Abstract –

Electrocardiogram (ECG) is a method of

measuring the electrical activities of heart. Every portion of ECG is very

essential for the diagnosis of different cardiac problems. But the amplitude

and duration of ECG signal is usually corrupted by different noises. Removing motion artifacts

from an electrocardiogram (ECG) is one of the important issues to be considered

during real-time heart rate measurements in health care. It is essential

to reduce these disturbances and improve the accuracy as well as reliability.

The noises that commonly disturb the ECG signals are Random noise, Gaussian

white noise, Power line interference, Baseline wander and Electromyography

(EMG) noise .These noises can be classified according to their frequency

content. The noise signals have been generated and added to the ECG signal

taken from MIT-BIH arrhythmia database. In this paper we have done a broader

study for denoising every types of noise involved with real ECG signal. One

adaptive filter, least-mean-square (LMS)

is applied to remove the noises. PSNR and MSE performance parameter

are estimated.

Keywords – ECG signal,

artifacts, adaptive filter, LMS algorithm

?.

INTRODUCTION

ECG

is generated by the heart muscle and measured on the skin surface of the body

This signal is ranging from 10microvolts to 5 mille-volts, frequency from 0.05

Hz to 100 Hz.ECG is helpful to detect

changes in cardiac muscles like myocardial infarction, conduction defects and

arrhythmia . When the electrical abnormalities of the heart occur, the heart

cannot pump and supply enough blood to the body and brain. As ECG is a

graphical recording of electrical impulses generated by heart, it is needed to

be done when chest pain occurred such as heart attack, shortness of breath,

faster heartbeats, high blood pressure, high cholesterol and to check the

heart’s electrical activity. It recognises the variability’s of heart activity,

so it is very important to get the ECG signal free from noise.

Basically ECG signal is characterized by five peak points – P, Q, R, S, and T. The waveform which

is repetitive and have various bumps and

parts of the waveform are designated as the P-wave, QRS complex and T-wave,

PR-segment, ST-segment, PR-interval and QT-interval as given in Figure 1. The

origins of these waves are:

i.

P wave: sequential activation (depolarization) of the right and left atria

ii.

QRS complex: right and left ventricular depolarization

iii.

T wave: ventricular repolarisation

iv.

U wave: repolarisation of the papillary muscles, rarely seen.

Fig.

1:ECG waveform

II. NOISES IN ECG SIGNAL

ECG

Signals generated from human body are often very weak so as to be easily

covered by background noise. The noise in the ECG signals occur due to various

reasons like electromagnetic interference due to ubiquitous supply lines and

plugs, movement of patient, signals generated by other organs and impedance mismatching

between electrodes. Hence the ECG signals can be corrupted by various types of noises

such as Power line interference, Electrode contact noise, Motion artifact, Muscle

contraction, Base line drift, Instrumental noise generated by electronic

devices . Power line interference noise is electromagnetic field from the

powerline which causes 50/60Hz sinusoidal interference. This noise causes

problem in interpreting low amplitude waveform like ECG. Various methods have

been employed for the removal of artifacts from ECG signals. Adaptive filtering

is one of the efficient method in the removal of noises in the ECG signals.

III.GENERATION

OF NOISES

The low frequency noise (base line

wander, i.e. electrode contact noise and motion artifact) has frequency less

than 1Hz, high frequency noise (EMG noise) whose frequency is more than 100Hz

and power line interference of frequency 50 Hz or 60Hz (depending on the

supply) can be generated as follows. These noises are generated in MATLAB based

on their frequency content, which is then added with the ECG signal to get the

noisy ECG.

A .Generation of

Random noise

High

frequency random noise signal is generated. The generated high frequency noise

is shown in Figure 2.

Fig. 2: Random

noise signal

B.

Generation of Gaussian White noise

Generated Gaussian white noise is

shown in Figure 3.

Fig. 3: Gaussian

white noise signal

C. Generation of

Baseline wander noise

Generated the baseline drift by which is shown in Figure 4.

Fig. 4: Baseline

wander noise signal

D. Generation of

Power line interference noise

We have

considered the 50 Hz power supply. So, we have taken a sine wave of 50 Hz

amplitude to represent the power line interference. The resulted power line

interference is shown in Figure 5.

Fig.

5: Powerline interference noise signal

IV. ADAPTIVE

FILTERING

ECG

signal has been a major diagnostic tool for the cardiologists and ECG signal

provides almost all the information about electrical activity of the heart. So

care should be taken while doing the ECG filtering, such that the desired

information is not distorted or altered in any way. The original ECG signal is taken from the MIT-BIH arrhythmia database. The different types of noise signal are generated by using MATLAB. The noise signal is then added with the real ECG signal. To remove the different types of noises, the noisy ECG signal is then pass through adaptive filter

algorithms (e.g., LMS). However, the basic block diagram of adaptive filtering is shown in figure 6.

Fig.

6: Adaptive filter

In

this paper, we have designed a multistage filter for cancellation of these

artifacts. The filter consists of two stages(1st stage and 2nd stage) as shown

in Fig 7.

Fig.

7: LMS two stage filtering

Least mean squares (LMS) algorithms are a class of

adaptive filter used to mimic a desired filter by finding the filter

coefficients that relate to producing the least mean squares of the error

signal (difference between the desired and the actual signal). It is a

stochastic gradient descent method in that the filter is only adapted based on

the error at the current time. The LMS Algorithm consists of two basic

processes

1. Filtering process -Calculate the output of FIR

filter by convolving input and taps. Calculate estimation error by comparing

the output to desired signal.

2.

Adaptation process:-Adjust tap weights based on the estimation error.

Consider

a length L LMS based adaptive filter, that takes an input sequence x(n)

and updates the weights as

w(n + 1) = w(n) + ? x(n) e(n)

where

w(n) = w0(n) w1(n)….wL?1(n)t is

the tap weight vector at the nth

index,

x(n) = x(n) x(n?1)….x(n?L+1)t

e(n) = d(n)?wt(n) x(n)

with

d(n) being the so-called desired response available during

initial training period and ? denoting so-called step-size parameter.

In

order to remove the noise from the ECG signal, the ECG signal s1(n) with

additive noise p1(n) is

applied as the desired response d(n) for the adaptive filter. If

the noise signal p2(n),

possibly recorded from another generator of noise that is correlated in some

way with p1(n) is

applied at the input of the filter, i.e.,

x(n) = p2(n)

The filter error becomes,

e(n) = s1(n) + p1(n) ? y(n)

The filter output y(n) is

given by,

y(n) = wt(n)x(n)

Since the signal and noise are

uncorrelated, the mean-squared error (MSE) is,

Ee2(n) = E{s1(n) ? y(n)2} + Ep21(n)

Minimizing the

MSE results in a filter output that is the best least-squares estimate of the

signal s1(n).

V.

SIMULATION RESULTS

To

show that LMS algorithm is really effective in clinical situations, the method

has been validated using several ECG recordings with a wide variety of wave

morphologies from MIT-BIH arrhythmia database. The arrhythmia data base

consists of 48 half hour sets of two channel ambulatory ECG recordings, which

were obtained from 47 subjects including 25 men aged 32-89 years and women aged

23-89 years. The recordings were digitized at 360 samples per second per

channel with 11-bit resolution over a 10mV range. The generated noises in

MATLAB based on their frequency content, which is then added with the ECG

signal to get the noisy ECG. The output of the first stage is coupled to the

second stage where the noise free ecg signal can be obtained. We have considered

four different types of noises to corrupt our signal namely Power line Interference,

Baseline Drift, guassian noise and white noise. Results of LMS along with mean

square error (MSE) and peak signal –to-noise ratio are also shown. Six ECG

signal were obtained from this database to validate the results. The following

figures are the ECG signals corrupted by various noises.

Fig.

8: Pure ECG signal

Fig. 9: Noisy

ecg signal(random noise)

Fig.

10: Noisy ecg signal (Gaussian white noise)

Fig. 11: Noisy ecg

signal(baseline wander noise)

Fig.

12: Noisy ecg signal (powerline interference noise)

The two stage

adaptive filtering of the various noises is obtained as follows.

Fig.

13: LMS Result(of random noise)

Fig.

14: LMS Result(of Gaussian white noise)

Fig.

15: LMS Result(of Baseline wander noise

Fig.

16: LMS Result(of Powerline interfernce noise)

TABLE I. VALUES OF PERFORMANCE PARAMETERS

OF TWO STAGE ADAPTIVE FILTER FOR DIFFERENT TYPES OF NOISE

NOISES

RECONSTRUCTED SIGNALS

MSE

PSNR

1ST

STAGE

2ND STAGE

1ST

STAGE

STAGE

S

2ND STAGE

Random

0.0154

6.09e-10

16.1990

90.2149

Gaussian white

0.0148

1.33e-08

16.3748

76.7935

Baseline wander

0.0247

1.35e-08

14.1409

76.7579

Powerline interference

0.0038

9.06e-07

22.2214

58.4863

VI. CONCLUSION

In this paper, the problem of noise

cancellation from ECG signal using adaptive filters are proposed and tested on

real signals with different artifacts obtained from the MIT-BIH database. For

this, the input and the desired response signals are properly chosen in such a

way that the filter output is the best least squared estimate of the original

ECG signal.LMS algorithm work effectively in removing the noises from the ECG

signal.

REFERENCES

1R.

Panda and U. C. Pati, “Removal of artifacts from electrocardiogram using digital filter,” in

Electrical, Electronics and Computer Science (SCEECS), 2012 IEEE Students’

Conference on IEEE, 2012, pp. 1–4

2S.

Mahmoodabadi, A. Ahmadian, M. Abolhasani, M. Eslami, and J. Bidgoli, “Ecg

feature extraction based on multiresolution wavelet transform,” in Engineering

in Medicine and Biology Society, 2005.IEEE-EMBS 2005. 27th Annual International

Conference of the. IEEE,2006, pp. 3902–3905.

3R.

Sameni, M. B. Shamsollahi, C. Jutten, and G. D. Clifford, “A non linear

bayesian filtering framework for ecg denoising,”IEEE Transactions on Biomedical

Engineering, vol. 54, no. 12, pp. 2172–2185, 2007

4C.H.

Chang, H.J. Ko, and K.-M. Chang, “Cancellation of Artifacts in ECG Signals

using Signbased Normalized Adaptive Filtering Technique,” in Biomedical Engineering

and Informatics (BMEI), 2010 3rd International Conference on, vol. 2. IEEE,

2010, pp. 787–790.

5 S. Singh, “Performance

evaluation of different adaptive filters for ecg signal processing 1,” 2010.

6

M. G. Khan, Rapid ECG interpretation. Springer Science & Business Media,

2008.

7

M. Z. U. Rahman, R. A. Shaik, and D. R. K. Reddy, “Efficient sign based

normalized adaptive filtering techniques for cancelation of artifacts in ecg

signals: Application to wireless biotelemetry,” Signal Processing, vol. 91, no.

2, pp. 225–239, 2011.

8

P. E. McSharry, G. D. Clifford, L. Tarassenko, and L. A. Smith, “A dynamical

model for generating synthetic electrocardiogram signals,” IEEE Transactions on

Biomedical Engineering, vol. 50, no. 3, pp. 289– 294, 2003.

9

V. Viknesh and P. R. Prashanth, “Matlab implementation of ecg signal

processing,” IOSR Journal of VLSI and Signal Processing, vol. 3, no. 1, 2013.

10

M. Alfaouri and K. Daqrouq, “Ecg signal denoising by wavelet trans-form

thresholding,” American Journal of applied sciences, vol. 5, no. 3, pp.

276–281, 2008.

11

S. Khan, S. M. Anwar, W. Abbas, and R. Qureshi, “A novel adaptive algorithm for

removal of power line interference from ecg signal,” Science International,

vol. 28, no. 1, 2016.