Applied Artificial Intelligence

An International Journal

ISSN: 0883-9514 (Print) 1087-6545 (Online) Journal homepage: www.tandfonline.com/journals/uaai20

Machine Learning Techniques Applied to Data

Analysis and Anomaly Detection in ECG Signals

Tomasz Andrysiak

To cite this article: Tomasz Andrysiak (2016) Machine Learning Techniques Applied to Data
Analysis and Anomaly Detection in ECG Signals, Applied Artificial Intelligence, 30:6, 610-634,
DOI: 10.1080/08839514.2016.1193720

To link to this article: https://doi.org/10.1080/08839514.2016.1193720

Published online: 21 Jul 2016.

Submit your article to this journal

Article views: 1306

View related articles

View Crossmark data

Citing articles: 6 View citing articles

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=uaai20
APPLIED ARTIFICIAL INTELLIGENCE
2016, VOL. 30, NO. 6, 610–634
http://dx.doi.org/10.1080/08839514.2016.1193720

Machine Learning Techniques Applied to Data Analysis

and Anomaly Detection in ECG Signals
Tomasz Andrysiak
Institute of Telecommunications and Computer Science, UTP University of Science and Technology,
Bydgoszcz, Poland

ABSTRACT
In this article Tomasz Andrysiak presents the use of sparse
representation of a signal based on overcomplete diction-
aries of base functions and a QRS detection method using
artificial neural network to detect anomalies in the ana-
lyzed ECG signals. Performance of the proposed method
was tested by means of a widely available database of
ECG signals, i.e., the MIT-BIH Arrhythmia Database, and
the obtained experimental results confirmed its effective-
ness for anomaly detection in the analyzed ECG signals. The
current dynamic and intensive development of information
technologies and more excellent methods of processing, ana-
lysis, and recognition of signals have enriched medicine with
new quality techniques in diagnosis and therapy. At present, a
special interest is put to noninvasive diagnostic methods for
rapid and objective determination of vital signs and, in parti-
cular, automatic electrocardiogram performed outside medical
facilities. The sparse representation of a signal was performed
in an adaptive manner by means of the matching pursuit
algorithm. In each step of the algorithm, there was implemen-
ted a linear signal decomposition of features belonging to the
dictionary with redundancy. Two kinds of dictionaries were
tested. The first type included elements that form the analytical
base functions, and the second represented the essential com-
ponents of the ECG signal searched with the use of the
K-Singular Value Decomposition (K-SVD) algorithm.
Recognition of the QRS was realized by means of a unidirec-
tional, multidimensional neural network with backpropagation
of an error. For this solution, tests were performed on a selec-
tion of appropriate neural network architectures and the
impact of a particular activation function on the recognition
results. Anomaly detection was realized with estimation of
sparse representation parameters of the tested ECG signal
within the field of the recognized QRS complex, and it was
compared with the reference values. Performance of the pro-
posed method was tested using a widely available database of
ECG signals—MIT-BIH Arrhythmia Database. The obtained
experimental results confirmed the effectiveness of the pro-
posed method of anomaly detection in the analyzed ECG
signals.

CONTACT Tomasz Andrysiak [email protected] Institute of Telecommunications and Computer

Science, UTP University of Science and Technology, Al. prof. S. Kaliskiego 7, 85-796 Bydgoszcz, Poland.
Color versions of one or more of the figures in the article can be found online at www.tandfonline.com/UAAI.
© 2016 Taylor & Francis
 APPLIED ARTIFICIAL INTELLIGENCE 611

Introduction
Intensive development of techniques used for acquisition, analysis, and
detection of signals gave grounds for searching new solutions in the field
of modern medical diagnostics. One of directions of this development is
automation of the process of analysis and detection of disturbances in ECG
signals (Tang and Lan 2014). This approach accelerates and facilitates the
processes of analyzing the electrocardiogram and enables keeping cardiolo-
gical supervision in real time outside medical centers. A visible effect of such
an intensive development is the currently emerging telemetric systems trans-
mitting in real time the results of processing the ECG signal to the monitor-
ing centers (Srinagesh, Sarala, and Aparna 2013).
The purpose of electrocardiography is to provide information about the
changing electric field on the outer surface of the human body as a result of
bioelectric processes (Besterman and Creese 1979). Observation and analysis
of those changes are the grounds of correct diagnosis of the heart rate
reflected in ECG sequences. These sequences describing the deflection of
an ECG signal were marked by Einthoven (1924) as the consecutive alphabet
letters P, Q, R, S, T, and are associated with polarization and depolarization
of the atria and ventricles. The most important information from the point of
view of analysis of the structure of the ECG is the QRS complex. This term
describes three, occurring next to each other, characteristic points Q, R, S of
an electrocardiogram, which form the deflection shape of a narrow pulse
(Figure 1). This complex is characterized by the highest deflection and the
shortest duration. Its shape and frequency supply the most diagnostic infor-
mation about the current state of the heart. Additionally, it is the basis of a
particular scheme of construction, i.e., the structure PQRST of the tested
person’s cardiac cycle. It is also used as an important message for automatic

Figure 1. An example of an ECG signal with marked folds.

612 T. ANDRYSIAK

determination of the heart rate (R-R interval). The correct detection of a QRS
complex in an ECG signal is essential information for most automatic ECG
analysis algorithms (Daskalov and Christov 1999).
Despite the practical application of systems that recognize the essential ele-
ments of the ECG, the problem of detection and identification of QRS is still a
subject of research of many scientific centers (Harikumar and Shivappriya
2011). The reason for this situation is that, besides the complexity of signal
and morphological variety of ECG folds, quite often noise and artifacts of
different kind appear, which significantly hinder its analysis (Alfaouri and
Daqrouq 2008). Therefore, algorithms of the QRS complex detection need to
fulfill strict requirements, i.e., they must have high resistance to distortion,
almost 100% accuracy of detection, and processing speed close to real time.
These algorithms are, in fact, the starting point for the overall analysis of ECG
and, in particular, for different kinds of anomalies appearing in it. Their
accuracy significantly influences the quality of functioning of the complete
system analyzing the ECG signal (de Chazal, O’Dwyer, and Reilly 2004).
One of the most often detected anomalyies of the heart rate is arrhythmia,
reflected in time changes of the R-R interval and fluctuations of QRS com-
plexes’ morphology in ECG records. There are two basic types of arrhythmia:
paroxysmal and chronic. Detection of chronic arrhythmia is relatively simple,
paroxysmal arrhythmia, though, requires recording the ECG signal over a
longer period of time. However, the diagnostic importance of paroxysmal
arrhythmia is immense because it represents deviation from regular sinus
rhythm caused by a progressive disease process or arrhythmogenic factors.
Arrhythmia detection is most often based on finding in the ECG signal the
QRS complexes correlated with the R-R interval and classification of their
morphology (Sanamdikar, Hamde, and Asutkar 2015).
Given the previous information, in this study, I suggest a system of data
analysis and anomaly detection in ECG signals based on machine learning
techniques. The main concept of the proposed approach is a skilful combina-
tion of the sparse decomposition of an ECG signal on the grounds of
redundant dictionaries of base functions with classifying properties of multi-
dimensional unidirectional neural network with backpropagation of an error.
The tasks of anomaly detection are realized by estimation of energy of the
components of ECG signal’s sparse representation in the field of the found
QRS structure. This approach allows rapid and effective anomaly detection in
the tested ECG signals.
This article is organized as follows. After the introduction, “Related Work”
describes related work on existing detection methods of the QRS complex and
anomaly detection in electrocardiographic records. Next, “Machine Learning
Techniques for Data Analysis and Anomaly Detection in ECG Signals” and its
subsections show the details of machine learning techniques for data analysis
and anomaly detection in ECG signals, the sparse representation of ECG signals
 APPLIED ARTIFICIAL INTELLIGENCE 613

based on the matching pursuit algorithm, outlines of the details of different

versions of dictionaries for sparse representation, the generation of the diction-
ary tree structure for the optimizing computation, and the QRS detection
method using artificial neural network is discussed. The section that follows
that presents the details of experimental results of the proposed anomaly detec-
tion approach. “Conclusions” are made in the final section.

Related work
Detection algorithms of a QRS complex are the key elements of every system
of automatic ECG signal analysis. Their accuracy and efficiency have sig-
nificant impact on the quality of the diagnostic data reported by the system
(Pan and Tompkins 1985). Detection of a QRS complex can be divided into
two basic stages (Augustyniak 2001). The first one implies performing
mathematical transformations of the ECG signal with sharpening the features
of the QRS complex and minimizing other, nonessential elements of the
signal (noise, disturbances). As a result of preprocessing the original signal,
we receive a function of detection, the measure of which is the steepness of
growing and falling of slopes on the QRS section borders. The second stage
consists of thresholding, which is used to find points classified as QRS
complexes. Most often, the value of a threshold is established in an adaptive
way, and the main objective of thresholding is to find the local maximum.
Algorithms of detection and classification of a QRS complex can be divided
into: (1) algorithms based on time domain or frequency domain; (2) algo-
rithms using elements of artificial intelligence based on neural network
models, Support Vector Machine classifiers, or genetic algorithms.
QRS complex detection in a time domain consists of finding local max-
imum in the ECG signal, most often by means of various signal differentia-
tion algorithms (Hsiung, Cheng, and Jan 1989), or filtration methods based
on principles introduced by Hamilton and Tompkins (1988). Frequency
domain analysis, however, involves the study of spectral properties of ECG
signals and, in practice, it uses discrete Fourier transform (Kwiatkowski
2003) or Gabor transform (Okajima 1998).
A development of the above methods is the time-frequency analysis per-
formed with the use of wavelet transform, proposed for the first time for the
ECG signal by Hassani and others in their work (Hassani, Fachrudin, and
Hubin 1994). The mentioned authors utilized the wavelet transform not only
to detect the QRS complex, but also to recognize tachycardia and long QT
syndrome. In the subject literature, the biggest set of algorithms connected to
wavelet transform used for detection of a QRS complex is based on the
approach by Mallat and Hwang (1992). The authors studied the correspon-
dence between the singularities of functions and local maxima in the wave
transform of the signal. The peak classification was performed by calculating
614 T. ANDRYSIAK

the degree of detail, so called Lipschitz’s local regularity, which was estimated
based on the distribution of wavelet coefficients. Both continuous and discrete
wavelet transforms may be successfully used not only for detection but also for
classification of the signal’s anomaly and recognizing arrhythmia. For instance,
Chazal the Daubechies family of wavelets of different orders.
Addison (Addison et al. 2000), on the other hand, utilized the wavelet trans-
form to study the ECG signal during ventricular fibrillation. The combination of
neural network and wavelet transform for the ECG signal’s diagnostics, however,
was first proposed by Celler and de Chazal (1998). For the ECG signal’s decom-
position into six levels, the authors used Daubechies and Symlet wavelets, and the
statistical parameters obtained in such a way constituted input data for a neural
network, which did not have any hidden layers. In turn, in Al-Fahour and Howitt
(1999), to automatically detect and classify arrhythmias, the authors proposed the
neural network with radial preprocessing, which uses the Daubechies wavelet of
the fourth order. Another approach in the field of artificial intelligence is the use of
Support Vector Machine techniques, which perform separation of components of
the ECG signal while maintaining fairly strict statistical requirements by seeking
global minimum of a convex cost function (Osowski, Tran Hoai, and Markiewicz
2004). A different approach is methods based on genetic algorithms for ECG
signal analysis, which were first used by Poli and Cagnoli (1995). They found that
genetic algorithms, in conjunction with polynomial filters, effectively handle noise,
artifacts, and diversity of morphology of the ECG in QRS complex detection tasks.

Machine learning techniques for data analysis and anomaly detection

in ECG signals
This article presents chosen methods of machine learning for detection of a QRS
complex and correlative anomalies reflected in ECG records. Selection of essential
features of the signal was realized by sparse decomposition of the ECG signal with
the use of redundant dictionaries. However, detection of the QRS complex in the
analyzed ECG signal was performed by means of a multilayer neural network with
backpropagation of an error. Novelty of the presented solution consists of com-
bining the sparse representation of the ECG signal with classification properties of
neural networks.

Sparse representation of an ECG signal

Representations of an ECG signal realized as linear expansions with regard to a
particular set of basis functions well located with respect to time and/or frequency
are not sufficiently precise and optimal in numerous cases. Depending on the
choice of such functions, the expansion coefficients represent different features of
the analyzed ECG signal. In the case of complicated signal structures, we are
unable to define the optimal scale parameters and frequencies for the particular
 APPLIED ARTIFICIAL INTELLIGENCE 615

basis functions. Then, a natural solution is introduction of more diverse, numer-

ous, character sets of functions that are adjusted to the ECG signal’s named
dictionaries with redundancy (Rubinstein, Bruckstein, and Elad 2010). By per-
forming the above, we obtain much more universal and flexible representations.
However, decomposition of an ECG signal with regard to a dictionary requires
then constant searching and matching of its suitable elements, which best reflect
the wanted features of the analyzed signal. The mentioned matching should be
realized in such a manner that it maximizes the correlations between the chosen
element of the dictionary and the rest of the processed signal. Therefore, there is a
need to determine the quality measure of such matching and an algorithm
ensuring finding the best, in the sense of that measure, the ECG signal’s
representation.
The nature of sparse representation is based on seeking the sparse solution
of coefficientsC that represent the signal S with the use of a dictionary with
redundancy, on condition that the remainder is smaller than the given
constant δ. It can be written as (Elad 2010):

X
K1
min k C k0 subjectto k S  ck dk k <δ; (1)
k¼0

where k k0 is the l0 norm counting the nonzero entries of a vector,

ck 2 Crepresenting a set of projection coefficients, and dk are the elements
of redundant dictionary D (Davis, Mallat, and Avellaneda 1997).
The optimal representation can be defined as such a subset of dictionary D
elements, the linear combination of which explains the highest percentage of
the signal’s S energy among all the subsets with the same number of
elements. The choice of such representation is NP-difficult (Troop 2004),
thus, in practice, we are satisfied with an adaptive iterative solution called the
matching pursuit algorithm. The outcome of this algorithm’s work is a
projection of the signal’s structural elements onto the chosen dictionary’s
elements.

The matching pursuit algorithm

To realize the S signal’s decomposition, we will set a linear expansion with
regard to a set of atoms selected from the dictionary D so that they are most
properly adjusted to the structural elements of the analyzed signal. These
operations will be realized by means of the matching pursuit algorithm (MP)
proposed in Mallat and Zhang (1993).
Owing to the MP algorithm, we obtain three pieces of information:
 
● the set of projection coefficients C ¼ c0 ; c1 ; . . . ; cp ; . . . ; cn1
● the set of residua RS ¼ fr0 s; r1 s; . . . ; rp s; . . . ; rn1 sg
616 T. ANDRYSIAK

● the list of elements of the dictionary selected  as approximate of S and

represented as D ¼ d0 ; d1 ; . . . ; dp ; . . . ; dn1 .

C, RS, and D are vectors which thoroughly denote the discrete signal S,
which can be written as
X
n1
S¼ ci di þ rn s; (2)
i¼0

where rn s is residuum in an n–term sum.

In every step p of the MP algorithm we will perform further decomposi-
tions of signal S by means of projections onto the elements of dictionary D.
Then, in the pth decomposition step, we obtain
rp s ¼ rp1 s  cp dφp ; (3)

and
D E
cp ¼ rp1 s; dφp ; (4)

where h; i is a scalar product, and rp s is the residuum resulting from

decomposition of S in direction dp . For residuum of zero order there is a
clear link r0 sn¼ S. The selected povectors’ indices are gathered in the index
vector Φp ¼ φ1 ; φ2 ; . . . ; φp1 ; φp , Φ0 ¼ ;, and the vectors are collected as
n o
columns of the matrix Dp ¼ dφ1 ; dφ2 ; . . . ; dφp and D0 ¼ ;.

Selection of φp is the core function of the MP algorithm. This task is

performed by finding the vector that is best adjusted to the residuum
achieved by projection of rp s onto the elements of the dictionary, i.e.,
φp ¼ arg maxjrp s; di j; φp Þφp1 ; φp 2 Φp : (5)
i2Φp

For the atom dp , chosen by means of dependency Equation (5), we minimize

the residuum of signal S in the next step of the MP algorithm. The number of
iterations N, in which we perform limitation of residua, depends on the
required preciseness of the S signal’s representation and is given by depen-
dency, e.g., Equation (5), being at the same time the criterion of the MP
algorithm’s stop,

rN s < βS; (6)

where the parameter β 2 ð0; 1Þ. For practical realization of the MP algo-
rithm’s stop criterion we take parameter value β ¼ 0:1, which gives 90%
accuracy of the S signal’s representation.
 APPLIED ARTIFICIAL INTELLIGENCE 617

The pace of residuum norm’s fall depends on the correlation between

consecutive residua of the signal and the selected atoms of the dictionary. If,
in the ECG signal, there is a sum of high energy components constituting
dictionary atoms, then the signal correlation coefficient and its residua are
significant. Consequently, their norm falls rapidly because the high-energy
components are the signal’s structural elements well correlated with the
chosen atoms of the dictionary.
The MP algorithm can be stated as follows:

● Step 1. Initialize the residuum r0 s ¼ S and initialize the set of selected

variables Φ0 , D0 , C0 . Let the iteration counter p ¼ 1.
● Step 2. Find in dictionary D a proper atom dp solving the maximization
problem given by the dependence Equation (5). Next, update the index
set Φp ¼ Φp1 [φp and the set of the dictionary Dp ¼ Dp1 [dφp .
● Step 3. Compute new coefficient cp , e.g., Equation (4) and new residuum
rp s, e.g., Equation (3). Then update the set of coefficients Cp ¼ Cp1 [cp
p
and the set of residua Rp s ¼ Rp1 s [r s.
● Step 4. If the stopping condition is achieved, e.g., Equation (6), stop the
algorithm. Otherwise, setp ¼ p þ 1 and return to Step 2.

To differentiate the adaptive decompositions of signal S, there has been

introduced an additional limitation φp Þφp1 in dependence Equation (5). It
modifies the MP algorithm and consists of a ban on the selection of the same
atoms in two successive iterations of the decomposition process. This con-
dition results in a decrease of correlation in set Dp containing exploited
atoms dφp , and it improves distinctiveness of the S signal decomposition
outcomes.

The orthogonal matching pursuit algorithm

The orthogonal matching pursuit (OMP) algorithm proposed in (Pati,
Rezaiifar, and Krishnaprasad 1993) presents an enhanced version of the
regular MP algorithm. Both algorithms have a greedy structure; however,
for the OMP algorithm, every chosen atom must be orthogonal in every
decomposition stage.
The φp in the pth iteration is selected by the algorithm by searching for a
vector that best matches the residuum achieved by projection of rp s onto the
elements of the dictionary, i.e.,
φp ¼ arg maxjrp s; di j; φp ‚Φp1 : (7)
i2Φp

The reselection issue is not present in the case of the stored dictionary. If
φp ‚Φp1 , then the index set is updated as Φp ¼ Φp1 [φp and
618 T. ANDRYSIAK

Dp ¼ Dp1 [dφp . Otherwise, Φp ¼ Φp1 and Dp ¼ Dp1 .The residuumrp s is

computed by means of
 1
rp s ¼ rp1 s  Dp DTp Dp DTp rp1 s; (8)

where DTp Dp is the Gram matrix (Gharavi-Alkhansari and Huang 1998).

The OMP algorithm can be presented as follows:

● Step 1. Initialize the residuum r0 s ¼ S and initialize the set of selected

variables Φ0 , D0 ,C0 . Let the iteration counter p ¼ 1.
● Step 2. Find in dictionary D a proper atom dp solving the maximization
problem given by the dependence Equation (7). If φp ‚Φp1 , then the
index set is updated as Φp ¼ Φp1 [φp and Dp ¼ Dp1 [dφp . Otherwise,
Φp ¼ Φp1 and Dp ¼ Dp1 .
● Step 3. Compute the new coefficient cp Equation (4) and new residuum
rp s Equation (8). Then, update the set of coefficients Cp ¼ Cp1 [cp and
p
the set of residua Rp s ¼ Rp1 s [r s.
● Step 4. If the stopping condition is achieved Equation (6), stop the
algorithm. Otherwise, set p ¼ p þ 1 and return to Step 2.

Dictionaries for sparse representation

Bases are the smallest complete dictionaries. The greatest dictionary for L2
space consists of all integral functions with the square of the unit standard
(Natarajan 1995). A precise representation of the analyzed ECG signal in a
dictionary larger than a base introduces redundancy. The expected concise-
ness can be obtained by accepting the signal’s reconstruction inaccuracy,
however, with the use of possibly the smallest amount of functions, which are
the dictionary’s elements. Optimal representation can be defined as such a
subset of the dictionary’s components, the linear combination of which
explains the highest percentage of the signal’s energy among all the subsets
of the same size (Liu, Wang, and Wu 2004).
For realization of sparse representation of a signal, we used two types of
dictionaries created on the basis of distinct approaches. One of them is a
dictionary whose elements (time-frequency atoms) are computed by means
of the properly parameterized Gabor function. The other, however, is created
adaptively on the basis of the class of the processed ECG signals with the use
of the K-SVD algorithm. Due to such an approach, we minimize the signal’s
approximation error for a particular size of the dictionary.
 APPLIED ARTIFICIAL INTELLIGENCE 619

Dictionary of gabor function

In the described solution, we proposed a wavelet form. A time-frequency
dictionary can be expressed as a translation ðaÞ, a dilation ðbÞ, and a
modulation ðcÞ of a window function g ðtÞ 2 L2 ðRÞ
 
1 t  a ict
dðtÞ ¼ pffiffiffi g e ; (9)
b b
where g ðtÞ is a standardized Gaussian function.
Time-frequency atoms defined by means of this function are called a
Gabor function, and a dictionary consisting of those atoms is a Gabor
dictionary. According to Mallat and Zhang‘s statement (Mallat and Zhang
1993), with proper discretization of a Gabor dictionary, we are able to obtain
fast numerical realization of the MP algorithm.
The optimal time-frequency resolution is achieved for the Gaussian win-
dow g ðtÞ, which for the analysis of real valued discrete signals gives a Gabor
functions dictionary (Gabor 1946),
x  q 2πk 
di ðxÞ ¼ Kði;ϕÞ g cos xþϕ ; (10)
b M
where M is the size of the signal for which the dictionary is built, Kði;ϕÞ is a
normalizing constant used for the achieved atom’s unit energy di ¼ 1, and i ¼
fb; q; kg constitutes parameters of the dictionary’s function (Durka 1996) .
We assume that the analyzed signal is composed of M ¼ 2L samples,
where L is an integer. Then, entering the dyadic sampling, we achieve
b ¼ 2j , 0  j < log2 M. For parameters q and k we assume the same sampling
period 2j . In this way, for the chosen parameters, we obtain a Gabor
dictionary characterized by great computational simplicity.
If resolution is defined as distance in time ðΔtÞ or frequency ðΔf Þ between
the neighboring atoms of the so defined subset of the Gabor dictionary, and
assuming that fs is a sampling frequency of the analyzed signal S, we get
2j fs
Δt ¼ ; Δf ¼ j ; (11)
fs 2
This means that resolution depends on the octave j, which defines the size of
atoms in time and frequency. The time variation interval of the dictionary
atoms determines their ability to represent the structural element of the
signal S in time. A modification in frequency is achieved by changing the
sampling frequency of parameters q and k.
We implemented the modified dictionary, originally proposed by Mallat
and Zhang (1993); the parameters of the atoms are chosen from dyadic
sequences of integers. Scale b, which corresponds to an atom’s width in
time, is derived from the dyadic sequence b ¼ 2j . However, parametersu
620 T. ANDRYSIAK

and ω, which correspond to an atom’s position in time and frequency, are

sampled for each octave j with an interval b ¼ 2j þ 1.
In order to create an overcomplete set of Gabor functions, the dictionary
D was built by varying subsequence atom parameters: scale ðbÞ, translation
ðaÞ, modulation ðcÞ, and phase ðϕÞ. The base functions of dictionary D were
created by using 25 different scales and 100 different modulations.

The adaptive method of a dictionary creation on the basis of an ECG signal

Creation of a dictionary containing analytical functions is relatively quick and
simple, but it might lead to dependence of the approximation error (in the
process of signal decomposition) on the nature and count of functions in the
dictionary. The adaptive method of a dictionary creation on the basis of K-SVD
algorithm, though, is deprived of the aforementioned disadvantage (Aharon,
Elad, and Bruckstein 2006). In every iteration of the algorithm, there is a
searching and updating of the dictionary’s components in order to, for particular
criterion, maximize the correlation between the found atom (an element of the
processed signal) and the remaining representation of the ECG signal.
The prime function of the K-SVD algorithm is finding the best dictionary,
which would represent the signal S as sparse composition, by solving
 
min S  DCF2 subjectto"i ci 0  T; (12)
D;C

where k  k2F is the Frobenius norm and T is a fixed and predetermined

number of nonzero entries.
The algorithm is divided into two stages:

● Sparse Coding Stage: In this phase, the coefficients ci are set on the
basis of the redundant dictionary D and signal si . In every step, we seek
the minimum number of coefficients allowing us to satisfy the Equation
(13). The given D is fixed. The orthogonal matching pursuit algorithm,
described previously, is used to calculate M sparse coefficients ci for
every sample of signal si , by approximating the solution of

 
min si  Dci 2F subject to ci0  Ti ¼ 1; 2; . . . ; M: (13)
C

● Dictionary Update Stages: In this phase, we update the D dictionary

atoms and vector of coefficients C. The alternative values of an atom and
coefficients are computed in order to lower the possibility of an error
between the signal S and the sparse representation D  C. To achieve the
above the SVD method is used. The given C and D are fixed. The
 APPLIED ARTIFICIAL INTELLIGENCE 621

attention is on the dictionary’s atom dk and its matching sparse vector

ckT (i.e., row k of ckT ). The corresponding objective function in Equation
(12) can be presented as:
0 1
X K X
dj cTF ¼ @S  dj cT A  dk ckT F2 ¼ Ek  dk ck2
j2 j
S  DCF2 ¼ S  TF ; (14)
j¼1 jÞk

where Ek depicts the representation error of the signal’s sample after the kth
atom. The following stages of the dictionary’s update are:

● We define the index set that uses the atom dk , and means that the case
ckT ðiÞ is nonzero as,

 
ωk ¼ ij1  i  M; ckT ðiÞÞ0 : (15)

● In the next step, we define the matrix Ωk with values equal to 1 in

positions ðωk ðiÞ; iÞ and compute ckR ¼ ckT Ωk .
● We compute Ek which represents an error between all the S signal’s
samples and the sparse representation D  C, provided that the atom dk
is removed from dictionary D,
X j
Ek ¼ S  dj cR ; (16)
jÞk

and restrict Ek by selecting the columns matchingωk so that we obtain ERk .

● In the last step, we use SVD to decompose ERk ¼ UΔV T and update dk so
that it is the first column of U; and the coefficient vector ckR to be the
first column of V multiplied by Δð1; 1Þ.

All dictionary atoms are updated is this manner. Iterating according to

thee two steps will create a dictionary that approximates the provided signal
S in a sparse and accurate way. The outcome of the K-SVD algorithm will be
a dictionary D containing often correlated atoms that best represent the
analyzed signalS with regard to its sparse representation D  C. To solve
this problem, we propose a modification of the K-SVD algorithm through
introduction of a decorrelation condition.
We add the following optimization problem at each iteration of the
dictionary learning algorithm that includes approximation followed by a
dictionary update
622 T. ANDRYSIAK

D  2F ;
^ ¼ arg min D  D (17)
D2Γ

whereμis the measure of the maximal correlation of different atoms

ci cj
μðDÞ ¼ max ; ; (18)
di ;dj 2DiÞj jc i j2 c j 2

and
 
Γ ¼ DjμðDÞ  μ0 ^ ci 2 ¼ 1; i 2 f1; ::; Mg ; (19)
where the objective μðDÞ is a cost function that describes the approximation
feature of the dictionary, and μ is the level of a fixed target coherence.
A possible solution to this issue might be to apply a decorrelation step to
the iterative scheme in the K-SVD algorithm. To a dictionary learning
algorithm implying a sparse approximation that is followed by a dictionary
update, we add, at every iteration, the following optimization problem:
^ ¼ arg min μðDÞ;
D (20)
D2Γ
 
 2F  θ ^ ci2 ¼ 1; i 2 f1; ::; Mg ;
Γ ¼ DjD  D (21)
where θ constitutes the unknown minimum value achieved by the criterion
Equation (17).
In our practice, we proposed to solve the problem by applying a decorr-
elation step in the K-SVD loop after the dictionary update.

A searching in the dictionary of atoms

In the basic matching pursuit algorithm, atoms are selected in every step
from the entire dictionary, which has a flat structure. In this case, the
algorithm causes significant processor burden by performing full search of
the dictionary. In the presented solution, for the purpose of optimization of
the searching process, it is proposed to use the construction of tree structures
of the dictionary on the basis of a K-means algorithm (Berkhin 2006).
For the obtained atoms, the clustering algorithm has been used to group
them into clusters. By means of this method, the clusters are created in the
initial dictionary, where they are organized into a hierarchical tree structure, in
which the elements of the dictionary form the leaves. Consequently, the cluster-
ing process creates levels of a tree structure. Every branch of the tree hasNchil-
dren and is completely described by the list of the atom indexes Wn . A centroid
gn is assigned to the branch of a tree that represents atoms of the dictionary in
the corresponding subtree (Jost, Vandergheynst, and Frossard 2005).
The clustering algorithm can be stated as follows:
 APPLIED ARTIFICIAL INTELLIGENCE 623

● Step 1. Initialize the set of centroids Gð0Þ ¼ ; and initialize the set of
n counter i ¼ 1. Then, othe choice
indices of each cluster. Let the iteration
ði1Þ ði1Þ ði1Þ
ofN initial cluster centroids Gði1Þ ¼ g1 ; g2 ; . . . ; gN .
● Step 2. Assign each atom to the group that has the closest centroid for
n ¼ 1; 2 . . . ; N
n o
ði1Þ
Wnði1Þ ¼ lj"kÞn ; dl  g ðni1Þ 2 <dl  g k 2 :

● Step 3. Calculate the new centroids for n ¼ 1; 2 . . . ; N

1 X
gnðiÞ ¼ dl :
ði1Þ
Wn ði1Þ
l2Wn

● Step 4. Calculate the error of the clustering

X
N X
ði1Þ
U ¼ di  g ðni1Þ 2:
n¼1 l2W ði1Þ
n

● Step 5. If the stopping condition is achieved U ðj1Þ <%, stop the algo-
rithm. Otherwise, set and return to Step 2.

As a result of operation of the above algorithm, we obtain a tree structure

of a dictionary, which enables a significant acceleration of the searching
process of a particular atom that best corresponds the analyzed signal S. In
practical implementation of the clustering algorithm, we assume the para-
meter values N ¼ 25 and % ¼ 10:

The concept of neural network for sparse representation of an ECG signal

The original idea of neural network realization for sparse representation
of a signal consists of finding a linear decomposition of the analyzed ECG
signal by means of MP or OMP greedy algorithms. As a result of the
realized decomposition, with regard to atoms belonging to a dictionary
with redundancy, we obtain essential structural properties of the signal,
which further undergo a selection process. So, a prepared vector of
features, emerged from finding a sparse representation of the ECG signal,
constitutes an optimal input vector for a properly matched structure of
the neural network. It is worth mentioning that a similar concept may be
found in the work of Shyu, Wu, and Hu (2004), describing detection of
624 T. ANDRYSIAK

various types of QRS complex morphologies; however, they use the

wavelet transform, not the sparse representation of the signal.
In the view of the discussed solution, the neural network structure is
conditioned by both the way of use of the sparse representation of the
ECG signal and the type of the utilized neural network. Therefore, it is
impossible to propose one specific and optimal structure. To solve this
problem, a unidirectional multilayer neural network was taken as the basis
for the research model.
Assuming that L is the number of layers, and in every layer there is
ðkÞ
Mk ðk ¼ 1; 2; . . . ; LÞ of neurons, xi ðtÞ is the value of the i-output of the
ðkÞ
k-hidden layer, and wij is the weight of thei-neuron of the k-layer, we can
ðkÞ
define the output signal of the yi neuron (i-output of k-layer) as
!
ðkÞ
X
M k1
ðkÞ ðkÞ
yi ¼f wij xi ðtÞ : (22)
j¼0

For experimental reasons, we will use two f activation functions described by

means of dependence Equation (23). One of them is a unipolar function
taking values in the range (0,1) and the other is bipolar determined on the
range of (−1.1).
1 1  eαx
f1 ð x Þ ¼ and f2 ð x Þ ¼ ; (23)
1 þ eαx 1 þ eαx
where α is a parameter allowing us to determine the appropriate shape of the
activation function and taking the values α > 0.
In the case of the presented solution, the process of neural network
learning is realized as learning with supervision, i.e., a defined set of input
data corresponds to an assigned output pattern. The network’s learning
consists of modification of weight connections in order to match particular
input components to the desired pi patterns. The weight adaptation comes
down to finding the minimum of the objective function modeled as a
function of the mean square error. Then, the complete mean square error
at the outputs of neurons of the neural network last layerL is described by the
dependence
ML 
X 2
ðLÞ ðLÞ
EðtÞ ¼ pi ðtÞ  yi ðtÞ : (24)
i¼1

The minimization process of EðtÞ error function is most often realized

by modifications of weight vector values by means of gradient methods,
for which the gradient vector points the direction of the quickest fall. In
the presented solution, the backpropagation error method (BP) was used
as the gradient learning algorithm of unidirectional multilayer neural
 APPLIED ARTIFICIAL INTELLIGENCE 625

networks. However, the neural network learning BP process algorithm

might be of low speed. Additionally, the learning process might termi-
nate not in the moment of finding the global minimum of the error
function, but in one of its local minimum. The proper selection of input
weight values is, therefore, of great importance. A significant improve-
ment of the BP method’s convergence can be obtained by using the
moment variation of the method. Then, changes of weights happen in a
smoother way, often causing acceleration of convergence. A more
detailed description can be found in the work Muller, Reinhardt, and
Strickland (1995).
In the presented solution, for the process of neural network learning,
we utilized the following method of initial weight creation: weights, the
benchmark of which i ¼ j, adopts values of þ 0:1, the other weights—the
value of 0. Such a manner of selection ensures repetition of the conducted
learning process. To complete the neural network training, the hold-out
method was used, which considers the minimum of testing error as the
stopping criterion. In our case, it was the value of 0.001. A very important
issue is also proper selection of the input dataset. Division of datasets into
the learning set and testing set was realized by means of the cross-
validation method, which consists of segmentation of the prepared set
according to equinumerous k–subsets, of which ðk  1Þ were involved in
the learning process.

Experimental results
In the conducted tests, we used the Massachusetts Institute of
Technology–Beth Israel Hospital Arrhythmia Database (MIT-BIH) com-
posed of eight datasets with various disturbances detectable with Holter’s
method (MIT-BIH Arrhythmia Database 2010). The test was performed
on 48 records from MITDB catalogue, which includes a file with signal, a
heading file (information about the patient) and a diagnostic information
file (with a binary form of the signal). The sampling frequency of signals
was 360 Hz, and resolution 12 bits. The process of determining the
parameters of the proposed solution (i.e., the number of levels of sparse
decomposition, the threshold detection value, the selection of the learning set and
neural network architecture, and the kind of activation function) was performed
by means of available signals in the MIT-BIH base. The detailed outcomes,
however, were presented exclusively for four selected signals. The mentioned
signals have consecutive numbers, and are characterized by (respectively) 105—
very big noise, 108—high P wave amplitude (often mistakenly detected as R), 203
—high T wave amplitude (often mistakenly detected as R), and 222—the noise
very similar to the QRS complex.
626 T. ANDRYSIAK

Exemplary fragments of those signals are presented in Figure 2 as follows:

(a) signal 105, (b) signal 108, (c) signal 203, and (d) signal 222.
While conducting research, great attention was drawn to signals contain-
ing a significant amount of noise, artifacts and morphological changes. As
stated in the field literature, the aforementioned signals from the MIT-BIH

Figure 2. The fragments of ECG signals’ records from the MIT-BIH base: (a) signal 105, (b) signal
108, (c) signal 203, (d) signal 222.
 APPLIED ARTIFICIAL INTELLIGENCE 627

base are the most problematic and authors obtained the lowest results for
them. For every signal, the tests included both the number of false negative
and false positive detections, and the following parameters were determined:
sensitivity measure, detection error, positive predictive value, and the per-
centage of correct detection.

The setup of experiments

Generally, detection of the QRS complex can be divided into two basic
stages. The first consists of performing mathematical transformations of
the ECG signal with sharpening the QRS complex features and minimiza-
tion of the remaining signal’s features together with noise and distur-
bances. The second stage includes thresholding, the aim of which is to
find the local maximum further used for discovering points classified as
the QRS complexes.
The herewith presented concept of the QRS complex detection consists of
performing preprocessing of the tested signal by means of a particular
number of sparse decomposition layers of the ECG signal. As a result, we
obtain the wanted characteristic features deprived of noise and disturbances,
which undergo further selection. The prepared features’ vector, emerging as
an effect of the utilized decomposition, constitutes an optimal input vector
for a properly attuned neural network structure. It is worth noting that in
the analyzed context, the neural network structure is conditioned by both
the manner of the use of the analyzed signal’s sparse representation, and the
applied neural network. Due to the above, it is impossible to propose a
priori its definite and optimal structure. One possibility for determining the
parameters of such a solution is to conduct experiments of the basis of a
significantly large set of the ECG signal in order to estimate the number of
sparse decomposition layers, the value of the detection threshold, the
adjustment of the learning set, architecture of the neural network, and the
type of activation function.
It should be emphasized that the presented solution is not aligned with
the classic wavelet-neural networks theory, where the essence of the
conception is substitution of the neural networks activation functions
with wavelet functions. The learning algorithm is reduced to learning para-
meters (the scale and/or shifts) of the wavelet transform. In the presented
approach, we seek optimal parameters and structures for the described method,
however, not by means of learning but through a research experiment.

The selection of the signal’s sparse decomposition level

The sparse representation of the ECG signal realized by means of the MP or
OMP is an approximated procedure, which, in every iteration, presents the
628 T. ANDRYSIAK

Table 1. The energy of signal’s residuum depending on decomposition level.

Computational method
Decomposition level MP—Gabor MP—KSVD OMP—Gabor OMP—KSVD
1 28.31 34.31 35.13 39.75
2 32.15 40.15 42.08 48.66
3 47.34 51.46 55.89 61.77
4 59.67 69.04 72.42 80.54
5 78.98 83.25 88.34 92.79
6 86.26 88.94 90.66 94.03
7 88.57 90.07 92.09 94.88

signal at smaller and smaller levels of detail. Depending on the selection of a

dictionary in which this representation in taking place (complying with the
rule of energy behavior) we observe different dynamics of the decrease of the
signal residuum’s norm. Therefore, with the increase of the sparse decom-
position level, the energy of the already computed representation increases
(the sum of decomposition coefficients and corresponding dictionary atoms).
This process is presented by the results in Table 1 (percentage of energy of
the decomposed signal).
It is easy to notice that the choice of the dictionary has an impact on the
size of the sparse representation’s energy in the decomposition process. The
best results are obtained for the dictionary created by means of K-SVD, the
elements of which are the particular structures of the ECG signal, and not
time-frequency Gabor atoms. Assuming that we wish to present the ECG
signal at least at the level of 90% of its energy, and on the lowest possible level
of decomposition, the best solution is to use the OMP method with a
dictionary created by means of K-SVD. For this option, five is the suitable
level of decomposition. As it is easy to notice, the dynamics of the decrease of
the sparse representation’s energy at the following levels is not that big.
However, the manner of searching the elements in the dictionary (a full
searching or the use of tree structure) did not have any significant impact on
the results in Table 1. It influenced only the pace of computation.

The selection of the detection threshold value

In order to define the detection threshold value for the analyzed solution, we
tested the dependence of the number of the detected QRS complexes on the
change in the threshold values belonging to the set (0.4, 0.5, 0.6, 0.7, 0.8). On
the basis of the conducted research, it was noticed that the number of the
detected QRS complexes increases to the threshold value of 0.6 and then falls,
adapting the shape similar to quadratic function parabola. It was also
observed that in cases where the number of the detected QRS complexes
falls with the increase of the threshold, it approximately takes the shape of a
linear function.
 APPLIED ARTIFICIAL INTELLIGENCE 629

Table 2. The maximum and minimum detection error for the given threshold value.
Detection threshold value
Detection error 0.4 0.5 0.6 0.7 0.8
max 9.35 8.10 6.84 6.71 10.55
min 7.61 3.20 1.95 4.99 8.08

However, in most cases (91% of the tested ECG signals) the detection
threshold value at the level of 0:6  0:025 showed the best results. The lowest
sensitivity and the largest detection error were obtained for the remaining
two extreme thresholds; that is, threshold values of 0.4 and 0.8. Exemplary
detection error values are presented in Table 2.

The selection of a training set and architecture of the neural network

It should be noted that in the case of medical diagnostic systems, it is important
that the part of the set that participates in the learning process is well selected
and contains all the characteristic signal structures. In order to conduct the
learning process of a neural network, a division of the datasets was made onto
the training set and the test set with the use of methods based on the split-sample
and cross-validation. The achieved results of correct detections (98.40% for split-
sample, and 98.89% for cross-validation) confirm the advantage of cross-valida-
tion of a dataset in the learning processes of a neural network.
To determine the architecture of a neural network, a test was performed
on networks containing from one to five hidden layers. For every one of the
structures, the number of the hidden neurons was changed, according to the
consecutive set’s elements (10,20,30,40,50,60). Additionally, the accepted rule
was that for networks with p number of hidden layers, if the first layer is
constructed of q neurons, then the second will change, respectively, accord-
ing to the sequence 5, 10, . . ., q.
Analyzing five types of neural networks, the best results for the most signals
were achieved in the case of three hidden-layer structures and the optimal
number of neurons equals 40 (this number constitutes the average value of the
discussed set of hidden layers). It can be stated that the conducted research did
not confirm the literature’s accepted point of view concerning a better ability to
generalize the network with the smallest number of neurons (Table 3).

Table 3. The accuracy of detection in relation to the number of neurons in the hidden layer.
The number of The number of hidden layers
neurons of
hidden layers 1 2 3 4 5
20 97.68 98.12 98.21 97.32 97.34
30 97.70 98.46 98.53 97.82 97.82
40 97.78 98.57 98.66 97.70 97.70
50 97.85 98.45 98.48 97.67 97.69
60 97.67 98.24 98.27 97.56 97.54
630 T. ANDRYSIAK

Table 4. The accuracy of the QRS complex detection in relation to the activation function.
Number of hidden layers
Type of activation function 1 2 3 4 5
Sigmoidal 98.36 98.55 98.63 98.70 98.71
Tangesoidal 98.22 98.44 98.79 98.71 98.72

The accuracy of the QRS complex detection and the choice of activation
function
Selection of suitable architecture of the neural network is also connected to
the choice of a relevant activation function. For this reason, five types of
neural networks were tested for parameter values αf1; 5; 10; 15; 20; 25; 30g.
It was noticed that for the analyzed networks, by means of two activation
functions, the growth of the QRS detection level happens simultaneously
with the increase of parameter α value (the worst results, 91% of correct
classification, were achieved for α ¼ 1, and the best—over 97% for the
parameter value α ¼ 20). Averaged outcomes for particular types of neural
networks and the given activation function when α ¼ 20 are included in
Table 4.
The analysis of the outcomes presented in Table 4 points to the purpose-
fulness of using the sigmoidal function for neural networks with one or two
hidden layers. However, for networks with more layers, it is advisable to use
the tangesoidal function.

The detection outcomes of the presented solution

As a result of the conducted experiments, five decomposition levels realized
by means of the OMP algorithm and a dictionary found with K-SVD method
were established for the sparse representation of an ECG signal. A unidirec-
tional three-layer neural network with a 40-neuron hidden layer was used as
a model network. The detection threshold was fixed at the level of
0:6  0:025, and tangesoidal activation function was accepted. In Table 5,
the average measures of detection error and sensitivity were presented in
terms of time domain, frequency, time and frequency, neural networks, and
the presented solution (SR-NN).
Table 6 presents the size of detection error and sensitivity to the chosen
signals from the MIT-BIH base.

Table 5. Comparison of average measure values of detection and sensitivity for different
solutions.
Type of method
Average detection measures Time Frequency Time -Frequency Neural Networks SR-NN
Detection error 4.70 3.85 1.10 1.14 0.97
Sensitivity 98.60 98.44 99.48 98.99 98.98
 APPLIED ARTIFICIAL INTELLIGENCE 631

Table 6. The comparison of detection error and sensitivity for signals: 105, 108, 203, and 222.
Type of signal
Detection measures 105 108 203 222
Detection error 1,24 1.48 1.19 0.71
Sensitivity 99.64 98.92 99.41 98.99

Analyzing the results presented in Tables 5 and 6, and comparing them to

those published in the field literature, a conclusion can be drawn that the
largest detection error is obtained for signal 109, containing high R wave
amplitude. This error was smaller in relation to the time and frequency
methods, but comparable to time-frequency methods and those utilizing
neural networks. However, the smallest detection error was noted for signal
222, containing noise similar to the QRS complex. So low detection error
value is probably a result of the signal’s sparse representation feature con-
nected with elimination of noise and artifacts. Analyzing sensitivity, however,
it is easy to notice that the highest value is obtained for signal 105, which is
also characterized by a significant amount of noise.

Conclusions
The aim of the presented solution is to create a simple and effective method
joining features and advantages of the sparse representation of the analyzed
ECG signal with classification attributes of the modeled neural network. The
sparse representation ensures elimination of minor distortions, noise, and
artifacts in the process of adaptive decomposition with the use of defined
redundant dictionaries. It also constitutes the basis for optimal selection of
the features. However, the parameters modeled in the experimental process
and the neural network structure allow obtaining satisfactory detective attri-
butes of the QRS complexes in the analyzed ECG signal. The solution
emerged in the aforementioned way can be accepted as a new approach
not present in the field literature yet.
This article presents the use of machine learning techniques for analysis and
anomaly detection in ECG signals. This process was performed in two steps.
The first stage was sparse decomposition of signal on the grounds of redundant
dictionaries of base functions. Two types of dictionaries were tested. The first
included elements that were analytical base functions, and the second presented
essential components of an ECG signal searched by means of the K-SVD
algorithm. The sparse representation of a signal, however, was performed in
an adaptive manner, with the use of the matching pursuit algorithm. In each
step of the MP or OMP algorithm, we made a linear signal decomposition in
respect of features belonging to the dictionary with redundancy in order to find
the relevant structural characteristics of the processed ECG signal. In the second
632 T. ANDRYSIAK

stage, we used the method of detection of the QRS complex using multivariate
unidirectional neural network with backpropagation of error. Elements from
sparse decomposition in the first stage constitute data for the exploited neural
network. To facilitate such a solution, we tested the selection of the right
architecture of the neural network and the impact of specific activation func-
tions on the recognition results. Anomaly detection was realized by estimation
of the elements’ energy of the ECG signal’s sparse representation in the area of
the structure of the found QRS.
As the result of the conducted experiments, the elements that were defined are
the required parameters’ set, the optimal structure, and essential features of the
proposed solution. At the preprocessing step, the signal’s sparse representation
was realized as adaptive decomposition to the fifth level, copying over 90% of the
ECG signal. Due to such an approach, the noise, disturbances, and artifacts were
eliminated, and the QRS structure was highlighted in the analyzed ECG signal.
The best results of the adaptive decomposition were obtained for the OMP
method and for a dictionary formed by means of the modified (through
decorrelation of selected atoms) K-SVD algorithm. So, the created set of features
(coefficients and their corresponding dictionary atoms) constituted the optimal
input vector for the properly adjusted neural network structure. As the network
model, in the experimental process, we adopted the unidirectional, three-layer
neural network with 40 neurons in the hidden layer. The detection threshold was
established at the level of 0:6  0:025, and the tangesoidal activation function
was adopted.
It is worth noticing that the conducted research did not confirm the
viewpoint functioning in the field literature that a network with a smaller
number of neurons has a better generalization ability, because, in the pre-
sented solution, number 40 constitutes the middle value of the analyzed set of
hidden layers.
Performance of the proposed method was tested with the use of a public
free base of ECG signals from the MIT-BIH Arrhythmia Database. The
obtained experimental results confirmed high efficiency of the presented
method of anomaly detection in the analyzed ECG signals.

References
Addison, P. S., J. N. Watson, G. R. Clegg, M. Holzer, F. Sterz, and C. E. Robertson. 2000.
Evaluating arrhythmias in ECG signals using wavelet transforms. IEEE Engineering in
Medicine and Biology 19 (5):104–109. doi:10.1109/51.870237.
Aharon, M. L., M. Elad, and A. Bruckstein. 2006. K-SVD: An algorithm for designing
overcomplete dictionaries for sparse representations. IEEE Transactionson Signal
Processing 54:4311–4322. doi:10.1109/TSP.2006.881199.
Al-Fahour, A. S., and I. Howitt. 1999. Combine wavelet transform and radial basic neural
networks for classifying life-threatening cardiac arrhythmias. Medical and Biological
Engineering Computational 37:566–573. doi:10.1007/BF02513350.
 APPLIED ARTIFICIAL INTELLIGENCE 633

Alfaouri, M., and K. Daqrouq. 2008. ECG signal denoising by wavelet transform thresholding.
American Journal of Applied Sciences 5 (3):276–281.
Augustyniak, P. 2001. Przetwarzanie sygnałów elektrodiagnostycznych. AGH Uczelniane
Wydawnictwo Naukowo – Dydaktyczne. AGH Kraków.
Berkhin, P. 2006. A survey of clustering data mining techniques. In Grouping multidimen-
sional data, eds. J. Kogan, C. Nicholas, M. Teboulle, Springer, 25–71.
Besterman, E., and R. Creese. 1979. Waller – pioneer of electrocardiography. British Heart
Journal 42 (1):61–64. doi:10.1136/hrt.42.1.61.
Celler, B. G., and P. de Chazal. 1998. Low computational cost classifiers for ECG diagnosis
using neural networks. In Proceedings of the annual international cong. IEEE engineering in
medicine and biology society, 1337–1340. IEEE.
Daskalov, I. K., and H. Christov. 1999. Electrocardiogram signal preprocessing for automatic
detection of QRS boundaries. Medical Engineering and Physics 21:37–44. doi:10.1016/
S1350-4533(99)00016-8.
Davis, D., S. Mallat, and M. Avellaneda. 1997. Adaptive greedy approximations. Journal of
Constructive Approximation 13:57–98. doi:10.1007/BF02678430.
de Chazal, P., M. O’Dwyer, and R. B. Reilly. 2004. Automatic classification of heartbeats using
ECG morphology and heartbeat interval features. IEEE Transactions on Biomedical
Engineering 51 (7):1196–1206. doi:10.1109/TBME.2004.827359.
de Chazal, P., and R. B. Reilly. 2000. A comparison of the use of different wavelet coefficients
for the classification of the electrocardiogram. Paper presented at Proceedings of the 14th
International Conference on Pattern Recognition, September 3–8, 2000, Barcelona.
Durka, P. 1996. Time-frequency analyses of EEG. Ph.D. dissertation, Institute of Experimental
Physic, Warsaw University.
Einthoven, W. 1924. The string galvanometer and the measurement of the action currents of
the heart. Nobel Lecture. Republished in 1965 in Nobel Lectures, Physiology or Medicine
1921–1941. Amsterdam: Elsevier.
Elad, M. 2010. Sparse and redundant representations: From theory to applications in signal and
image processing. Springer New-York.
Gabor, D. 1946. Theory of communication. Journal of Institution Electrical Engineering 93
(26):429–457.
Gharavi-Alkhansari, M., and T. S. Huang. 1998. A fast orthogonal matching pursuit algo-
rithm. The presented at International Conference on Acoustics Speech and Signal
Processing, May 15, Seattle, Washington. 1389–1392.
Hamilton, P. S., and W. J. Tompkins. 1988. Adaptive matched filtering for QRS detection,
engineering in medicine and biology society. In Proceedings of the Annual International
Conference of the IEEE 4–7:147–148. IEEE.
Harikumar, R., and S. N. Shivappriya. 2011. Analysis of QRS detection algorithm for cardiac
abnormalities – A review. International Journal of Soft Computing and Engineering 1
(5):2231–2307.
Hassani, M., I. Fachrudin, and M. Hubin. 1994. Pattern recognition of ECG wavesforms using
wavelet transform experimental results in the sudden infant death syndrome monitoring.
MPBE 94 Conference 2:212–216.
Hsiung, H. S. H., C. K. Cheng, and G. J. Jan. 1989. A real-time QRS detection and delineation
algorithm based on the inflection angle of the ECG waveform. In Proceedings of the annual
international conference of the IEEE engineering 9-12, 138–139. IEEE.
Jost, P., P. Vandergheynst, and P. Frossard. 2005. Tree-based pursuit: algorithm and proper-
ties. Signal Processing Institute Technical Report, TR-ITS-2005.013, Swiss Federal Institute
of Technology Lausanne (EPFL), Switzerland.
634 T. ANDRYSIAK

Kwiatkowski, W. 2003. Wstęp do cyfrowego przetwarzania sygnałów. Wydawnictwo

Wojskowej Akademii Technicznej, Warszawa, Poland.
Liu, Q., Q. Wang, and L. Wu. 2004. Size of the dictionary in matching pursuit algorithm. IEEE
Transactions on Signal Processing 52 (12):3403–3408. doi:10.1109/TSP.2004.837423.
Mallat, S., and W. L. Hwang. 1992. Singular detection and processing with wavelets. IEEE
Transaction on Information Theory 38:6. doi:10.1109/18.119727.
Mallat, S., and Z. Zhang. 1993. Matching pursuit with time-frequency dictionaries. IEEE
Transactions on Signal Processing 41:3397–3415. doi:10.1109/78.258082.
MIT-BIH Arrhythmia Database. 2010. [Online]. www.physionet.org/physiobank/database/
mitdb (accessed on 15.12.2015).
Muller, B., J. Reinhardt, and M. T. Strickland (1995). Neural networks: An introduction. vol. 1
in Physics of Neural Networks, Berlin, Heidelberg: Springer-Verlag.
Natarajan, B. K. 1995. Sparse approximate solutions to linear systems. SIAM Journal of
Computation 24:227–234. doi:10.1137/S0097539792240406.
Okajima, K. 1998. The Gabor function extracts the maximum information from input local
signals input local signals. Neural Networks 11:435–439. doi:10.1016/S0893-6080(98)00008-2.
Osowski, S., L. Tran Hoai, and T. Markiewicz. 2004. Support vector machine-based expert
system for reliable heartbeat recognition. IEEE Transaction on Biomedical Engineering 51
(4):582–589. doi:10.1109/TBME.2004.824138.
Pan, J., and W. J. Tompkins. 1985. A real-time qrs detection algorithm. IEEE Transactions on
Biomedical Engineering 32 (3):230–236. doi:10.1109/TBME.1985.325532.
Pati, Y. C., R. Rezaiifar, and P. S. Krishnaprasad. 1993. Orthogonalmatching pursuit:
Recursive function approximation with applications to wavelet decomposition. Asilomar
Conference on Signals, Systems and Computers 1:40–44.
Poli, R., and S. Cagnoli. 1995. Genetic design of optimum linear and non-linear QRS
detection. IEEE Translation Biomedical Engineering 42:1137–1141. doi:10.1109/10.469381.
Rubinstein, R., M. Bruckstein, and M. Elad. 2010. Dictionaries for sparse representation
modeling, Proceedings of the IEEE 98:1045–1057.
Sanamdikar, S. T., S. T. Hamde, and V. G. Asutkar. 2015. A literature review on arrhythmia
analysis of ECG Signal. International Research Journal of Engineering and Technology 2
(3):307–312.
Shyu, L. Y., Y. H. Wu, and W. Hu. 2004. Using wavelet transform and fuzzy neural networks
for VPC detection from the Holter ECG. IEEE Transaction on Biomedical Engineering 51
(7):1269–1273. doi:10.1109/TBME.2004.824131.
Srinagesh, M., P. Sarala, and K. D. Aparna. 2013. ECG wireless telemetry. International
Journal of Engineering and Innovative Technology 2 (8):75–78.
Tang, X., and S. Lan. 2014. A frame work of automatic analysis system of electrocardiogram
signals. International Journal of Signal Processing. Image Processing and Pattern Recognition
7 (2):211–222. doi:10.14257/ijsip.2014.7.2.20.
Troop, J. A. 2004. Greed is good: Algorithmic results for sparse approximation. IEEE
Transactions on Information Theory 50 (10):2231–2242.