17-03-2014, 10:06 PM
Abstract— The Electrocardiogram (ECG) has been established as
a powerful diagnostic tool in medicine which provides important
information about the patient’s heart condition. The correct
identification of the QRS complexes is a fundamental step in
every automated or semi-automated ECG analysis method. A
major problem that is often encountered in automatic QRS
detection is the presence of artifacts in the ECG data, which
cause considerable alterations to the signal. In this work, the
objective was to develop a method, based on Time-Frequency
Analysis (TFA), which would be able to automatically detect and
remove artifacts in order to increase the reliability of automatic
QRS detection. The TFA method used for the analysis of the
ECG data, was based on a time-varying Autoregressive (AR)
model whose solutions were obtained using Burg’s method. The
algorithm could detect and remove 95.6% of the artifact areas
and correctly identify 92.0% of QRS complexes (322 out of 335
annotated QRS complexes). The proposed method was compared
with one of the most commonly used methods in ECG analysis,
which is based on the use of wavelets. The wavelet-based method
resulted in an accuracy of QRS detection of 65.3% mainly due to
the large number of false positive detections in the regions of
artifact.
Keywords— Electrocardiogram, ECG, Time-Frequency
Analysis, spectrogram, artifact detection, QRS detection
I. INTRODUCTION
Heart disease is one of the main causes of death in the western
world and much effort is expended on its diagnosis and
treatment. Electrocardiography is considered to be one of the
most powerful diagnostic tools in medicine that is routinely
used for the assessment of the functionality of the heart.
Different waves reflect the activity of different areas of the
heart. A normal electrocardiogram (ECG) consists of a P wave,
a QRS complex, and a T wave. The P wave is caused by
electric currents produced by the depolarization of the atria
before their contraction, while the QRS complex is caused by
electric currents produced by the depolarization of the
ventricles prior to their contraction, during the extending of
the depolarization in the ventricular myocardium. The
detection of the QRS complex is the crucial first step in every
automated algorithm for ECG analysis. Due to their
characteristic shape, the QRS complexes serve as the
reference point for automated heart rate monitoring and as the
starting point for further evaluation [1].
Numerous approaches have been proposed for
automatically finding the QRS complexes in as ECG [2]. Such
algorithms include artificial neural networks [3-6] and genetic
algorithms [2]. Other approaches included signal derivatives,
for detection of the steep slope of the QRS complex, [7-11],
cross-correlation methods, where an initial template was
aligned to the current ECG signal [4,5], and syntactic
approaches, where the ECG signal was represented as a
piecewise linear approximation and was analyzed using
syntactic rules. The wavelet transform method is currently
considered to be a state-of-the-art method for automatic ECG
analysis and QRS detection [12-16]. Almost all of the
proposed algorithms so far, share a common algorithmic
structure, that is, a preprocessing stage, including filtering, a
feature extraction stage, and a decision stage in which peak
detection and decision logic are included [2,17-19].
a powerful diagnostic tool in medicine which provides important
information about the patient’s heart condition. The correct
identification of the QRS complexes is a fundamental step in
every automated or semi-automated ECG analysis method. A
major problem that is often encountered in automatic QRS
detection is the presence of artifacts in the ECG data, which
cause considerable alterations to the signal. In this work, the
objective was to develop a method, based on Time-Frequency
Analysis (TFA), which would be able to automatically detect and
remove artifacts in order to increase the reliability of automatic
QRS detection. The TFA method used for the analysis of the
ECG data, was based on a time-varying Autoregressive (AR)
model whose solutions were obtained using Burg’s method. The
algorithm could detect and remove 95.6% of the artifact areas
and correctly identify 92.0% of QRS complexes (322 out of 335
annotated QRS complexes). The proposed method was compared
with one of the most commonly used methods in ECG analysis,
which is based on the use of wavelets. The wavelet-based method
resulted in an accuracy of QRS detection of 65.3% mainly due to
the large number of false positive detections in the regions of
artifact.
Keywords— Electrocardiogram, ECG, Time-Frequency
Analysis, spectrogram, artifact detection, QRS detection
I. INTRODUCTION
Heart disease is one of the main causes of death in the western
world and much effort is expended on its diagnosis and
treatment. Electrocardiography is considered to be one of the
most powerful diagnostic tools in medicine that is routinely
used for the assessment of the functionality of the heart.
Different waves reflect the activity of different areas of the
heart. A normal electrocardiogram (ECG) consists of a P wave,
a QRS complex, and a T wave. The P wave is caused by
electric currents produced by the depolarization of the atria
before their contraction, while the QRS complex is caused by
electric currents produced by the depolarization of the
ventricles prior to their contraction, during the extending of
the depolarization in the ventricular myocardium. The
detection of the QRS complex is the crucial first step in every
automated algorithm for ECG analysis. Due to their
characteristic shape, the QRS complexes serve as the
reference point for automated heart rate monitoring and as the
starting point for further evaluation [1].
Numerous approaches have been proposed for
automatically finding the QRS complexes in as ECG [2]. Such
algorithms include artificial neural networks [3-6] and genetic
algorithms [2]. Other approaches included signal derivatives,
for detection of the steep slope of the QRS complex, [7-11],
cross-correlation methods, where an initial template was
aligned to the current ECG signal [4,5], and syntactic
approaches, where the ECG signal was represented as a
piecewise linear approximation and was analyzed using
syntactic rules. The wavelet transform method is currently
considered to be a state-of-the-art method for automatic ECG
analysis and QRS detection [12-16]. Almost all of the
proposed algorithms so far, share a common algorithmic
structure, that is, a preprocessing stage, including filtering, a
feature extraction stage, and a decision stage in which peak
detection and decision logic are included [2,17-19].