Upload
others
View
30
Download
0
Embed Size (px)
Citation preview
Journal of Engineering, Computers & Applied Sciences (JEC&AS) ISSN No: 2319-5606 Volume 2, No.8, August 2013 _________________________________________________________________________________
www.borjournals.com Blue Ocean Research Journals 55
Wavelet Based EMG Artifact Removal From ECG Signal
Josy Joy, Department of Electronics and Instrumentation Engineering, Karunya University, Coimbatore,
Tamilnadu, INDIA
P.Manimegalai, Department of Electronics and Instrumentation Engineering, Karunya University, Coimbatore,
Tamilnadu, INDIA
ABSTRACT Electrocardiogram recordings (ECG) are obtained from the heart. Some sections of the recorded ECG may be
corrupted by electromyography (EMG) noise from the muscle. In real situations, exercise test ECG recordings and
long term recordings, are often corrupted by muscle artifacts. These EMG noise needs to be filtered before data
processing. In this paper, wavelet transform is applied to remove the EMG noise from ECG signal. In this work, an
improved thresholding is proposed for removing EMG noise in ECG signal. The proposed method selects the
best suitable wavelet function based on DWT at the decomposition level of 5, using SNR. The advantage of the
improved thresholding de-noising method is that it retains both the geometrical characteristics of the original
ECG signal and variations in the amplitudes of various ECG waveforms effectively.
Keywords- ECG, EMG, DWT, THRESHOLDING, REALTIME, LabVIEW, MATLAB
Introduction Electrocardiogram (ECG) signal, the electrical
interpretation of the cardiac muscle activity, is very
easy to interfere with different noises while gathering
and recording. The most troublesome noise sources
are the Electromyogram (EMG) signal. Such noises
are difficult to removing typical filtering procedures.
The EMG, a high frequency component, is due to the
random contraction of muscles, while the abrupt
transients are due to sudden movement of the body.
Furthermore, the non-stationarybehaviour of the
ECG signal, that becomes severe in the cardiac
anomaly case, incites researchers to analyze the ECG
signal. Wavelet thresholding de-noising method
based on discrete wavelet transform (DWT) proposed
by Donoho et al. is often used in de-noising of ECG
signal [4, 5]. In 1999, Agnate used it in de-noising of
ECG signal.
In real situations, exercise test ECG recordings
and long term recording are often corrupted by
muscle artifacts due to restlessness of patients. In
such cases, it is not possible to ensure relaxed
conditions for the patient and muscular activity is
reflected.
Automatic interpretation, which is strongly
dependent on accurate detection of characteristics
ECG points and waves and measurement of signal
parameters, becomes an extremely difficult and often
virtually impossible task. EMG artifacts so obtained
from the same electrodes as the ECG are difficult to
remove, due to considerable overlapping of the
frequency spectra of these two types of signals.
EMG artifacts in ECG are quite common with
uncontrollable tremor, in disabled persons having to
exert effort in maintaining a position of their
extremities or a body posture, in children .These
EMG noise needs to be filters before data processing.
Adequate ECG denoising algorithms and
procedures should
a) Improve the signal to noise ratio (SNR) to obtain
clean and readily observable recordings, allowing the
subsequent use of straightforward approaches for
correct automatic detection of its specific waves and
complexes.
b) Preserve the original shape of the signal and
especially the amplitudes of sharp Q, R and S peaks,
without introducing distortions in the low- amplitude
ST-segment and P- and T-waves.
Wavelet thresholding de-noising methods deals
with wavelet coefficients using a suitable chosen
threshold value in advance. The wavelet coefficients
at different scales could be obtained by taking
DWT of the noisy signal. Normally, those wavelet
coefficients with smaller magnitudes than the
preset threshold are caused by the noise and are
replaced by zero, and the others with larger
magnitudes than the preset threshold are caused by
original signal mainly and kept (hard-thresholding
case) or shrunk (the soft-thresholding case). Then the
denoised signal could be reconstructed from the
resulting wavelet coefficients. These methods are
simple and easy to be used in de-noising of ECG
signal. But hard-thresholding de-noising method may
Journal of Engineering, Computers & Applied Sciences (JEC&AS) ISSN No: 2319-5606 Volume 2, No.8, August 2013 _________________________________________________________________________________
www.borjournals.com Blue Ocean Research Journals 56
lead to the oscillation of the reconstructed ECG
signal and the soft-thresholding de-noising method
may reduce the amplitudes of ECG waveforms, and
especially reduce the amplitudes of the R waves. To
overcome the above said disadvantages an improved
thresholding de-noising method is proposed.
Methods ECG signal is easy to be contaminated by random
noises uncorrelated with the ECG signal, such as
EMG which can be approximated by a white
Gaussian noise source.
a) Decomposing of the noisy signals using
wavelet transform
Using the discrete wavelet transform by selecting
mother wavelet (db8), the noisy signal is
decomposed, at the decomposition level of 5. As a
result approximate coefficients and detail coefficients
are obtained.
b) Apply thresholding: It is done in order to obtain the estimated wavelet
coefficients For each level a threshold value is found,
and it is applied for the detailed coefficients.
1) Hard-Thresholding Method:
Where preset threshold is
Tj=σ√2log||dj|| ......... (a)
The σ can be estimated by the wavelet coefficients
with
σ = median d / 0.6745 .......... (b)
Here median denotes the median value of the
absolute values of wavelet detailed coefficients.
c) Reconstruction: Reconstructing the de-noised ECG signal x (n) by
inverse discrete wavelet transform (IDWT).
The same steps are to be followed for soft
thresholding, and improved thresholding de-
noising methods.
2) Soft-thresholding method:
3) Improved thresholding method:
Where β > 1 and ∈ R . Because the magnitudes of the
wavelet coefficients related to the Gauss white noise
decreases as the scale j increases, hence the threshold
value will be chosen as
Tj=σ√21log||dj||/log (j+1) ........ (c)
For each level, find the threshold value that gives the
minimum error between the detailed coefficients of
the noisy signal and those of original signal. The
improved wavelet thresholding denoising method has
the following characteristics:
It makes the reconstructed ECG signal remain the
characteristics of the original ECG signal and keep
the amplitudes of R waves effectively.
Equation (3) will be equivalent to hard-thresholding
when β → ∞ and will be equivalent to soft-
thresholding, when β →1. This shows that the
improved threshold denoising method can be adapted
to both hard- and soft-thresholding de-noising
methods. Therefore, the improved thresholding de-
noising method could be regarded as a compromise
between the hard- and soft-thresholding denoising
methods. So, that the improved thresholding
denosing method presented in this paper could
choose an appropriate β by trail –and-error to satisfy
the request of de-noising of the ECG signal from
EMG artifacts.
Evaluation Criteria We utilize output SNR value between the
constructed de-noised ECG signal and the original
ECG signal with DC offset rejected x(n) (the
reference ECG signal to evaluate our method.
Determination of SNR Criteria:
The output SNR is given by
Journal of Engineering, Computers & Applied Sciences (JEC&AS) ISSN No: 2319-5606 Volume 2, No.8, August 2013 _________________________________________________________________________________
www.borjournals.com Blue Ocean Research Journals 57
Results To validate the superiority of the proposed improved
thresholding de-noising method, ECG signal in MIT-
BIH database is intercepted to be the original ECG
signal. The length of the original signal ECG signal
(i.e., the number of the sample points) is N=1000.
Gauss white noise is added to the original ECG
signal, the noisy ECG signal. Noise reduction
Procedures were implemented in Mat lab 7.0.1.
Thresholding was performed by trial and error
method.
The proposed method is based on choosing threshold
value by finding output SNR of denoised signal and
original wavelet sub signal (coefficients). Therefore,
high quality denoised signal can be accomplished.
Our study establishes particular approach to fit ECG
signal that has nonstationary clinical information. To
preserve the distinct ECG waves and different low
pass frequency shapes, the method thresholds
detailed wavelet coefficients only.
Fig1: De-noising of ECG signal from EMG noise using threshold methods with DB8:
Fig.2 shows the SNR values for different threshold
values:
Conclusion The wavelet transforms allow processing of non-
stationary signals such as ECG signal. The proposed
method shows a new experimental threshold value
for each decomposition level of wavelet detailed
coefficients.
This improved thresholding de-noising method in this
paper is superior to other traditional thresholding
denoising methods in many aspects. It retains
geometrical characteristics. This can be done in Lab
View.
REFERENCES
[1] Filtering of electromyogram artifacts from the
electrocardiogram, Ivaylo I. Christov,center of
BME,bulgarian academy of sciences, accepted 6
Jan 2000.
[2] Suppression of electromyogram interference on
the electrocardiogram by transforms denoising.
0 500 1000 1500-8
-6
-4
contaminated ecg sinal with emg noise
0 200 400 600-0.05
0
0.05
1st level decomposition
0 100 200 300-0.5
0
0.5
2nd level decomposition
0 50 100 150-1
0
1
2
3rd level decomposition
0 20 40 60 80-2
0
2
4th level decomposition
0 10 20 30 40 50-5
0
5
5th level decomposition
0 500 1000 1500-2
-1
0
1
Detail D5
0 500 1000 1500-2
-1
0
1
ecg signal after thresholding
Mother wavelet
Thresholding method
Threshold value
SNR
Db8 Improved
thresholding .05 59.517
Db8 Improved
thresholding .07 59.533
Db8 Improved
thresholding .13 59.574
Db8 Improved
thresholding .16 59.753
Journal of Engineering, Computers & Applied Sciences (JEC&AS) ISSN No: 2319-5606 Volume 2, No.8, August 2013 _________________________________________________________________________________
www.borjournals.com Blue Ocean Research Journals 58
N.nikolaev, institute of information technologies,
med.biol, 2001.
[3] Denoising the electrocardiogram from
electromyogram artifacts by combined transform
and dynamic approximation method.
[4] Atanas Gotchev, centre of biomedical engg,
Bulgarian academy of sciences,0-703-7402-9,
2002 IEEE
[5] Application of ICA in removing artifacts from
the ECG
[6] Taigang He, department of engineering Sciences
University of oxford, 2006
[7] ECG signal interference removal using wavelet
based CSTD technique. R.shantha selva kumari,
from 0-7695-3050-8, 2007 IEEE computer
society
[8] A mathematical algorithm for ECG signals
denoising using window analysis. Hamid
sadabadi, biomed pap med fac univ palacky
Olomouc Czech republic.2007
[9] ECG signal denoising by wavelet transforms
thresholding. Mikhled alfauori ,American journal
of
applied sciences 5(3):276-281, 2008
[10] ECG denoising with adaptive bionic wavelet
transforms. O.Sayadhi, MSc BME, Sharif
University of technology, 2003
[11] A cardio electro-physiological model based
approach for filtering high frequency noise.
[12] ECG denoising by sparse wavelet shrinkage.