IRJET- Design Simulation and Analysis of Efficient De-Noising of ECG Signals with Adaptive Filtering Algorithms & Patch based Method

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 07 Issue: 01 | Jan 2020 www.irjet.net p-ISSN: 2395-0072
© 2020, IRJET | Impact Factor value: 7.34 | ISO 9001:2008 Certified Journal | Page 1616
Design Simulation and Analysis of Efficient De-noising of ECG Signals with
Adaptive Filtering Algorithms & Patch Based Method
Anita Yadav(M.TECH Scholar)1, Manish Mukhija(Asst. prof.)2
1,2Department of Computer Science, Modern Institute of Technology and Research Center, Alwar, Rajasthan, India
-------------------------------------------------------------------------***-----------------------------------------------------------------------
Abstract— ECG signals have been proven a very versatile
tool for detection of cardiovascular diseases. But during
recording of these signals, the ECG data gets contaminated by
various noise signals caused by power line interference, base
lone wander, electrode movement, muscle movement (EMG)
etc. These noise signals are known as artifacts. These artifacts
mislead the diagnosis of heart which is not desired. To avoid
this problem caused by artifacts, removal of these artifacts
has become essential. There are various techniques which
have been used for artifacts rejection from ECG. Conventional
filters remove the artifacts up to some extent but these filters
are static filters. These filters cannot update their coefficients
with change in environment.
Hence adaptive filters, now days, are used for artifact removal
from ECG signals. Adaptive filters update their coefficients
according to the requirement. There are various adaptive
algorithms such as Leas Mean Square (LMS), Recursive Least
Square (RLS), Normalized Least Mean Square (NLMS) etc are
present. Moreover, there is one more method is described
which is patch based and used for artifact rejection from ECG
signals. This method was previously used only for image
denoising but now it has been using for artifact rejection from
biomedical signals.
Indexed Terms- ECG, EMG, LMS, NLMS, RLS
I. INTRODUCTION
ECG records carry information about abnormalities or
responses to certain stimuli in the heart. Some of the
characteristics of these signals are the frequency and
morphology of their waves. These components are in the
order of just a few up to 1mV and their frequency content
within 0.5Hz and 100Hz depending on individual [1]. The
morphology and frequency are analyzed by physicians in
order to detect heart disorders and heart related pathologies.
However, the ECG signal is generally with other biological
signals like electroencephalogram, electromyogram, baseline
Wander and power line interference. Due to the presence of
artifacts, it is difficult to analyze the ECG, for they introduce
spikes which can be confused with cardiological rhythms.
Thus, noise and undesirable signals must be eliminated or
attenuated from the ECG to ensure correct analysis and
diagnosis [1].
The removal of artifacts in ECG signals is an essential
procedure prior to further diagnostic analysis in many
clinical applications, e.g., detection of QRS complexes,
classification of ectopic beats, analysis of asymptomatic
arrhythmia, extraction of the fetal ECG signal from the
maternal abdominal ECG, diagnosis of myocardial
ischemia, diagnosis of atrial fibrillation, and ECG data
compression.
The goal of ECG signal enhancement is to separate the
valid signal components from the undesired artifacts,
so as to present an ECG that facilitates easy and
accurate interpretation[2].
In recording a heart beat (an ECG), which is being
corrupted by a PLI (50Hz/60Hz) noise (the frequency
coming from the power supply(50Hz) in many countries)
[3]. We remove the noise is to filter the signal with a
notch filter at 50 Hz. However, due to slight variations
in the power supply to the hospital, the exact
frequency of the power supply might (hypothetically)
wander between 47 Hz and 53 Hz. A static filter
would need to remove all the frequencies between 47 and
53 Hz, which could excessively degrade the quality of the
ECG since the heart beat would also likely have
frequency components in the rejected range [5]. To
circumvent this potential loss of information, an adaptive
noise cancellation filter [4] has been used. The adaptive
filter would take input both from the patient and from
the power supply directly and would thus be able to
track the actual frequency of the noise as it fluctuates.
Such an adaptive technique generally allows for a filter
with a smaller rejection range, which means, in our
case, that the quality of the output signal is more
accurate for medical diagnoses[5].
II.LEASTMEANSQUAREADAPTIVEALGORITHM
The LMS algorithm is extensively used in different application of
adaptive filtering due to low computational complexity, stability
andunbiasedconvergence.In anysignal’sprocessestherecan be
error occurred in the required output. There must be suitable
algorithm needed to manipulate this problem. The least mean
square (LMS) algorithm is introduced to minimize the error
betweenagivenpreferredsignalandoutputofthelinearfilterby
adjusting recursively the parameters of a linear filter [6]. The
more suitable and basic algorithm for the adaptive filtering is
LMS, which is also famous for the stability of the system [7].LMS
is the most important algorithms in whole family of algorithms,
which has been developed for minimizing the error [8]. Least
mean square algorithm has lots of benefits in different
applications; it has been productively used in many applications
duetothefollowingperformanceaspects.LMShavetheabilityto
reject noisy data due to minute step size parameter μ. LMS
demonstrate slowly time varying system [7]. In general LMS

adaptive filter removes noise or obtains a desired signal by adapting
thefiltercoefficientwithleast-squarealgorithmbasedongiven filter.
TheperformanceoftheLMSalgorithmisveryhighanditissimplein
implementationfortheremovaloflowfrequencynoise.
The suitable value for step size parameter μ can be selected
accordingtotheapplicationsrequirement[10].
LMSisusedforthesimplificationofgradientvector computation.[6]
The overview of the structure and operation of the LMS algorithm
can be discussed according to LMS algorithm’s properties and its
processes. [9] The main property of LMS algorithm is its
convergence behavior in a stationary environment. [6] LMS is a
linear adaptive filtering algorithm and is consists of two basic
processes.
FilteringProcess:Filtering processisusedtocalculatetheoutputof
linear filter and to generate an estimated error by comparing this
outputwithadesireresponse.[9]
AnAdaptiveProcess:Anadaptiveprocessisusedfortheautomatic
adjustment of the filter’s parameters in accordance with the
estimatederror.[9]
The overview of least mean square (LMS) algorithm is shown in
figure 2.1. The primary input has been taken, where ‘X’ is the
reference input. The error signal occurs for the desired output, there
LMSadaptivefilterhasemployedtomanipulatetheerror.
Fig:2.1:AdaptiveFilter
The equation above shows the desired signal and the filter output,
where d(n) is the desired signal and y(n) is the filter output. For the
minimization of error signal the input vector x(n) and e(n) are
employed. Here it needs to work according to the criterion that is
supposedtominimize.
III.PATCHBASEDNONLOCALMEANSMETHOD
NL is an image de-noising process based on non-local averaging
ofallthepixelsinanimage.Inparticular,theamountofweighting
for a pixel is based on the degree of similarity between a small
patchcenteredonthatpixelandthesmallpatchcenteredaround
the pixel being de-noised [14]. If compared with other
well-known denoising techniques, such as the Gaussian
smoothing model, the anisotropic diffusion model, the total
variation denoising, the neighborhood filters and an elegant
variant, the Wiener local empirical filter, the translation invariant
wavelet thresholding, the NL-means method noise looks more
like white noise [13]. Image denoising has been a subject of
interest in the field of image processing for many years. Noise is
inherent during image acquisition. Reducing the amount of noise
in an image makes the image more pleasing to the eye and it is
also an important pre - processing step since it improve s the
performanceofhighleveltaskssuchasedgedetectionandobject
trackingTherearemanydifferentdenoisingalgorithms,butmost
belongtooneofthefollowingthreeclasses:
1.Filters that act on a local region within an image, like mean,
medianorGaussianfilters.
2. Filters that take the entire image into consideration, such as
frequency do main filters which reduce noise in the Fourier or
wavelet domain; and Neighborhood filters , which act on pixels
withsimilargraylevelvalues.
A.BasicNonlocalMeansAlgorithm
Nonlocal means denoising [11] addresses the problem of
re-coveringthe true signal ugivena setof noisyobservations,v=
u+n,where nisadditive noise. Fora givensamples,theestimate
ˆu(s) is a weighted sum of values at other points t that are within
some“searchneighborhood’N(s)
Where, λ is a bandwidth parameter, while Δ represents a
local patch of samples surrounding s, containing LΔ
samples; a patch of the same shape also surrounds t.
Although a variety of patch weightings are possible [11],
square patches centered on the points of interest are most
common [12.] The novelty of NLM is that the weighting
w(s, t) depends on the patch similarity, not on the physical
distance between the points s and t.
Averaging similar patches helps to preserve edges in
contrast to more typical filtering (cf., convolution by a
Gaussian smoothing kernel) [12].
Assuming self-similarity extends throughout the signal,
)()(2)()1( nxnenwnw 

the neighborhood N (s) is ideally taken to be the
entire signal, so the averaging process is fully nonlocal.
However, the computation scales linearly with the size of N
(s), so N (s) is usually limited to reduce computation [12].
IV. SIMULATION & RESULTS
To denoise the ECG data with LMS adaptive filtering
algorithm, the ECG signal is generated in MATLAB. To
contaminated the ECG signal 50 Hz power Line Interference is
also generated in MATLAB. Then to validate the denoising
process, the generated power line interference noise is added
to generated ECG signal. The generated ECG signal, noise
signal along with noise added ECG signal is shown below:
Fig 4.1: ECG Signal generated in MATLAB
Fig 4.2: Power Line Interference Noise generated in
MATLAB
Fig 4.3: ECG Signal + Noise
This ECG signal mixed with noise is then filtered by LMS
adaptive algorithm. The ECG signal free from power line
interference is obtained successfully and shown below in
figure 4.4.
Fig 4.4: Filtered ECG Signal from 50 Hz Power Line
Interference Noise
Fig4.5:OriginalSignal,50HzNoiseSignal,ECGSignal+
Noise&FilteredSignalwithLMSAdaptiveFiltering
0 100 200 300 400 500 600 700 800 900 1000
-1
-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
0.8
1
original signal
0 100 200 300 400 500 600 700 800 900 1000
-1.5
-1
-0.5
0
0.5
1
1.5
noise signal
0 100 200 300 400 500 600 700 800 900 1000
-2
-1.5
-1
-0.5
0
0.5
1
1.5
2
ECG + Power Line Interference Noise
0 1000 2000 3000 4000 5000 6000
-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
0.8
1
Filtered Signal
0 1000 2000 3000 4000 5000 6000
-1
0
1
original ecg signal
0 100 200 300 400 500 600 700 800 900 1000
-2
0
2
noise signal
0 1000 2000 3000 4000 5000 6000
-5
0
5
signal after adding noise
0 1000 2000 3000 4000 5000 6000
-1
0
1
Filtered Signal

In this method, The ECG data from MIT-BHE database is taken. Then
10 db SNR noise is created in MATLAB environment. The noise is
added in ECG signal. The denoising of this mixed signal is done using
Patch Based Non Local Means algorithm. The results are shown as
follow:
Fig4.6:ECGwaveformfromMIT-BHEDatabase
Fig 4.7: ECG Signal + Noise
Fig 4.8: Filtered ECG Signal with Patch Based Method
Fig 4.9: OriginalSignal,NoiseSignal,ECGSignal+Noise&
FilteredSignalwithPatchBasedMethod
V.CONCLUSION&FUTURESCOPE
In this paper, the ECG signal contaminated with power line
artifact is denoised with Least Mean Square (LMS) adaptive
algorithm. The ECG signal is created. The power line noise with
50 Hz frequency is created and added with ECG signal. This ECG
signal contaminated with power line noise is then filtered using
Least Mean Square (LMS) adaptive algorithm. The denoised ECG
signal is then recovered. The Patch Based Non Local means
method has also removed the noise from the ECG signal
successfully. The LMS algorithm is simple, robust and easy
to implement.
In future, algorithm with fast convergence can be developed
for denoising of ECG signal since the LMS adaptive
algorithm has slow convergence.
REFERENCES
1.SUPPRESSION OF POWERLINE INTERFERENCE IN ECG
USING ADAPTIVE DIGITAL FILTER, Mbachu C.B,
International Journal of Engineering Science and
Technology (IJEST).
2.Respiration Wander removal from Cardiac Signals using
an Optimized Adaptive Noise Canceller Sowmya. I,
Canadian Journal on Biomedical Engineering & Technology
Vol. 2 No. 3, June 2011.
3.Y. Der Lin and Y. Hen Hu, “Power-line interference
detection and suppression in ECG signal processing,” IEEE
Trans. Biomed. Eng., vol. 55, pp. 354-357, Jan.2008.
4.B. Widrow, J. Glover, J. M. McCool, J. Kaunitz, C. S.
Williams, R. H. Hearn, J. R. Zeidler, E. Dong, and R.
Goodlin, “Adaptive noise cancelling: Principles and
applications,” Proc. IEEE, vol. 63, pp. 1692-1716, Dec.
1975.
5..CONTROL AND ESTIMATION OF BIOLOGICAL SIGNALS
(ECG) USING ADAPTIVE SYSTEM, SUSHANTA MAHANTY,
International Journal of Electrical and Electronics
1000 1200 1400 1600 1800 2000 2200 2400 2600 2800 3000
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
0.8
1
Time, samples
Original Signal
1000 1200 1400 1600 1800 2000 2200 2400 2600 2800 3000
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
0.8
1
Time, samples
Denoised Signal
1000 1200 1400 1600 1800 2000 2200 2400 2600 2800 3000
-1
0
1
Time, samples
Original Signal
1000 1200 1400 1600 1800 2000 2200 2400 2600 2800 3000
-1
0
1
Time, samples
Original Signal + noise
1000 1200 1400 1600 1800 2000 2200 2400 2600 2800 3000
-1
0
1
Time, samples
Denoised Signal + noise
1000 1200 1400 1600 1800 2000 2200 2400 2600 2800 3000
-1
0
1
Time, samples
Denoised Signal
1000 1200 1400 1600 1800 2000 2200 2400 2600 2800 3000
-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
0.8
1
1.2
Time, samples
Original Signal + noise

Engineering (IJEEE) ISSN (PRINT): 2231 – 5284, Vol-2, Iss-1,
2012.
6. Paulo S.R. Diniz, Adaptive Filtering Algorithm and Practical
Implementation,
KluwerAcademicPublishers,1997
7. N. Kalouptsidis. Adaptive System Identification and Signal
Processing Algorithm (University of Athens) and S. Theodoridis
(UniversityofPatras)PrenticeHallInc.,1993.
8.L. Guo, L. Ljung, G.J. Wang, Necessary and Sufficient Condition for
Stability of LMS, Department of Electrical Engineering, Linkoping
University,Sweden,1995
9. Simon Haykin, Adaptive Filter Theory, Fourth Edition, Prentice Hall,
Inc2002
10. International Journal of Control, Automation, and Systems,
vol. 3, no. 1, March 2005
11.[6] A. Buades, B. Coll, and J. M. Morel, “A review of
image denoising algorithms, with a new one,” Multiscale
Modeling and Simulation, vol. 4, no. 2, pp. 490-530
12. Ieee Nonlocal Means Denoising of ECG Signals, Brian H.
Tracey, IEEE TRANSACTIONS ON BIOMEDICAL
ENGINEERING, VOL. 59, NO. 9, SEPTEMBER 2012
13.http://123seminarsonly.com/Seminar-Reports/029/42184313
-10-1-1-100-81.pdf
14.http://en.wikipedia.org/wiki/Non-local_means_filter

IRJET- Design Simulation and Analysis of Efficient De-Noising of ECG Signals with Adaptive Filtering Algorithms & Patch based Method

More Related Content

What's hot (18)

Similar to IRJET- Design Simulation and Analysis of Efficient De-Noising of ECG Signals with Adaptive Filtering Algorithms & Patch based Method (20)

More from IRJET Journal (20)

Recently uploaded (20)

IRJET- Design Simulation and Analysis of Efficient De-Noising of ECG Signals with Adaptive Filtering Algorithms & Patch based Method