Вы находитесь на странице: 1из 5

2017 2nd IEEE International Conference On Recent Trends In Electronics Information & Communication Technology, May 19-20, 2017,

India

Comparative Study of ECG Feature Extraction


Methods
Akanksha Agrawal, Dhanashri H.Gawali,
NBN Sinhgad School of Engineering, NBN Sinhgad School of Engineering,
Savitribai Phule Pune University, Pune,India, Savitribai Phule Pune University, Pune,India,
akanksha.893@gmail.com dhanashree.gawali@gmail.com

Abstract—Electrocardiogram (ECG) signals are used for the for the production of electrical signals that spreads through
detection of heart related diseases from more than a century both the atria and makes them contract. The AV node, which
now. Electrocardiogram (ECG) is an electrical activity of the is located on the opposite side of the right atria, serves as an
heart that is recorded by placing many electrodes on the body. electrical gateway to the ventricles. The signal spreads both
Hence, it is important to identify an appropriate method for the atria causing the muscle cells to depolarize and contracts
extracting features in ECG signal. This paper provides insight to induce a phase known as atrial systole on the ECG this
into different efficient methods for this purpose. The atrial depolarization is represented by P wave. The atrial
prominent methods presented in this paper are Pan-Tompkins, systole starts after 100 millisecond's after the P wave begins.
Hilbert transform, Histogram approach, Wavelet transform,
The periods of conduction that follows the atrial systole and
Auto-regression(AR), Independent Component Analysis(ICA),
follows the contraction of the ventricles is depicted on the
Linear prediction(LP), Adaptive threshold. A comparative
analysis is done for these feature extraction methods based on ECG by the PR segment. A flat line following the P-wave
various performance parameters like sensitivity, predictivity when the signal leaves the atria, it entries the ventricles via
and accuracy. The above methods are reviewed on lines of the AV node located in the inter-atrial septum. It enters the
these parameters, which can be used to identify suitable bundle of His and spreads through the bundle branches and
method for ECG feature extraction. the large diameter Purkinje's fibers along the ventricle
values. As the signal spreads through the ventricles, the
Keywords—ECG; AR; Wavelet; Hilbert; Histogram; Adaptive contractile fiber depolarizes and contract very rapidly
threshold; ICA; LP; Pan-Tompkins; inducing ventricular systole. In ECG, the QRS complex
represents these rapid ventricular depolarizations. Atrial
I. INTRODUCTION repolarization also occurs at this time but any; atrial activity
The heart is a vital organ of the human body. It pumps is hidden on the ECG by the QRS complex. Finally, as the
oxygenated blood into different parts of the body with the signal passes out of the ventricles, the ventricular walls start
help of arteries. The proper functioning of the heart is very to relax and recover a state described at ventricular diastole.
important for the whole body to function smoothly. The dome shape T wave on the ECG marks this ventricular
However, with time, a number of people are now suffering repolarization. On the ECG, the ST segment depicts the
from different types 6% of men suffer from heart-related period when the ventricles are depolarized.
problems in the year 2016 by heart attacks and strokes
mainly. Hence, to measure the performance rate of the heart,
ECG signal is used extensively. It is the most reliable, safe
and accurate method to estimate the working of the heart.
The frequency of the ECG signal lies in the range of 0.05 to
100 Hz. and it has a dynamic range is of 1-10mV.The
different clinical values of ECG signal and its associated
peaks are mentioned in Table 1[1].
Table I: Clinical values of ECG signal
Sr.No Feature Amplitude(mV) Duration(ms)
1 P wave 0.25 60-80
2 PR interval - 120-200
3 R wave 1.60 80-120
4 Q wave 25% of R wave 90-100
5 QT interval - 360-440
6 ST segment - 100-120
7 T wave 5-8 120-160
8 RR interval - (0.6-1.2)sec
9 PR segment - 50-120

II. CARDIOVASCULAR SYSTEM OF HEART


Each beat begins in the right atrium with an action Fig 1.(a) Structure of heart (b) ECG peaks (c) Peaks corresponding
potential signal from SA node. The SA node is responsible with the functioning of heart [2]

978-1-5090-3704-9/17/$31.00 © 2017 IEEE


246
2017 2nd IEEE International Conference On Recent Trends In Electronics Information & Communication Technology, May 19-20, 2017, India

The QT interval represents the time it takes for both x Shifting


depolarization and repolarisation of the ventricles to occur. x Linearity.
The sequence of the events just mentioned and the Hilbert transform has a few advantages over the other
associated ECG trace repeats with every heartbeat. ECG is methods of detecting QRS. It can minimize the unwanted
an amalgamation of many action potential that constitutes effect of the large peaks T and P by creating envelope
the electrical activity of the heart. around the R peak. Also, it is an odd function [5] impelling
In this paper, a few methods that are presented to extract that Hilbert transformed signal will cross the X axis
the features of the PQRST peaks of the ECG signal. These whenever the signal is disturbed by any means. When the
methods are used in time domain, frequency domain and ECG signal is transformed with the help of Hilbert
inboth time –frequency domain. The structure of the heart transform, the waveform then obtained is shown in Fig 2.
with its labeling and how each function of the heart
corresponds to the different peaks in the ECG signal is
depicted in fig.1
III. METHODOLOGY
A. Time Domain

1) Pan Tompkins Algorithm


The Pan-Tompkins algorithm is most widely used and
highly acknowledged algorithm. It is a simple yet powerful
algorithm for QRS detection. It is a real-time QRS detection
algorithm based on slope, amplitude, and width of the QRS Fig. 2. Waveform after Hilbert transforms [7]
complex[3]. It is quiet robust and efficient in nature. The
algorithm uses a special digital band-pass filter. It is used to
reduce the noises and false detection caused by faulty 3) Linear Prediction
signals that contaminate the ECG signal. In this case, the
Linear prediction method is used in audio signal and
sensitivity increase as this method allows the use of low
speech processing. It is used to reproduce the original signal
thresholds. The algorithm is simple because, it automatically
accurately [8]. FIR filter of appropriate order is the basic
adapts and adjusts the threshold value for next peak
block of the system. The forward coefficient of the signal
detection from the previous smallest peak.
and the solution viz the difference equation presented which
Boiling it down to, the steps of the procedure are stated expresses a linear combination of each sample with the
below[4]: previous sample has to be determined [9]. It is accurate in
nature also; it has faster computation than others. The actual
x Band-pass filtering ECG sequence P(i) is approximated by another sequence
x Applying a derivative operator Ṕ(i) which is further determined by a distinctive set of
x Square and integrate signal predictor coefficients and the past S samples as follows[8]
x Fiducial mark determination by thresholding
x QRS wave detection Ṕ(i) = ∑ ( )∗ ( − )+ ( ) (3)
x Average R_R interval and rate limits adjustment
2) Hilbert Transform Where Ṕ (i) is the predicted signal value, p(i-k) is the
Bolton and Westphal proposed the Hilbert Transform previous observed value and a is the predictor coefficient
method. This algorithm and the Hilbert transform utilize the and e(i) is the residual error.
first differentiation of the ECG signal. It is used to identify
and detect the R-peaks in the ECG waveform. The Hilbert
transform develops an envelope of ECG signals where the
maximum peaks are obtained at the zero crossing
point.Given that u(t) is a real time function. The Hilbert
transform is [5]
ȗ(t)= Ƕ[u(t)] (1)


ȗ(t)= ∫ ∞
( ) (2) Fig 3. Linear prediction model [10]

Here, ‘Ƕ’ is the Hilbert transform coefficient. Hilbert 4) Histrogram Method


transform has a few properties that are stated below [6]:
A histogram is a display of statistical information that
x Scaling and time reversal uses rectangles to show the frequency of data items in
x Convolution successive numerical intervals of equal size[11]. This
x Orthogonal method is used as an estimator for the waves of ECG signal.
This approach gives the measure of the orientation the
247
2017 2nd IEEE International Conference On Recent Trends In Electronics Information & Communication Technology, May 19-20, 2017, India

object points in a few quantized directions. The histogram This algorithm is simple, accurate, real-time performance
can be divided into a horizontal and vertical histogram. Both and stability. It can also effectively solve the baseline drift
these directional histograms are used for determination of problems. When the ECG signal is exposed to adaptive
separate aspects of the ECG signal. This technique does not thresholding method, the waveform hence obtained is shown
employ the usage of any mathematical complexity. This is in Fig 5
combined with an adaptive threshold value to calculate the
maximum bin of the histogram for the detection of different
peaks. In this method, first the determination of R peaks of
the ECG signal by setting a threshold value and by
computing the maximum value within the marked
histogram. Subsequently, the T and the P peaks of the ECG
signal is marked and thereby, the whole ECG signal is
converted into an array of rectangular peaks corresponding
to different peaks of the ECG. Amongst those rectangular
peaks, the R peaks, the T wave, and the P peak are marked
for the extraction of features present in ECG signal.

Fig 5: Waveform after applying Adaptive Threshold [7]

C. Time-Frequency Domain

1) Wavelet Transform
Wavelet transform method is a mixture of two types,
continuous and discrete. Discrete wavelet is sort off called
wavelet filter banks, as it uses 2 filters, a Low pass Filter
(LPF) and a high pass filter(HPF) to decompose the signal
into different frequency scales[14].These functions of
mathematics are used for decomposing data into different
components of frequency[1]. Wavelet transform is suitable
for both high and low frequencies varying its wavelet band
depending on the range of frequency value. With the help of
Fig 4. Peak Detection using Histogram Approach [11] this technique, the transform divides the whole wave i.e. the
This proposed method is accurate in nature for the mother wavelet into different wavelet transform bases. This
detection of peaks purpose [11].The ECG signal is is displayed in Fig 6, where a wavelet is obtained from the
converted to the form of histogram and the P-R-T peaks are mother wavelet.
detected. This is shown in Fig 4

B. Frequency Domain

1) Adaptive -Threshold method


One of the significant methods of detection of the R wave Fig 6: Wavelet Signal [16]
peak and others is the usage of adaptive threshold. The main
idea behind this algorithm is that the value of threshold
changes with time and it gets updated time to time in An important role in analysis and synthesis is of the
correspondence to the changes in the QRS morphology and proper selection of the wavelet basis function. The wavelet
also the levels of noise and artifacts present in the ECG transform can be represented by [15]
signal[12]. Hence, to detect the peaks and the baseline drift
and power interference, this method is widely used, as the
method is much simpler and more effective. It can also do W(a,b) = ∫ ( ) ( ) (5)
,
the real-time processing. It is important to fix an initial
threshold value to detect the peaks. Nevertheless, fixing of
threshold value will not be able to adapt the changing

situations and will lead to false positivity. As stated above, , = (6)

this method operates on real-time processing; hence, the
threshold value would automatically be updated to the amp-
thr0 value and the value in the new crest value. The formula Where * represents the complex conjugation and the ,
for the same is [13] indicates the mother wavelet of the signal.

Amp-thr1= 0.6*amp-thr0+0.3*V (4)


2) Independent Component Analysis(ICA)

248
2017 2nd IEEE International Conference On Recent Trends In Electronics Information & Communication Technology, May 19-20, 2017, India

Independent component analysis (ICA) is an application B. Predictivity (+P), it gives the balance between the count
method used in signal processing. It has its main usage in of events that are correctly detected, TP and the total
separation of multivariate signal into its subcomponents count of events that the analyzer detects, and its
underlying it, where these signals are mutually independent formulated as
signals in nature[17]. ICA is used for the removal of hidden
layers of noise in the ECG signal, but it needs to be assumed
that these hidden signals are non-Gaussian in nature. +P(%) = % (11)
Assume that x1,x2,x3….xn be n linear mixtures of n
independent signals s1,s2,s3…sn[11] .In vector-matrix Here FP (false positive) gives the count of events that
notations, can be written as follows[18] detection contains flaws.
C. Accuracy (%) is calculated by percentage of detected
X= A*S (7) peaks among total number of peaks.

Where X denotes a column vector having random


variables x1,x2,x3..xn be n as its elements.S denotes a Accuracy (A) = 100% (12)
column vector having random variables s1,s2,s3…sk as its
elements and A denotes the mixing mixture with elements aij
[11].
There is different ICA algorithms are Jade ICA, Fast These parameters are the factors that help us evaluate
ICA, constrained ICA, and Kernel ICA and extended ICA which feature extraction method is better than the others are.
[17].

3) Auto Regression(AR) V. PERFORMANCE OF METHODS


Auto-regression method is linear in nature. AR is a two-
lead ECG model to improve the cardiac problems, extract With the help of above mentioned parameters, deduction
the needed features from the acquired ECG signal, and of which feature extraction method has better performance
classify them related to its disease.[19] The AR coefficients role in comparison to the other proposed methods can be
computed from the ECG signal are classified using a found out. Each method has been evaluated for a set of data
generalized linear model and a multi-layer feed forward base from MIT-BIH Arrhythmia and its parameters are
neural network. The AR model of order M is given by [19]: calculated. In table II, each method and its calculated
parametric values are mentioned.
1( ) = ∑ 1( − ) 1( ) + 1( ) (8)
2( ) = ∑ 2( − ) 2( ) + 2( ) (9)
Table II: Performance analysis of different Feature Extraction Methods
Where; Method of Applicatio Sensitivi Predictiv Accura
HR1 (n) and HR2 (n) denote ECG time series feature n Domain -ty i-ty -cy
[19],e1(n)represents values of unknown zero means extraction
[19],e2(n) represents incorrect random variables [19],a1 (j) Auto Time- 97.28 97.3 96.6
and a2 (j) are the AR model coefficients[19]. regression[19] Frequency
AR analyses has a lead over the other methods in the
Wavelet Time- 99.89 99.86 99.75
matters of its simplicity and also how it is well suited for the transform[1,14,1 frequency
real-time applications at hospitals especially in the ICU . AR 5, 16]
modeling has greater accuracy however, as AR modeling is
Linear Time 96.9 80.4 93.2
linear in nature. prediction[8,9,10
]
IV. PERFORMANCE PARAMETERS
ICA[17,18,19] Time- 97.8 99 90.13
All the above-discussed methods are compared with one frequency
another based on three statistical parameters to evaluate.
Hilbert Time 99.93 99.91 99.83
These parameters will determine which methods give more transform[5,6]
accurate, more precise and more determined results while it
detects the peaks of the ECG signal .They are[11]: Adaptive Frequency 99.55 99.28 99.70
threshold[12,13,7
A. Sensitivity(Se), it is given by the ratio by count of events ]
that are correctly detected ,TP(true positive) to the total Histogram Time 99.86 99.76 99.8
count of events is formulated by Approach[11]

Pan Tompkins Time 99.83 98.65 86


[3,4]
Se(%)= % (10)

Here FN(false negative) is the count of missed events. Our methods are functioning all in 3 different domains-
time, frequency and time-frequency domain both.

249
2017 2nd IEEE International Conference On Recent Trends In Electronics Information & Communication Technology, May 19-20, 2017, India

In the time domain, the Hilbert and histogram methods Conference on Control, Instrumentation, Energy & Communication
giving out the optimum required result in terms of all the 3 (CIEC), Kolkata, 2016, pp. 168-172
[12] K. Akazawa, K. Motoda, A. Sasamori, T. Ishizawa and E. Harasawa,
parametric values. Hilbert transform requires a tedious "Adaptive threshold QRS detection algorithm for ambulatory
amount of mathematical computation work and its method is ECG," [1991] Proceedings Computers in Cardiology, Venice, 1991,
more complex than histogram method. The histogram pp. 445-448
method however, requires less computational time & the [13] G. Chen, X. Wang and W. Wan, "An ECG R-wave detection
method is less mathematically complex than another algorithm based on adaptive threshold," 2015 International
Conference on Smart and Sustainable City and Big Data (ICSSC),
method. Shanghai, 2015, pp. 145-149.
In the mixed domain, the wavelet transform method [14] A. Mukherjee and K. K. Ghosh, "An efficient wavelet analysis for
ECG signal processing," 2012 International Conference on
performs fairly well above the other stated methods in Informatics, Electronics & Vision (ICIEV), Dhaka, 2012, pp. 411-415.
relation to the calculated parametric terms. The advantage of [15] C. L. P. Lim, W. L. Woo and S. S. Dlay, "Enhanced wavelet
this method is it can be used even for very small waveforms. transformation for feature extraction in highly variated ECG
This helps to expand its horizon of workspace. signal," 2nd IET International Conference on Intelligent Signal
Processing 2015 (ISP), London, 2015, pp. 1-6.
VI. CONCLUSION [16] M. J. d. Silva, "Characterization of QRS Complex in ECG Signals
Applying Wavelet Transform," 2015 International Conference on
AR, LP, Hilbert transform, Histogram, Adaptive Mechatronics, Electronics and Automotive Engineering (ICMEAE),
threshold, Pan-Tompkins, wavelet transform, ICA are Prague, 2015, pp. 86-89
discussed throughout this paper for the extraction of [17] M. Phegade and P. Mukherji, "ICA based ECG signal
different features from the ECG signal. Accuracy is a very denoising," 2013 International Conference on Advances in
Computing, Communications and Informatics (ICACCI), Mysore,
importance aspect of the ECG feature extraction system. 2013, pp. 1675-1680.
The sensitivity and predictivity is also compared for these [18] H. P. Kasturiwale and C. N. Deshmukh, "Quality Assessment of ICA
methods. It is found that amongst these methods, Hilbert and for ECG Signal Analysis," 2009 Second International Conference on
histogram methods give the best results in terms of accuracy Emerging Trends in Engineering & Technology, Nagpur, 2009, pp.
to detect the PQRST peaks of the ECG signal. The 73-75.
histogram method has an upper hand on Hilbert transform in [19] F. M. Vaneghi, M. Oladazimi, F. Shiman, A. Kordi, M. J. Safari and
F. Ibrahim, "A Comparative Approach to ECG Feature Extraction
terms of mathematical complexity and it can be used for Methods," 2012 Third International Conference on Intelligent
online analysis of it. These 2 methods are apt for detecting Systems Modelling and Simulation, Kota Kinabalu, 2012, pp. 252-
the peaks correctly and not giving any false determination of 256.
any peaks. These approaches provide proper modeling of the
ECG feature extraction methods.
REFERENCES
[1] G.S. Kumari, S. K. Kumar, J. Anusha and M. P. Rao,
"Electrocardiographic signal analysis using wavelet transforms," 2015
International Conference on Electrical, Electronics,Signals,
Communication and Optimization (EESCO), Visakhapatnam, 2015,
pp. 1-6.
[2] R. Gupta, M. Mitra and J.N. Bera, ECG acquisition and automated
remote processing, Springer, New Delhi, India, 2014, pp. 1-6.
[3] J. Pan and W. J. Tompkins, "A Real-Time QRS Detection
Algorithm," in IEEE Transactions on Biomedical Engineering, vol.
BME-32, no. 3, pp. 230-236, March 1985.
[4] I. L. b. Ahmad, M. b. Mohamed and N. A. b. Ab Ghani,
"Development of a concept demonstrator for QRS complex detection
using combined algorithms," 2012 IEEE-EMBS Conference on
Biomedical Engineering and Sciences, Langkawi, 2012, pp. 689-693
[5] Hahn, Stefan L., Hilbert Transforms, in Transforms and Applications
Handbook, Ed. Alexander D.Poularikas, CRC Press Inc., Boca Raton,
FL, 1996.
[6] P. Valluraiah and B. Biswal, "ECG signal analysis using Hilbert
transform," 2015 IEEE Power, Communication and Information
Technology Conference (PCITC), Bhubaneswar, 2015, pp. 465-469.
[7] A. Farahabadi, E. Farahabadi, H. Rabbani and M. P. Mahjoub,
"Detection of QRS complex in electrocardiogram signal based on a
combination of hilbert transform, wavelet transform and adaptive
thresholding," Proceedings of 2012 IEEE-EMBS International
Conference on Biomedical and Health Informatics, Hong Kong, 2012,
pp. 170-173
[8] E. K. Zaghouani, A. Jemai, A. Benzina and R. Attia, "ELPA: A new
key agreement scheme based on linear prediction of ECG features for
WBAN," 2015 23rd European Signal Processing Conference
(EUSIPCO), Nice, 2015, pp. 81-85.
[9] P. Sarma, S. R. Nirmala and K. K. Sarma, "Classification of ECG
using some novel features," 2013 1st International Conference on
Emerging Trends and Applications in Computer Science, Shillong,
2013, pp. 187-191.
[10] C. C. Lin, W. Hu and C. M. Yang, "Automatic classification of
arrhythmic heartbeats using the linear prediction model," Computing
in Cardiology 2013, Zaragoza, 2013, pp. 971-974.
[11] B. Halder, S. Mitra and M. Mitra, "Detection and identification of
ECG waves by histogram approach," 2016 2nd International
250

Вам также может понравиться