You are on page 1of 14

See

discussions, stats, and author profiles for this publication at: http://www.researchgate.net/publication/265016066

Heart monitoring systems-A review


ARTICLE in COMPUTERS IN BIOLOGY AND MEDICINE NOVEMBER 2014
Impact Factor: 1.48 DOI: 10.1016/j.compbiomed.2014.08.014

CITATION

DOWNLOADS

VIEWS

24

63

2 AUTHORS, INCLUDING:
Puneet Jain
Indian Institute of Technology Jodhpur
4 PUBLICATIONS 6 CITATIONS
SEE PROFILE

Available from: Puneet Jain


Retrieved on: 20 June 2015

Computers in Biology and Medicine 54 (2014) 113

Contents lists available at ScienceDirect

Computers in Biology and Medicine


journal homepage: www.elsevier.com/locate/cbm

Heart monitoring systemsA review$


Puneet Kumar Jain n, Anil Kumar Tiwari
Center of Excellence in Information and Communication Technology, Indian Institute of Technology Jodhpur, Rajasthan, India

art ic l e i nf o

a b s t r a c t

Article history:
Received 15 April 2014
Accepted 12 August 2014

To diagnose health status of the heart, heart monitoring systems use heart signals produced during each
cardiac cycle. Many types of signals are acquired to analyze heart functionality and hence several heart
monitoring systems such as phonocardiography, electrocardiography, photoplethysmography and
seismocardiography are used in practice. Recently, focus on the at-home monitoring of the heart is
increasing for long term monitoring, which minimizes risks associated with the patients diagnosed with
cardiovascular diseases. It leads to increasing research interest in portable systems having features such
as signal transmission capability, unobtrusiveness, and low power consumption. In this paper we intend
to provide a detailed review of recent advancements of such heart monitoring systems. We introduce the
heart monitoring system in ve modules: (1) body sensors, (2) signal conditioning, (3) analog to digital
converter (ADC) and compression, (4) wireless transmission, and (5) analysis and classication. In each
module, we provide a brief introduction about the function of the module, recent developments, and
their limitation and challenges.
& 2014 Elsevier Ltd. All rights reserved.

Keywords:
Heart monitoring system
Cardiovascular diseases
Cardiography
Electrocardiography
Phonocardiography
Photoplethysmography
Seismocardiography

1. Introduction
Worldwide, the number of patients of cardiovascular diseases
(CVD) is huge [1]. Mortality caused by CVD in 2008 was 17.3
million which represents 30% of global deaths. In the U.S. alone,
2200 persons lose their life due to CVD each day [2]. According to
American Heart Association (AHA) report, the total cost of CVD
and stroke in the U.S. for 2008 is estimated to be 298 billion dollar
[1]. 80% of the total mortality caused by CVD occur in low and
middle-income countries.
These gures indicate need of systems that should be (1) sensitive to detect CVD at early stage, (2) capable of continuous
monitoring, (3) light weight for portability, (4) cost effective. Lack
of early stage detection and hence delay in medication causes
heart diseases to extent at a level where it is difcult to cure [3].
Persons diagnosed with CVD need continuous monitoring of
health status of their heart as they are at a higher risk to their
lives as compared to the normal persons. According to the Heart
Association, people diagnosed with CVD have 46 times higher
mortality than normal one [4]. Portability of such systems makes it
highly useful for elderly patients as this minimizes visits to clinics
or hospitals. A cost effective system will emphasis the use of heart
monitoring systems in low and middle income countries. Proper

This paper was not presented at any IFAC meeting.


Corresponding author. Tel.: 91 9252903393.
E-mail addresses: puneetjain@iitj.ac.in (P.K. Jain), akt@iitj.ac.in (A.K. Tiwari).

http://dx.doi.org/10.1016/j.compbiomed.2014.08.014
0010-4825/& 2014 Elsevier Ltd. All rights reserved.

diagnosis reduces the mortality caused due to CVD which ensues


economic up-lift of the country [5].
Due to the problems mentioned above, a lot of work has been
done in development of a diagnostic efcient system [610].
Keeping in view that heart monitoring systems would be used in
different socio-economic conditions, rural-urban population, and
deciency of availability of cardiac experts [11], recent research is
focused towards system features such as cost effective, portability,
easy diagnose process, and signal transmission capability. In view
of these developments, we propose a review of various work done
in this area.
Portable heart monitoring systems are used in two manners, as
shown in (Fig. 1), one is on-site and other is off-site. In on-site
monitoring, the acquired heart signal is processed on the patient
site, without transmitting it to the remote site. While in off-site,
the acquired heart signal is transmitted to a remote site using
a wireless module. On-site heart monitoring system has advantages in the case of low latency feedback is required or wireless
access is not accessible. Furthermore, it eliminates data transmission and hence eliminates the radio power consumption. However,
the on-site monitoring has limitation that it has only a set of
general diagnosis steps and thus unable to perform a detailed
diagnosis. On the other hand, in off-site monitoring, diagnosis is
performed at remote location with high computation capable
processors and supports input from a cardiologist. This makes it
suitable for accurate and detailed diagnosis. It is attractive because
of higher processing capability and less power restrictions on such
remote computation. Off-site monitoring also reduces the false
alarm rate and thus reduces visits to clinics or hospitals. In view of

P.K. Jain, A.K. Tiwari / Computers in Biology and Medicine 54 (2014) 113

Fig. 1. Heart monitoring system.

these advantages of off-site monitoring, this paper is intended to


provide a detailed review of recent research in the off-site
monitoring system.
A typical off-site heart monitoring system consists of ve
modules as shown in (Fig. 2). The system with rst four modules
body sensor, signal conditioning, ADC and compression, and
wireless module is situated at the patient site. While the fth
module that is analysis and classication module is situated at
remote site which can be any computational device with high
computational ability.
Heart monitoring system uses signals produced by heart to
diagnose its health status. It extracts diagnostic features from the
acquired signal which caries information of heart functionality
such as re-polarization, depolarization, and valve movements.
Analysis of these features leads to specic health status of heart
such as normal, arrhythmic, myocardial infarction, regurgitation,
and stenosis. However, extraction of diagnostic features from the
heart signal is challenging due to its non-stationary nature and the
presence of noises such as muscles movement noise and environment noise in the signal. In this paper, we reviewed recent
developments in the area of heart monitoring systems which are
portable and have good diagnostic efciency.
Rest of this paper is organized as follow. Physiology of heart
and cardiac cycle in Section 2 and brief introduction about
important cardiovascular diseases in Section 3 are given for reader's simplicity. Section 4 describes the recent developments in the
body sensors. While different approaches of signal conditioning
have been reviewed in Section 5. In Section 6, analog to digital
conversion and compression techniques are presented. Section 7
discusses various wireless transmission technologies and Section 8
gives comprehensive review of noise removal algorithms, analysis
and classication techniques for heart signals. Conclusion and
potential research area have been presented in Section 10 followed
by important references.

2. Physiology of heart and cardiac cycle


Since understanding of various components of heart monitoring systems needs knowledge of heart functioning, a relevant
physiology of heart is described in this section.

Fig. 2. Off-site heart monitoring system.

Heart is a prominent organ of human body. It supplies


replenish oxygen to each part of the body and removes waste of
each cell. Physiologically, heart comprises of four chambers named
as left and right ventricles and left and right atrium as shown in
Fig. 3. There are two atrioventricular valves namely tricuspid valve
and mitral valve. As can be seen in Fig. 3, tricuspid valve separates
right atrium and right ventricle while mitral valve separates left
atrium and left ventricle. Aortic valve and pulmonary valve jointly
called as semilunar valves separate left and right ventricles from
aorta and pulmonary artery respectively. At rest, each cell of the
heart muscle has a negative charge, called the membrane potential. Due to rapid change of membrane potential towards zero, due
to inux of positive cations (Na and Ca ), an electrical
impulse is generated at sinoatrial node. From sinoatrial node, the
impulse spreads over both the atrium and both the ventricles.
Presence of the impulse causes the contraction of atrium and
ventricles sequentially.
Contraction of both atrium pushes the blood into respective
ventricles. Then the impulse spreads all over left and right
ventricles which causes the contraction of both the ventricles.
This contraction results to closing of both atrioventricular valves
and opening of both semilunar valves. During this contraction
phase, oxygenated blood from left ventricle ows into the body

P.K. Jain, A.K. Tiwari / Computers in Biology and Medicine 54 (2014) 113

signal. Fig. 4 shows signals of one cardiac cycle acquired using


electrodes,
stethoscope,
accelerometer,
and
diode
as
described below.

4.1. Electrodes (electrocardiography)

Fig. 3. Anatomy of the human heart.

through aorta while deoxygenated blood from right ventricle ows


into the lungs through pulmonary artery for oxygenation. Since
body and lungs receive blood due to contraction process, pressure
in body lungs becomes higher than pressure in atrium. Due to this
pressure difference, now, blood ows into left and right atrium
from lungs and body respectively. This process completes a
cardiac cycle.

Initial electrocardiography (ECG) was based on string galvanometer and was invented by Willem Einthoven in 1903. As
discussed in Section 2, an electrical impulse originates at sinoatrial
node and then travels through atria and ventricles. ECG measures
the electrical activity of heart using electrodes placed on both side
of the heart. The measured signal consists of different waves
named as P, Q, R, S, T and U as shown in Fig. 4(a). P wave
represents atria contraction while Q, R and S waves (called as QRS
complex) reect contraction of both left and right ventricles. T
wave represents relaxation of ventricles and U wave is caused by
the relaxation of inter-ventricular septum. Thus duration and
amplitude of these waves provide signicant information for
diagnosis of health status of heart.
Extraction of duration and amplitude of some of the waves (P, T
and U) is difcult due to very weak amplitude, typically in the
range of 100300 V [14]. ECG signal lies in frequency band of 1
250 Hz, where icker noise is dominant and common-mode
interference from the main power line is likely to interfere with

3. Important cardiovascular diseases


Heart beats 100,000 times and pumps 2000 gallon blood in
a day [12]. Heart function gets affected due to many factors such as
psychosocial stress, smoking, excessive use of alcohol, malnutrition, lack of physical activity, and congenital diseases [13]. These
factors may affect electrical activity of heart, structure of heart,
and arteries. Due to these defects, different heart diseases occur.
Dysfunction of electrical conduction system causes diseases such
as sinus arrhythmia, atrial uttering, atrial brillation, ventricular
brillation, atrioventricular block, and bundle branch block.
Defects in structure of the heart cause regurgitation, stenosis,
enlargement of chambers, ventricular septum defect, etc. Defects
in arteries cause hypertension, stroke, myocardial infarction, etc.
According to American Heart Association report [2], following
are the most common heart diseases. Myocardial infarction (heart
attack), which occurs due to blockage in coronary arteries which
supply blood to the heart cells. Insufciency of blood supply
causes dying of the heart cells and consequentially heart muscles
lose pumping capacity. Hypertensive heart diseases occur when
the blood pressure in arteries is much higher than the normal.
High pressure causes stiffness of arteries, consequentially blood
ow gets affected. Arrhythmia is fast and irregular beating of the
heart. Congenital heart diseases are those which are present at the
time of birth. Heart failure is a condition which indicates inability
of the heart to pump the blood. Mortality caused by myocardial
infarction was 21.6% of all mortality due to CVD in U.S., while 9.2%
by hypertensive heart disease, 5.7% by arrhythmia, 8.0% by congenital heart diseases, and 6.9% by heart failure was accounted [2].
Since heart problems need diagnostic system, the following section discusses about body sensors used to acquire heart signal for
this purpose.

4. Body sensors
Different body sensors acquire heart signals in different forms
such as electrical signal, acoustic signal, seismic signal, and optical

Fig. 4. Signals of one cardiac cycle: (a) electrocardiography signal, (b) phonocardiography signal, (c) seismocardiography signal, and (d) photoplethysmography signal.

P.K. Jain, A.K. Tiwari / Computers in Biology and Medicine 54 (2014) 113

ECG signal [15]. To overcome these problems signal conditioning,


discussed in Section 5, is essential.
Major problem with ECG is to establish good electrical conductivity between skin and electrodes. Some of the ECG sensors,
widely used in practice, can be classied into following three
categories.
4.1.1. Wet sensors
In these type of sensors, AgAgCl electrodes are attached to the
skin using gel which provide a conducting medium for charge
transfer between the electrodes and the body [6]. These sensors
provide good signal quality, but it is inconvenient in terms of long
term wear-ability due to use of gel which creates irritation and
etching problem [6]. Moreover, For signal acquisition, attachment
of electrodes to different points on the body restricts patient's
mobility. The acquired signal quality may deteriorate due to sweat
[16] and due to gel dehydration [17].
4.1.2. Dry sensors
These sensors use a metal plate direct placed on the skin
without the use of gel. Thus the problem of irritation and etching
caused by gel has been eliminated [18]. Although, it still has
a direct contact with the skin. Dry sensors are robust to environment noises and sweat noise but more vulnerable to motion noise
compare to wet sensors. Quality of the signal acquired using these
sensors depends on the composition of the materials and the size
of the electrode [17]. Increasing size of the electrode gives better
capacitance and consequentially good signal quality but it
decreases the patient's convenience.
4.1.3. Capacitive coupled sensors
Capacitive coupled (CC) sensors avoid direct contact with the
skin that minimizes patients' inconvenience as stated in the
previous subsections, (a) and (b). A thin layer of insulator is placed
between the body and metal-plate sensing electrode [6]. The
electrode, together with the skin and insulator, forms a capacitance that conveys the ECG signal from the body to the sensor.
Sensitivity of such sensors increases with the value of capacitance,
which can be increased by increasing electrode area, by reducing
thickness of insulator, and by using insulator with high dielectric
constant. General expression of capacitance is given as follows:
C

0 A
d

where 0 is the dielectric constant, A is the electrode area, and d is


the thickness of insulator. CC sensors have been developed on
chair [19,20], on bed [21] and textiles [22]. Development of sensors
on chair, bed and textiles supports continuous heart monitoring
even when working in ofce and sleeping. CC sensors are highly
sensitive to motion noise as in case of dry sensors. This is because,
a movement of electrode changes the coupling capacitance and
consequentially the acquired signal [6].
A comparative study between three types of electrodes is
presented in (Table 1). For clinical use, where simplicity of
operation, less processing time, and good signal quality are
preferred, wet sensors are suitable. Additionally, the availability,
relative cheapness, and disposability of wet electrodes overcome
hygiene concerns. While, the dry and capacitive electrodes are
convenient in use, and consistent in performance. These features
make these sensors suitable for long term, and unsupervised
monitoring. However, the performance of these types of electrodes
depends on the electrode geometry. Furthermore, these electrodes
require proper shielding and settling time to perform comparable
to, or better than wet electrode. Researchers in the past have made
numerous attempts to overcome these problems [16].

4.2. Stethoscope (phonocardiography)


Stethoscope was invented by Ren Laennec in 1816. It is
basically a transducer which converts vibration signal into acoustic
signal. A phonocardiogram (PCG) is a plot of acoustic signal,
acquired by stethoscope. Stethoscope makes PCG a highly portable, low cost, and non-invasive cardiography technique [10]. PCG
signal, as shown in Fig. 4(b), consists of two classical heart sounds
known as S1 (Lub) and S2 (Dub). S1 is generated due to closing of
the tricuspid and mitral valves. It is composed of energy in 40
65 Hz frequency band and 130200 ms time duration [10]. S2 is
generated due to closing of aortic and pulmonary valves and lies
within 4565 Hz frequency band and 100150 ms time duration
[10]. Period between S1 and S2 is called as systole, while S2S1
phase is called as diastole. There are two more components in PCG
signal called S3 and S4 and they rarely occur. Murmurs are
additional sounds that lie within a frequency band of 100
500 Hz [10]. They indicate diseases in heart such as aortic stenosis,
pulmonary stenosis, mitral regurgitation, and mitral stenosis.
Stenosis and regurgitation are valvular diseases caused by stiffness
of valves and improper closing of valves respectively. Stenosis
restricts proper blood ow, while regurgitation causes blood to
ow in opposite direction to the normal [10]. PCG has high
potential to detect these valvular diseases and important cardiovascular diseases except myocardial infarction and congenital
heart diseases. However it indicates the abnormality caused by
myocardial infarction and congenital heart diseases. PCG signal
may get contaminated by noises occurring due to breathing,
movement of stethoscope while recording, ambient sources, etc.
As a result of these noise problems, many ltering techniques have
been developed to minimize affect of noise on PCG signal. Detailed
discussion about the same is given in Section 8.
4.3. Accelerometer (seismocardiography)
Seismocardiography (SCG) is another non-invasive technique
because it works using accelerometer, a non-invasive device.
It measures mechanical vibrations which are generated by heart
movement and transmitted to the chest wall [23]. As shown in
Fig. 4(c), SCG contains waves corresponding to atrial contraction
(ATC), mitral valve closing (MC), aortic valve opening (AO), point of
maximal acceleration in the aorta (MA), aortic valve closure (AC),
mitral valve opening (MO), and rapid lling of left ventricle (RF)
[24,25]. Shapes of these waves give signicant information about
health status of heart. SCG is convenient to patients, as there is no
need of multiple electrode contacts as in ECG. However, obtaining
a clean SCG signal from an accelerometer is difcult, because of
interference due to breathing.
4.4. Diode (photoplethysmography)
In photoplethysmography (PPG) uorescent body parts such as
earlobe and nger are illuminated with lights of different wavelengths emitted from light emitting diodes (LEDs). Then intensity
of transmitted or reected light is measured by photo-diode [26].
The measured intensity varies in time with the heart beat because
blood vessels expand and contract with each heartbeat. The PPG
waveform is composed of an ac component and a quasi-dc
component [27], as shown in Fig. 4(d). The ac component is
associated with heart beat and has fundamental frequency typically around 1 Hz. The quasi-dc component, superimposed with ac
component, relates to the respiration system. As the ac component
of the PPG signal is in synchronization with heart beat, information about heart rate can be extracted from it. PPG uses LEDs and
photo-diode which makes it low cost, non-invasive, easy to use
and portable system. Since it operates optically, it is not

P.K. Jain, A.K. Tiwari / Computers in Biology and Medicine 54 (2014) 113

Table 1
Three types of electrodes.
Characteristics

Wet electrodes

Dry electrodes

Capacitive electrodes

Signal acquisition

AG/AGcl electrodes, uses electrolyte

Signal quality

Low contact impedance ensues good


signal quality
Gel dehydrate with time,
which reduces quality of the signal

Benign metal (stainless steel), no


electrolyte
Depends on electrode geometry

Metal or semiconductor, no direct


contact
Depends on electrode geometry
After settling time good performance
due to reduction in skin/electrode
interface
No direct contact, fabricated in cloth,
chair, bed, that increases convenience
to user

Size

Use of electrolyte cause etching


problem, removal of gel is unpleasant
and time consuming, and toxicological
concern
Lightweight

After settling time good performance


due to reduction in skin/electrode
interface
Direct contact with skin, which may
cause irritation

Heavy due to shielding

Noise vulnerable

Moving charge sensitivity

Movement artifact

cost

Lower cost

Expensive

Bulky due to required circuitry for


buffers and extra cables for power
Motion and environment artifact,
electric eld problem
Expensive

Consistency

Convenience

intrinsically susceptible to capacitive coupling interference as in


ECG [28]. However, photo-diodes are sensitive to natural and
articial light sources. PPG based heart monitoring systems are
unobtrusive as size and weight of the device of PPG is low. PPG is
generally used in direct contact with the patient's skin as in the
case of other systems (ECG, PCG, SCG). But in the case of neonates,
skin-damage, direct contact to skin is not feasible and hence
Huelsbusch and Blazek [29] proposed a remote PPG (rPPG) that
can acquire PPG signal without the contact with skin. The main
concern with rPPG is its sensitivity to the subject motion.
Portability of the system is induced by small size and low
weight sensor, used by the system. ECG, PCG, SCG and PPG are
portable as they satisfy these conditions. Portable diagnostic
system is useful in many scenarios. Most use of the portable
system is for long term monitoring, which is required for acute
symptom detection and early diagnosis of the heart diseases.
However the acceptance of a portable system depends on the
diagnostic efciency, performance of the system in different
environmental conditions, comfort to the user, easy to operate,
and cost.
ECG signal contains information about the electrical activity of
the heart. Thus provides better insight on the issues related to
electrical conduction abnormality of the heart. While the PCG
signal acquires acoustic sounds produced by the heart valves
(mechanical action) and thus useful in diagnosis of the valvular
diseases. Due to different source of producing these signals (ECG
and PCG), the existence of a problem (e.g. structural abnormalities)
in PCG signal does not imply the existence of the same problem in
ECG signal and vice versa. SCG signal is also produced by the
mechanical action (acceleration of the heart), measures compression waves. Acceleration is a second derivation of the displacement and that is why SCG signal contains more information
compared to the PCG signal. PPG signal provides only limited
information about the heart. It measures the blood variation in the
blood vessels. Although, with the combination of ECG or PCG it has
been used for the calculation of pulse transit time (PTT), which is
an important diagnostic parameter in the case of obstructive sleep
apnea detection and blood pressure measurement.
For clinical use, all methods are suitable as signal to noise ratio
(SNR) remains high. At home, in the presence of the environmental and motion noise, robustness of the sensor to the noises is
a major issue. In case of ECG, as discussed in (Table 1), wet sensors
are robust to noise while dry and capacitive sensors are vulnerable
to noise. PCG is more vulnerable to patient's motion noise and
environmental noise compared to the ECG. On the other hand, SCG
is robust to the motion noise and environmental noise.

PPG is the most suitable technique in terms of portability and


used widely for heart rate calculation. Problem with the ECG is that
it uses a gel which causes etching problem, and reduces comfort of
the patient and hence reduces acceptability for long term monitoring. PCG has advantages over ECG in terms of comfort of the patient
and easy to operate. SCG is superior to the PCG in terms of comfort
because of the low weight (o3 g) accelerometers.

5. Signal conditioning
Heart signals, acquired by different body sensors, often get
contaminated by noise components such as icker noise,
common-mode interference, power-line interference, and baseline
wandering [15]. Also, amplitude of the acquired signal is typically
low. A signal conditioning module typically consists of algorithm
for noise minimization and amplier to amplify low amplitude
signals. This module operates on signals in analog domain. Power
consumption of this module used to be low so as to support long
term operability of heart monitoring systems. For the same
purpose, Rieger [30] proposed a variable gain circuit consisting
of a continuous-time input stage using lateral bipolar transistors.
Spinelli et al. [31] proposed a driven right leg circuit to reduce
common-mode interference. Gomez-Clapers and Casanella [18]
used dual ground conguration to reduce the noise caused by
power line interference and base line wandering. Since most of the
heart monitoring systems are digital in nature and need communication for remote monitoring, the following section discuses
about analog to digital conversation (ADC) and compression
algorithms.

6. Analog to digital conversion and compression


Heart signals (analog) are converted into digital signals for its
processing on digital computers. This is done by sampling the heart
signal and quantizing the sampled values. This process is done on
Digital Signal Processor (DSP) called Analog to Digital Converter (ADC).
Selection of the DSP depends on the desired sampling rate, number of
bits to be used for quantisation, operating frequency and power
consumption. As described in Table 2, MSP based processors have
lower power consumption compared to the PIC based processors,
while the PIC based processors have higher operating frequency
compared to the MSP based processors. Both types of processors
provide multiple working and idle modes according to the required
computational power, to optimize the power consumption. DSP with

P.K. Jain, A.K. Tiwari / Computers in Biology and Medicine 54 (2014) 113

low power consumption supports long term monitoring of the heart.


To optimize the power consumption, Bachmann et al. [32] proposed a
DSP with the capability to perform in different power modes according to the required accuracy and available computational power.
Power consumption was optimized at different abstraction layers from
application optimization and mapping to system.
Conventional sampling techniques, sample signals at or above
Nyquist rate, which ensues perfect reconstruction of the signal.
Nyquist rate is twice the maximum frequency component present
in the signal to be sampled. Typically, heart signal components are
below 1 kHz frequency and hence, as per Nyquist rate, 2 K samples
per second are sufcient to avoid aliasing error. However, even 2 K
samples per second sampling rate generates a huge number of
samples, if the heart is monitored for a long time. Consequentially,
the power requirement of DSP increases as the number of samples
to be processed is huge. In spite of it, compressed sensing (CS)
enables sub-Nyquist sampling of signals.
Compressed sensing (CS) is a data acquisition approach that
requires only a few incoherent measurements to compress signals
that are sparse in some domain [38]. Let [] be an original input
vector of dimension N  1 and [] is the N  N sampling basis or
sparsifying matrix containing orthonormal basis (such as a wavelet
basis) [ 1 ; 1 ; ; n ]. Then [X], sparse in [] domain with length
N can be found as
X 

Then output compressed vector is dened as


Y X

where [] is the M  N measurement or sensing matrix. So, we get


an output vector [Y] with length M, where M oN. CS captures M
measurements from N samples using random linear projections.
Now, as the lower number of measurements were taken than the
original signal, non-linear optimization techniques are used to
reconstruct the original signal [38]. Reconstruction of the signal
can be achieved as
b J1
Min J X

b
subject to Y X

Perfect reconstruction of the signal depends on the incoherence


between [] and [] matrices. Thus, random matrices can be used
as a measurement matrix because random matrices are, with high
probability, highly incoherent with any xed basis []. CS is

considered to be non-adaptive because measurement matrix []


remains constant.
Several measurement matrix design considerations and reconstruction algorithms have been presented in [39] and found that
using Bernoulli measurement matrix, compression ratio of 16 is
achievable. Mamaghanian et al. [40] compared CS and the DWTbased compression algorithms and found that CS was inferior to
DWT-based algorithm in compression performance. Despite of it,
CS-based compression outperforms in terms of energy efciency
due to its lower complexity and reduced CPU execution time.
After digitization of analog signals, digital signals are compressed to reduce amount of data. The basic purpose of data
compression is to represent the original signal with a smaller
number of bits than that is needed for the original signal. The
compression is typically achieved by removing redundancy from
the signal to be compressed. Since power requirement of wireless
module directly depends on the amount of data to be transmitted,
one of the major advantages with compression is a reduction in
power requirement by wireless module. However, there is a loss of
information, in general, when signal is reconstructed from the
compressed data. A proper balance is maintained with compression ratio and the requirement that the information of diagnostic
importance is preserved.
Heart monitoring systems have additional requirement for
compression algorithms to be computationally efcient to support
long term monitoring. Various compression algorithms for heart
signals have been reported in the literature. Wavelet transform
[41,42], Walsh transform [43], Hermite function [44] and discrete
cosine transform [45] based compression algorithms rst decompose the signal into coefcients, by projecting the signal onto basis
functions of transforms. Then compression is achieved by retaining only a small number of coefcients which typically preserves
essential information of the heart signal. As stated before, computational complexity is crucial for the compression algorithm as
these have to be implemented on the patient side. Computational
complexity of DWT is O(N), DCT is O(NlogN), Walsh transform is O
(NlogN), and Hermite function is O(Nlog2N). Thus DWT has lowest
computational complexity. However, compression performance
of the Hermite function based method is better than DCT and
DWT based basic compression algorithms [44]. The performance
of these algorithms depends on the following parameters:
(a) choice of basis function, (b) decomposition level, and (c) the

Table 2
Digital signal processors.
Processor

No. of bits

Operational frequency (MHz)

Power consumption

Used in

Characteristics

PIC24FJ64GA

10

32

 Run mode: 650 A at 2.0 V, 1 MHz


 Idle mode: 150 A at 2.0 V
 Sleep mode: 0.1 A at 2.0 V

[6]

 Low operating voltage range


 On-the-y clock switching

PIC18f2423

12

40

 Run mode: 330 A at 2.0 V, 1 MHz


 Idle mode: 5:8 A at 2.0 V
 Sleep mode: 0:1 A at 2.0 V

[33]

 Multiple idle and run modes


 Nano watt technology
 On-the-y clock switching

PIC16F877

20

 Run mode: 600 A at 3 V, 4 MHz


 Idle mode: 20 A at 3 V, 32 kHz
 Sleep mode: 1 A

[34]

High performance RISC CPU

MSP430f2274

10

16

 Run mode: 270 A at 1 MHz, 2.2 V


 Idle mode: 0:7 A
 Sleep mode: 0:1 A

[18,35]

 Ultra-fast wake-up
 Ultra-low power
 RISC mixed-signal microprocessors

MSP430F1611

12

 Run mode: 330 A at 1 MHz, 2.2 V


 Idle mode: 1:1 A
 Sleep mode: 0:2 A

[36]

 Ultra-fast wake-up from stand-by


 Five power saving modes

MSP430F2410

12

16

 Run mode: 270 A at 1 MHz, 2.2 V


] Idle mode: 0:3 A
 Sleep mode: 0:1 A

[37]

 Ultra-fast wake-up
 Ultra-low power

P.K. Jain, A.K. Tiwari / Computers in Biology and Medicine 54 (2014) 113

percentage of retained energy (number of coefcients). In these


approaches, trade-off between the percentage of retained energy
and compression ratio is crucial. Increment in the percentage of
retained energy reduces compression ratio and enhances reconstructed signal quality and vice versa. Retained coefcients are
compressed using conventional compression algorithms such as
zero-removal [41], Huffman coding [41,42], dead zone quantization [45], and run length coding [42,46]. While, Sharma et al. [47]
applied multi-scale principal component analysis (MSPCA) on
wavelet transform coefcients and then MSPCA coefcients are
uniformly quantized and encoded by Huffman coding. All the
above algorithms compressed the entire frame with same compression ratio. On the other hand, researchers have been proposed
approaches to use different compression ratio for different block of
signals [43,46,48,49]. Different statistical parameters have been
calculated to identify signicance of the segment such as Wang
et al. [48] calculated kurtosis, Kim et al. [49] calculated mean
deviation (MD), Ma et al. [50] calculated wavelet coefcient
energy. In [43], signicance of the segment is calculated based
on the energy of the Walsh coefcients. While, Rajoub [46] applied
DWT and then divided the coefcients into three groups based on
magnitude of coefcients and then applied different thresholding
for each group.
Researchers have also proposed compression algorithms that
preserve features of heart signal (rather than preserving the
waveform) [51,52]. In [51] Alvarado et al. proposed a compression
algorithm based on integrator and re sampler. Similarly, Kim
et al. [52] proposed an algorithm based on curvature points, which
calculated the important information from the signal.
Compression algorithms which require low computation are
suitable for long term heart monitoring. DWT based compression
algorithms have lower computational complexity than other
algorithms and thus have been used extensively. On the other
hand, feature preserving compression algorithms have high compression ratio. They are suitable for heart signals because diagnosis
can be performed based on these features. However, selection
process of the diagnostic features from the heart signal is complex
process. Furthermore, the performance of these types of algorithm
deteriorated in the presence of noises.

7. Wireless module
Digitized and compressed heart signals are transmitted to
remote site. In off-site monitoring, analysis and classication of
the heart signals are performed at the remote site. Transmitter
consists of wireless module which helps to transmit heart signals
to remote site. Low power consumption, convenient connection
process, and low latency are some important features of wireless
modules that promote wide acceptance of heart monitoring

systems. In the literature, various wireless communication techniques and protocols have been proposed for transmission purpose (Table 3). Bluetooth 4.0 [37] wireless system supports
24 Mbps data rate, has working range up to 100 m, and consumes
low power. Bluetooth devices with these features are suitable to be
integrated with heart monitoring systems. But Bluetooth wireless
systems require initial connection setup that has to be done
manually. Patient's intervention is not desirable in a heart monitoring system as it reduces convenience. To overcome this
problem an approach was proposed by Morak et al. [37] using
radio-frequency identication (RFID) and near eld communication (NFC). In this approach, the connection establishes by bringing two NFC enabled devices closer and using RFID information of
both devices. The drawback of this approach is that it requires
permanent activation of Bluetooth which results in extra power
consumption. Moreover, NFC can support data rate up to 424 kbps
only. Since data rate is lower than Bluetooth 4.0, it takes long time
to transmit data. A good review of the state-of-the art technologies
for wireless network was presented by [32].
Keeping in view the desired features of wireless module,
various protocols have been proposed [6,36]. Chen et al. [36]
proposed a reliable protocol based on any-cast routing algorithm.
This algorithm automatically selects nearest hop (sink), in case of
failure in original path, instead of rebuilding the path from the
source node. Thus, it provides a reliable communication as well as
reducing trafc overhead and transmission latency. However,
selection of the hop process increases the complexity of routing
algorithm and the complexity increases power consumption. To
optimize power consumption, Nemati et al. [6] proposed an ANT
protocol. The ANT protocol was used as a low-data-rate wireless
module to reduce the power consumption and size of the sensor.
ANT is an adaptive isochronous ad hoc wireless protocol based on
master slave model. It consumes from 1 mA to 6.3 mA current and
supports many topologies such as peer-to-peer, star, tree, and
mesh. SimpliciTI is also a low power radio frequency (RF) network
protocol used in heart monitoring systems [18,35]. SimpliciTI was
designed by Texas Instruments for easy implementation and
deployment on RF platforms. It is low data rate and low duty
cycle protocol and supports star and peer-to-peer network
topology.
Ma et al. [50] proposed an unequal-error protection approach
for heart signals to reduce transmission distortion and to reduce
power consumption of wireless transmission. In this approach
more protection is provided to the segment of heart signal which
contains diagnostic important features compared to the other
segments. Results showed that nearly 40% of transmission energy
can be saved compared to the equal error protection.
In a different approach, Atakan et al. [53] introduced the concept
of a body area network (BAN) with molecular communication,
where the messenger molecule is used as a communication carrier

Table 3
Wireless modules.
Module

CC2420 (zigbee)
Bluescense (blutooth)
nRF24E1 (Eco-wireless)
ANT-AP2
cc2500 (zigbee)
UZ2400 (zigbee)
Zebra (zigbee)
BlueNiceCom-4 (bluetooth class-2)
Xbee (Emosense)

Power consumption
Receive

Transmission

18.8
33
22
17
13.3
18

17.4

65
50

10
15
21.2
22

45

Size (mm)

Transmission range (m)

Manufacturer

Used in

77
37  21
13  11
20  20
44
66
16  33
27  16
24  27

70

Texas instruments
Corscience
UC Irvine
Dynastream Innovations
Texas instruments
Uniband Electronics Corp.
senTec Elektronik
AMBER wireless
Digi International Inc.

[54]
[55]
[56]
[6]
[18,35]
[36]
[57]
[37]
[58]

10
30
30
10500
20
3090

P.K. Jain, A.K. Tiwari / Computers in Biology and Medicine 54 (2014) 113

from a sender to receiver. However, the communications at the


molecular scale are subject to numerous problems, some similar to
the ones faced on a larger scale in existing wireless networks.

8. Analysis and classication


Analysis and classication module performs automatic
machine diagnosis that enhances diagnostic accuracy. It is very
helpful in the present scenario, where the number of cardiologists
is low when compared to the number of cardiac patients [5].
Typically, analysis and classication is performed in two steps:
noise suppression, and analysis and classication. In noise suppression step, noises are suppressed from heart signals. In the next
step, the heart signals are classied in normal and different CVDs.
8.1. Noise suppression
Noise suppression from heart signals is essential as its presence
may lead to imprecise or inaccurate classication of the signals.
For noise suppression, classical lters such as Gaussian lter,
Chebyshev lter, Butterworth lter, and Weiner lter have been
used extensively. Because, heart signals lie in the 20500 Hz
frequency band and these lters are able to suppress noise in
the selected frequency band (below 20 Hz and above 500 Hz). But
noises, which overlap with spectral contents of heart signals, are
not easy to suppress from the signals. Hence, sophisticated lters
have been proposed in the literature for suppression of these types
of noises (in-band noise) [59,60]. Filters have been developed
based on wavelet transform [59]. In these lters, signals are
transformed into wavelet coefcients, as discussed in Section 6.
Then noise suppression is achieved by discarding the coefcients
which are correlated to noises, by applying threshold. Although,
wavelet based lters are able to suppress the in-band noise, but
the threshold value plays a crucial role in this approach. If the
threshold is selected high, signal information will be lost, while
small value will not have a signicant effect on the signal. To
obtain optimal denoising parameter for DWT based denoising,
Messer et al. [61] performed experiments and found that level
5 for the signal decomposition and soft thresholding with rigrsure
threshold selection rule gives the best result.
Almasi et al. [60] introduced model-based Bayesian denoising
framework which combined the extended Kalman lter and
dynamic model of the heart signal. Results demonstrate that the
proposed method has the superiority over wavelet based denosing. However, the requirement of a model of the heart signal limits
the use of this framework.
Researchers have proposed many ltering approaches which
analyze diversity between characteristics of heart signal components and characteristics of noises [7,15,62]. Lee et al. [7] used rst
order-intrinsic mode function (F-IMF) to minimize motion noise
from the heart signals. F-IMF of the clean signal has periodic
patterns, whereas noise contaminated signal has highly varying
irregular dynamics with lower magnitudes. Thus, noisy segment
can be classied from the clean heart signal. Liu et al. [15] removed
the noises from heart signal components based on the characteristic of wavelet coefcient that the signal coefcients with large
magnitude at a ner scale will also be large in magnitude at
coarser scales. However, for coefcients which are caused by
noises, magnitude will decay rapidly along the scales. Manikandan
and Soman [62] calculated lag-1 auto-correlation coefcients,
which give positive values for heart signal components and
negative values for spurious noise.
Quasi-cyclostationary nature of heart signals also has been
considered to lter noise from the signals [63,64]. Quasicyclostationary means that the morphology of the heart signals

does not change abruptly from a cardiac cycle to consecutive


cardiac cycle. Thus, noise suppression can be achieved by correlating the consecutive cycles of the signal because noise components,
in general, are uncorrelated. However, quasi-cyclostationary nature of the heart signal may not be fullled due to variation in
waveform, presence of murmurs, and variation in the timing of the
heart sound components. Furthermore, the performance of this
approach depends on the segmentation of cycle.
Adaptive noise cancellation (ANC) techniques are also found
suitable for heart signals as they can detect dynamic variation in
the signal [65]. Least mean square (LMS) is an ANC technique
which calculates lter coefcients that relate to producing the
least mean squares of error signal (difference between the desired
signal and the ltered signal). Estimation of lter coefcients
requires high computation. To reduce the computation of LMS
algorithms, various variations in LMS algorithms have been
proposed in the literature and reported in [65]. To further reduce
the computational complexity, the author proposed [65] sign and
error non-linear sign based LMS.
Respiratory system also affects heart signals signicantly. To
overcome this problem, Chen et al. [66] proposed a zero crossing
method. It calculates the time interval between two consecutive
upward and downward points (IBI) in the signal. Then the inverse
of IBI gives the frequency of breathing signal, which can be
removed by notch ltering.
Noise often appears in parts of the heart signal recordings.
In some part, noise affects severely to the heart signal while in
others it affects mild. In the case of severe contamination, that
part of the signal can be eliminated from diagnostic consideration while in the case of mild contamination, noise suppression
algorithm can be used. This approach will improve diagnosis
efciency as well as optimize complexity of denoising algorithms. This approach will be also helpful in home care systems
for alarming to the user for the bad signal quality. Thus, it is of
interest to obtain a signal quality index to nd out a subsequence with better signal quality with respect to the rest of the
cycle. Li et al. [67] proposed an optimum heart sound selection
scheme based on cycle frequency spectral density. In this
approach, the quality of the heart sound signal depends on the
periodicity of the heart signal. In [68], the quality index was
calculated using the Cepstral distance between homogeneous
cardiac sounds. In this algorithm, rst, the heart signal was
segmented into separate cardiac cycle using wavelet based
approach. After segmentation, Mel frequency Cepstral coefcient (MFCC) was calculated for each cycle. Finally, the reciprocal of distance between MFCC coefcients of consecutive cycle
gives the quality score. The performance of this algorithm
depends on the segmentation of the heart signal into cardiac
cycle. Naseri [69] described an approach to identify the level of
noise in the heart signal cycle. In this approach, rst, the signal
is segmented into separate heart cycle. Then, cycles are clustered into a nite number of groups based on geometrical
parameter and spectral content. Next, median of these clusters
is correlated to the test cycle features. Finally, by applying a
threshold, the cycle is prescribed as clean or noisy. Although,
requirement of a test cycle features limits the use of this
approach.
8.2. Analysis and classication
Analysis and classication of heart signals are challenging tasks
due to non-stationary nature of the heart signals. Moreover, time
to time varying morphology of heart signals from intra- and interpatient needs sophisticated classication algorithms. Classication
of heart signals is performed by analyzing diagnostic features
present in the signal as follows.

P.K. Jain, A.K. Tiwari / Computers in Biology and Medicine 54 (2014) 113

8.2.1. Electrocardiography
ECG signal consists of different waves, as discussed in Section 4
(A). Each wave is associated with particular functionality of the
heart. Analysis of the shape of these waves leads to diagnosis of
important cardiovascular diseases which includes MI, hypertensive heart diseases, arrhythmia, CHD. The impact of CVDs can be
seen on the waves in ECG signal. Myocardial infarction causes ST
elevation or depression depending on the severity of the infarction. Location of the infarction can be identied by analyzing ECG
signals of different leads. In the case of hypertension, QRS voltage
increases due to both thickening of wall (pressure overload) and
dilatation of chamber (volume overload) of the left ventricle. The
RR interval is critical in the diagnosis of many arrhythmia such as
premature ventricular contractions, left and right bundled branch
blocks, and paced beats [70].
Classication of ECG signals is performed by analyzing shape of
the waves presents in the signal. Parameters of the shape of the
waves act as features for classication algorithms. Computational
requirement of classication algorithms depends directly on the
number of the features used and accuracy of classication depends
on quality of the features. Thus, feature selection plays a prominent role in the classication of ECG signals. In the literature, many
approaches have been proposed to select optimal features. Bashir
et al. [70] calculated QRS, P and T waves morphological parameters
as features to detect different arrhythmia. Then a parameter score
was calculated for an adaptive selection of feature subset for
particular arrhythmia. Accordingly, there will be a different feature
set for each arrhythmia, which enhances the accuracy, and at the
same time reduces the computational burden. While, Llamedo and
Martinez [71] calculated interval features and morphological
features for classication of arrhythmia. Interval features were
calculated from R peaks, and morphological features were calculated from three sources, RR interval, 2-D vectocardiogram loop
and DWT of the ECG signal. Then outliers form the feature set
were removed based on Kurtosis coefcients. Mar et al. [72]
applied sequential forward oating search algorithm with a new
criterion function index. Drawback of the proposed method is that
in many cases the subset with highest criterion value has a very
large number of features. Kamath [73] selected mean of Teager
energy operator (TEO) in the time domain and frequency domain
as features set. Key characteristic of the TEO is that it models
energy of the source that generated signal rather than the energy
of the signal itself. Hence, any deviations in the regular rhythmic
activity of the heart get reected in the TEO. Most of the above
algorithms face the same challenge, requirement of a large
number of the feature set. Large number of feature set is required
for diagnosis of the different types of diseases, but it results in
large computational complexity. Another challenge is due to
variation in morphological descriptors of the heart signal
with time.
Since mathematical operators work in the time domain, these
are computationally efcient and hence consume low power.
Mathematical morphological operators have been used [74] to
extract structural information of the ECG signals. However, computational requirement increases as increment in order of the
operators. To optimize computation requirement, Zhang and Bae
[8] proposed 1 dilation and 1 erosion based morphology operator
sets. However, effectiveness of these algorithms depends on the
selection of three structural components of the operator, shape,
length, and amplitude.
T wave delineation is crucial as prolongation of T wave to end
of the T wave is associated with ventricular pre-arrhythmicity and
sudden cardiac death. Noriega et al. [75] analyzed respiration
effect on T wave. Atrial brillation (AF) is associated with an
increased risk of cardiovascular and coronary artery disease,
hypertension, etc. AF is typically diagnosed by analyzing irregular

RR intervals. Huang et al. [76] proposed an algorithm to classify AF


by analyzing RR interval.

8.2.2. Phonocardiography
As discussed in Section 4.2, PCG signal consists of four sound
components. Characteristics (intensity, frequency, and duration) of
these sound components change due to the presence of CVDs.
Additional murmur sounds may also be present in PCG signal due
to the presence of CVDs. Although PCG can indicate abnormalities
caused by important CVDs, it is used extensively for diagnosis of
valvular diseases as sound components are produced by the
valvular activity [10]. Valvular diseases cause systolic and diastolic
murmurs. Aortic stenosis (AS), pulmonary stenosis (PS), and mitral
regurgitation (MR) cause systolic murmurs. On the other hand,
diastolic murmurs occur due to aortic regurgitation (AR), mitral
stenosis (MS) etc. Systolic murmurs, AS lies in the frequency band
of 120250 Hz, PS lies in 200250 Hz, and MR lies in 300400 Hz.
MR can be classied from other two systolic murmurs (PS and AS),
as with wider duration and higher frequency band. S1 also
becomes quieter than normal, in case of MR. Whereas PS causes
longer duration between aortic and pulmonary components of the
S2, called as split S2. Splitting of S2 also may occur due to atrial
septal defect and right bundle branch block. Diastolic murmurs lie
in 100250 Hz. MS causes mid-diastolic murmur with louder S1
and causes a high frequency opening snap of 90130 ms after the
aortic component of S2. While AR is relatively louder than other
diastolic murmurs. But more severe AR causes a lower intensity
murmur with longer duration. Two other sound components, S3
and S4, in PCG signal rarely occur and may indicate abnormalities.
Presence of S3 in a child is normal, while in adults, it represents
diastolic overload or cardiomyopathy [77]. S4 occurs just after
atrial contraction and it may be due to ventricular hypertrophy or
pulmonary arterial hypertension [77]. Analysis of the sound
components and the murmurs leads towards to the classication
of the PCG signals.
Heart sound classication algorithms rst partition the PCG
signal into S1, S2, systole, and diastole intervals, by emphasizing
them. To emphasize the heart signal components, envelop based
parameters such as Hilbert transform, Shannon energy, cardiac
sound characteristic waveform (CSCW) and timefrequency
domain analysis such as STFT and wavelet have been presented
in the literature. After segmentation of components, classication
of PCG signal is performed by analyzing the characteristics of these
components.
Envelope extraction based classication algorithms are able to
detect fundamental heart sound components and to classify the
signal as normal or abnormal. Choi and Jiang [78] compared three
envelope extraction algorithms, normalized average Shannon
energy; envelope information of Hilbert transform, and the CSCW.
As shown in the results, CSCW gives a more uniform representation of the fundamental components. The main challenge for the
envelope extraction based algorithm is the selection of the threshold value. A higher value of threshold missed the S1 and S2, while
the lower value of threshold detects spurious components and
inaccurate S1 and S2. To resolve this problem, Atbi et al. [79]
proposed a two step thresholding scheme. In the rst step threshold is selected to detect S1 and S2 and in the next step, to detect
murmurs. Envelop extraction based algorithms are computationally low complex. However, the performance of these algorithms
depends on the morphology of the PCG signal. Furthermore, it
becomes difcult to detect S1 and S2, where murmurs are merged
with them.
Frequency domain transformation techniques such as Fourier
transform, discrete cosine transform, and auto-regressive based
spectral analysis techniques provide frequency characteristics of

10

P.K. Jain, A.K. Tiwari / Computers in Biology and Medicine 54 (2014) 113

PCG signal components. However, timefrequency domain analysis is more suitable for PCG signal analysis due to the diagnostic
signicance of timing and frequency of PCG components. Time
frequency analysis of PCG signal has been done using short time
Fourier transform (STFT) [10,80], wavelet transform [80,81]. Boutana et al. [10] classied murmurs from PCG sound components by
analyzing the Renyi marginal entropy of STFT coefcients. Renyi
marginal entropy remains high for murmurs and low for sound
components. While, author [80] implement the PCG analysis
algorithm using STFT and wavelet on digital signal processing
board. In [81], rst heart sound signal is segmented into intervals
associated with cardiac cycle. Then intervals were grouped
together based on similarities between their STFT coefcients. In
STFT, a trade off between time resolution and frequency resolution
arises. Increment in size of time window increases the frequency
resolution, but reduces time resolution and vice versa. Thus,
selection of the optimal size of time window is crucial. For time
frequency analysis, wavelet offers a better compromise in terms of
resolution. Its main difference with STFT is that the size of the
window is not constant. It varies in inverse proportion to
frequency in such a way that good time and poor frequency
resolution obtain at high frequencies while good frequency and
poor time resolution obtain at low frequencies.
PCG signals have been classied using articial intelligence
algorithms such as Hidden Markov model [9] and neural network
[82,83]. Extracted features from PCG signals using time and
frequency analysis tools such as wavelet are used as feature points
for these articial intelligence techniques [82,83]. Use of the
machine learning algorithms reduce tedious envelop analysis
and its disadvantage in case of murmurs can be avoided, but at
the cost of having to prepare the training dataset. To prepare the
training dataset for PCG signal, Ahlstrom et al. [84] proposed a
feature subset selection algorithm from features of different
domain, including Shannon energy, wavelet, fractal dimension,
and recurrent quantication analysis.
PCG signal modelling is required to generate test data to
analyze efcacy of the developed algorithm. Modelling of PCG
signals has been done using exponential damped sinusoidal model
[85], matching pursuit method [86]. These methods provide
complete parameterization of the signal, but require a large
number of components. Whereas linear chirp signal modelling is
not suitable for PCG signal because components of PCG signal do
not have a linear relationship with time. To achieve better
accuracy, Xu et al. [87] proposed non-linear chirp signal modelling
of the heart sound components.

8.2.3. Seismocardiography
Seismocardiography (SCG) measures mechanical vibrations
produced by heart during each cardiac cycle. As discussed in
Section 4.3, SCG signal is composed of many waves. Each wave is
associated with a particular event of the cardiac cycle. Thus,
analysis of these waves provides diagnostic information related
to the health of the heart. In [88], author studied the relation
between the cardiac event position in SCG with ultrasound signal
and showed SCG as an accurate indicator for cardiac events. Thus
SCG signal can be used to detect cardiac cycle boundary, heart rate
[89], heart rate variability [90]. SCG signals have been also used to
obtain systolic blood pressure (SBP) [24]. It was shown that SBP
has the correlation with starting point of the SCG signal in the
x-axis to the midpoint of the z-axis. However, SCG has been used
for the heart monitoring purpose, but its sensitivity to motion
noise imposes limitation on its wide use.
Characterization of the relation between SCG signal and ECG
signal provides signicant information related to heart functionality. Wick et al. [88] analyzed relation between the R wave of ECG

and the AC wave of SCG signal. The R-AC period varies across two
individual and also for the same person at different heart rates.
This study strongly suggested the cardiac events also vary in the
same manner. Tavakolian et al. [91] analyzed period between the R
wave of the ECG and the AO wave of SCG to analyze the myocardial
contractility. This period is called as pre-ejection period (PEP),
increment in the PEP indicates reduction in contract-ability of
myocardial. Moreover, SCG has potential benets over ECG such as
better specicity and sensitivity for detection of coronary artery
diseases [92].
8.2.4. Photoplethysmography
PPG signal contains sufcient parameters to measure heart
rate, arterial oxygen saturation, and information related to
respiratory system [27,93,94]. As discussed in Section 4.4, PPG
measures variation in intensity of light, reected or transmitted,
induced by variation in the amount of blood in blood vessels.
Respiration information can be extracted using three vital
parameters: PPG amplitude, variation in SpO2, and respiratory
sinus arrhythmia [27]. Now a day, pulse oximeters (variant of PPG)
are being used extensively for heart monitoring [27,9395]. It
measures multiple PPG signals at different wavelengths viz., red
(660 nm) and infrared (940 nm). Pulse oximeters have been used for
sleep apnea detection [27], pulse wave velocity calculation [93],
hypoxia detection [94], and heart rate turbulence analysis [95]. Pulse
oximeters have been developed as an in-ear sensor for cardiovascular
monitoring [27,94]. This setup of sensor could offer three important
advantages: (1) comfortable to wear and hence, suitable for longterm monitoring, (2) the tight-tting could reduce interference from
motion artifacts, and (3) robustness to conditions such as temperature or skin perfusion.
However, PPG signals get contaminated primarily due to
ambient light, motion artifacts and other physiological process.
To extract information from the contaminated PPG signals, Madhav et al. [96] proposed a multi scale principal component analysis
based algorithm. In this algorithm, noise suppression from the PPG
signals was achieved using wavelet decomposition and reconstruction. Selection of coefcients to reconstruct relatively clean
signal was done based on two measures, energy contribution level
(ECL) and Kurtosis. After reconstruction of the clean signal,
principal component (PC) analysis was performed to extract
information about the respiratory system. Li and Warren [93]
developed a sensor circuit in which photodetecters are radially
distributed around the LED to increase the sensing area. This setup improved the signal quality without ltering algorithm.
Whereas, Stuban and Niwayama [97] analyzed optimal corner
frequency of low pass lter for PPG signal. Setting the corner
frequency to the fundamental frequency of the PPG signal resulted
in decreased noise, and consequently, decreased standard deviation. de Haan and Jeanne [98] analyzed robustness of the chrominance based algorithms to separate motion induced distortion
from rPPG signals in case of modest and vigorous motion.

9. Use of mobile
The latest generation of mobile phones (smartphones) is
increasingly used for health monitoring, due to their powerful
on-board computing capability, large memory, large screens and
open operating systems that encourage application development.
Technical features of mobile phone including text messaging,
camera, internet access, inbuilt sensors, make it an appropriate
platform for improving health care service [99]. Wireless technologies, including GPRS, GSM, 3GSatelite, Wireless, Lan networks,
have been used for wireless transmission of the heart signal [100].

P.K. Jain, A.K. Tiwari / Computers in Biology and Medicine 54 (2014) 113

Mobile phones are also suitable platform to develop a heart


monitoring system because of (1) the widespread adoption of
phones, (2) peoples tendency to carry their phones with them
everywhere, and (3) context awareness features [99]. Furthermore,
visible representation of the health status of the patient on mobile
encourages to be attentive for health promoting behaviour.
Mobile phones are using for long term heart monitoring at home
for, both manner, on-site monitoring and off-site monitoring. Home
monitoring supports to reduce the rates of admission to hospital for
chronic heart failure, improve the quality, reduce excessive travel
time and reduce the cost [101103]. While long term monitoring
supports in early identication of deteriorations in patient condition
and symptom control, and timely intervention of a medical team
[99,104]. Smartphones based software application can help clinicians
in identifying acute symptoms, decreasing unnecessary tests, to
understand principles of disease diagnosis, and communication
facility among clinicians [99,102]. In addition, mobile applications
have been used for remote coaching, reminder to patient for
appointments and health related information, public health research,
primary care, emergency care, health information for self, drug
reference, medical training, to encourage for primary care check-up,
etc. Mobile phone based health monitoring system have been
discussed in [102]. Some of them for heart monitoring are as follows:
(1) cardiomobile is comprised of a heart and activity monitor, single
lead ECG, GPS receiver, and programmed smartphone. The smartphone sends ECG rate, walking speed, heart rate, elapsed distance,
and patient location to a secure server for real-time monitoring by
a qualied exercise scientist. (2) Pulmonary rehabilitation is an
application for chronic obstructive pulmonary disease (COPD) rehabilitation and self-management, developed for smartphones. (3) mVisum is a specialized application for cardiology communications that
monitor ECG data, alarm the user in abnormal case, and transmit
data to clinician. (4) iCPR is a cardiopulmonary resuscitation (CPR)
training application. This application measures the chest compression rate and gives audiovisual feedback, improving the performance
of chest compression by helping the user to achieve the correct chest
compression rate. Another smart-phone based health monitoring
system, BioSign, is reviewed in [104]. BioSign system alerts the
patient in case of abnormality. It represents the health status of the
patient as patient status index (PSI), which is calculated based on ve
vital parameters: heart rate, breathing rate, blood pressure, arterial
oxygen saturation, and skin temperature.
To enhance the user acceptability, Scully et al. [105] proposed
a reection photoplethysmography based on imaging by mobile
phones. In this approach, the palmer side of the left index nger
was placed over the camera lens of mobile with its ash turned on.
Then variation of intensity in captured video indicates the heart
beat. This approach does not require any extra hardware. However
sensitivity of device get affects due to motion and pressure
variation of the nger. Another approach proposed by Poh et. al
[106] integrates reective photo-diode into earphone which are
unobtrusive, and low in size and weight. Then acquired signal was
sent to mobile phone for monitoring purpose.
Major challenges for smartphone-based health care system
include cost, network bandwidth, battery power efciency, small
screen size, computer viruses, etc. [102]. Other issue is that it must
be seamless and autonomous in its operation. It would be
benecial for those who are non-familiar with the technology.

10. Conclusion
In this paper we have provided a detailed review of recent
advancements for portable heart monitoring systems. We considered ECG, PCG, PPG, and SCG based systems and portability of such
systems are feasible due to use of light weight and small size body

11

sensors. ECG has diagnostic superiority of important CVDs as


compared to other portable systems. However, it has limitations
in long term monitoring of heart due to the requirement of skin
contact of electrodes and the use of gel. PCG also contains
sufcient diagnostic features, but it is vulnerable to motion noise.
Although, PCG has advantage over ECG in terms of easy to operate.
This feature makes PCG useful in scenarios where the number of
cardiac experts is low. PPG has found its wider use for long term
monitoring as it is more comfortable in terms of wearability than
other systems. But it has limited diagnostic features related to
important CVDs as it acquires signals far from the heart. SCG has
not been used for important cardiovascular diseases. However, it
provides some additional diagnostic parameters than those obtain
from PCG. Its higher sensitivity to motion noise than PCG and PPG
limits its wider use as portable system.
Recently, focus on at-home monitoring of heart is increasing for
long term monitoring, which is advantageous in cases where heart
abnormalities are detected and suspicious to have CVDs. It leads to
increasing research in development of portable systems having
features of low power consumption, signal transmission capability,
and unobtrusiveness. Computationally efcient algorithms are
developed that ensures low power consumption, Signal transmission capability ensures remote monitoring, and unobtrusiveness of
the system helps in its use for long term monitoring. Power
consumption and user friendly connection process are main issues
of signal transmission module. Advanced compression algorithms
for heart signals will reduce power consumption of this module.
To increase the unobtrusiveness, in case of ECG and PPG, researchers have proposed sensors that do not require skin contact. Use of
the cellular phones for monitoring purpose eliminates extra hardware possession and hence increases user convenience. Nowadays,
cellular phones are equipped with considerable computational
power and hence can play a crucial role for heart monitoring.

Conict of interest statement


None declared.
References
[1] W.H. Organization, Global Status Report on Noncommunicable Diseases
2010. http://www.who.int/nmh/publications/ncd_report2010/en/, 2010
(Online, accessed 19.07.13).
[2] A.H. Association, Heart Disease and Stroke Statistic 2012 Update. circ.
ahajournals.org/content/125/1/e2, 2012 (Online, accessed 19.07.13).
[3] Center for Disease Control and Prevention, State Specic Mortality From
Sudden Cardiac Death, www.cdc.gov/heartdisease/facts.htm, 2002 (Online,
accessed 19.07.13).
[4] J. Choi, J. Park, J. Chung, B. Min, An intelligent remote monitoring system for
articial heart, IEEE Trans. Inf. Technol. Biomed. 9 (December) (2005)
564573.
[5] WHO, Global Health Observatory, http://www.who.int/gho/health_nan
cing/en/index.html, 2011 (Online, accessed 19.07.13).
[6] E. Nemati, M. Deen, T. Mondal, A wireless wearable ecg sensor for long-term
applications, IEEE Commun. Mag. 50 (January) (2012) 3643.
[7] J. Lee, D. McManus, S. Merchant, K. Chon, Automatic motion and noise
artifact detection in holter ecg data using empirical mode decomposition and
statistical approaches, IEEE Trans. Biomed. Eng. 59 (June) (2012) 14991506.
[8] C. Zhang, T.-W. Bae, Vlsi friendly ecg qrs complex detector for body sensor
networks, IEEE J. Emerg. Sel. Top. Circuits Syst. 2 (March) (2012) 5259.
[9] C. Kwak, O.W. Kwon, Cardiac disorder classication by heart sound signals
using murmur likelihood and hidden Markov model state likelihood, IET
Signal Process. 6 (4) (2012) 326334.
[10] D. Boutana, M. Benidir, B. Barkat, Segmentation and identication of some
pathological phonocardiogram signals using timefrequency analysis, IET
Signal Process. 5 (September) (2011) 527537.
[11] Deloitte, Cardiovascular Diseases in India Challenges and Way Ahead, www.
deloitte.com/in, 2011 (Online, accessed 19.07.13).
[12] V.K. Nelso Fausto, Abul Abbas, Robbins and Cotran Pathologic Basis of
Disease. Saunders, 2004.
[13] M. Clinic, Heart Disease, www.mayoclinic.com/health/heart-disease/
DS01120/DSECTION=causes, 2013 (Online; accessed 19.07.13).

12

P.K. Jain, A.K. Tiwari / Computers in Biology and Medicine 54 (2014) 113

[14] J.G. Webster, Medical Instrum.: Appl. Des., Wiley, New York, 1995.
[15] X. Liu, Y. Zheng, M. Phyu, F. Endru, V. Navaneethan, B. Zhao, An ultra-low
power ecg acquisition and monitoring asic system for wban applications,
IEEE J. Emerg. Sel. Top. Circuits Syst. 2 (March) (2012) 6070.
[16] L. Searle, A. Kirkup, Direct comparison of wet, dry and insulating bioelectric
recording electrodes, Physiol. Meas. 21 (May) (2000) 271283.
[17] H.-C. Jung, J.-H. Moon, D.-H. Baek, J.-H. Lee, Y.-Y. Choi, J.-S. Hong, S.-H. Lee,
Cnt/pdms composite exible dry electrodesfor long-term ecg monitoring,
IEEE Trans. Biomed. Eng. 59 (May) (2012) 14721479.
[18] J. Gomez-Clapers, R. Casanella, A fast and easy-to-use ecg acquisition and
heart rate monitoring system using a wireless steering wheel, IEEE Sens. J. 12
(March) (2012) 610616.
[19] L.S.A. Aleksandrowicz, Wireless and non-contact ecg measurement system
the aachen smartchair, Acta Polytech. 2 (2007) 6871.
[20] H.J. Baek, G.S. Chung, K.K. Kim, K.S. Park, A smart health monitoring chair for
nonintrusive measurement of biological signals, IEEE Trans. Inf. Technol.
Biomed. 16 (January) (2012) 150158.
[21] Y.G. Lim, K.K. Kim, K.-S. Park, Ecg recording on a bed during sleep without
direct skin-contact, IEEE Trans. Biomed. Eng. 54 (4) (2007) 718725.
[22] G. Cho, K. Jeong, M.J. Paik, Y. Kwun, M. Sung, Performance evaluation of
textile-based electrodes and motion sensors for smart clothing, IEEE Sens.
J. 11 (December) (2011) 31833193.
[23] J. Zanetti, D. Salerno, Seismocardiography: a technique for recording precordial acceleration, in: Proceedings of the Fourth Annual IEEE Symposium
on Computer-Based Medical Systems, 1991, pp. 49.
[24] M. Imtiaz, R. Shrestha, T. Dhillon, K. Yousuf, B. Saeed, A. Dinh, K. Wahid,
Correlation between seismocardiogram and systolic blood pressure, in: 26th
Annual IEEE Canadian Conference on Electrical and Computer Engineering
(CCECE), 2013, pp. 14.
[25] M. Di Rienzo, E. Vaini, P. Castiglioni, G. Merati, P. Meriggi, G. Parati, A. Faini,
F. Rizzo, Wearable seismocardiography: towards a beat-by-beat assessment
of cardiac mechanics in ambulant subjects, Autonom. Neurosci. 178 (12)
(2013) 5059. http://dx.doi.org/10.1016/j.autneu.2013.04.005.
[26] R. Krishnan, B. Natarajan, S. Warren, Two-stage approach for detection and
reduction of motion artifacts in photoplethysmographic data, IEEE Trans.
Biomed. Eng. 57 (8) (2010) 18671876.
[27] B. Venema, J. Schiefer, V. Blazek, N. Blanik, S. Leonhardt, Evaluating
innovative in-ear pulse oximetry for unobtrusive cardiovascular and pulmonary monitoring during sleep, IEEE J. Transl. Eng. Health Med. 1 (2013)
18.
[28] K. Sweeney, T. Ward, S. McLoone, Artifact removal in physiological signals
x2014;practices and possibilities, IEEE Trans. Inf. Technol. Biomed. 16 (May)
(2012) 488500.
[29] M. Huelsbusch, V. Blazek, Contactless mapping of rhythmical phenomena in
tissue perfusion using ppgi, Proc. SPIE 4683 (2002) 110117.
[30] R. Rieger, Variable-gain, low-noise amplication for sampling front ends,
IEEE Trans. Biomed. Circuits Syst. 5 (June) (2011) 253261.
[31] E. Spinelli, N. Martinez, M. Mayosky, A transconductance driven-right-leg
circuit, IEEE Trans. Biomed. Eng. 46 (12) (1999) 14661470.
[32] C. Bachmann, M. Ashouei, V. Pop, M. Vidojkovic, H. Groot, B. Gyselinckx,
Low-power wireless sensor nodes for ubiquitous long-term biomedical
signal monitoring, IEEE Commun. Mag. 50 (January) (2012) 2027.
[33] H. Tang, T. Li, T. Qiu, Cardiac cycle detection for heart sound signal based on
instantaneous cycle frequency, in: 2011 4th International Conference on
Biomedical Engineering and Informatics (BMEI), vol. 2, 2011, pp. 676679.
[34] D.-H. Shih, H.-S. Chiang, B. Lin, S.-B. Lin, An embedded mobile ecg reasoning
system for elderly patients, IEEE Trans. Inf. Technol. Biomed. 14 (May) (2010)
854865.
[35] R. Dilmaghani, H. Bobarshad, M. Ghavami, S. Choobkar, C. Wolfe, Wireless
sensor networks for monitoring physiological signals of multiple patients,
IEEE Trans. Biomed. Circuits Syst. 5 (August) (2011) 347356.
[36] S.-K. Chen, T. Kao, C.-T. Chan, C.-N. Huang, C.-Y. Chiang, C.-Y. Lai, T.-H. Tung,
P.-C. Wang, A reliable transmission protocol for zigbee-based wireless
patient monitoring, IEEE Trans. Inf. Technol. Biomed. 16 (January) (2012)
616.
[37] J. Morak, H. Kumpusch, D. Hayn, R. Modre-Osprian, G. Schreier, Design and
evaluation of a telemonitoring concept based on nfc-enabled mobile phones
and sensor devices, IEEE Trans. Inf. Technol. Biomed. 16 (January) (2012)
1723.
[38] E. Candes, M. Wakin, An introduction to compressive sampling, IEEE Signal
Process. Mag. 25 (March) (2008) 2130.
[39] A. Dixon, E. Allstot, D. Gangopadhyay, D. Allstot, Compressed sensing system
considerations for ecg and emg wireless biosensors, IEEE Trans. Biomed.
Circuits Syst. 6 (April) (2012) 156166.
[40] H. Mamaghanian, N. Khaled, D. Atienza, P. Vandergheynst, Compressed
sensing for real-time energy-efcient ecg compression on wireless body
sensor nodes, IEEE Trans. Biomed. Eng. 58 (September) (2011) 24562466.
[41] W.-C. Kao, W.-H. Chen, C.-K. Yu, C.-M. Hong, S.-Y. Lin, Portable real-time
homecare system design with digital camera platform, IEEE Trans. Consum.
Electron. 51 (4) (2005) 10351041.
[42] J. Martinez-Alajarin, R. Ruiz-Merino, Wavelet and wavelet packet compression of phonocardiograms, Electron. Lett. 40 (August) (2004) 10401041.
[43] A. Kadrolkar, R. Gao, R. Yan, W. Gong, Variable-word-length coding for
energy-aware signal transmission, IEEE Trans. Instrum. Meas. 61 (April)
(2012) 850864.

[44] A. Sandryhaila, S. Saba, M. Puschel, J. Kovacevic, Efcient compression of qrs


complexes using hermite expansion, IEEE Trans. Signal Process. 60 (February) (2012) 947955.
[45] A. Bendifallah, R. Benzid, M. Boulemden, Improved ecg compression method
using discrete cosine transform, Electron. Lett. 47 (January) (2011) 8789.
[46] B. Rajoub, An efcient coding algorithm for the compression of ecg signals
using the wavelet transform, IEEE Trans. Biomed. Eng. 49 (4) (2002)
355362.
[47] L. Sharma, S. Dandapat, A. Mahanta, Multichannel ecg data compression
based on multiscale principal component analysis, IEEE Trans. Inf. Technol.
Biomed. 16 (July) (2012) 730736.
[48] H. Wang, D. Peng, W. Wang, H. Sharif, H. hwa Chen, A. Khoynezhad,
Resource-aware secure ecg healthcare monitoring through body sensor
networks, IEEE Wirel. Commun. 17 (February) (2010) 1219.
[49] H. Kim, Y. Kim, H.-J. Yoo, A low cost quadratic level ecg compression
algorithm and its hardware optimization for body sensor network system,
in: 30th Annual International Conference of the IEEE Engineering in
Medicine and Biology Society (EMBS 2008), 2008, pp. 54905493.
[50] T. Ma, P. Shrestha, M. Hempel, D. Peng, H. Sharif, H.-H. Chen, Assurance of
energy efciency and data security for ecg transmission in basns, IEEE Trans.
Biomed. Eng. 59 (April) (2012) 10411048.
[51] A. Alvarado, C. Lakshminarayan, J. Principe, Time-based compression and
classication of heartbeats, IEEE Trans. Biomed. Eng. 59 (June) (2012)
16411648.
[52] T.-H. Kim, S.-Y. Kim, J.-H. Kim, B.-J. Yun, K.-H. Park, Curvature based ecg
signal compression for effective communication on wpan, J. Commun. Netw.
14 (February) (2012) 2126.
[53] B. Atakan, O. Akan, S. Balasubramaniam, Body area nanonetworks with
molecular communications in nanomedicine, IEEE Commun. Mag. 50 (January) (2012) 2834.
[54] S.H. Lee, S.M. Jung, C.K. Lee, K.S. Jeong, G. Cho, S.K. Yoo, Wearable ecg
monitoring system using conductive fabrics and active electrodes, Hum.Comput. Interact. Ambient, Ubiquitous Intell. Interact. 5612 (2009) 778783.
[55] G. Gargiulo, P. Bifulco, M. Cesarelli, M. Ruffo, M. Romano, R.A. Calvo, C. Jin,
A. Schaik, An ultra-high input impedance ecg amplier for long-term
monitoring of athletes, Med. Devices (Auckl) 3 (July) (2010) 19.
[56] C. Park, P. Chou, Y. Bai, R. Matthews, A. Hibbs, An ultra-wearable, wireless,
low power ecg monitoring system, in: Biomedical Circuits and Systems
Conference (BioCAS 2006), IEEE, November 2006, pp. 241244.
[57] S.L.A. Aleksandrowicz, Wireless and non-contact ecg measurement system
the aachen smartchair, Acta Polytech. 47 (4) (2007).
[58] B. Massot, N. Baltenneck, C. Gehin, A. Dittmar, E. McAdams, Emosense: an
ambulatory device for the assessment of ans activity x2014; application in
the objective evaluation of stress with the blind, IEEE Sens. J. 12 (March)
(2012) 543551.
[59] S.M. Debbal, F. Bereksi-Reguig, Filtering and classication of phonocardiogram signals using wavelet transform, J. Med. Eng. Technol. 32 (February)
(2008) 5365.
[60] A. Almasi, M.B. Shamsollahi, L. Senhadji, Bayesian denoising framework of
phonocardiogram based on a new dynamical model, IRBM 34 (3) (2013)
214225.
[61] S.R. Messer, J. Agzarian, D. Abbott, Optimal wavelet denoising for phonocardiograms, Microelectron. J. 32 (12) (2001) 931941.
[62] M. Sabarimalai Manikandan, K. Soman, Robust heart sound activity detection
in noisy environments, Electron. Lett. 46 (2010) 11001102.
[63] H. Tang, T. Li, Y. Park, T. Qiu, Separation of heart sound signal from noise in
joint cycle frequency x2013;time x2013; frequency domains based on fuzzy
detection, IEEE Trans. Biomed. Eng. 57 (October) (2010) 24382447.
[64] H. Tang, T. Li, T. Qiu, Noise and disturbance reduction for heart sounds in
cycle-frequency domain based on nonlinear time scaling, IEEE Trans.
Biomed. Eng. 57 (2) (2010) 325333.
[65] M. Rahman, R. Shaik, D. Reddy, Efcient and simplied adaptive noise
cancelers for ecg sensor based remote health monitoring, IEEE Sens. J. 12
(March) (2012) 566573.
[66] S. Chen, M. Bolic, V. Groza, H. Dajani, I. Batkin, S. Rajan, Extraction of
breathing signal and suppression of its effects in oscillometric blood
pressure measurement, IEEE Trans. Instrum. Meas. 60 (May) (2011)
17411750.
[67] T. Li, T. Qiu, H. Tang, Optimum heart sound signal selection based on the
cyclostationary property, Comput. Biol. Med. 43 (6) (2013) 607612.
[68] F. Beritelli, A. Spadaccini, Heart sounds quality analysis for automatic cardiac
biometry applications, in: First IEEE International Workshop on Information
Forensics and Security (WIFS 2009), December 2009, pp. 6165.
[69] H. Naseri, M. Homaeinezhad, H. Pourkhajeh, Noise/spike detection in
phonocardiogram signal as a cyclic random process with non-stationary
period interval, Comput. Biol. Med. 43 (9) (2013) 12051213.
[70] M. Bashir, D.G. Lee, M. Li, J.-W. Bae, H. Shon, M.C. Cho, K.H. Ryu, Trigger
learning and ecg parameter customization for remote cardiac clinical care
information system, IEEE Trans. Inf. Technol. Biomed. 16 (July) (2012)
561571.
[71] M. Llamedo, J. Martinez, Heartbeat classication using feature selection
driven by database generalization criteria, IEEE Trans. Biomed. Eng. 58 (3)
(2011) 616625.
[72] T. Mar, S. Zaunseder, J. Martnez, M. Llamedo, R. Poll, Optimization of ecg
classication by means of feature selection, IEEE Trans. Biomed. Eng. 58 (8)
(2011) 21682177.

P.K. Jain, A.K. Tiwari / Computers in Biology and Medicine 54 (2014) 113

[73] C. Kamath, Ecg beat classication using features extracted from teager
energy functions in time and frequency domains, IET Signal Process. 5 (6)
(2011) 575581.
[74] F. Zhang, Y. Lian, Qrs detection based on multiscale mathematical morphology for wearable ecg devices in body area networks, IEEE Trans. Biomed.
Circuits Syst. 3 (August) (2009) 220228.
[75] M. Noriega, J. Martinez, P. Laguna, R. Bailon, R. Almeida, Respiration effect on
wavelet-based ecg t-wave end delineation strategies, IEEE Trans. Biomed.
Eng. 59 (7) (2012) 18181828.
[76] C. Huang, S. Ye, H. Chen, D. Li, F. He, Y. Tu, A novel method for detection of the
transition between atrial brillation and sinus rhythm, IEEE Trans. Biomed.
Eng. 58 (4) (2011) 11131119.
[77] F. Beritelli, S. Serrano, Biometric identication based on frequency analysis of
cardiac sounds, IEEE Trans. Inf. Forens. Secur. 2 (September) (2007) 596604.
[78] S. Choi, Z. Jiang, Comparison of envelope extraction algorithms for cardiac
sound signal segmentation, Expert Syst. Appl. 34 (2) (2008) 10561069.
[79] A. Atbi, S.M. Debbal, F. Meziani, A. Meziane, Separation of heart sounds and
heart murmurs by hilbert transform envelogram, J. Med. Eng. Technol. 37 (6)
(2013) 375387 (doi: 23875931).
[80] D. Balasubramaniam, D. Nedumaran, Efcient computation of phonocardiographic signal analysis in digital signal processor based system, Int.
J. Comput. Theory Eng. 2 (4) (2010) 660664.
[81] Z. Syed, D. Leeds, D. Curtis, F. Nesta, R. Levine, J. Guttag, A framework for the
analysis of acoustical cardiac signals, IEEE Trans. Biomed. Eng. 54 (4) (2007)
651662.
[82] T.R. Reed, N.E. Reed, P. Fritzson, Heart sound analysis for symptom detection
and computer-aided diagnosis, Simul. Model. Pract. Theory 12 (2) (2004)
129146 (Advances in modelling and simulation in biology and medicine).
[83] D. BarschdorfF, U. Femmer, E. Trowitzsch, Automatic phonocardiogram signal
analysis in infants based on wavelet transforms and articial neural networks, in: Computers in Cardiology, 1995, pp. 753756.
[84] C. Ahlstrom, P. Hult, P. Rask, J.-E. Karlsson, E. Nylander, U. Dahlstrm, P. Ask,
Feature extraction for systolic heart murmur classication, Ann. Biomed.
Eng. 34 (11) (2006) 16661677.
[85] H. Sava, P. Grant, J.T.E. McDonnell, Spectral characterization and classication
of CarpentierEdwards heart valves implanted in the aortic position, IEEE
Trans. Biomed. Eng. 43 (10) (1996) 10461048.
[86] X. Zhang, L. Durand, L. Senhadji, H. Lee, J.-L. Coatrieux, Analysis-synthesis of
the phonocardiogram based on the matching pursuit method, IEEE Trans.
Biomed. Eng. 45 (August) (1998) 962971.
[87] J. Xu, L. Durand, P. Pibarot, Nonlinear transient chirp signal modeling of the
aortic and pulmonary components of the second heart sound, IEEE Trans.
Biomed. Eng. 47 (October) (2000) 13281335.
[88] C. Wick, J.-J. Su, J. McClellan, O. Brand, P. Bhatti, A. Buice, A. Stillman, X. Tang,
S. Tridandapani, A system for seismocardiography-based identication of
quiescent heart phases: implications for cardiac imaging, IEEE Trans. Inf.
Technol. Biomed. 16 (5) (2012).
[89] H. Nguyen, J. Zhang, Y.-H. Nam, Timing detection and seismocardiography
waveform extraction, in: 2012 Annual International Conference of the IEEE
Engineering in Medicine and Biology Society (EMBC), 2012, pp. 35533556.
[90] J. Ramos-Castro, J. Moreno, H. Miranda-Vidal, M. Garcia-Gonzalez,
M. Fernandez-Chimeno, G. Rodas, L. Capdevila, Heart rate variability analysis
using a seismocardiogram signal, in: 2012 Annual International Conference
of the IEEE Engineering in Medicine and Biology Society (EMBC), 2012,
pp. 56425645.
[91] K. Tavakolian, G. Portacio, N. Tamddondoust, G. Jahns, B. Ngai, G. Dumont, A.
Blaber, Myocardial contractility: a seismocardiography approach, in: 2012
Annual International Conference of the IEEE Engineering in Medicine and
Biology Society (EMBC), 2012, pp. 38013804.
[92] R.A. Wilson, V.S. Bamrah, J. L. Jr., M. Schwaiger, J. Morganroth, Diagnostic
accuracy of seismocardiography compared with electrocardiography for the

[93]

[94]

[95]

[96]

[97]
[98]
[99]
[100]

[101]

[102]
[103]
[104]
[105]

[106]

13

anatomic and physiologic diagnosis of coronary artery disease during


exercise testing, Am. J. Cardiol. 71 (7) (1993) 536545.
K. Li, S. Warren, A wireless reectance pulse oximeter with digital baseline
control for unltered photoplethysmograms, IEEE Trans. Biomed. Circuits
Syst. 6 (3) (2012) 269278.
B. Venema, N. Blanik, V. Blazek, H. Gehring, A. Opp, S. Leonhardt, Advances in
reective oxygen saturation monitoring with a novel in-ear sensor system:
results of a human hypoxia study, IEEE Trans. Biomed. Eng. 59 (7) (2012)
20032010.
E. Gil, P. Laguna, J. Martinez Cortes, O. Barquero-Perez, A. Garcia-Alberola,
L. Sornmo, Heart rate turbulence analysis based on photoplethysmography,
IEEE Trans. Biomed. Eng. PP(99) (2013) 11.
K. Madhav, M. Ram, E. Krishna, N. Komalla, K. Reddy, Robust extraction of
respiratory activity from ppg signals using modied mspca, IEEE Trans.
Instrum. Meas. 62 (5) (2013) 10941106.
N. Stuban, M. Niwayama, Optimal lter bandwidth for pulse oximetry, Rev.
Sci. Instrum. 83 (10) (2012), pp. 104708104708-5.
G. de Haan, V. Jeanne, Robust pulse-rate from chrominance-based rppg, IEEE
Trans. Biomed. Eng. PP(99) (2013) 11.
P. Klasnja, W. Pratt, Healthcare in the pocket: mapping the space of mobilephone health interventions, J. Biomed. Inf. 45 (1) (2012) 184198.
E. Kyriacou, C. Pattichis, M. Pattichis, An overview of recent health care
support systems for eemergency and mhealth applications, in: Annual
International Conference of the IEEE Engineering in Medicine and Biology
Society (EMBC 2009), September 2009, pp. 12461249.
A. Martnez, E. Everss, J.L. Rojo-lvarez, D.P. Figal, A. Garca-Alberola,
A systematic review of the literature on home monitoring for patients with
heart failure, J. Telemed. Telecare 12 (5) (2006) 234241.
A.S. Mosa, I. Yoo, L. Sheets, A systematic review of healthcare applications for
smartphones, BMC Med. Inf. Decis. Mak. 12 (1) (2012) 67.
A.G. Ekeland, A. Bowes, S. Flottorp, Effectiveness of telemedicine: a systematic review of reviews, Int. J. Med. Inf. 79 (11) (2010) 736771.
L. Tarassenko, A. Hann, D. Young, Integrated monitoring and analysis for early
warning of patient deterioration, Br. J. Anaesth. 97 (1) (2006) 6468.
C. Scully, J. Lee, J. Meyer, A. Gorbach, D. Granquist-Fraser, Y. Mendelson,
K. Chon, Physiological parameter monitoring from optical recordings with a
mobile phone, IEEE Trans. Biomed. Eng. 59 (February) (2012) 303306.
M.-Z. Poh, K. Kim, A. Goessling, N. Swenson, R. Picard, Cardiovascular
monitoring using earphones and a mobile device, IEEE Pervasive Comput.
11 (4) (2012) 1826.

Puneet Kumar Jain was born in Rajasthan, India, in 1988. He received his bachelor
of technology degree in computer science and engineering from the University of
Rajasthan in 2009 and master of technology degree in information system from
Delhi Technological University in 2011. He is currently a research scholar at Indian
institute of technology Jodhpur. His research interest includes image processing,
biometric system and biomedical signal processing.

Anil Kumar Tiwari received the master of technology degree in 2001 from the
Indian institute of technology Kharagpur and the Ph.D. degree in electronics and
communication from the Indian institute of technology Kharagpur in 2005. He is
currently an assistant professor at Indian institute of technology Jodhpur. His
research interest are in Image Processing, Video Processing and Signal Processing
application in Bio-Medical.