Sensors: Anomaly Detection Using Electric Impedance Tomography Based On Real and Imaginary Images

sensors
Article
Anomaly Detection Using Electric Impedance
Tomography Based on Real and Imaginary Images
Imam Sapuan 1 , Moh Yasin 1 , Khusnul Ain 2, * and Retna Apsari 1
1 Department of Physic, Faculty of Science and Technology, Universitas Airlangga, Surabaya 60115, Indonesia;
i_sapuan@fst.unair.ac.id (I.S.); yasin@fst.unair.ac.id (M.Y.); retna-a@fst.unair.ac.id (R.A.)
2 Biomedical Engineering, Faculty of Sciences and Technology, Universitas Airlangga, Surabaya 60115,
Indonesia
* Correspondence: k_ain@fst.unair.ac.id

Received: 21 January 2020; Accepted: 15 March 2020; Published: 30 March 2020
Abstract: This research offers a method for separating the components of tissue impedance, namely
resistance and capacitive reactance. Two objects that have similar impedance or low contrast can
be improved through separating the real and imaginary images. This method requires an Electrical
Impedance Tomography (EIT) device. EIT can obtain potential data and the phase angle between
the current and the potential measured. In the future, the device is very suitable for imaging organs
in the thorax and abdomen that have the same impedance but different resistance and capacitive
reactance. This device consists of programmable generators, Voltage Controlled Current Source
(VCCS), mulptiplexer-demultiplexer potential meters, and phase meters. Data collecting was done by
employing neighboring, while reconstruction was used the linear back-projection method from two
different data frequencies, namely 10 kHz and 100 kHz. Phantom used in this experiment consists of
distillated water and a carrot as an anomaly. Potential and phase data from the device is reconstructed
to produce impedance, real, and imaginary images. Image analysis is performed by comparing the
three images to the phantom. The experimental results show that the device is reliable.
Keywords: Electrical Impedance Tomography; detection of anomaly; resistive and capacitive image;
real and imaginary image
1. Introduction
Mapping the electrical impedance distribution of an object is the basic principle of Electrical
Impedance Tomography (EIT) technology. The impedance of human body tissue is able to provide
information about the physiological and pathological properties of the tissue [1–3]. Both of these
properties are related to the information of medical applications [1,2,4–12]. This technology is
carried out by injecting electrical currents and measuring the voltage generated through the electrodes.
The measurement of voltage with forward-problem results in voltage distribution. An inverse problem is
the reconstruction process to obtain an impedance distribution [3,4,13–16]. The body tissue impedance
consists of two parts, which are real resistance and imaginary capacitive reactance [3,14,16,17].
The present study separates the distribution of those two impedance components.
The EIT imaging technique is noninvasive, portable, simple, fast, easy to use, economical,
comfortable for the patient’s body, it also does not cause ionizing radiation [18,19]. This technology
has been applied in many clinical applications such as detection for breast cancer [6], nerve activity in
the brain [7], respiratory disorders [10,20], and lung function detection [4].
The basic principle of EIT in detecting abnormalities is based on the impedance differences
between normal tissue and anomalies. The EIT is able to detect abnormalities in body tissue if both
impedances are different [4–7]. For the benefit of EIT researches, normal tissue objects and anomalies
Sensors 2020, 20, 1907; doi:10.3390/s20071907 www.mdpi.com/journal/sensors

Sensors 2020, 20, 1907 2 of 18
use phantoms that are made of materials with different impedances. Some of EIT applications include
detection of breast cancer, normal breast tissue, and cancer using different impedances [3,14,21].
Multi-frequency Electrical Impedance Tomography (Mf-EIT) [3,14,16] is a type of EIT which
produces many structural images for different frequencies. The resulting image is an impedance
distribution as a function of frequency. The images are impedance images from 5, 10, 25, 50, 75,
and 100 kHz [16]. These results show the image of resistivity inhomogeneity is caused by differences.
The weakness of the impedance image generated from the MF-EIT system [8,13] is its inability to
distinguish objects that have the same impedance but are different tissues. Some tissues that have
the similar impedance include liver and lungs with the impedance value of 1.622 Ω and 1.624 Ω,
heart and deflated lung with the impedance value of 1.001 Ω and 1.077 Ω at the frequency of 10 khz.
While the impedance value of colon and cervix are respectively 3.3168 Ω and 3.2875 Ω at the frequency
of 100 kHz. The data was obtained from a simulation developed by C. Gabriel [22]. The impedance
of these organs has different conductivity and permittivity even with the similar impedance. As a
result, the Mf-EIT system does not show the differences in these kind of tissue. However, if real and
imaginary images are produced, it will be possible to show the difference between the two objects
which have the same impedance.
To overcome the disadvantage of the Mf-EIT system, we have built EIT which can break down
two components of the body tissue impedance, namely resistive resistance and capacitance [23–25].
The concept used to break down two impedance components is by using the resistor capasitor RC series
model [23,25], which has been applied to Bioelectrical Impedance Analysis (BIA) devices. This method
uses the properties of the circuit resistance, capacitance, impedance, and phase shift in the alternating
current AC circuit. The whole series and or parallel circuits consisting of resistors R and capacitors C
has an equivalency to RC series [23,26]. The circuit consists of RC components in series and or parallel
relationships, regardless of the number and value of RC it could be simplified in the final form of RC
series with a certain value. Body tissue composed of unknown R and C, which is analogous to the
concept, could be replaced with RC series with the equivalent R and C values.
The practice of this method begins by assuming the body tissue as a RC series. Furthermore,
the body tissue is injected with AC current. At some point voltage and "phase difference" will be
measured. The study used the Neighboring data collection method with 16 electrodes. This method
works by injecting current and measuring the voltage between the two closest electrodes [4].
The intended voltage is the difference in electrical potential at two specific points when current
is injected into the object. Meanwhile, the phase difference is the phase difference between the voltage
and the injection electric current. From these two data, impedance, resistive and capacitive reactance
can be calculated. This concept is applied to build EIT in mapping the impedance, resistive and
capacitive reactance of body tissues. In the reconstruction process, the resistive reactance potential
produces a real image, capacitive reactance produces a capacitive image, and the impedance produces
an impedance image.
The study has designed and developed the EIT device without involving a high-speed data
acquisition system but it can produce impedance image components that include real and imaginary
images. To show anomalies in normal tissues, EIT utilizes a functional image [27]. The image uses two
different frequencies of injection current [4,27]. The research used current sources at the frequency
10 kHz and 100 kHz. EIT produces three functional images, namely impedance image, resistive image,
and capacitive image. These three images can be used in the processes of identifying and analyzing
tissues. Moreover, with these three images, tissue diagnostics and analysis are improved to be more
accurate than by using only one impedance image.
2. Materials and Methods

Electrical Impedance Tomography (EIT) focuses on the separation of electrical impedance
components, which are the resistance (part of the real impedance) and the capacitive reactance
(part of the imaginary impedance) [3]. The experimental EIT flowchart in this study was shown in
Sensors 2020, 20, 1907 3 of 18
Figure
Sensors 1. The
2020, 20, study
x FOR PEER began
REVIEWwith
the design of the EIT data acquisition system. Next, we examined3 of 18 the
performance of the equipment tested in each section, and carried out integration and synchronization.
Synchronization
Synchronization andand integration
integration combine
combine allall equipmentincluding
equipment includinghardware
hardwareandandsoftware
software into one
one EIT
EIT system.
system. The last
last stage
stage was
was an
an experiment
experiment byby testing
testing the
the performance
performance of
of the
the EIT
EITsystem
systemto
toobtain
obtainreal
realand
andimaginary
imaginary images.
images.
Figure 1. Schematic diagram of the experimental procedure.

Figure 1. Schematic diagram of the experimental procedure.
The real and imaginary part of impedance can be measured using Fourier transforms. The method
isThe real however
simple, and imaginary part
we try to of impedance
avoid can be
data acquisition measured
systems using Fourier
or high-speed transforms.
sampling systemsThe in this
method is simple, however we try to avoid data acquisition systems or high-speed sampling
study. The method for obtaining resistance and reactance of an object is to inject AC into the object. systems
in this
We study. The method
then measure for obtaining
the voltage and theresistance and reactance
phase difference between of the
an object
currentisinjection
to inject and
AC into the
the voltage.
object. We then measure the voltage and the phase difference between the current injection
These steps are performed for all electrodes attached to the object. This phase difference measurement and the
voltage. These
uses the AC steps
series are performed
circuit for all
concept [23]. electrodes
Various attached
combined to theand
RC series object. This
parallel phasecan
circuits difference
be replaced
measurement uses the AC series circuit concept [23]. Various combined RC series and parallel
with a value equivalent to a single RC series value [23,26]. Equations (1) and (2) express the relationship circuits
can between
be replaced with a value
impedance, phase equivalent
difference, to a single and
resistance, RC capacitive
series value [23,26]. in
reactance Equations (1) and
an RC series (2)
circuit.
express the relationship between impedance, phase difference, resistance, and capacitive reactance in
q
an RC series circuit.
Z = R2 + X2c (1)
𝑍 = R +X (1)
Xc
tan = (2)
X R
tan  = (2)
where Z is the impedance, R is the resistance, XC isRcapacitive reactance, and θ is the phase difference.
These
where Z equations can be used
is the impedance, R to
is determine the value
the resistance, XC isof capacitive
R and XC . The value ofand
reactance, R and θ X are expressed
isC the phase
by Equations
difference. These (3) and (4). can
equations Theberelation
used toof determine
R, Xc, and the
Z isvalue
shownofinRFigure
and XC2.. The value of R and XC
are expressed by Equations (3) and (4). The relation of R, Xc, and Z is shown in Figure 2.
R = Z cos θ (3)
Xc = Z sin θ (4)
Figure 2. The relation of R, Xc, and Z.
R = Z cos θ (3)
X
tan  = (2)
R
where Z is the impedance, R is the resistance, XC is capacitive reactance, and θ is the phase
difference. These equations can be used to determine the value of R and XC. The value of4Rof and
Sensors 2020, 20, 1907 18
XC
are expressed by Equations (3) and (4). The relation of R, Xc, and Z is shown in Figure 2.
Sensors
Sensors 2020,
2020, 20,20, x FOR
x FOR PEER
PEER REVIEW
REVIEW 4 of
4 of 1818
XcX=c Z
=Z sin
sin θθ (4)(4)
Figure3 3shows
showsthat thatresistance
resistance and capacitive
The reactance areobtained
obtained by measuring the voltage
Figure Figure
and
Figure 2.relation
2.capacitive of R, Xc,
reactance
The relation and
Xc,Z.
are
of R, and Z. by measuring the voltage
at a pair of adjacent electrodes, and the phase difference
at a pair of adjacent electrodes, and the phase difference (θ) between the (θ) between the injection
injection current
current (I)(I)
andand the
the
Figure 3 shows
resultingvoltage.
voltage.The that resistance
Theimpedance and
impedancebetween capacitive
betweenthe thetwo reactance
twoelectrodes are obtained
electrodesis isassumed by measuring
assumedtotobebethe theRC the
RCseries voltage
seriesvalue.
value.
resulting
at The
a pair of adjacent
result of electrodes,
measuring the and the phase
potential between R =two
difference
the Z cos
(θ) θbetweenisthe
electrodes theinjection
voltage current
(V). To (I) and
get a the (3)
current
The result of measuring the potential between the two electrodes is the voltage (V). To get a current
resulting
injection,voltage. Thetoimpedance betweencircuit
the two required.
electrodesThis
is assumed totobeconvert
the RCcurrent
series value.
injection, a acurrent
current voltageconverter
to voltage converter circuit is isrequired. circuitis isto
This circuit convert current signals
signals
The result
intovoltageof measuring
voltagesignals
signalsVIV the potential
. Thevoltage between
voltagesignals the
signalsVIVand two electrodes is the voltage (V).
I and V were connected to the phase detector inputTo get a current
into . IThe V were connected to the phase detector input
injection,
circuit. aThe
current
output to voltage
of the converter
phase circuit
detector is required.
displayed the This circuit
direct current is to
DC convert
voltage current signals into
circuit. The output of the phase detector displayed the direct current DC voltage asas a function
a function ofof the
the
voltage
phase signals
differenceVI . between
The voltage the signals
two VI and
signals. The Vphase
were connected
difference to the phase
detected by detector
the phase input circuit.
detector is the
phase difference between the two signals. The phase difference detected by the phase detector is the
The output
same of output
the phase detector displayed the direct current DC voltage as a function of the phase
same asas
thethe voltage.
output voltage.
difference between the two signals. The phase difference detected by the phase detector is the same as
the output voltage.
Figure
Figure
Figure 3.The
3.3.The The
schemescheme inthe
themeasurement
measurement
in theinmeasurement
scheme ofofimpedance
of impedance impedance
components components separately
separately through
components througha a
a measurement
separately through
of measurement
voltage and of
measurement of
phasevoltage and
difference.
voltage phase difference.
and phase difference.
TheThe
The injection
injection current
injection measurement
current
current measurement
measurement results (I) and
results
results voltage
(I)(I)
and (V)
andvoltageare used
voltage toare
(V)are
(V) determine the magnitude
usedtotodetermine
used determine the
the
of the impedance
magnitude of (Z).
the The phase
impedance difference
(Z). The (θ) is
phase measured
difference using
(θ) a
is phase detector.
measured
magnitude of the impedance (Z). The phase difference (θ) is measured using a phase detector. Resistance
using a phase (R) and
detector.
capacitive
Resistance
Resistance reactance
(R)
(R) and
and (Xc) could
capacitive
capacitive be calculated
reactance
reactance (Xc)
(Xc) using
could
could beEquations
be (3)
calculated
calculated and Equations
using
using (4). The EIT
Equations data
(3)(3) and
and acquisition
(4).
(4). The
The EIT
EIT
system
data is needed to
dataacquisition
acquisition apply is
system
system the concept
isneeded
neededto oftoapply
resistance
applythe and
the capacitive
concept
concept reactanceand
ofofresistance
resistance measurement.
capacitiveThe
andcapacitive block
reactance
reactance
diagram of the
measurement. data
The acquisition
block system
diagram of was
the shown
data in Figure
acquisition 4.
system
measurement. The block diagram of the data acquisition system was shown in Figure 4. was shown in Figure 4.
. .
Figure 4. Block diagram of Electric Impedance Tomography System for real and imaginary
Figure4. 4.
functional
Figure Blockdiagram
images.
Block diagramofofElectric
ElectricImpedance
ImpedanceTomography
TomographySystem
Systemforforreal
realand
andimaginary
imaginary
functional
functional images.
images.
Thissystem
This systemconsisted
consistedofofoscillator,
oscillator,buffer,
buffer,and
andVoltage
VoltageControlled
ControlledCurrent
CurrentSource
Source(VCCS),
(VCCS),
multiplexer of current injection, object (phantom), de-multiplexer of voltage measurement,
multiplexer of current injection, object (phantom), de-multiplexer of voltage measurement, current current
totovoltage
voltageconverter,
converter,differential
differentialamplifier,
amplifier,ADAD8302
8302phase
phasedetector
detectorbybyATI
ATITechnologies
TechnologiesInc
Inc
(Markham, Canada), and Arduino
(Markham, Canada), and Arduino Mega. Mega.
Sensors 2020, 20, 1907 5 of 18
This system consisted of oscillator, buffer, and Voltage Controlled Current Source (VCCS),
multiplexer of current injection, object (phantom), de-multiplexer of voltage measurement, current to
voltage converter, differential amplifier, AD 8302 phase detector by ATI Technologies Inc (Markham,
Canada),
Sensors and
2020, 20, Arduino
x FOR Mega.
PEER REVIEW 5 of 18
2.1.2.1. Voltage
Voltage Controlled
Controlled Current
Current Source
Source (VCCS)
(VCCS)
VCCS
VCCS hashas a very
a very important
important role
role as as a constant
a constant current
current source.
source. Figure
Figure 5 shows
5 shows a constant
a constant current
current
source generator scheme consisting of AD 9850 modules by Digi-Key Electronics
source generator scheme consisting of AD 9850 modules by Digi-Key Electronics (Thief River Falls, (Thief River Falls,
MN,MN, USA),
USA), DCDC blocks,
blocks, amplifier
amplifier circuits,
circuits, buffers,
buffers, andand VCCS
VCCS double
double op-amps.
op-amps. TheThe working
working principle
principle
of VCCS starts from the sinusoidal DC signal generated by the AD 9850
of VCCS starts from the sinusoidal DC signal generated by the AD 9850 modules. Then themodules. Then the sinusoidal
DC signalDC
sinusoidal is passed
signalthrough the DC
is passed block the
through circuit
DCtoblock
generate an AC
circuit to signal. Thean
generate non-inverting
AC signal.op-amp
The
amplifies theop-amp
non-inverting DC block outputthe
amplifies signal
DC as VCCS
block input.
output Theas
signal VCCS
VCCS input voltage
input. is proportional
The VCCS to the
input voltage
is proportional to the current generated by VCCS. The signal amplifier circuit is used to increase theby
current generated by VCCS. The signal amplifier circuit is used to increase the current generated
VCCS.generated
current This VCCS by generates
VCCS. This sinusoidal currents sinusoidal
VCCS generates of 1 mA, 10 to 100 of
currents kHz.
1 mA,To keep
10 to the
100AC current
kHz. To
constant, the buffer is placed before the VCCS double op-amp.
keep the AC current constant, the buffer is placed before the VCCS double op-amp
Figure 5. A signal generator and constant current.

Figure 5. A signal generator and constant current.
The maximum voltage generated by the AD 9850 module is 0.313 volts. The non-inverting op-amp
The maximum voltage generated by the AD 9850 module is 0.313 volts. The non-inverting
amplifier functions to amplify the signal output of the 9850 AD module. The magnitude of this
op-amp amplifier functions to amplify the signal output of the 9850 AD module. The magnitude of
amplification will affect the amount of current generated by VCCS. The resistor in VCCS dual op-Amp
this amplification will affect the amount of current generated by VCCS. The resistor in VCCS dual
Amp RF , Rs, and R1 have the same value of 2 kΩ. The amount of gain depends on the amount of
op-Amp Amp RF, Rs, and R1 have the same value of 2 kΩ. The amount of gain depends on the
current needed. The VCCS circuit is loaded RL = 1, 2, and 3 kΩ. VCCS characterization is related to
amount of current needed. The VCCS circuit is loaded RL = 1, 2, and 3 kΩ. VCCS characterization is
the stability of the current generated with respect to frequency and load.
related to the stability of the current generated with respect to frequency and load.
2.2. Electric Mechanism of Current Injection and Voltage Measurement, Current to Voltage Circuit to Measure
2.2.Phase
Electric Mechanism of Current Injection and Voltage Measurement, Current to Voltage Circuit to
Difference
Measure Phase Difference
This study used 16 electrodes namely electrodes E1 , E2 , E3, . . . , E16 . The process began with
This study
injecting used current
an electric 16 electrodes
through namely electrodes
the adjacent E1, EE2,1 Esignal
electrode 3,..., E-16E . 2The process
ground. began
At the same with
time,
injecting an electric
the voltage in the current throughelectrodes
pair of adjacent the adjacent (E2electrode
–E3, E3 –E4,E1Esignal
4 –E 5 , . .
- .E 2Eground.
, 16 –E 1 ) At
were the same
measured. time,
Next,
thethe
voltage
current inwas
the injected
pair of adjacent
into the Eelectrodes
2 signal - E(E –E3, E3–E
3 2ground 4, E4–E5,...and
electrode , E16–Ethe1)voltage
were measured.
in the pairNext, the
of adjacent
current was injected
electrodes (E3 –E4, Einto
4 –E5,the
E5 –E , . . . , E- 1E–E
E2 6signal 3 ground electrode
2 ) were again and theThis
measured. voltage in the
process waspair of adjacent
repeated until all
electrodes
adjacent (E 3–E4, E4–E
electrodes 5, E5–E6,current
became ..., E1–Einjection
2) were again measured. This process was repeated until all
electrodes.
adjacent electrodes became current injection electrodes.
2.2.1. Electronic mechanism for Current Injection
2.2.1. Electronic
The processmechanism for Current
of EIT current Injection
injection was carried out by an electronic mechanism. This electronic
mechanism is shown in Figure 6. (current part). out
The process of EIT current injection was carried Theby selection of themechanism.
an electronic current injector
This electrode
electronicon
the objectisused
mechanism shownonein
IC4052
Figurede-multiplexer and
6. (current part). two
The IC 4051of
selection multiplexers. The electrode
the current injector mechanism
electrode on the
for measuring potential differences between two adjacent electrodes was controlled
object used one IC4052 de-multiplexer and two IC 4051 multiplexers. The electrode mechanism by two ICfor
4051
de-multiplexer.
measuring Thedifferences
potential designed electric
between mechanism scheme
two adjacent is shownwas
electrodes in Figure 6. by two IC 4051
controlled
de-multiplexer. The designed electric mechanism scheme is shown in Figure 6.
Sensors
Sensors 20,20,
2020,
2020, 1907PEER REVIEW
x FOR 6 of6 18
of 18
Figure
Figure 6. The
6. The electric
electric mechanism
mechanism forfor selecting
selecting a current
a current injector
injector electrode
electrode and
and measuring
measuring voltage.
voltage.
For For technical

technical reasons,
reasons, thethe process
process ofof current
current injection
injection and and voltage
voltage measurement
measurement in inthisthis system
system
was
was divided
divided into
into two two mechanisms.
mechanisms. In In
thethe
firstfirst mechanism,
mechanism, oddodd electrodes
electrodes werewere used used asas thethe signal
signal
while even electrodes used as ground. On the other
while even electrodes used as ground. On the other hand, in second mechanism, odd electrodes hand, in second mechanism, odd electrodes were
used as ground and even electrodes
were used as ground and even electrodes used as signal. used as signal.
The The firstfirst mechanism
mechanism process
process was was started
started byby injecting
injecting electric
electric current
current toto the the adjacent
adjacent electrodes
electrodes
which are
which are E1 signal–E E 1 signal–E 2 ground,
2 ground,
followed by measuring the voltage
followed by measuring the voltage of the adjacent electrodes pair of the adjacent electrodes pair
(E(E
2-E
–E
23, E3,
3–E4,
E3 –E 4, E
E4–E –EE,16. –E
5,4 ..., 5
. . 1,).E16
The–E1next
). The next processes
processes were injecting
were injecting currentcurrent to E3 signal–Eto E3 signal–E 4 ground 4 ground
and
measuring the voltage of the adjacent electrode (E4–E5, E5–E6, E6–E7,..., E2–E3). This processprocess
and measuring the voltage of the adjacent electrode (E 4 –E 5, E 5 –E 6, E 6 –E 7 , . . . , E 2 –E 3 ). This was
was repeated
repeated until all until
adjacentall adjacent
electrodes electrodes
became became
currentcurrentinjection injection
electrodes. electrodes.
InIn Figure
Figure 6, the6, the process
process of injecting
of injecting electric
electric currents
currents waswas startedstarted fromfrom the in-signal
the in-signal connectedconnected to
4052 through x1 and the signal exited through via Z1. The relationship between x1 and Z1Zwas
to 4052 through x 1 and the signal exited through via Z 1 . The relationship between x 1 and 1 was
controlled
controlled byby ananelectrical
electricalmechanism
mechanismbybyControl ControlSelect SelectCS, CS,namely
namelyINH INH==0,0,BB 0=0 = 0, 0,andandAA 0 =0 1.
= 1.
Furthermore,
Furthermore, Z1Z was 1 was connected
connected to input
to input Y on Y onIC IIC4051,I 4051, through
through outputsoutputs (X1,(X X2,1 ,XX32,, XX4,3X,X XX
5, 4,
, X6,8)X7 ,
6, 5X7, X
X
which ) which then were connected to electrodes for current
8 then were connected to electrodes for current injection to objects (E1, E3, E5, E7, E9, E11, E13, injection to objects (E1, E3, E5, E7, E9, E11,
E13,and
E15) E15) thenandpaired
then paired with electrodes
with electrodes (E2, E4,(E2,E6, E4,E8,E6,E10,E8, E10,E12,E12, E14,E14, E16)E16) which which were were forwarded
forwarded to to
ground. The connection with ground through de-multiplexer IC II 4051 (X1, X2, X3, X4, X5, X6, X7, X8) 8is) is
ground. The connection with ground through de-multiplexer IC II 4051 (X 1 , X 2, X 3 , X 4, X 5 , X 6, X 7 , X
thethe input
input whilewhile Y is Y isthethe output.
output. The Thesetset
upupof of
thethe electrical
electrical mechanism
mechanism onon ICIC I 4051
I 4051 was was done
done through
through
CSCS (A(A1, B1 ,1 BC11)Cwhile
1 ) while IC IC II 4051
II 4051 was was done
done through
through CSCS (A(A2, 2B,2BC22). C2Y). IC Y IC II 4051
II 4051 pinpin was was connected
connected withwith
Z
Z2 IC IC 4052 as input that was connected to output y .
2 4052 as input that was connected to output y1. 1Finally, the current through the resistor was Finally, the current through the resistor was
connected to ground. The resistor which was passed by the current injection was then connected in in
connected to ground. The resistor which was passed by the current injection was then connected
parallel
parallel with with thethe op-ampop-amp amplifier,
amplifier, soso that
that atatthe the output
output ofof the the op-amp
op-amp thethevalue value and and thetheshapeshape of of thethe
injection current were obtained. The op-amp amplifier circuit in this position functions as a current to
injection current were obtained. The op-amp amplifier circuit in this position functions as a current
tovoltage
voltage(C (Cto toV) V)circuit.
circuit.
The second
The second mechanism mechanism process
process was
was started
started byby injecting
injecting electric
electric current
current to to thethe adjacent
adjacent electrodes
electrodes
which are
which are E2 signal–E E 2 signal–E 3 ground,
3 ground, then measuring the voltage of the adjacent electrodes pair (E3–E4, E(E
then measuring the voltage of the adjacent electrodes pair 4–E 3 –E
5, 4,
E –E , E –E . . . , E –E ). The process of current injection
E5-E6..., E1–E2). The process of current injection was continued on the E4 signal–E5 ground, and followed 5
4 5 5 6 1 2 was continued on the E 4 signal–E
byground,
measuring and the followedvoltage byon measuring
the pair of theadjacent
voltageelectrodes
on the pairwhich of adjacentwere (E electrodes which were (E –E
5–E6, E6–E7, E7–E8, ..., E1–E52). 6,
E –E
This6 process E
7, 7 was –E , . .
8 repeated. , 1E –E ). This process was repeated until all
2until all adjacent electrodes became current injection electrodes. adjacent electrodes became current
injection
In Figure electrodes.
6, the process of injecting electric currents started from in-signal that was connected to
4052 through y0 6,
In Figure and theexited
process of injecting
through Z2. Theelectric currents started
relationship betweenfrom in-signal
y0 with Z2 was thatcontrolled
was connected by
electrical mechanism by CS, namely INH = 0, B0 = 0, and A0 = 0. Furthermore, Z2 was connected to theby
to 4052 through y 0 and exited through Z 2 . The relationship between y 0 with Z 2 was controlled
Y electrical
input on IC mechanism
II 4051, through by CS, namely the outputsINH (X = 10, =3,0,Xand
, XB2,0 X A = 0. Furthermore, Z was connected to
4, X5, X06, X7, X8) which were 2connected to the
the Y input on IC II 4051, through
electrodes for the current injection to the objects alternately the outputs (X 1 , X X
2, 3 (E , X24,, EX45, ,EX6,6,EX , X ) which were connected
8, 7E10, 8E12, E14, E16) and paired
with electrodes (E1, E3, E5, E7, E9, E11, E13, E15) which then were forwarded to ground. The connection
Sensors 2020, 20, 1907 7 of 18
to the electrodes for the current injection to the objects alternately (E2 , E4 , E6 , E8, E10 , E12, E14 , E16 )
and paired with electrodes (E1 , E3 , E5 , E7, E9 , E11, E13 , E15 ) which then were forwarded to ground.
The connection with ground via IC I 4051 (X1 , X2, X3 , X4, X5 , X6, X7 , X8 ) as inputs and Y as output.
The setting of the electrical mechanism on IC I 4051 was done through CS (A1 , B1 C1 ) and on IC II
4051 was done through CS (A2 , B2 C2 ). Next, Y IC I 4051 pin was connected to Z1 IC 4052 as input
connected to output x0 , then the current through the resistor was connected to ground. Similar to the
first mechanism, the resistor which was passed by the current injection current was then connected in
parallel with the op-amp amplifier, so that on the output the value and form of the current injection
were obtained. The op-amp amplifier circuit in this position functions as a current to voltage (C to V)
circuits. This series of C to V generated value and form of current signals.
2.2.2. Electronic Mechanism of Voltage Measurement

The voltage measurement mechanism of the EIT system was electrically controlled. The electric
circuit scheme of the voltage measurement process is shown in Figure 6. (voltage part). The circuit used
was an AD 8039 op-amp buffer and two IC 4051 multiplexers. All electrodes to measure the voltage
were connected to the buffer circuit first before being connected to a multiplexer. Odd electrodes
(E1 , E3 , E5 , E7, E9 , E11, E13 , E15 ) were each connected to a buffer then to the IC III 4051 multiplexer,
respectively, at the inputs (X1 , X2, X3 , X4, X5 , X6, X7 , X8 ). Next, the even electrodes (E2 , E4 , E6 , E8, E10 ,
E12, E14 , E16 ) were also connected with a buffer then to the IC IV 4051 multiplexer, respectively, at the
inputs (X1 , X2, X3 , X4, X5 , X6, X7 , X8 ).
Voltage measurement on the electrode pair was carried out after injecting the current. The selection
of the active odd electrode was controlled by IC III 4051 through CS (A3 , B3 , C3 ) while the selection of
the active even electrode was controlled by IC IV 4051 through CS (A4 , B4 , C4 ). In order to measure the
voltage between E1–E2, electrode 1 and electrode 2 were activated by giving CS values on IC III 4051
(A3 = 0, B3 = 0, C3 = 0) and IC IV 4051 (A4 = 0, B4 = 0, C4 = 0). Same methods were done to activate
other electrodes and measure the voltage, which was to change both CS inputs simultaneously.
Output pin Y IC III 4051 and Y IC IV 4051 were connected with differentiator amplifiers which
produced voltage value and form between the two adjacent electrodes. The value and form of this
signal is called the voltage signal.
2.2.3. Measurement of Voltage and Phase Difference

Figure 7 shows a circuit schematic for the measurement of voltage and phase difference. The output
voltage signal in the scheme is a sinusoidal signal differential between two adjacent electrodes. These two
electrodes were used to measure the potential difference in an object caused by an injection current.
In the application, the output voltage signals were connected to two connections. First, it was connected
to the root mean square RMS to DC input. This RMS to DC circuit is also called an RMS to DC converter
circuit. The input signal of this circuit is sinusoidal, so the output voltage is a DC voltage of Vrms from
the potential difference measured on the object.
Second, the output voltage signal was connected to input A in AD8302 phase detector.
This connection functions as a comparison of current signals output which was connected to input
B in the AD8302 phase detector. Current signals and voltage signals provide information of phase
difference between current injection and voltage. The phase detector output is a DC voltage, the value
of which is proportional to the phase difference of the two signals. The phase difference of the two
signals which was detected by the AD8302 phase detector is expressed by Equation (5). The value of
this phase corresponds to the scheme and data-sheet components of AD8302. In the data sheet, VPSET
is connected to the voltage reference in the module. The equation is to generate an AD 8302 module
output voltage, as a measured phase.

900mV − VPSET (V ) ◦ ◦
◦ + 90 = Phase sp( ) (5)
10mV/

Sensors 2020, 20, x FOR PEER REVIEW 8 of 18

Sensors 2020, 20, 1907 8 of 18
Figure 7. Measurement scheme of voltage and phase difference.

The circuit
The circuitscheme scheme (see(seeFigure
Figure 7) 7)hashastwotwofunctions
functions that havehavetotomeetmeet certain
certain criteria.
criteria. To
To measure
measure
the the voltage
voltage of of RMS
RMS to DC to circuit,
DC circuit, it hasit to
hasmeet
to meet
the the voltage
voltage stability
stability function
function at the atrequired
the required frequency.
The circuit scheme (see Figure 7) has two functions that have to meet certain criteria. To
frequency. To achieve
Tomeasure
achieve thisthe
thisvoltage
goal, goal, the experiment wastodone to observe the stability of theand output and
the of experiment
RMS to DC was done
circuit, it hasobserve
to meetthethestability
voltage of the output
stability function at input
the voltage
required
inputatvoltage
a certainatvaluea certain
of the value
RMSgoal, toofDC thecircuit
RMSattovarious DC circuit at various frequencies. Then the were
frequency. To achieve this the experiment wasfrequencies.
done to observe Thenthe thestability
experimentalof the data
output and
experimental
presented databy were
a graphpresented by a graph of the relationship between input andatoutput voltages
input voltage at a of the relationship
certain value of the between
RMS input
to DC and output
circuit atvoltages various
various frequencies. frequencies.
Then the
at various
The frequencies. The tools used in this performance test include a signal generator, RMS to DC
experimental data were presented by a graph of the relationship between input and output storage
tools used in this performance test include a signal generator, RMS to DC circuit and digital voltages
circuitoscilloscope
and digital BK storage oscilloscope
Precision 2542 B. BK Precision 2542 B.
at various frequencies. The tools used in this performance test include a signal generator, RMS to DC
The second
Theand function
second functionis toismeasure
to measure thethephase
phase difference
difference of two sinusoidal
sinusoidalinput inputsignals.
signals.The The
circuit digital storage oscilloscope BK Precision 2542of B. second
second parameter
parameter that needsto betoknown
be known theislinearity
the linearity of thedifference
phase difference conversion to thephase
The that
secondneeds function is to is measure of the phase
the phase difference of two conversion
sinusoidal to the
inputAD8302
signals. The
AD8302 phaseoutput
detector detector output
voltage. The voltage.
tools used Theto tools usedthe
determine tocharacteristics
determine the of characteristics
this component of
are this
MyDAQ
second parameter that needs to be known is the linearity of the phase difference conversion to the
component
module, are MyDAQ
AD8302 module,
phase AD8302
detector, phase
digital detector,
storage digital
oscilloscope storage
BK oscilloscope
Precision 2542 BK
B, Precision
and voltmeters.
AD8302 phase detector output voltage. The tools used to determine the characteristics of this
2542 This
B, and voltmeters.
experiment used This experiment
themodule,
MyDAQAD8302 usedgenerator
signal the MyDAQ signal generator
to produce to produce twophase
component are MyDAQ phase detector, digital two
storagesinusoidal
oscilloscope signalsBKofPrecision
sinusoidal signals
difference. Theof experiment
phase difference. process Thewas experiment
started process
with was started
generating signals with
1 andgenerating
2 at the signals
same 1
frequency,
2542 B, and voltmeters. This experiment used the MyDAQ signal generator to produce two
and 2while
at thein same
sinusoidal
frequency,
signal while
2, theofvarious
signals
in signal 2,
phase difference
phase difference.
the various phase
ϕ was added.
The experiment
difference
At signal
process
φ was
1, V(t)with
was started = A sin(ωt)
added. At signal
generating
1,
serves as A 1
signals
V(t) =input
A sin(ωt)
and serves
at signal as2A V(t) =
input A and
sin(ωt +
at signal
ϕ) 2 V(t) =asABsin(ωt
served input + φ)
in the served
phase
and 2 at the same frequency, while in signal 2, the various phase difference φ was added. At signal 1,
as B inputThe
detector. in the phasephase
AD8302
detector. The AD8302
detector aphase detector output is a DC with voltagemulti-meter.
that can be measured with output
a
V(t) = Aoutput
sin(ωt)isservesDC voltage
as A input thatand can atbesignal
measured 2 V(t) = Aasin(ωt + φ) servedThe phase detector
as B input in the phase
multi-meter.
voltage The phase
is aThe
linear detector
function output
in phase voltage
difference is a linear
of the function in phase difference of the two
detector. AD8302 phase detector output is atwo DCsignals.
voltageThe that setup
can ofbethemeasured
Phase Detector
with a
signals. The setup
linearity test of the
tool is Phase
shown Detector
in Figure linearity
8. test tool is shown in Figure 8.
multi-meter. The phase detector output voltage is a linear function in phase difference of the two
signals. The setup of the Phase Detector linearity test tool is shown in Figure 8.
Figure 8. The set up for the linearity test of Phase Detector circuit.
2.3. Reconstruction
2.3. Reconstruction
The image reconstruction was done by using the reconstruction method of Linear Back Projection
The image
(LBP). reconstruction
The algorithm was done
of this method by using
is based the reconstruction
on a collection of neighboring method of Linear
data with circularBack
geometry
2.3. Reconstruction
Projection (LBP). The algorithm
and equipotential based. of this
The method is based Linear
equipotential-based on a collection
Back of neighboring
Projection algorithmdata
waswith
initially
The image reconstruction was done by using the reconstruction method of Linear Back
circular geometry
proposed by and equipotential
Barber-Brown [4] which based.
in The
matrix equipotential-based
notation can be expressedLinear
in Back
Equation Projection
(6).
Projection (LBP). The algorithm of this method is based on a collection of neighboring data with
algorithm was initially proposed by Barber-Brown [4] which in matrix notation can be expressed in
circular geometry and equipotential based. The equipotential-based Linear Back Projection
Equation (6). [δρ ](px1) = [F](pxp) [B](pxq2 ) [δVn ](q2 x1) (6)
algorithm was initially proposednby Barber-Brown [4] which in matrix notation can be expressed in
Equation [δρ ]( ) = [F]( ) [B]
(6). number of elements [δV ] (6)weight
where p is the and q is the number of electrodes, [B] is the back projection
[δρ ](of filters,
matrix, [F] is the representative matrix ) = [F][(δVn)][B] [δV ] change in normalized boundary,
is the potential (6)
and [δρn ] is a normalized change in impedance distribution.
Sensors 2020, 20, 1907 9 of 18
The linearization method assumes that potential boundary changes are a linear function of changes
in conductivity [4], so Equation (6) can be written in matrix as stated:
[δV](q2 x1) = [S](q2 xp) [δρ](pxp) (7)
where [δV] is the potential boundary change, [S] is the sensitivity matrix, and [δρ] is the resistivity
change. In order to obtain the sensitivity matrix [S], a variation of δρ on all ρ in forward problem model.
Equation (7) can be solved after the matrix [S] was measured by algebraic manipulation. However,
because the matrix [S] is not square, [δρ] cannot be obtained directly.
[S]T (pxq2 ) [δV](q2 x1) = [S]T (pxq2 ) [S](q2 xp) [δρ](pxp) (8)
−1
[δV](q2 x1) = [S]T [S] (pxp) [S]T (pxq2 ) [δρ](pxp)

(9)
Generally, [S]T [S] is a singular matrix so that the matrix does not have inverse. To solve these
problems, Tikhonov regularization can be used so that the matrix has inverse.
−1
[δp](q2 x1) = [S]T [S] + αI (pxp) [S]T (pxq2 ) [δV](pxp)

(10)
where α is the regulatory parameter and I is the identity matrix.

In this study, in order to get real and imaginary images, the impedance was divided into two
components, namely resistance components and capacitive reactance components. Equation (10)
connected the impedance with the resistance components and the capacitive reactance components.
In order to get a real image, Equation (10) becomes Equation (11):
−1
[δρR ](q2 x1) = [S]T [S] + αI (pxp) [S]T (pxq2 ) [δVR ](pxp)

(11)
Equation (10) connected the impedance to the capacitive reactance components. In order to get an
imaginary image, Equation (10) becomes Equation (12):
−1
= [S]T [S] + αI (pxp) [S]T (pxq2 ) [δVXc ](pxp)
h i
δρXC (12)
(q2 x1)
3. Results
The influence variables from the EIT system, as depicted in the Figure 3, consist of the stability of
the Voltage Controlled Current Source (VCCS), the accuracy of the voltage between the electrodes,
and the phase difference. The instability of the current will lead into invalid voltage measurements
which can affect the obtained impedance value. The invalid impedance will directly affect the image
of the reconstruction results. The AD 8302 phase detector is needed to measure the phase difference
between the voltage on the object and the injection current. The phase difference in this experiment
is required to determine the impedance components separately, which are resistive reactance and
capacitive reactance. In addition, invalid phase difference measurement will affect the accuracy of
the calculation result of resistive reactance and capacitive reactance. It will then directly influence the
image of the reconstruction results. Overall, the final results of the EIT system are greatly influenced
by the performance of VCCS and phase detector.
3.1. Voltage Controlled Current Source (VCCS)

The VCCS performance can be seen from the stability of the currents to the electrical loads and the
width of the frequency. Figure 9 shows a graph of the current stability of the experimental results of
VCCS test. The VCCS produces a constant electric current of 1 mA that is stable with a load of 1 kΩ at
detected. Anomalies in objects could be detected properly if the object impedance and anomalies are in
the loads area of current source stability with the condition that in electrical impedance systems,
anomalies and objects have different impedances. Meanwhile, the function of the current stability to
the frequency changes is to measure the change in impedance when there is a change in the frequency
Sensors
of 2020, 20,current.
injection 1907 One of these applications is for the breast tumor detection system. There 10 of
is 18
a
change in the characteristics of electrical conductivity of the tissue for different frequencies.
Conductivity
frequencies upand impedance
to 150 in frequencies
kHz. At the normal tissue are constant,
of 150–200 whileis conductivity
kHz, there a 1% decreaseand impedance
in amplitude fromof
breast tumors change significantly at the current injection of 10 kHz and 100 kHz [7].
the stable state. The 2 kΩ load on VCCS is stable until 100 kHz while for the 3 kΩ load reaches 80 kHz.
(a) (b)
Figure 9.
Figure (a) The
9. (a) The relationship
relationship between
between frequencies
frequencies and
and loads
loads in
in the
the Voltage
Voltage Controlled
Controlled Current
Current Source
Source
(VCCS) to the currents; (b) the graphic of current stability to loads at 10 kHz and 100
(VCCS) to the currents; (b) the graphic of current stability to loads at 10 kHz and 100 kHz. kHz.
The effect of VCCS loads of 1–5 kΩ on the stability of frequency currents of 10–100 kHz is shown
3.2. Voltage and Phase Difference Measurements
in Figure 8. The change in current value is quite small due to changes in loads. The stability of the
Theagainst
current circuit the
scheme
load used
is usedto to
measure
determine the voltage and theofphase
the suitability difference
the current of with
source this EIT
the is shown
object in
to be
Figure
detected.7. This circuit has
Anomalies two functions,
in objects namely toproperly
could be detected measureif thethe voltage of the object
object impedance and and to measure
anomalies are
the phase
in the loads difference between
area of current the injection
source stabilitycurrent
with theand the potential
condition that inproduced. Therefore, there
electrical impedance are
systems,
two mechanisms
anomalies usedhave
and objects in the testing.impedances.
different The first mechanism
Meanwhile, is the
thestability
functionby ofmeasuring
the currentvoltage
stabilityattoVthe
rms
to DC. This
frequency test was
changes done
is to to ensure
measure the accuracy
the change of thewhen
in impedance voltage measurement.
there is a change inThethestability
frequency wasof
observed at the input
injection current. One and output
of these voltages is
applications offor
10,the
20, breast
50, 100tumor
mV atdetection
frequencies of 10There
system. Hz tois 550 kHz.
a change
The
in thegraph stability ofofthe
characteristics relation
electrical between theofinput
conductivity voltage,
the tissue the output
for different voltage, and
frequencies. the frequency
Conductivity and
are displayed
impedance in in Figure
normal 10. are constant, while conductivity and impedance of breast tumors change
tissue
Figure 10
significantly atisthe
a difference of Vrms from
current injection DC output
of 10 kHz and 100 atkHz
100 kHz
[7]. to a lower frequency. The difference
is very obvious, especially for the amplitude of 100 mV. In addition, it could be seen that the higher
3.2. input
the Voltage and Phase of
amplitude Difference Measurements
Vrms to DC, the bigger the difference in output voltage to the input voltage,
which The circuit scheme used to measurevoltages
causes the measurement of same at 100
the voltage andkHztheand at adifference
phase lower frequency to be
of this EIT is different.
shown in
There
Figure 7. This circuit has two functions, namely to measure the voltage of the object and tothe
must be a solution to solve the difference in the results of voltage measurements for same
measure
input voltage at 100 kHz with a low frequency. In the process of image reconstruction,
the phase difference between the injection current and the potential produced. Therefore, there are two the problem
of differencesused
mechanisms in measurement
in the testing. has
Thebeen
firstovercome
mechanism by using the normalization
is the stability method.
by measuring voltage at V to
rms
DC. This test was done to ensure the accuracy of the voltage measurement. The stability was observed
at the input and output voltages of 10, 20, 50, 100 mV at frequencies of 10 Hz to 550 kHz. The graph
stability of2020,
Sensors the 20,
relation between
x FOR PEER the input voltage, the output voltage, and the frequency are11displayed
REVIEW of 18
in Figure 10.
Figure 10. The

Figure stability
10. The of of
stability root mean
root meansquare
square (RMS) todirect
(RMS) to directcurrent
current (DC)
(DC) circuit
circuit of output
of output voltage
voltage to to
DC against frequencies.
Related to the measurement of phase difference, the second activity is the linearity test of
different phase conversion voltages by inputting the AD 8302 phase detector circuit. The aim of the
test is to prove the feasibility of the circuit as a part of EIT system that acts to measure the phase
difference. From the detector phase test, two types of data were obtained which show the feasibility
Sensors 2020, 20, 1907 11 of 18
Figure 10 is a difference of Vrms from DC output at 100 kHz to a lower frequency. The difference
is very obvious, especially for the amplitude of 100 mV. In addition, it could be seen that the higher
the input amplitude of Vrms to DC, the bigger the difference in output voltage to the input voltage,
which causes the measurement of same voltages at 100 kHz and at a lower frequency to be different.
There must be a solution to solve the difference in the results of voltage measurements for the same
Figure 10. The stability of root mean square (RMS) to direct current (DC) circuit of output voltage to
input voltage at 100 kHz with a low frequency. In the process of image reconstruction, the problem of
differences in measurement has been overcome by using the normalization method.
Related
Relatedtotothethe measurement
measurement of phase difference,
of phase the second
difference, activityactivity
the second is the linearity test of different
is the linearity test of
phase
different phase conversion voltages by inputting the AD 8302 phase detector circuit.of
conversion voltages by inputting the AD 8302 phase detector circuit. The aim theaim
The testof
is the
to
prove the feasibility of the circuit as a part of EIT system that acts to measure
test is to prove the feasibility of the circuit as a part of EIT system that acts to measure the phase the phase difference.
From the detector
difference. From the phase test, two
detector phase types
test, of
two data were
types of obtained
data werewhich
obtained show the feasibility
which of using
show the feasibility
the circuit. The first data were obtained from testing using two sinusoidal
of using the circuit. The first data were obtained from testing using two sinusoidal signals with signals with the samethe
amplitude and frequency, but there was the phase difference between
same amplitude and frequency, but there was the phase difference between the two signals that can the two signals that can be
modified
be modified repeatedly.
repeatedly. TheThesamesametreatment
treatment forfor
this test
this for
test forsinusoidal
sinusoidalsignals’
signals’frequencies
frequenciesatat11kHz,
kHz,
33kHz, and 5 kHz.
kHz, and 5 kHz.
The
Thetest
testresults
resultsdatadataare
arepresented
presentediningraphgraphwhich
whichshows
showsthe therelationship
relationshipbetween
betweenthe thephase
phase
differences
differences of the two input signals of AD 8302 phase detector to the output voltage (see Figure 11).
of the two input signals of AD 8302 phase detector to the output voltage (see Figure 11).
This
Thisresearch
researchcan can only
only prove
prove the linearity of
the linearity of the
the phase
phasedifference
differenceto tothe
theoutput
outputvoltage
voltageup up5 5kHz
kHzduedue to
to the limitations of the tools. Figure 10 shows that there is no change in
the limitations of the tools. Figure 10 shows that there is no change in the linearity graph due to the linearity graph due to
changes
changesin infrequency.
frequency.However,However,the theresults
resultsofofthis
thisverification
verificationarearethe
thesame
sameasasthetheIC ICsheet
sheetdata.
data.InInthe
the
data
datasheet,
sheet,ititisisexplained
explainedthat thatthe
theIC
ICAD8302
AD8302has hasthe
theability
abilityto
todetect
detectchanges
changesofof10 10mV/degree.
mV/degree.ThisThisIC IC
phase detector is capable of reading stable phase differences up to frequency
phase detector is capable of reading stable phase differences up to frequency signals of 2 MHz. signals of 2 MHz.
Figure 11. (a) The relation of the linearity of phase differences and voltages from signals detector with
the same amplitude. (b) The linearity graph of phase detector’s voltages to the phase differences for
Figure 11. (a) The relation of the linearity of phase differences and voltages from signals detector
two different signal amplitudes.
with the same amplitude. (b) The linearity graph of phase detector’s voltages to the phase differences
for two different signal amplitudes.
The second experiment was related to the phase detector, this second data were obtained by
providing input of two sinusoidal signals with different amplitudes, same frequency, and different
The second experiment was related to the phase detector, this second data were obtained by
phase variations. The second experimental setup in Figure 8 was used to test the phase detector
providing input of two sinusoidal signals with different amplitudes, same frequency, and different
linearity which utilized two input of the sinusoidal signals (1 and 2) at different amplitudes with a
phase variations. The second experimental setup in Figure 8 was used to test the phase detector
ratio of V1 /V2 = 1:1, 1:2, 1:3 and 1:4. Signal 1 and signal 2 were generated at 3 kHz. From these results,
linearity which utilized two input of the sinusoidal signals (1 and 2) at different amplitudes with a
a linearity graph of the output voltage to the phase with the amplitude ratio mentioned was created
(see Figure 11). The graph explains the plotting result of the second experiment. The graph shows the
same linearity even though the two input amplitudes are different. This linear graph also illustrates
the difference in amplitude of the two signals would shift the linearity position but does not change the
linearity between the phase difference and the output voltage. The shift in the position of the linearity
gives the constant information that changes due to changes in the amplitude ratio. Data related to the
linearity characteristics of Figure 11 are shown in Table 1.
the same linearity even though the two input amplitudes are different. This linear graph also
illustrates the difference in amplitude of the two signals would shift the linearity position but does
not change the linearity between the phase difference and the output voltage. The shift in the
position of the linearity gives the constant information that changes due to changes in the amplitude
ratio. Data related to the linearity characteristics of Figure 11 are shown in Table 1.
Sensors 2020, 20, 1907 12 of 18
From the graph of the linearity data in Table 1, a plot of the relationship between the constants
and the ratio V1/V2 was made. The relationship between the ratios of the two input signal amplitudes
to the constants of the
Table 1. Data oflinearity shifting
the linearity is shown
shift and in Figure
the phase 12.ratio
detector’s Theof graph shows
2 signal that the relationship
amplitudes.
between the amplitude ratio and the shift constant is linear. The linear function of the relationship
No Ratio V /V Constant Slope
will provide a solution for the linearity shift1 in2 the phase detector. If the amplitude ratio of V1 and V2
has been obtained, then from 1 the linear equation
1 1.9455 12, the
in Figure 0.0104
constant value information can be
2 1 1.7338 0.0107
acquired. Thereafter, the linier graph of the 2 relationship between the phase detector output voltage
3 1/3 1.6747 0.0107
and the phase difference was 4 generated. 1/4Furthermore, from the
1.6334 phase detector output voltage, the
0.0107
phase from the two signals are obtained. This step is used to measure the phase difference between
voltage and current injection with AD 8302 phase detector.
From the graph of the linearity data in Table 1, a plot of the relationship between the constants
and the ratio VTable1 /V2 was made.
1. Data of theThe relationship
linearity shift andbetween
the phasethe ratios ofratio
detector’s theof
two inputamplitudes.
2 signal signal amplitudes
to the constants of the linearity shifting is shown in Figure 12. The graph shows that the relationship
between the amplitude ratio and the No shift
Ratio V1/V2 isConstant
constant linear. The slope
linear function of the relationship
will provide a solution for the linearity shift in the phase1.9455
1 1 detector. 0.0104
If the amplitude ratio of V1 and V2
has been obtained, then from the linear 2 ½ in Figure
equation 1.7338
12, the0.0107
constant value information can be
3 1/3 1.6747 0.0107
acquired. Thereafter, the linier graph of the relationship between the phase detector output voltage and
the phase difference was generated.4 Furthermore, 1/4 from1.6334 the phase 0.0107
detector output voltage, the phase
from the two signals are obtained. This step is used to measure the phase difference between voltage
and current injection with AD 8302 phase detector.
Figure 12. The relation of the shift constants and the amplitude ratio.
Figure 12. The relation of the shift constants and the amplitude ratio.
3.3. Simulation
3.3. Simulation
Simulations are carried out to solve forward problems using Finite Element Methods (FEM) so
that potential data isare
Simulations obtained
carriedfrom
out numerical objects.problems
to solve forward The numerical
using object
Finite is built from
Element the electrical
Methods (FEM) so
properties of the heart
that potential data and deflated lungs
is obtained from at Table 2. The
numerical simulation
objects. uses the neighbor
The numerical object isdata collection
built from the
method. Theproperties
electrical triangularofelement model
the heart used in this
and deflated study
lungs. Theconsists of 141
simulation nodes,
uses 248 elements,
the neighbor and 16
data collection
electrodes,
method.as Theillustrated
triangularinelement
Figure 13.
model used in this study consists of 141 nodes, 248 elements, and 16
electrodes, as illustrated in Figure 13.
Table 2. Data of the electrical properties many tissue from Gabriel.
Frequency
10 kHz 100 kHz
Heart Deflated Lung Distillated Water Heart Deflated Lung Distillated Water
Z (Ω) 1.001 1.078 2000 2.676 2.988 2000
Re (Ω) 0.245 0.084 2000 0.66 0.313 2000
Im (Ω) 0.971 1.074 100 2.593 2.972 100
Sensors 2020, 20, 1907 13 of 18
Figure 13. The triangular element model used in this study.
Frequency
10 kHz Frequency 100 kHz
Heart 10Deflated
kHz Lung Distillated Water Heart 100Deflated kHz Lung Distillated Water
Z (Ω) Heart
1.001 Deflated Lung Distillated
1.078 2000Water Heart 2.676 Deflated Lung Distillated
2.988 2000Water
ZRe
(Ω)(Ω) 1.001
0.245 1.078
0.084 2000
2000 2.676
0.66 2.988
0.313 2000
2000
ReIm(Ω)
(Ω) 0.245
0.971 0.084
1.074 2000
100 0.66
2.593 0.313
2.972 2000
100
Im (Ω) 0.971 1.074 100 2.593 2.972 100
From Figures 14a–c and 15a–c, it can be seen that the heart and deflated lung have the same
From Figures
impedance, 14a–cFigure
but differentand
real15a–c,
andThe
13. ittriangular
can be images,
imaginary seen that
element the heart
respectively.
model used inand
thisdeflated
study. lung have the same
impedance, but different real and imaginary images, respectively.
Frequency
10 kHz 100 kHz
Heart Deflated Lung Distillated Water Heart Deflated Lung Distillated Water
Z (Ω) 1.001 1.078 2000 2.676 2.988 2000
Re (Ω) 0.245 0.084 2000 0.66 0.313 2000
Im (Ω) 0.971 1.074 100 2.593 2.972 100
(a) (b) (c)

(a) (b) (c)
Figure 14. (a) Impedance image, (b) real image, (c) imaginary image.
Figure
Figure14.
14.(a)
(a)Impedance
Impedanceimage,
image,(b)
(b)real
realimage,
image,(c)
(c)imaginary
imaginaryimage.
image.
(a) (b) (c)

Figure 14. (a) Impedance image, (b) real image, (c) imaginary image.
(a) (b) (c)
(a) (b) (c)
Figure15.
Figure 15. Line
Line profile
profileof
of(a)
(a)impedance
impedanceimage,
image,(b)
(b)real
realimage,
image,(c)
(c)imaginary
imaginaryimage.
image.
Figure 15. Line profile of (a) impedance image, (b) real image, (c) imaginary image.
3.4.
3.4. Image
Image Scanning
ScanningandandReconstruction
Reconstruction
3.4. Image Scanning and Reconstruction
In this stage, we integrated several devices that had been built and the available modules.
This integration process produced a system of electrical impedance metallography (EIT), as shown in
Figure 4. The scanning process was carried out by using the neighboring method to obtain electric
potential data and the phase difference between the current injection current and the measured potential.
This EIT system was used to scan phantom objects.
The phantom object is a circle container filled with water as a medium and a carrot as an anomaly,
as depicted in Figure 16. Data from Tusarkanti Bera [5] shows that the carrot object at the frequency of
(a) (b) (c)
10 kHz has a Z value of 7.2 kΩ, real 6.2 kΩ, and imaginary = −3.2 kΩ. Whereas at the frequency of
100 kHz it has a value
Figure 15. Line = 2.2 kΩ,
of Z profile of (a)real = 1.3 kΩ,
impedance and imaginary
image, = −1.5
(b) real image, kΩ. The image.
(c) imaginary electrodes made of
printed circuit board PCB were attached around the edge of the phantom object. The scanning process
3.4. Image Scanning and Reconstruction
potential.
TheThis EIT system
phantom objectwasis aused to scan
circle phantom
container filledobjects.
with water as a medium and a carrot as an
The phantom
anomaly, object
as depicted is a circle
in Figure container
16. Data from filled with Bera
Tusarkanti water[5]asshows
a medium
that theand a carrot
carrot objectasatan
the
anomaly, as depicted in Figure 16. Data from Tusarkanti Bera [5] shows that
frequency of 10 kHz has a Z value of 7.2 kΩ, real 6.2 kΩ, and imaginary = −3.2 kΩ. Whereas at the the carrot object at the
frequency
frequencyofof 10100
kHzkHz has ita Z
hasvalue of 7.2ofkΩ,
a value Z =real
2.2 6.2
kΩ,kΩ,
realand imaginary
= 1.3 kΩ, and= imaginary
−3.2 kΩ. Whereas at the
= −1.5 kΩ. The
frequency 2020,of
electrodes
Sensors 100
1907kHz
made
20, it has circuit
of printed a value of ZPCB
board = 2.2were
kΩ, attached
real = 1.3 kΩ, and
around the imaginary
edge of the=phantom
−1.5 kΩ.14 The
object.
of 18
electrodes madeprocess
The scanning of printed
used circuit board PCB
AC current wereatattached
injection around
10 and 100 kHz.the edgethe
From of scanning
the phantom object.
process, two
The scanning
data process namely
were obtained, used ACthe current
electric injection
potential at and
10 and
the 100
phasekHz. From the
difference. scanning
The electricprocess,
potentialtwodata
used AC current injection at 10 and 100 kHz. From the scanning process, two data were obtained,
data
andwere obtained, namely
the difference the electric
in the results of thepotential
scanningand the phase
process difference.
are shown The 17.
in Figure electric
We canpotential
observedata
that
namely the electric potential and the phase difference. The electric potential data and the difference in
and theare
there difference
potential in differences
the results of the scanning
measured processfrequencies.
at different are shown in Figure
This 17. We
potential can observe
graph that
demonstrates
the results of the scanning process are shown in Figure 17. We can observe that there are potential
there are potential
a periodic patterndifferences measured
of measurements onatalldifferent
electrodesfrequencies.
and in the This potential
pairs graphinjection.
of electrode demonstrates
In the
differences measured at different frequencies. This potential graph demonstrates a periodic pattern of
a phase
periodic pattern of measurements on all electrodes and in the pairs of electrode
difference pattern (Figure 17) for the 100 kHz, it depicts the correct pattern but not quite injection. In the
measurements on all electrodes and in the pairs of electrode injection. In the phase difference pattern
phase
smoothdifference
due to the pattern (Figure
imperfect tool,17)
so for
thatthe 100tokHz,
needs it depictsinthe
be improved thecorrect
future.pattern but not quite
(Figure 17) for the 100 kHz, it depicts the correct pattern but not quite smooth due to the imperfect tool,
smooth due to the imperfect tool, so that needs to be improved in the future.
so that needs to be improved in the future.
Figure 16. Phantom object.

Figure16.
Figure Phantomobject.
16.Phantom object.
Figure 17. (a) Data of electric potential from the scanning using Neighboring method. (b) Data of
phase difference from the scanning using Neighboring method.
Figure 17. (a) Data of electric potential from the scanning using Neighboring method. (b) Data of
Figure
phase17. (a) Datafrom
of electric
From difference
the multiplication ofpotential
the scanning from
using
the electric the scanning
Neighboring
potential using Neighboring method. (b) Data of
method.
and the phase difference data from Figure 17
phase difference from the scanning using Neighboring method.
and by using Equations (3) and (4), real potential and imaginary potential were obtained for each
From the multiplication of the electric potential and the phase difference data from Figure 17
corresponding value, namely [VZ ], [VCosϕ ], [Vsinϕ ]. The reconstruction of the three potentials was
From the multiplication
and by using Equations (3) of the(4),
and electric potentialand
real potential andimaginary
the phase potential
differencewere
dataobtained
from Figure 17
for each
conducted to get the images. From the reconstruction process data impedance, resistive reactance,
and by using Equations
corresponding (3) and [V
value, namely (4),Z],real potential
[VCosφ ], [Vsinφand imaginary
]. The potential
reconstruction of were obtained
the three for each
potentials was
and capacitive were obtained. This reconstruction process uses a linear method. All three data were
corresponding
conducted to value,
get thenamely
images.[V Z], [V
From the ], [Vsinφ]. The reconstruction
Cosφreconstruction of the three
process data impedance, potentials
resistive was
reactance,
taken from two different frequencies. Thus, the resulting image is a functional image. The image of the
conducted to get the images. From the reconstruction process data impedance, resistive reactance,
reconstruction results from the impedance data is called the impedance image while the reconstruction
image of the resistive reactance data is called the real image, and the reconstructed image of the
capacitive reactance data is called the imaginary image.
The experiments obtained functional images for impedance, resistive (real), and capacitive
(imaginary). By using α = 0.01, potential and phase data are reconstructed to produce impedance
images, real images, and imaginary images as shown in Figure 18. These three images increase the
accuracy of the EIT system in determining the abnormality of objects that have the same impedance.
The statistical analysis used to show image quality is the standard deviation. It calculates the relative
image of the capacitive reactance data is called the imaginary image.
The experiments obtained functional images for impedance, resistive (real), and capacitive
(imaginary). By using α = 0.01, potential and phase data are reconstructed to produce impedance
images, real images, and imaginary images as shown in Figure 18. These three images increase the
Sensors 20,the
2020,of
accuracy 1907EIT system in determining the abnormality of objects that have the same impedance.
15 of 18
The statistical analysis used to show image quality is the standard deviation. It calculates the
relative
errors errors
of the of the impedance,
impedance, real, andimages.
real, and imaginary imaginary
The images. The value
value of each of each
impedance, impedance,
real, real,
and imaginary
and imaginary image are 0.01621, 0.02476, and
image are 0.01621, 0.02476, and 0.02123, respectively.0.02123, respectively.
Figure 18. (a) Impedance functional image, (b) real functional image, (c) imaginary functional image.
4. Discussion
Figure 18. (a) Impedance functional image, (b) real functional image, (c) imaginary functional image.
The noninverting op-amp added at the VCCS double op-amp [27] AD8039 can increase the current.
4. Discussion
VCCS performance is displayed in the frequency response graph. VCCS with a given load of 3 kΩ
generates
The anoninverting
steady current of 0.1–1.5
op-amp mA atatfrequencies
added the VCCS 500 Hz–80
double kHz, at[27]AD8039
op-amp a frequency of can90 increase
kHz current the
drops 0.03%, and at a frequency
current. VCCS performance is displayed in of 100 kHz the current drops 0.05% from a stable state.
frequency response graph. VCCS with a given load of VCCS with a
load
3 kΩofgenerates
2 kΩ producesa steady a steady
currentcurrent of 0.1–1.5
of 0.1–1.5 mA at mA for at a frequency
frequencies 500 Hz–80 of 100
kHz, kHz.
at aThe current used
frequency of 90
in
kHzthiscurrent
experiment
dropsis0.03%,
1 mA.and The at
current used on
a frequency ofthe
100EIT
kHz depends
the currenton the object.
drops Thefrom
0.05% current generated
a stable state.
by
VCCSVCCS depends
with a load of on2thekΩmagnitude
produces aofsteady
the gain. current of 0.1–1.5 mA for at a frequency of 100 kHz. The
Thisused
current research
in thisrequires two currents
experiment is 1 mA.withThe a frequency
current usedof 10 onkHzthe and
EIT100depends
kHz. The ondual
the frequencies
object. The
appertain to the method
current generated by VCCS used to detect
depends on theabnormalities,
magnitude ofwhich is using functional imagery [28,29].
the gain.
However,This this VCCSrequires
research can produce stable currents
two currents with afor a wide frequency.
frequency of 10 kHz VCCSand double
100 kHz. op-amps are
The dual
capable
frequenciesto support
appertain multi-frequency
to the method and usedprogrammable EIT. As in
to detect abnormalities, the Mf-EIT
which is usingstudy, the required
functional imagery
frequencies
[28,29]. However,are 5 10, this25,
VCCS50, 75,
canand 100 kHz
produce [5].currents
stable This VCCS for acan widegenerate a stable
frequency. VCCScurrent
doubleat
those
op-amps frequencies.
are capable to support multi-frequency and programmable EIT. As in the Mf-EIT study, the
This frequencies
required EIT has succeeded are 5 10,in25,
separating
50, 75, and the100
impedance
kHz [5]. distribution
This VCCS can intogenerate
two namely the resistance
a stable current at
distribution and
those frequencies. capacitive reactance. The method used to separate these components is indeed simple,
whichThis is byEITusing the analogyinofseparating
has succeeded a series ofthe RCimpedance
series withdistribution
body tissue. Bytwo
into measuring
namely the potential,
resistance
the phase difference
distribution between reactance.
and capacitive the potential Theand the injected
method used current,
to separate “forward
these problem”
components andis“inverse
indeed
problem”,
simple, which we can is byobtain
usingthethe
distribution
analogy ofofaimpedance,
series of RC resistance,
series with andbody
capacitive
tissue.reactance.
By measuring the
Distribution
potential, the phase of difference
the three between
components obtainedand
the potential threetheimages,
injected namely
current, the impedance
"forward problem" image,
and
resistance, and capacitive
"inverse problem", we canreactance.
obtain theThe resistanceofimage
distribution is takenresistance,
impedance, from the realandimpedance component
capacitive reactance.
so it isDistribution
called the real ofimage, whilecomponents
the three the capacitive reactancethree
obtained image is takennamely
images, from thethe imaginary
impedance impedance
image,
component
resistance, so andit iscapacitive
called an imaginary
reactance. image.
The resistance image is taken from the real impedance
The purpose
component so it of is the separation
called the real of image,
two components
while theiscapacitive
to detect the material image
reactance specifications
is taken of from
an object
the
in a more detailed
imaginary impedance manner. This research
component so it isis called
a preliminary result to
an imaginary distinguish inhomogeneous material.
image.
From the three images, it can be seen that there are differences in the image profiles of the same object.
Each material or network has an identical image impedance, resistance, and reactance image for each
tissue. These three images have improved the accuracy of the EIT system in detecting abnormalities
when compared to just one image [2–11,16]. The results obtained have been compared with data
from other researchers’ experiments [3], the results of these comparisons show a corresponding trend.
That the Z image is bigger than the real image and imaginary valued negative. Which is all in
accordance with the carrot impedance values both Z, real, and imaginary values.
Algorithms and data acquisition systems to separate resistance and capacitive reactance have
been successfully developed. Although EIT has a problem in image resolution, the LBP and Tikhonov
methods in EIT can produce high contrast images. The reference reconstruction method requires
Sensors 2020, 20, 1907 16 of 18
reference data that can be obtained from different frequency data so that this reconstruction method
requires two data from different frequencies. The reconstruction produces a relative image or also
called a functional image.
Jiang and Soleimani [12] have developed Capacitively Coupled Electrical impedance Tomography
CCEIT for brain imaging. Whereas Wang et al. have developed a suitable reconstruction method
for the CCEIT system using fusion images. Real, imaginary, and magnitude images from the
LBP method are summed and by using K-means as the filter threshold, a better image has been
obtained [30]. The advantage of the CCEIT system is that it requires no contact between the electrode
and the object, thus the system is safer when compared to a standard EIT. The CCEIT developed by
Wang [30] uses a total impedance of two-phase gas-liquid flow. This research produces real, imaginary,
and impedance images, and combines real and imaginary parts. Images reconstructed from various
impedances are investigated and discussed. The LBP reconstruction method is used to obtain object
images from various impedance parts and K-means are adopted to obtain the gray level threshold
automatically. The weighted combination method was introduced to combine images by combining
images reconstructed from real and imaginary impedances.
Our research obtains impedance, real, and imaginary images. All images are obtained from
potential data and phases that can be collected from the developed system. LBP reconstruction method
from two different frequency data can produce the three images. This is the similarity between our
research and CCEIT. The advantage of CCEIT compared to our study is that it is non-contact, so it can
be used for objects that have high resistivity on the surface.
Studies by Jiang and Wang [12,30] stated that the suitable object for CCEIT applications is head.
One of the requirements in using electrical modality to observe the brain is that it should be able
to penetrate the skull that have high resistivity therefore the EIT system is less suitable for this use.
However, for organs that do not have high resistivity such as the thorax and abdomen, the EIT system
is suitable. The system that we have developed can be used for imaging organs in the thorax or
abdomen in the future. As we explained earlier that there are organs in the thorax and abdomen that
have the same impedance but with a different conductivity and permittivity. The EIT system that
we have developed can produce images not only impedance but also of real and imaginary images.
The last two images are embodiments of the conductivity and permittivity of the object.
The study has conducted a design and development EIT device without involving a high-speed
data acquisition system but it can produce impedance image components that include real and
imaginary images. The EIT provides important information that tissue impedance components have
been successfully separated. These three images show a more detailed and specific identity of a tissue.
Three images are more accurate in distinguishing different objects with the same impedance composed
of different impedance components.
5. Conclusions
This research has succeeded to develop the EIT system that can separate impedance components.
The impedance components consist of resistive reactance as the real parts and capacitive reactance as
the imaginary parts This experiment utilized alternating current with frequencies of 10 and 100 kHz.
Two current sources from different frequencies was injected and produce voltage and phase data.
Two current sources from different frequencies was injected and produce voltage and phase data,
The linear back projection was used to reconstruct voltage and phase data to generate impedance,
real and imaginary images. It can be concluded that the EIT system can produce functional impedance,
real and imaginary images. The functional real and imaginary images more sensitive in detecting
anomalies compared to the functional impedance image. It can be seen from the capacitive reactance
image of carrots as an anomaly in the water that looks clearer than in the impedance functional image.
The anomaly in the resistive reactance image is still not visible which may be because the resistance of
carrots is worth the same as the water medium.
Sensors 2020, 20, 1907 17 of 18
Author Contributions: Developed the algorithm and instrumentation, I.S.; modeling and reconstruction, K.A.;
research methodology, K.A., M.Y., R.A.; formal analysis, interpreted the data and images, I.S., K.A; writing—original
draft, I.S.; writing—review and editing, I.S., K.A., M.Y., and R.A. All authors have read and agreed to the published
version of the manuscript.
Funding: This research was funded by the Ministry of Research, Technology and Higher Education, Doctoral
Dissertation Research Number: 200/UN3.14 LT/2018, Airlangga University of Indonesia.
Acknowledgments: Thank you to the Ministry of research, technology and higher education, the Directorate
General of higher education who provided research funding through the Doctoral Dissertation Program.
Conflicts of Interest: All authors declare that they have no conflicts of interest.
References
1. Hu, G.; He, B. Magnetoacoustic Imaging of Electrical Conductivity of Biological Tissues at a Spatial Resolution
Better than 2 mm. PLoS ONE 2011, 6, e23421. [CrossRef] [PubMed]
2. Affanni, A. Ruben Specogna, Francesco Trevisan Estimating the Volume of Unknown Inclusions in an
Electrically Conducting Body with Voltage Measurements. Sensors 2019, 19, 637. [CrossRef] [PubMed]
3. Bera, T.K.; Nagaraju, J.; Lubineau, G. Electrical impedance spectroscopy (EIS)-based evaluation of biological
tissue phantoms to study multifrequency electrical impedance tomography (Mf-EIT) systems. J. Vis. 2016,
19, 691–713. [CrossRef]
4. Ain, K.; Kurniadi, D.; Suprijanto, S.; Santoso, O. Dual modality electrical impedance and ultrasound reflection
tomography to improve image quality. J. Electr. Bioimpedance 2017, 8, 3–10. [CrossRef]
5. Choi, M.H.; Kao, T.J.; Isaacson, D.; Saulnier, G.J.; Newell, J.C. A Reconstruction Algorithm for Breast Cancer
Imaging With Electrical Impedance Tomography in Mammography Geometry. IEEE Trans. Biomed. Eng.
2007, 54, 700–710. [CrossRef]
6. Pak, D.D.; Rozhkova, N.I.; Kireeva, M.N.; Ermoshchenkova*, M.V.; Nazarov, A.A.; Fomin, D.K.; Rubtsova, N.A.
Diagnosis of Breast Cancer Using Electrical Impedance Tomography. Biomed. Eng. 2012, 46, 154–157.
[CrossRef]
7. Aristovich, K.Y.; Packham, B.C.; Koo, H.; dos Santos, G.S.; McEvoy, A.; Holder, D.S. Imaging fast electrical
activity in the brain with electrical impedance tomography. NeuroImage 2016, 124, 204–213. [CrossRef]
8. Yang, L.; Liu, W.; Chen, R.; Zhang, G.; Li, W.; Fu, F.; Dong, X. In Vivo Bioimpedance Spectroscopy
Characterization of Healthy, Hemorrhagic and Ischemic Rabbit Brain within 10 Hz–1 MHz. Sensors 2017, 17,
791. [CrossRef]
9. Visentin, F.; Fiorini, P.; Suzuki, K. A Deformable Smart Skin for Continuous Sensing Based on Electrical
Impedance Tomography. Sensors 2016, 16, 1928. [CrossRef]
10. Karsten, J.; MBohlmann, K.; Sedemund-Adib, B.; Wnent, J.; Paarmann, H.; Iblher, P.; Meier, T.; Heinze, H.
Electrical impedance tomography may optimize ventilation in a postpartum woman with respiratory failure.
Int. J. Obstet. Anesth. 2013, 22, 67–71. [CrossRef]
11. Shetiye, P.C.; Ghatol, A.A.; Ghate, V.N.; Patil, S.R. Detection of Breast Cancer Using Electrical Impedance and
RBF Neural Network. Int. J. Inf. Electron. Eng. 2015, 5, 5. [CrossRef]
12. Jiang, Y.D.; Soleimani, M. Capacitively Coupled Electrical Impedance Tomography for Brain Imaging.
IEEE Trans. Med. Imaging 2019, 38, 2104–2113. [CrossRef] [PubMed]
13. Liang, G.; Ren, S.; Zhao, S.; Dong, F. A Lagrange-Newton Method for EIT/UT Dual-Modality Image
Reconstruction. Sensors 2019, 19, 1966. [CrossRef]
14. Santos, S.A.; Robens, A.; Boehm, A.; Leonhardt, S.; Teichmann, D. System Description and First Application
of an FPGA-Based Simultaneous Multi-Frequency Electrical Impedance Tomography. Sensors 2016, 16, 1158.
[CrossRef] [PubMed]
15. Lionheart, W.R. EIT reconstruction algorithms: Pitfalls, challenges and recent developments, review article.
Physiol. Meas. 2004, 25, 125–142. [CrossRef] [PubMed]
16. Bera, T.K.; Nagaraju, J. Electrical impedance spectroscopic study of broiler chicken tissues suitable for the
development of practical phantoms in multifrequency EIT. J. Electr. Bioimpedance 2011, 2, 48–63. [CrossRef]
17. Mari, L.-P.; Olli, K.; Aku, S.; Marek, R.; Raul, L.; Paul, A.; Mart, M.; Edite, F.; Jari, H. EIT in hybrid imaging
setup for tissue engineering applications. In Proceedings of the 20th International Conference on Biomedical
Applications of Electrical Impedance Tomography (EIT2019), London, UK, 1–3 July 2019.
Sensors 2020, 20, 1907 18 of 18
18. Singh, G.; Anand, S.; Lall, B.; Srivastava, A.; Singh, V. Development of a Microcontroller based Electrical
Impedance Tomography System. In Proceedings of the 2015 Long Island Systems, Applications and
Technology, Farmingdale, NY, USA, 1 May 2015.
19. Khalighi, M.; Vahdat, B.V.; Mortazavi, M.; Soleimani, M. Practical Design of Low-cost Instrumentation for
Industrial Electrical Impedance Tomography (EIT). Ijst Trans. Electr. Eng. 2014, 38, 1–20.
20. Seward, R.; Ethan, M.; Courtney, M.; Fu, Z.; Hilda, G.; Badria, M.; Ryan, H. Using EIT to assess Pulmonary
Function in ALS Patients. In Proceedings of the 20th International Conference on Biomedical Applications of
Electrical Impedance Tomography (EIT2019), London, UK, 1–3 July 2019.
21. Qiao, G.; Wang, W.; Wang, L.; He, Y.; Bramer, B.; Al-Akaidi, M. Investigation of biological Phantom for 2D
and 3 D Breast EIT images. IMBE Proc. 2007, 17, 328–331.
22. Gabriel, C. Compilation of the Dielectric Properties of Body tissue at RF and Microwave Frequencie; Report N.
AL/OE -TR-1996-0037; Occupational and Environmental Health Directorate, Radiofrequency Radiation
Division Broooks Air Force Base: San Antonio, TX, USA, June 1996.
23. Mylott, E.; Kutschera, E.; Widenhor, R. Bioelectrical impedance analysis as a laboratory activity: At the
interface of physics and the body. Am. J. Phys. 2014, 82, 521–528. [CrossRef]
24. Kyle, U.G.; Bosaeus, I.; De Lorenzo, A.D.; Deurenberg, P.; Elia, M.; Gómez, J.M.; Heitmann, B.L.; Kent-Smith, L.;
Melchior, J.C.; Pirlich, M.; et al. Biolelectrical Impedance Analysis-Part 1: Review of Principles And Metods.
Clin. Nutr. 2004, 23, 1226–1243. [CrossRef]
25. Pandya, A.S.; Arimoto, A.; Agarwal, A.; Kinouchi, Y. Anovel Approach for measuring Electrical Impedance
Tomography for Local Tissue with Artificial Intelligent Algorithm. Int. J. Biom. Bioinform. 2010, 3, 66.
26. Rizzoni, G. Principles and Applications of Electrical Engineering, 5th ed.; Mc Graw-Hill: New York, NY, USA;
Singapore, 2007; ISBN 007-125444-7.
27. Li, Z.; Xu, Z.; Ren, C.; Wang, W.; Zhao, D.; Zhang, H. Study of Voltage Control Current Source in Electrical
Impedance Tomography System. In Proceedings of the 4th International Conference on Bioinformatics and
Biomedical Engineering, Chengdu, China, 18–20 June 2010. [CrossRef]
28. Sapuan, I.; Ain, K.; Suryanto, A. Dual frequency electrical Impedance Tomografy to Obtain Functional Image.
J. Phys. Conf. Ser. 2017, 853, 012002. [CrossRef]
29. Gong, B.; Krueger-Ziolek, S.; Moeller, K.; Schullcke, B.; Zhao, Z. Electrical impedance tomography: Functional
lung imaging on its way to clinical practice? J. Expert Rev. Respir. Med. 2015, 9, 721–737. [CrossRef] [PubMed]
30. Wang, Y.; Ji, H.; Huang, Z.; Wang, B.; Li, H. Study on image reconstruction of capacitively coupled electrical
impedance tomography (CCEIT). Meas. Sci. Technol. 2019, 30, 094002. [CrossRef]
© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).

Sensors: Anomaly Detection Using Electric Impedance Tomography Based On Real and Imaginary Images

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Sensors: Anomaly Detection Using Electric Impedance Tomography Based On Real and Imaginary Images

Uploaded by

Copyright:

Available Formats

sensors

Sensors 2020, 20, 1907; doi:10.3390/s20071907 www.mdpi.com/journal/sensors

2. Materials and Methods

Figure 1. Schematic diagram of the experimental procedure.

Figure 2. The relation of R, Xc, and Z.

Figure 5. A signal generator and constant current.

For For technical

2.2.2. Electronic Mechanism of Voltage Measurement

2.2.3. Measurement of Voltage and Phase Difference

Sensors 2020, 20, x FOR PEER REVIEW 8 of 18

Figure 7. Measurement scheme of voltage and phase difference.

[δV](q2 x1) = [S](q2 xp) [δρ](pxp) (7)

where α is the regulatory parameter and I is the identity matrix.

3.1. Voltage Controlled Current Source (VCCS)

Figure 10. The

(a) (b) (c)

(a) (b) (c)

Figure 16. Phantom object.

You might also like