only on a single point, C2 is equal to
J.Therefore, it now is possible to annotate beating on signal H, described signal H is provided as the output of part 2 420.
Example signal H having in the annotated heart beating indicated at 451 is illustrated at 450., this annotation is explaining warp
Cardiologist can be provided to together with signal 450 during the ECG signal 450 filtered use.
Heart rate can be calculated as below, as indicated by by equation (7).
Wherein:
BnbIt it is the detected number of times beated;
FsIt it is sample frequency;
N is the number of obtained sampling;And
T is the number of seconds of observation window
Fig. 6,7A to 7D illustrate the waveform produced during the analysis carried out ECG signal by real-time ECG wave filter 400.
Curve 602 represent with reference to Fig. 4 at the point 441 of the part 1 410 of real time filter 400 by postpone 414 output original
Delayed ECG data.Curve 604 represents the ECG signal through bandpass filtering exported at point 442 by band filter 413.
Curve 606 represents the signal 443 through medium filtering produced by subtraction 415, and illustrates baseline drift and be suppressed and T
Ripple is also suppressed.
Fig. 7 A illustrates the representation signal B in the part 3 430 being input to real time filter 400.In figure 7b, curve
704 represent signal C2, and curve 705 represents signal D, and curve 706 represents signal D.In fig. 7 c, curve 704 represents signal
C2, curve 707 represent signal E, and curve 708 represents signal J.In fig. 7d, curve 710 represents signal H, and 711
Indicated at point represent the heart beating that detects.As described above, signal H can be passed to cardiologist in real time to enter
Row is analyzed.
Fig. 8 to 9 be a diagram that by the ECG system of Fig. 1 carry out to by ECG data exception indicate early-warning conditions
The sequential chart of detection.Fig. 8 illustrates the detection to heart rate variability (HRV), and described heart rate variability elapses in time
Time difference between twice continuous chattering (R-R interval).Curve 802 is can be through analyzing to detect warning further with reference to Fig. 4
The example ECG signals H of condition.As indicated at 803, according to described above, detected heart beating is annotated.
Simple HRV detection function can detect HRV by the timing of relatively each annotated heart beating 803.If HRV illustrates aobvious
Work and change drastically, then it may be an indicator that be referred to as the condition of premature ventricular beat.
Premature ventricular beat refers to slightly abnormal heart rhythm, and described slightly abnormal heart rhythm is usual in the case of there is not other cardiovascular disease
There is no serious problem.This condition describes wherein alternans of heart and carries out the shape that " normally " beats and " too early " beats
State.Term " regular sinus rhythm " describes normally beating of heart, and wherein electric pulse is initial in sinuatrial node (SA knot);Advance logical
Cross atrium and atrioventricular node (AV knot);And terminate in ventricle.When atrium or ventricle are receiving the arteries and veins from sinuatrial node
There will be when starting the electric pulse of himself before punching and beat too early.Term " premature ventricular beat " is generally used for describing when normal sinus is jumped
Move situation when beating alternately with premature ventricular.
Fig. 9 illustrates another condition being referred to as bigger ST section disappearance as indicated by 902, and described bigger ST section lacks
Mistake can be easily detected and be used for trigger in real time early warning.With reference to Fig. 4, ECG signal H can be through analyzing further
Be determined by S point after the heart beating detected the most regularly be reduced to specify negative threshold value get off detection bigger ST lack
Lose.
Although Fig. 8 to Fig. 9 illustrates two distinct types of exception, it should be appreciated that there is the most eurypalynous exception, described
Extremely the analysis to ECG signal after ECG signal has been filtered and detected as described above heart beating can be passed through
Detect in real time.
Figure 10 be a diagram that the flow chart of the operation being carried out ECG analysis by ECG system as described above.As the most more
Add and describe in detail, from AFE (analog front end) subsystem, receive 1002 raw ECG data sample streams, described AFE (analog front end) subsystem pair
ECG signal carries out sampling and convert thereof into ECG data sample streams.ECG data sample streams represents the cycle sequence of PQRST pattern
Row.
By using nonlinear filter to filter 1004 raw ECG data to form filtered ECG data, in order to
Littleization baseline drift, thus suppress the T ripple part of each PQRST pattern.Can be by using median filter and low pass filtered
The cascade filtering raw ECG data of ripple device, make raw ECG data postpone, and deduct low from delayed raw ECG data
The output signal of bandpass filter, in order to transmit the filtering 1004 performing raw ECG data through the data of high-pass filtering.
Filtered ECG data is performed 1006 nonlinear operations with formed expand each PQRST pattern R peak through putting
Big signal.Can be by forming amplified signal to filtered ECG data is squared at least one times.In this embodiment
In, ECG data is sought twice square.
Amplified signal use mobile maximum filter form 1008 first threshold signals.Similarly, to warp
The signal amplified uses mobile maximum filter to form 1010 the 3rd threshold signals.By first window width being used for
One threshold signal and different window widths, for the 3rd threshold signal, can obtain first threshold signal and the 3rd threshold
Value.
Amplified signal use mobile mean filter obtain 1009 Second Threshold signals.
Wave filter is used to filter 1014 the 3rd threshold signals to form filtered threshold signal, at described wave filter
In, cut-off frequency is dynamically selected 1012 for first threshold signal and the function of Second Threshold signal.Can be by utilizing one
3rd threshold signal is filtered by rank infinite impulse response (IIR) wave filter, and by dynamic for the cut-off frequency that is used for IIR
Selection 1012 is the linear function of the difference between the logarithm of first threshold signal 1008 and the logarithm of Second Threshold signal 1008
Perform the 3rd threshold signal is filtered.
In another embodiment, the 3rd threshold signal is filtered 1014 to be performed by following: utilize one
3rd threshold signal is filtered by rank infinite impulse response (IIR) wave filter, and is dynamically selected by the cut-off frequency being used for IIR
Select 1012 for first threshold signal 1008 and the linear function of the ratio of Second Threshold signal 1009.
Only when sampled value is equal to 1020 first threshold signals 1008 and the threshold signal 1014 filtered more than 1022
Sampling identification 1024 in amplified signal is heart beating point by.
Nonlinear filter can also be used to filter 1030 raw ECG data, in order to minimize each PQRST pattern
Baseline drift, thus form visual ECG signal.Visual ECG signal can be with annotated 1032 to indicate each heart beating point.Such as,
Annotated visual ECG signal can be shown 1034 subsequently and or be sent to remote location for heart on locally displayed device
Sick expert watches.Raw ECG data is filtered 1030 to be performed by following: use median filter and low pass
Raw ECG data is filtered, makes raw ECG data postpone by the cascade of wave filter, and from delayed raw ECG number
Deduct low pass filter according to outputs signals to the transmission data through high-pass filtering.
Figure 11 illustrates another embodiment of the mobile ECG system 1100 utilizing cell phone 1120.Analog front-end module
1110 can include that the high impedance for various electrode leads such as operational amplifier (op-amps) 1112,1113 inputs.Modulus
Low-pass filtered analogue signal is converted into numeral expression by transducer 1114.In this embodiment, ADC 1114 has 24
Precision.In other embodiments, it is possible to use there is the ADC of greater or lesser precision.Illustrate three electricity in this example
Leading in pole, described electrode lead can be used for motion and simply monitors purposes.As discussed the most in further detail, other is implemented
Example can include more electrode lead.
Analog portion 1111 may be implemented within single integrated circuit such as ADS1294/6/8/4R/6R/8R, described
ADS1294/6/8/4R/6R/8R is family's multichannel, simultaneously sampling, the delta sigma analog-digital converter (ADC) of 24, described Δ-
∑ analog-digital converter has built-in programmable gain amplifier (PGA), internal reference, and airborne agitator.ADS1294/
6/8/4R/6R/8R is incorporated in medical treatment electrocardiogram (ECG) and electroencephalogram (EEG) application all features typically required.
The Serial No. that the ECG produced by ADC is sampled can be sent to the USB on mobile phone 1120 by USB function 1116
Input.Alternatively, can include bluetooth transmitters circuit 1115 with will be produced by ADC ECG sampling Serial No. with nothing
Line mode is sent to the Bluetooth Receiver on mobile phone 1120.
Mobile phone 1120 represents any one in the some smart mobile phones being purchased from multiple supplier.Smart mobile phone generally includes
The signal processing hardware similar to the signal processing 140 of Fig. 1 and ability.Application software, also referred to as " app ", can
Download in mobile phone to use to download for the normal app of this smart mobile phone.Described app is configured to perform storage
Software instruction in memory performs filtering the most described in more detail, threshold process, and QRT and detects function, institute
It is addressable by dsp processor or the other type of processor that is included in smart mobile phone 1120 for stating memorizer.As above
Literary composition is described in more detail, and described application program can be with detection various alert if (such as, HRV, premature ventricular beat, and bigger ST
Section disappearance etc.) monitoring of coming together process after ECG signal and heartbeat message is provided.
Electrode may be coupled to the performance of the heart in the main body such as people or domestic animal with supervision subjects in real time.Such as, ECG
App can show ECG signal on the display screen of smart mobile phone.Such as, ECG app can show ECG on the display screen
Early warning, or use the speaker being included in smart mobile phone to send audible warning.Such as, ECG app can also use and include
Data transmission capabilities in smart mobile phone will be equivalent to the filtered ECG signal of the signal H on Fig. 4 in real time and is sent to remotely
Monitoring system.The early-warning conditions that it has been detected by can also be sent to long distance control system by ECG app.
Long distance control system may be located in (such as) hospital or in the office of doctor.Cardiologist can the most again
Check the ECG signal in long distance control system and the early warning information being associated.Cardiologist can use smart mobile phone
Voice channel is come and Parties ' Mutual, and ECG data uses the data channel of smart mobile phone to continue to be sent to remote site simultaneously.
Figure 12 A illustrates the other embodiments of mobile ECG system.In this example, analog front-end module 1210 can be with coupling
Close one group and lead on 1204, described in lead in the end point on the shirt 1202 being coupled in turn to include electrode, described electrode
It is fixed on shirt and is arranged to when shirt 1202 is put on by main body carry out contact skin with main body.The example of shirt 1202
It is NuMetrex shirt or the motion bra being purchased from Tyke Si Zhuoni Ces Co., Ltd (Textronics).
Front-end module 1210 is similar to front-end module 1110, and can be implemented to support that (such as) 12 ECG lead
Connection or the ECG Lead of lesser number.Front-end module can use ADS 1294 analog-digital converter 1212 to implement, as the most more
Describe in detail.Microcontroller is alternatively coupled on ADC 1212 and controls the data transmission from ADC 1212 to USB output.Micro-
Controller is also coupled in Bluetooth circuit 1214 and thus controls being wirelessly transferred of the data from ADC 1212.Micro-control
Device 1213 processed can be the MSP430 device that (such as) is purchased from Texas Instruments.Bluetooth circuit 1214 can be that (such as) can
CC2560 device purchased from Texas Instruments.
Such as, USB line may be used for being coupled on laptop computer 1230 front-end module 1210.Laptop computer
1230 represent any number of known portable computer, and described portable computer includes that enough disposal abilities are with in execution
Literary composition Real-Time Filtering described in more detail and QRS identify.Laptop computer 1230 is configured to from front-end module 1210
Receive ECG data stream.Laptop computer 1230 is also configured with software application, and described software application utilizes and is included in
Processor in laptop computer 1230 processes the ECG data stream sent from the main body wearing shirt 1202.
Described app is configured to perform storage software instruction in memory and performs the most described in more detail
Filtering, threshold process, and QRT detect function, described memorizer is included in the dsp processor in computer 1230 and can visit
Ask.As the most described in more detail, described application program can with detect various early-warning conditions (such as, HRV, premature ventricular beat, with
And bigger ST section disappearance etc.) monitoring of coming together process after ECG signal and heartbeat message is provided.
Electrode may be coupled to the performance of the heart in the main body such as people or domestic animal with supervision subjects in real time.Such as, ECG
App can show ECG signal on the display screen of laptop computer 1230.Such as, ECG app can be on the display screen
Display ECG early warning, or use the speaker being included in laptop computer to send audible warning.Such as, ECG app also may be used
To use the data transmission capabilities being included in laptop computer to will be equivalent to the filtered ECG of the signal H on Fig. 4 in real time
Signal is sent to long distance control system.The alert if that it has been detected by can also be sent to remotely monitor system by ECG app
System.Transmission to remote supervisory station point can be carried out by the Internet, such as, uses and is included in laptop computer 1230
Wired or wireless connection.
Long distance control system may be located in (such as) hospital or in the office of doctor.Cardiologist can the most again
Check the ECG signal in long distance control system and the warning message being associated.Cardiologist can use calculating on knee
Voice channel (such as, the speech through the Internet) on machine is come and Parties ' Mutual, and ECG data uses laptop computer simultaneously
Data channel continue to be sent to remote site.
Figure 12 B illustrates another embodiment of mobile ECG system.In this example, analog front-end module 1210 can be attached
Link on shirt 1203 and be coupled to include that in one group of end point on the shirt 1203 of electrode, described electrode is fixed to shirt
Go up and be arranged to carry out contact skin when shirt 1203 is put on by main body with main body.The example of shirt 1203 is through suitably repairing
Change to carry the NuMetrex shirt of analog front-end module 1210 or motion bra.
Such as, analog front-end module 1210 be configured to by from the main body wearing shirt 1203 collect ECG data with
It is transmitted wire-lessly to the neighbouring Bluetooth Receiver being positioned in smart mobile phone 1220.Smart mobile phone 1220 is configured with ECG app, as
The most described in more detail, and thus can process and show ECG data in real time and also ECG data is sent to remotely
Website is analyzed for cardiologist or other healthy ward.
Therefore, it has been described that low cost mobile system, described mobile system provides the existing research than announcing more preferable
QRS detects performance in real time.Need not any study stage needed for many existing algorithms.Various embodiments can be by analyzing
PQRST waveform provides the other heart information outside only heart rate;This can enable various embodiment provide for various bars
The real-time early warning of part.These low cost mobile systems can be useful to domestic use, body-building purposes etc..
Other embodiments
AFE (analog front end) part can be in the various modes of the currently known or later exploitation by providing protection and ease for use
Pack.Such as, packaging can be waterproof and firm so that AFE (analog front end) can be permanently attached to shirt or motion bra
Go up and stand to use normally and cleaning process.
AFE (analog front end) can be battery powered, or it can packaged have from the motion of trial body, solar energy etc.
In obtain the energy extraction units of electric power.
Although it have been described that wired USB link or wireless blue tooth link, but other embodiments can use any classification
Known or wired (metal or the optics) of later exploitation or radio interconnected agreement, described agreement support required transfer rate with
Hold the real-time Transmission of filtered ECG data.Certainly, required transfer rate will depend upon which and selects the ADC essence for described embodiment
Degree and sampling rate.
Although there is described herein personal computer and smart mobile phone for providing required signal processing to perform
Filtering and detection function, but other type of mobile consumption device can be used for implementing filtering described herein and
Detection function, such as tablet PC, personal digital assistant etc..In general, as used herein, consumption device is can be by
Consumer provides the ECG fexible unit of monitoring for other task with other.
Although there is described herein scaling by logarithmic scale of threshold value, but another embodiment can using another form of non-
Linear bi-directional scaling.Another embodiment can use linear expression and usage rate to be substituted for comparing the difference of threshold value
Value.
In other embodiments, in order to perform the required signal processing of filtering and detection function can be by through clear and definite
It is designed for the digital processing system of ECG purposes, non-moving system, medical grade system etc. to provide.
Although filtered ECG data may can be sent to remote site by system described herein in real time, but
Other embodiments can provide storage function to store a part of filtered ECG data to be delivered to remote system after a while.
If the embodiment of wave filter described herein and method can be provided in the digital display circuit of dry type
On any one: digital signal processor (DSP), general purpose programmable processors, special circuit, or SOC(system on a chip) (SoC), such as,
DSP and the combination of reduced instruction set computer (RISC) processor together with various special accelerators.SoC can be containing one or many
Individual megacell, described megacell each includes setting with the customization of the pre-designed functional circuit combination provided by design function storehouse
Meter functional circuit.
Technology described in the present invention can use hardware, software, firmware, or its any combination is implemented.If using software
Implement, then described software can perform in one or more processors, such as microprocessor, special IC
(ASIC), field programmable gate array (FPGA), or digital signal processor (DSP).The software performing described technology can be just
It is stored in computer-readable media (such as compact disk (CD), disk, tape, file, memorizer, or other meter any with beginning
Calculation machine readable storage devices) in, and load within a processor and perform.In some cases, software can also calculate
Selling in machine program product, described computer program includes computer-readable media and for described computer-readable matchmaker
The packaging material of body.In some cases, software instruction can be by can removable computer readable media (such as, floppy disk, light
Dish, flash memories, usb key), by coming from the transmission path in the computer-readable media in another digital display circuit etc.
Distribution.
It is understood by those skilled in the art that, in the range of the invention advocated, many other embodiments and change
It is all possible for changing.
Claims (12)
1., for the method processing ECG data, described method includes:
Receive the stream of the raw ECG data sampling including PQRST pattern;
By using nonlinear filter to be filtered forming filtered ECG data to described raw ECG data, in order to
Minimize baseline drift, and thus suppress the T ripple part of each PQRST pattern;
Described filtered ECG data is performed nonlinear operation to form the amplified of the R peak of each PQRST pattern of expansion
Signal;
Described amplified signal use mobile maximum filter obtain first threshold signal and the 3rd threshold signal;
Described amplified signal use mobile mean filter obtain Second Threshold signal;
Wave filter is used to be filtered forming filtered threshold signal to described 3rd threshold signal, at described wave filter
In, cut-off frequency is dynamically selected for described first threshold signal and the function of described Second Threshold signal;And
Only when the value of sampling is equal to described first threshold signal and when exceeding described filtered threshold signal, by described through putting
The big sampling in signal is identified as heart beating point,
Wherein it is filtered including to described raw ECG data:
Described raw ECG data is filtered by the cascade using median filter and low pass filter;
Postpone described raw ECG data;And
The output signal of described low pass filter is deducted, in order to transmit the data of high-pass filtering from the raw ECG data postponed.
Method the most according to claim 1, it farther includes:
Use nonlinear filter that described raw ECG data is filtered, in order to minimize the baseline drift of each PQRST pattern
Move, thus form visual ECG signal;And
Annotate to indicate each heart beating point to described visual ECG signal.
Method the most according to claim 2, wherein use nonlinear filter described raw ECG data is filtered with
Just minimize the baseline drift of each PQRST pattern thus form visual ECG signal and include:
Described raw ECG data is filtered by the cascade using median filter and low pass filter;
Postpone described raw ECG data;And
The output signal of described low pass filter is deducted, in order to transmit and filter through high pass from described delayed raw ECG data
The data of ripple.
Method the most according to claim 1, wherein by being formed described filtered ECG data is squared at least one times
Described amplified signal.
Method the most according to claim 1, wherein is filtered including to described 3rd threshold signal:
Utilize single order infinite impulse response i.e. iir filter that described 3rd threshold signal is filtered;And
The cut-off frequency being used for described IIR is dynamically chosen as the logarithm of described first threshold signal and described Second Threshold signal
Logarithm between the linear function of difference.
Method the most according to claim 1, wherein is filtered including to described 3rd threshold signal:
Utilize single order infinite impulse response i.e. iir filter that described 3rd threshold signal is filtered;And
The cut-off frequency being used for described IIR is dynamically chosen as the ratio of described first threshold signal and described Second Threshold signal
Linear function.
Method the most according to claim 1, wherein uses mobile maximum filter to described amplified signal
It is that first window width is used for described first threshold signal to first threshold signal and the 3rd threshold signal, and by difference
Different window width in described first threshold signal is used for described 3rd threshold signal.
8. being used for processing in real time a computer implemented system for ECG data, described system includes:
For receiving the raw ECG data sampling including PQRST pattern from the AFE (analog front end) being coupled on described computer
The device of stream;
For by using nonlinear filter to be filtered forming filtered ECG data to described raw ECG data
To minimize baseline drift and thus suppressing the device of the T ripple part of each PQRST pattern;
For described filtered ECG data is performed nonlinear operation with formed the R peak that expands each PQRST pattern through putting
The device of big signal;
For described amplified signal uses mobile maximum filter obtain first threshold signal and the 3rd threshold value
The device of signal;
For using mobile mean filter to obtain the device of Second Threshold signal described amplified signal;
For using wave filter to be filtered being formed the device of filtered threshold signal to described 3rd threshold signal, in institute
Stating in wave filter, cut-off frequency is dynamically chosen as described first threshold signal and the function of described Second Threshold signal;With
And
For only when sampling value equal to described first threshold signal and when exceeding described filtered threshold signal by described warp
Amplify the sampling in signal and be identified as the device of heart beating point,
Wherein the device for being filtered described raw ECG data is configured to:
Described raw ECG data is filtered by the cascade using median filter and low pass filter;
Postpone described raw ECG data;And
The output signal of described low pass filter is deducted, in order to transmit the data of high-pass filtering from the raw ECG data postponed.
Computer implemented system the most according to claim 8, it farther includes:
The baseline of each PQRST pattern is minimized for using nonlinear filter that described raw ECG data is filtered
Drift about thus form the device of visual ECG signal;
For described visual ECG signal is annotated to indicate the device of each heart beating point;And
For annotated visual ECG signal is shown the device on the display being coupled on computer.
Computer implemented system the most according to claim 8, wherein by least one times to described filtered ECG number
Described amplified signal is formed according to squared.
11. computer implemented systems according to claim 8, wherein for being filtered described 3rd threshold signal
Device be configured to:
Utilize single order infinite impulse response i.e. iir filter that described 3rd threshold signal is filtered;And
The cut-off frequency being used for described IIR is dynamically chosen as the logarithm of described first threshold signal and described Second Threshold signal
Logarithm between the linear function of difference.
12. computer implemented systems according to claim 8, wherein for being filtered described 3rd threshold signal
Device be configured to:
Utilize single order infinite impulse response i.e. iir filter that described 3rd threshold signal is filtered;And
The cut-off frequency being used for described IIR is dynamically chosen as the ratio of described first threshold signal and described Second Threshold signal
Linear function.
Applications Claiming Priority (6)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
EPEP12290088.9 | 2012-03-12 | ||
EP12290088.9 | 2012-03-12 | ||
EP12290088 | 2012-03-12 | ||
US13/434,725 | 2012-03-29 | ||
US13/434,725 US8755877B2 (en) | 2012-03-12 | 2012-03-29 | Real time QRS detection using adaptive threshold |
PCT/US2013/030586 WO2013138372A1 (en) | 2012-03-12 | 2013-03-12 | Real time qrs detection using adaptive threshold |
Publications (2)
Publication Number | Publication Date |
---|---|
CN104203091A CN104203091A (en) | 2014-12-10 |
CN104203091B true CN104203091B (en) | 2016-11-30 |
Family
ID=
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US8755877B2 (en) | Real time QRS detection using adaptive threshold | |
EP2676604B1 (en) | Real time QRS duration measurement in electrocardiogram | |
Preejith et al. | Wearable ECG platform for continuous cardiac monitoring | |
US9078577B2 (en) | Circuit for heartbeat detection and beat timing extraction | |
Joo et al. | Prediction of spontaneous ventricular tachyarrhythmia by an artificial neural network using parameters gleaned from short-term heart rate variability | |
Gawali et al. | Implementation of ECG sensor for real time signal processing applications | |
Mukti et al. | Performance analysis of twelve lead ECG based on delivery distance using bluetooth communication | |
Prasath | Wireless monitoring of heart rate using microcontroller | |
KR20160008368A (en) | Method and apparatus for processing biosignal using recursive estimation | |
US20100145205A1 (en) | Analyzing alternans from measurements of an ambulatory electrocardiography device | |
Cristea et al. | A wireless ECG acquisition device for remote monitoring of heart rate and arrhythmia detection | |
CN104203091B (en) | Use the real-time QRS detection of adaptive threshold | |
US8165665B2 (en) | Chip for sensing a physiological signal and method for sensing the same | |
Riyadi et al. | Development of FPGA-based three-lead electrocardiography | |
Anas et al. | On-line monitoring and analysis of bioelectrical signals | |
Xing et al. | A portable neuECG monitoring system for cardiac sympathetic nerve activity assessment | |
Tsau et al. | A new approach to the diagnostic quality ambulatory ECG recordings | |
Sakli et al. | Internet of Things (IoT) System with Matlab interface for Multi Patient ECG's Monitoring | |
Iskandar et al. | Blue electrocardiogram (ECG): ECG with bluetooth feature integrated with smartphone | |
CN101558994B (en) | Method and device for testing heart rate deceleration capacity | |
Yadav et al. | Designing of portable and low cost ECG signal acquisition system | |
Dhande | LABVIEW BASED ECG SIGNAL ACQUISITION AND ANALYSIS | |
Bharadwaj et al. | Design and Analysis of Electrocardiogram Heart Rate Variability Monitoring Systems | |
Liu et al. | Chaotic phase space differential (CPSD) algorithm for real-time detection of VF, VT, and PVC ECG signals | |
Uddin | Non-conventional Energy Operated ECG System |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant |