2024年5月4日发(作者：汤宛丝)
I.J. Information Engineering and Electronic Business, 2017, 6, 43-50
Published Online November 2017 in MECS (/)
DOI: 10.5815/ijieeb.2017.06.06
A Robust Approach for R-Peak Detection
Amana Yadav
Department of Electronics and Communication Engineering, Manav Rachna International University, India
Email: @
Dr. Naresh Grover
Department of Electronics and Communication Engineering, Manav Rachna International University, India
Email: ics@
Received: 23 June 2017; Accepted: 01 August 2017; Published: 08 November 2017
Abstract—Electrocardiogram (ECG) is very crucial and
important tool to detect the cardiac problems. For ECG
analysis, it is essential to measure ECG parameter
accurately. It is very critical in all types of ECG
application. The accurate R Peaks detection is starting
step in extracting ECG features which is necessary for the
other ECG performance stages. It is very essential to
detect these R-peaks accurately and efficiently to detect
heart diseases or anomalies which create primary source
of death in the universe. Hence automatic R-peaks
detection in a lengthy duration ECG signal is very
meaningful to diagnose the cardiac disorders. Here a
latest R-peak exposure algorithm depended on Shannon
energy envelope estimator and logic to find peaks has
been proposed which uses the simple threshold of
Shannon energy.
Fig.1. ECG signal made of a P wave, a QRS complex and a T wave [5]
Index Terms—ECG, R-peak detection, QRS complex, P-
During last few years, many systems have been
QRS-T waves, sampling frequency, Cardiac arrhythmia,
designed for QRS detection. Numerous QRS detection
MATLAB
algorithms based on the derivatives [6], filtering
techniques [7-10], wavelet transform [11-13],
mathematical morphology [14,15], empirical mode
I.
I
NTRODUCTION
decomposition (EMD) [16], geometrical matching [17],
The ECG is widely used to find out the cardiac
artificial neural networks [18] and hybrid approach [19],
problems, since it measures the contractile electrical
genetic algorithms [20], syntactic methods [21], Hilbert
activity of the heart developed over the cardiac rhythm
transform [22], Markov models [23] etc. reported in
through different electrodes placed at different places on
literature have been developed for R-peaks detection. The
the human body [1]. ECG is an exclusive signal which is
filtering techniques and decision rules based methods are
based on the physical composition especially chest of a
very efficient so best for all ECG analysis [9]. Many
particular person. Therefore ECG can be used to
approaches comprise of a preprocessing or extraction of
features and then a decision block [24]. To accentuate the
distinguish any individuals [2].
A normal ECG signals characterized by a P wave, a T
QRS complex various signal processing techniques are
wave and a QRS complex are shown in figure 1. R wave
applied in preprocessing stage which also suppresses the
has the highest amplitude in heart signal than the other
noises but they have some drawbacks. In [7], there is a
tradeoff between absence and false identification of peaks
portions.
The performance of R Peak detection systems depends
based on the choice of filter’s bandwidth and capacity of
on how accurate the R-peak detector is. Therefore it is
the moving-window integrator. The Empirical mode
very essential to identify the R-peaks in QRS complex
decomposition in [16] can defeated the selection problem
accurately and efficiently. Automatic R-peaks detection
of mother wavelet of Wavelet based QRS detector but
in a large duration ECG signal is very meaningful to
under noisy environments it is very difficult to select the
diagnose the cardiac disorders [3]. R Peaks are pointed
set of intrinsic mode functions (IMF). By introducing
towards the positive side ever and hence can’t have
further useful filtering technique and threshold alteration
method performance can be improved [16]. To study the
negative values [4].
ECG, a faithful approach to detect QRS which depends
Copyright © 2017 MECS I.J. Information Engineering and Electronic Business, 2017, 6, 43-50
44 A Robust Approach for R-Peak Detection
on the highest slope identification was introduced [25]. In
subsequent years many changes have been done on the
method of Pan Tompkins. Baseline correction is required
to get the complete details of QRS complex, so instead of
using adaptive thresholding a fixed thresholding is
required [26]. Instead of using the HPF and differentiator
in Pan Tompkins’ algorithm a novel Pan Tompkins
algorithm by adopting a Savitzky-Golay filter is
generated [27].
In this paper, an advanced preprocessor depending on
the Shannon energy envelope estimator and logic to
detect peaks using simple threshold of Shannon energy
has been proposed which is straightforward with better
accuracy and takes minimum computation time. The
threshold T can be evaluated by using universal threshold
suggested by Donoho.
The major purpose of proposed work is to check this
technique on ECG of human being who is not well; hence
ECG signal is a personal recognition of different patients.
Algorithm is implemented with the help of MATLAB.
Almost all the R peak recovering techniques utilizes the
standard Massachusetts Institute of Technology-Beth
Israel Hospital (MIT-BIH) record from for
the analysis of an ECG signal [8]. The arrhythmia
displays the irregularities of heart which are observed as
Tachycardia and Bradycardia can be simply obtained.
The proposed R-peak detector has 99.60% accuracy,
99.84% sensitivity and 99.75% positive predictivity. The
results prove that the given R Peak detector works better
than other conventional techniques for pathological or
noisy signals.
The paper is arranged in such a way:
In Section 2, the five-stage technique to obtain the R-
peak is explained precisely. This section introduces
proposed preprocessor and a new automatic technique to
find peaks in detail. Section 3, shows the empirical
results to present the standard of the proposed technique.
Finally, Section 4 concludes our study and also present
future scope.
II.
P
ROPOSED METHODOLOGY
The target of this paper is to propose a new algorithm
to detect R-peak using a novel approach [28]. All existing
R-peak detection methods are suffered from the noise.
Therefore to analyze the ECG signals, firstly we have to
remove the noise from the signals in preprocessing stage.
The architecture to obtain R Peak is represented in Fig.
2 which includes five stages, as preprocessing and
filtering (band pass filter), first-order forward difference
operation to highlight the QRS complex, an amplitude
normalization, Shannon energy envelope withdrawal
stage, peak-detection logic stage and exact R-peak finder.
Filter is used to pre-process and filter the signal. Once the
signal is free from the noise then Shannon energy
envelope (SEE) estimator is used to find the R Peak
position and their amplitude. In this stage to find Shannon
energy (SE) envelope, this method uses Shannon energy
assessment and zero-phase filtering which plays an
important role. It can be observed that in the SE envelope
major local maxima detect the approximate R-peaks
locations in ECG. Then we applied a peak finding logic
where the proposed technique is developed based on
simple threshold of Shannon energy. Actual positions of
the local maxima are identified using the proposed
technique. Finally, to find accurate R peak locations in
ECG signal, these positions of local maxima are used as
guides. The architecture of proposed methodology is
represented in Fig 2.
Fig.2. R-peak Detection Technique
The detailed discussions of each stage are as follows:
A. Preprocessing and Filtering Stage
In the realistic environments the ECG signal obtain
from the patient gets corrupted by external noises; hence
to make ECG signal proper noise free is essential.
Various types of noise are frequency interference, power
line interference, polarization noise, baseline drift,
muscle noise, muscle contraction, electrode contact noise,
internal amplifier noise, and motion artifacts and have
large T and P waves. In ECG signal, Artifacts are the
noises which induced due to movements of electrodes.
Useful information from the ECG signal can be extracted
after the processing of raw ECG signal. Therefore, Band
pass filter stage and first-order differentiation stage are
used for accentuating the QRS complex. This will also
reduces the noise and the effect of T and P waves. To
avoid the phase distortion the filter is enforced in two
forward as well as reverse.
B. First Order Forward Differencing (FOFD)
After preprocessing and filtering stage, data of the
ramp of the QRS complexes can be finding by
differentiating the output signal of filter, f[n]. Filtered
ECG signal is differentiated by implementing
2024年5月4日发(作者：汤宛丝)
I.J. Information Engineering and Electronic Business, 2017, 6, 43-50
Published Online November 2017 in MECS (/)
DOI: 10.5815/ijieeb.2017.06.06
A Robust Approach for R-Peak Detection
Amana Yadav
Department of Electronics and Communication Engineering, Manav Rachna International University, India
Email: @
Dr. Naresh Grover
Department of Electronics and Communication Engineering, Manav Rachna International University, India
Email: ics@
Received: 23 June 2017; Accepted: 01 August 2017; Published: 08 November 2017
Abstract—Electrocardiogram (ECG) is very crucial and
important tool to detect the cardiac problems. For ECG
analysis, it is essential to measure ECG parameter
accurately. It is very critical in all types of ECG
application. The accurate R Peaks detection is starting
step in extracting ECG features which is necessary for the
other ECG performance stages. It is very essential to
detect these R-peaks accurately and efficiently to detect
heart diseases or anomalies which create primary source
of death in the universe. Hence automatic R-peaks
detection in a lengthy duration ECG signal is very
meaningful to diagnose the cardiac disorders. Here a
latest R-peak exposure algorithm depended on Shannon
energy envelope estimator and logic to find peaks has
been proposed which uses the simple threshold of
Shannon energy.
Fig.1. ECG signal made of a P wave, a QRS complex and a T wave [5]
Index Terms—ECG, R-peak detection, QRS complex, P-
During last few years, many systems have been
QRS-T waves, sampling frequency, Cardiac arrhythmia,
designed for QRS detection. Numerous QRS detection
MATLAB
algorithms based on the derivatives [6], filtering
techniques [7-10], wavelet transform [11-13],
mathematical morphology [14,15], empirical mode
I.
I
NTRODUCTION
decomposition (EMD) [16], geometrical matching [17],
The ECG is widely used to find out the cardiac
artificial neural networks [18] and hybrid approach [19],
problems, since it measures the contractile electrical
genetic algorithms [20], syntactic methods [21], Hilbert
activity of the heart developed over the cardiac rhythm
transform [22], Markov models [23] etc. reported in
through different electrodes placed at different places on
literature have been developed for R-peaks detection. The
the human body [1]. ECG is an exclusive signal which is
filtering techniques and decision rules based methods are
based on the physical composition especially chest of a
very efficient so best for all ECG analysis [9]. Many
particular person. Therefore ECG can be used to
approaches comprise of a preprocessing or extraction of
features and then a decision block [24]. To accentuate the
distinguish any individuals [2].
A normal ECG signals characterized by a P wave, a T
QRS complex various signal processing techniques are
wave and a QRS complex are shown in figure 1. R wave
applied in preprocessing stage which also suppresses the
has the highest amplitude in heart signal than the other
noises but they have some drawbacks. In [7], there is a
tradeoff between absence and false identification of peaks
portions.
The performance of R Peak detection systems depends
based on the choice of filter’s bandwidth and capacity of
on how accurate the R-peak detector is. Therefore it is
the moving-window integrator. The Empirical mode
very essential to identify the R-peaks in QRS complex
decomposition in [16] can defeated the selection problem
accurately and efficiently. Automatic R-peaks detection
of mother wavelet of Wavelet based QRS detector but
in a large duration ECG signal is very meaningful to
under noisy environments it is very difficult to select the
diagnose the cardiac disorders [3]. R Peaks are pointed
set of intrinsic mode functions (IMF). By introducing
towards the positive side ever and hence can’t have
further useful filtering technique and threshold alteration
method performance can be improved [16]. To study the
negative values [4].
ECG, a faithful approach to detect QRS which depends
Copyright © 2017 MECS I.J. Information Engineering and Electronic Business, 2017, 6, 43-50
44 A Robust Approach for R-Peak Detection
on the highest slope identification was introduced [25]. In
subsequent years many changes have been done on the
method of Pan Tompkins. Baseline correction is required
to get the complete details of QRS complex, so instead of
using adaptive thresholding a fixed thresholding is
required [26]. Instead of using the HPF and differentiator
in Pan Tompkins’ algorithm a novel Pan Tompkins
algorithm by adopting a Savitzky-Golay filter is
generated [27].
In this paper, an advanced preprocessor depending on
the Shannon energy envelope estimator and logic to
detect peaks using simple threshold of Shannon energy
has been proposed which is straightforward with better
accuracy and takes minimum computation time. The
threshold T can be evaluated by using universal threshold
suggested by Donoho.
The major purpose of proposed work is to check this
technique on ECG of human being who is not well; hence
ECG signal is a personal recognition of different patients.
Algorithm is implemented with the help of MATLAB.
Almost all the R peak recovering techniques utilizes the
standard Massachusetts Institute of Technology-Beth
Israel Hospital (MIT-BIH) record from for
the analysis of an ECG signal [8]. The arrhythmia
displays the irregularities of heart which are observed as
Tachycardia and Bradycardia can be simply obtained.
The proposed R-peak detector has 99.60% accuracy,
99.84% sensitivity and 99.75% positive predictivity. The
results prove that the given R Peak detector works better
than other conventional techniques for pathological or
noisy signals.
The paper is arranged in such a way:
In Section 2, the five-stage technique to obtain the R-
peak is explained precisely. This section introduces
proposed preprocessor and a new automatic technique to
find peaks in detail. Section 3, shows the empirical
results to present the standard of the proposed technique.
Finally, Section 4 concludes our study and also present
future scope.
II.
P
ROPOSED METHODOLOGY
The target of this paper is to propose a new algorithm
to detect R-peak using a novel approach [28]. All existing
R-peak detection methods are suffered from the noise.
Therefore to analyze the ECG signals, firstly we have to
remove the noise from the signals in preprocessing stage.
The architecture to obtain R Peak is represented in Fig.
2 which includes five stages, as preprocessing and
filtering (band pass filter), first-order forward difference
operation to highlight the QRS complex, an amplitude
normalization, Shannon energy envelope withdrawal
stage, peak-detection logic stage and exact R-peak finder.
Filter is used to pre-process and filter the signal. Once the
signal is free from the noise then Shannon energy
envelope (SEE) estimator is used to find the R Peak
position and their amplitude. In this stage to find Shannon
energy (SE) envelope, this method uses Shannon energy
assessment and zero-phase filtering which plays an
important role. It can be observed that in the SE envelope
major local maxima detect the approximate R-peaks
locations in ECG. Then we applied a peak finding logic
where the proposed technique is developed based on
simple threshold of Shannon energy. Actual positions of
the local maxima are identified using the proposed
technique. Finally, to find accurate R peak locations in
ECG signal, these positions of local maxima are used as
guides. The architecture of proposed methodology is
represented in Fig 2.
Fig.2. R-peak Detection Technique
The detailed discussions of each stage are as follows:
A. Preprocessing and Filtering Stage
In the realistic environments the ECG signal obtain
from the patient gets corrupted by external noises; hence
to make ECG signal proper noise free is essential.
Various types of noise are frequency interference, power
line interference, polarization noise, baseline drift,
muscle noise, muscle contraction, electrode contact noise,
internal amplifier noise, and motion artifacts and have
large T and P waves. In ECG signal, Artifacts are the
noises which induced due to movements of electrodes.
Useful information from the ECG signal can be extracted
after the processing of raw ECG signal. Therefore, Band
pass filter stage and first-order differentiation stage are
used for accentuating the QRS complex. This will also
reduces the noise and the effect of T and P waves. To
avoid the phase distortion the filter is enforced in two
forward as well as reverse.
B. First Order Forward Differencing (FOFD)
After preprocessing and filtering stage, data of the
ramp of the QRS complexes can be finding by
differentiating the output signal of filter, f[n]. Filtered
ECG signal is differentiated by implementing