See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/279201814

Analysis of Dispersive Effects of Breast Phantom Model on Ultra Wideband

Microwave Imaging of Breast Cancer Tumor

Conference Paper · February 2013

CITATIONS READS

2 237

3 authors:

İlhami Ünal Umut Bulus

The Scientific & Technological Research Council of Turkey Antenom Antenna Technologies
42 PUBLICATIONS 189 CITATIONS 31 PUBLICATIONS 68 CITATIONS

SEE PROFILE SEE PROFILE

Cahit Canbay
Yildiz Technical University
35 PUBLICATIONS 86 CITATIONS

SEE PROFILE

All content following this page was uploaded by İlhami Ünal on 04 January 2020.

The user has requested enhancement of the downloaded file.

 Proceedings of the IASTED International Conference
Biomedical Engineering (BioMed 2013)
February 13 - 15, 2013 Innsbruck, Austria

ANALYSIS OF DISPERSIVE EFFECTS OF BREAST PHANTOM MODEL ON

ULTRA WIDEBAND MICROWAVE IMAGING OF BREAST CANCER TUMOR
İlhami Ünal1, Bahattin Türetken2, Umut Buluş2, and Cahit Canbay1
1
Yeditepe University, Department of Electrical and Electronics Engineering
34755, İstanbul, Turkey
ilhamiunal@yeditepe.edu.tr, canbay@yeditepe.edu.tr
2
TUBITAK – BILGEM – UEKAE, Antenna Test and Research Center (ATAM)
PO 74 Gebze, Kocaeli, Turkey
bahattin.turetken@tubitak.gov.tr, umut.bulus@tubitak.gov.tr

ABSTRACT breast tissues at X-ray frequencies [1]. It also requires

In this paper, effects of frequency dispersive dielectric
properties of a half-spherical breast phantom model on
breast cancer tumor detection capability of the microwave
radar-based imaging system, currently being developed at
the Scientific and Technological Research Council-
BILGEM-UEKAE, are investigated by simulation studies.
An ultra wideband (UWB) bow-tie antenna array
surrounding the breast has been designed in the presence
of the breast phantom. The designed antenna operates
efficiently across the band from 1 GHz to 8 GHz and it’s
assumed to be immersed in a coupling medium to get a
good impedance matching with the breast. Fidelity factor
of the case with the dispersive breast phantom is found to
be 4.14% less than the non-dispersive case, due to
increasing pulse distortion in dispersive dielectric tissues.
Images are formed by using delay-and-sum (DAS)
algorithm for the detection of spherical tumor with 2 mm
diameter. Imaging results of the case with the dispersive
breast phantom present signal reduction of 0.628 dB in
the tumor response level as well as faulty locating of the
tumor in the image, as compared to the successful results
of the non-dispersive case. A better imaging algorithm
compensating the dispersive effects will be used in the
future.
KEY WORDS
Medical Imaging (non MRI/S), Electromagnetic Field
Simulation, Breast Cancer, Dispersive Effects.
1. Introduction
Breast cancer is the most frequent cause of cancer related
to death for women in both developed and developing
countries. Early diagnosis and treatment are the hot keys
to survive from breast cancer. The present “golden”
standard screening technology for detecting early-stage
breast cancer is X-ray mammography. It has several
limitations, especially when dealing with younger women
who have dense breast tissues. The difficulty arises from
the small intrinsic contrast between healty and diseased
painful and uncomfortable breast compression and
exposes the patient to ionizing radiation.
Electromagnetic waves and antennas have a huge
application area, and one of the challenging area is remote
sensing systems and detection systems using microwaves,
today. Increasing demand on non-destructive sensing or
detecting breast tumor keeps this subject hot in this field.
There are various passive and active microwave
techniques which have been proposed as an alternative
especially to the most widely used X-ray mammography
for early detection of breast cancer, such as microwave
radiometry [2], hybrid microwave-induced acoustic
imaging [3], microwave tomography [4] and UWB
microwave radar technique [1], [5]-[7].
The passive methods (microwave radiometry) use
radiometers to measure the temperature differences in the
breast as temperature increases in the presence of tumor
than with healthy breast tissue. The hybrid methods
(microwave-induced acoustic imaging) use microwave
energy to heat tumors. They expand and generate pressure
waves which are detected by ultrasound transducers. In
microwave tomography, an inverse scattering problem is
solved to reconstruct an image of the dielectric properties
in the breast. In contrast to the image reconstruction aim
of the microwave tomography technique, UWB radar-
based imaging approach solves a simpler computational
problem faster dealing with only to identify the presence
and location of significant scattering obstacles such
malignant breast tumors.
There is a significant contrast in the electrical
properties of the normal and the malignant breast tissues
[1], which exists in the earliest stage of tumor
development, at microwave frequencies. Another
advantage of the microwave imaging technique is that it
would be nonionizing and it doesn’t require painful breast
compression. Furthermore, microwave attenuation in
normal breast tissue is low enough to make signal
propagation through even large breast volumes quite
feasible [1]. Other available screening techniques such as
ultrasound and MRI are either less effective or are too
costly.

DOI: 10.2316/P.2013.791-133 401

 Because of its excellent advantages, recent years have
shown a dominant interest in UWB microwave imaging
technique, for a particular technique to detect and locate a
breast tumor [5]-[15]. This technique specifically involves
transmitting and receiving short duration pulses for
various locations of UWB probe antenna or alternatively
by an antenna array controlled with switches. The UWB
imaging technique offers creation of an image, which can
be formed by combining all of the signals (Sii ve Sij, i≠j)
coming from different antennas. Well-known delay-and-
sum (DAS) algorithm is used to create microwave images
of breast cancer tumors [1], [8].
In this study, UWB microwave imaging technique is
used, and we’re interested in the 1–8 GHz frequency
range; which guarantees balance between resonable
contradictory needs of better spatial resolution, better
penetration depth [12], less attenuation of electromagnetic
waves through the breast and smaller dimensions of a
multi-function active imaging system. The selected
frequency range is expected to provide reasonable tumor
detection capabilities. In particular, since the skin
reflections back to the antenna adversely affect imaging
results, better penetration of electromagnetic waves into
the breast tissue will be determined by operating the
antenna in a coupling medium whose dielectric properties
are close to the breast tissue. Antenna size would also be
selected to be smaller since the wavelength in the
coupling medium will be smaller than air.
On the other hand, it’s been reported that frequency
dispersive dielectric properties of surrounding coupling
medium affect antenna performance as well as tumor
detection capability of the microwave imaging system
[16]. Dielectric properties of human body tissues vary
over a considerable UWB operation frequency band [5],
[17]. As electromagnetic pulse propagates in a dispersive
medium, it’s distorted inside the medium depending on
the frequency dispersive behaviour of the dielectric
properties of the medium [18], [19]. Therefore, dispersive
properties of breast phantom as well as surrounding
coupling medium should also be considered for the
antenna design.
Several different types of antennas have been
considered and reported over the past decade by research
groups involved in radar-based UWB breast imaging;
such as UWB bow-tie [9], ridged pyramidal horn [10],
cross-polarized bow-tie [11], U-slot [12], stacked
microstrip patch [13], antipodal vivaldi [14] and tapered
slot [15]. These antennas were generally used and tested
in planar, cylindrical or spherical scanning surfaces, as a
single element or in an antenna array. There are a lot of
advantages of using circular or hemispherical antenna
array configurations compared to planar antenna arrays;
such as increased tumor detection sensitivity,
enhancement on signal reception [20], increased
illuminated coverage area inside the breast [21], better
signal-to-clutter (S/C) ratio [22], etc.
Therefore, ultra wideband (UWB) bow-tie antennas
are used in an half-spherical conformal array encircling
the breast phantom model to investigate effects of
402
frequency dispersive dielectric properties of the phantom
on tumor detection capability of the microwave radar-
based imaging system, currently being developed at the
Scientific and Technological Research Council-BILGEM-
UEKAE.
Seven UWB bow-tie antennas are located 4.71 mm
above the dispersive homogeneous half-spherical breast
phantom that is modeled on a full-wave electromagnetic
simulator (CST Microwave Studio®). All of the time-
domain signals (Sii and Sij, i≠j) coming from different
antennas are obtained from the simulation model with and
without 2 mm diameter tumor, by feeding each antenna
sequentially. Recorded datas are processed on DAS
algorithm, and then images of the computed backscattered
signal energies for each pixel are created as a function of
position.
This paper is organized as follows: The layered half-
spherical breast phantom model, antenna design results
and antenna array structure are introduced with
comparative antenna simulations in Section 2, followed
by pulse distortion analysis. In Section 3, after briefly
outlining the DAS imaging algorithm, imaging results and
discussions are given. Finally, the findings of the study
and future work are briefly given in Section 4.
2. Simulation Models
2.1 Layered Half-Spherical Breast Phantom Model
The layered breast phantom model consists of a skin
tissue layer with thickness of 2 mm and a half-spherical
breast fat tissue model with radius of 58 mm under the
skin (Figure 1). The model does not include spatial
inhomogeneity of the breast. The homogeneous breast
phantom is surrounded by a coupling medium in which
antennas would be placed. Non-dispersive relative
dielectric permittivity εr and conductivity σ values of
skin tissue, breast fat tissue and coupling medium are
selected as; εr=36, σ=4 S/m [15]; εr=9, σ=0.4 S/m [9],
[22] and εr=9, σ=0 S/m [22], respectively. Moreover, non-
dispersive dielectric properties of the spherical tumor
model with diameter of 2 mm are selected as εr=50, σ=7
S/m [22].
Figure 1. Illustration of layered half-spherical breast
phantom model
 To include the frequency dependence of dielectric
properties of the breast phantom in the CST model, the
first Debye dispersion model has been applied assuming
the values such as ε∞, εs and σs, as in [5] valid from 100
MHz to 20 GHz and shown in Table 1: (a) (b)

εs − ε∞ σ
εr = ε∞ + −j s (1)
1 + j ωτ ω ε0
where εs is the relative low-frequency permittivity, ε∞ is
the relative high-frequency permittivity, τ is the relaxation
time, ε0 is the permittivity of free-space, and ω is the
angular frequency and σs is the ionic conductivity.

Table 1
Debye Model Parameters
τ σs
Tissue ε∞ εs (c)
(ps) (S/m) Figure 2. (a) Bow-tie antenna, (b) Simulation model of
Skin 4 48 8 0.5 bow-tie antenna array designed in front of the breast
Breast Fat 7 10 7 0.15 phantom, (c) Simulation results of return loss of the bow-
Tumor 4 54 7 0.7 tie antenna in the half-spherical antenna array encircling
the non-dispersive and dispersive breast phantom
2.2 Antenna Design
The performance of the antenna while immersed in
The bow-tie antenna (26 mm × 40 mm) presented in this the coupling medium is also tested for cases of antenna in
paper is aimed to be used as an UWB probe element of a the array operating with non-dispersive phantom, antenna
half-spherical antenna array as a part of the microwave alone operating with non-dispersive phantom and antenna
imaging system which is operating between 1 and 8 GHz alone operating without non-dispersive phantom,
(Figure 2(a)). The array is assumed to be immersed in a respectively (Figure 3).
coupling medium to achieve the best possible matching
with the breast tissue. The intented use of a matching
medium is required to reduce adverse effects of signal
reflections at the antenna-air-breast interface [15]. Non-
dispersive dielectric properties of the coupling medium
will be used in this study.
The bow-tie antenna is designed when it operates
with other antenna elements in the half-spherical antenna
array and also in front of the non-dispersive breast
phantom model (Figure 2(b)). The design objective was
optimizing the antenna geometry and size to obtain return
Figure 3. Comparative simulation results of return loss of
loss less than -10 dB over the bandwidth from 1 to 8 GHz.
the bow-tie antenna in the array operating with non-
Comparative results of return loss of the bow-tie antenna
dispersive phantom, antenna alone operating with non-
in the half-spherical antenna array encircling the non-
dispersive phantom and antenna alone operating without
dispersive and dispersive breast phantom are shown in
non-dispersive phantom
(Figure 2(c)). Simulation results are obtained for the 1.
antenna located at the center of the half-spherical antenna
Slightly better matching results have been obtained
array, as shown in Figure 2(b).
for the antenna operating with dispersive breast phantom,
The results show that the -10 dB bandwidth of the
in all, at different resonance frequencies. According to the
antenna which is operating in the half-spherical antenna
results in Figure 2 and Figure 3, these resonance
array surrounding the breast phantom extends from
frequencies are slightly affected by dispersive properties
around 1 GHz to above 8 GHz, with the exception of a
of breast phantom as well as other antenna elements in the
small (around 0.7 dB) mismatch centered at 7.37 GHz,
half-spherical array configuration. The presence of the
which has not been found to be important in imaging
phantom also affects them, too. However, the overall
results.
UWB performance of the bow-tie antenna in the presence
of the phantom and other array elements has been kept
same over the bandwidth from 1 to 8 GHz.
The wavelength at the center frequency (4.5 GHz) is
22 mm inside the coupling medium. The antenna-skin

403
distance (h) has been obtained as 4.71 mm, by optimizing
the distance with parametric sweep for the purpose of best
matching over the bandwidth. This value is close to the
expected distance which is theoretically quarter-
wavelength at the center frequency [23].
As the UWB microwave imaging system operates in
the time domain by sending a narrow pulse to penetrate
the breast and measure the scattered pulses, the bow-tie
antenna used in our system should introduce low
distortion in the time domain [15]. It’s important to study
distortion when the radiated pulse propagates through the
breast tissues. For this purpose, the transmitted pulse from
the 1. antenna located at the center of the half-spherical
antenna array is monitored at different distances normal Figure 5. Calculated fidelity factor with respect to
from the antenna aperture. The time domain performance distance from the antenna, in the presence of non-
of the antenna when it operates in front of the phantom is dispersive and dispersive breast phantoms
shown in Figure 4 for the cases with dispersive and non-
dispersive breast phantom models, respectively. 30 mm depth inside the breast. However, the fidelity
factor is kept at slightly higher values when the breast
phantom is non-dispersive. It’s more than 68.10% at 30
mm depth inside the breast. The fidelity factor of the case
with dispersive breast phantom is less than the case with
the non-dispersive one, as expected, due to increasing
pulse distortion in dispersive dielectric tissues [18]. If the
phantom model would be heterogeneous including gland
tissues inside the breast, the fidelity factor is expected to
decrease sharply due to the multiple reflections in the
heterogeneous phantom [15]. On the other hand, for the
bow-tie antenna presented in this paper, the fidelity factor
is within reasonable values (more than 60%) even inside
the breast phantom [25].

3. Imaging Results and Discussion

Figure 4. The excited and transmitted pulses inside the
dispersive and non-dispersive breast phantoms
In order to find out the distortion level in the
transmitted pulses inside the breast phantom model, the
fidelity factor is calculated at different locations within
the breast phantom. The fidelity factor is defined as the
maximum magnitude of the cross correlation between the
observed pulse ( sR ( t ) ) at a certain distance and the
excitation pulse ( sT ( t ) ) [24]:
When one of the seven UWB bow-tie antennas in the
array is excited by Gaussian pulse, back-scattered time-
domain signals (Sii and Sij, i≠j) are recorded. This
procedure is repeated by feeding each antenna
sequentially, for the cases with and without 2 mm
diameter tumor. 49 time-domain signals coming from
different antennas are recorded for each case. Tumor
response signals SijT are obtained by calibrating the
recorded signals as in Eq. (3):

∞
SijT = Sij − Sij (3)
with tumor without tumor

∫ s ( t ) s ( t − τ ) dt
d
T R

F = maxτ d −∞
(2) The tumor response signals are additionally
∞ ∞
compensated for 1/r attenuation of electric fields inside
∫ s (t ) dt ⋅ ∫ sR ( t ) dt
2 2
T
the breast and electromagnetic scattering from spherical
−∞ −∞
tumor with different signal levels at different angles [26].
where τ d is the required time delay for obtaining When 1. antenna is fed and signal is received from 3.
maximum magnitude of the cross correlation. antenna, one can easily compute time delay for the
The result in Figure 5 indicates an increasing pulse possible tumor location depicted in Figure 4.
distortion as the signal propagates through the breast Accordingly, time delay between the transmitted signal
phantom due to the reflections in the layered phantom from the 1. antenna and the received signal by the 3.
model. For the case with dispersive breast phantom, the antenna can be computed as in Eq. (4), regarding
fidelity factor decreases more and it becomes 63.93% at

404
velocities of electromagnetic fields in different media,
individually.
Figure 6. Illustration of breast phantom model and
antenna placement for computing time delays
K K
K d131 ( r ) d 231 ( r )
τ 31d ( r ) = + +
ν coupling ν skin
K observe the feasibility of the microwave imaging system
d (r )
31
d 31
+ 4
+ 5
ν breast fat ν skin
Total tumor response for each pixel is obtained, as in
Eq. (5), regarding computed time delays between each
antenna and pixel points, one by one [8]. Then, images of
the computed backscattered signal energies for the total
tumor response for each pixel, are created as a function of
position.
7 7
K K
T ( r ) = ∑∑ SijT (τ ijd ( r ) )
i =1 j=1
Imaging results of breast cancer tumor with 2 mm
diameter are obtained by using the DAS algorithm which
has been briefly described above (Figure 7). The tumor
with 2 mm diameter is succesfully detected at its actual
location, for the non-dispersive case, as shown in Figure 7
(a). On the other hand, it is also detected at a distance
around 6 mm far away from its actual location, with
decreased signal levels for the dispersive case, as shown
in Figure 7 (b). Frequency dispersive velocities of
electromagnetic waves had been kept constant as it was
done in the non-dispersive case. Therefore, the reason of
detection of the tumor at a faulty location obviously
depends on wrong computation of time delays used in the
DAS algorithm.
Moreover, if one compares the dispersive breast
phantom to the non-dispersive phantom from calculated
total tumor response levels point of view, signal reduction
of 0.628 dB is observed in the dispersive case. This
reduction is possibly related with slightly worse pulse
distortion performance of the antenna operating with the
dispersive breast phantom, and also incoherent summing
of the inaccurate delayed signals, because of the
frequency dispersive velocities of electromagnetic fields.
Therefore, new imaging algorithms compensating the
dispersive effects, such as "Microwave Imaging via
Space-Time Beamforming" (MIST) [10], should be used
(a) (b)
Figure 7. (a) Images of breast cancer tumor with 2 mm
diameter for the cases with (a) non-dispersive and (b)
dispersive breast phantoms
in the design of the microwave imaging system. On the
K other hand, more realistic values of dispersive dielectric
d331 ( r ) properties [27] and spatial inhomogeneity [12], [15], [28]
ν breast fat should be included in the breast phantom model to
(4)
( rK ) K
d 631 ( r )
+ better.
ν coupling
4. Conclusion and Future Work
An ultra wideband (UWB) bow-tie antenna array
surrounding the breast has been designed and tested on a
full-wave electromagnetic simulator, in order to
investigate the effects of frequency dispersive dielectric
properties of the half-spherical breast phantom on breast
cancer tumor detection capability of the microwave radar-
(5)
based imaging system, currently being developed at the
Scientific and Technological Research Council-BILGEM-
UEKAE. Obtained simulation results are reasonably
reliable. Time domain behaviour of the antenna operating
with the dispersive breast phantom has indicated slightly
worse pulse distortion performance through the breast.
Images of the detected spherical tumor with 2 mm
diameter have been formed by using the DAS algorithm
which is not found to be suitable for the frequency
dispersive breast tissues. Therefore, better imaging
algorithms compensating the dispersive effects would be
used with more realistic heterogeneous breast phantom
models in the future studies.
References
[1] E.C. Fear, X. Li, S.C. Hagness, & M.A. Stuchly,
Confocal microwave imaging for breast tumor detection:
Localization in three dimensions, IEEE Trans. Antennas
Propag., 49(8), 2002, 812-822.
[2] K.L. Carr, Microwave radiometry: Its importance to
the detection of cancer, IEEE Trans. Microwave Theory
and Tech., 37(12), 1989, 1862-1869.
[3] L.V. Wang, X. Zhao, H. Sun, & G. Ku, Microwave-
induced acoustic imaging for biological tissues, Rev. Sci.
Instrum., 70(9), 1999, 3744-3748.
[4] A.E. Bulyshev, S.Y. Semenov, A.E. Souvorov, R.H.
405
 Svenson, A.G. Nazarov, Y.E. Sizov, & G.P. Tatsis,
Computational modelling of three-dimensional
microwave tomography of breast cancer, IEEE Trans.
Biomed. Eng., 48(9), 2001, 1053-1056.
[5] E.J. Bond, X. Li, S.C. Hagness, & B.D. Van Veen,
Microwave imaging via space-time beamforming for
early detection of breast cancer, IEEE Trans. Antennas
Propag., 51(8), 2003, 1690-1705.
[6] M. Klemm, J.A. Leendertz, D. Gibbins, I.J.
Craddock, A. Preece, & R. Benjamin, Microwave radar-
based differential breast cancer imaging: Imaging in
homogeneous breast phantoms and low contrast scenarios,
IEEE Trans. Antennas Propag., 58(7), 2010, 2337-2344.
[7] M. Klemm, D. Gibbins, J. Leendertz, T. Horseman,
A.W. Preece, R. Benjamin, & I.J. Craddock, Development
and testing of a 60-element UWB conformal array for
breast cancer imaging, Proc. 5th European Conf. on
Antennas and Propagation, Rome, Italy, 2011, 3077-
3079.
[8] Y. Zhao, W. Shao, & G. Wang, UWB microwave
imaging for early breast cancer detection: Effect of two
synthetic antenna array configurations, IEEE Int.
Conference on Systems, Man and Cybernetics, The
Hague, Netherlands, 2004, 4468-4473.
[9] S.C. Hagness, A. Taflove, & J.E. Bridges, Wideband
ultralow reverberation antenna for biological sensing,
Electronics Letters, 33(19), 1997, 1594-1595.
[10] X. Li, S.K. Davis, S.C. Hagness, D.W. van der
Weide, & B.D. Van Veen, Microwave imaging via space-
time beamforming: Experimental investigation of tumor
detection in multilayer breast phantoms, IEEE Trans.
Microw. Theory Tech., 52(8), 2004, 1856-1865.
[11] X. Yun, E.C. Fear, & R.H. Johnston, Compact
antenna for radar-based breast cancer detection, IEEE
Trans. Antennas Propag., 53(8), 2005, 2374-2380.
[12] N. Tavassolian, S. Nikolaou, & M.M. Tentzeris, A
flexible UWB elliptical slot antenna with a tuning uneven
U-shape stub on LCP for microwave tumor detection,
Asia-Pasific Microwave Conference, Bangkok, Thailand,
2007, 1-4.
[13] İ. Ünal, B. Türetken, K. Sürmeli, & C. Canbay, "An
experimental microwave imaging system for breast tumor
detection on layered phantom model", URSI GASS 2011,
İstanbul, Turkey, 2011, 1-4.
[14] J. Bourqui, M. Okoniewski, & E.C. Fear, Balanced
antipodal vivaldi antenna for breast cancer detection, The
Second European Conference on Antennas and
Propagation, Edinburgh, UK, 2007, 1-5.
[15] B.J. Mohammed, A.M. Abbosh, & M.E. Bialkowski,
Design of tapered slot antenna operating in coupling
liquid for ultrawideband microwave imaging systems,
APSURSI, Washington, USA, 2011, 145-148.
[16] H.M. Jafari, M.J. Deen, S. Hranilovic, & N.K.
Nikolova, A study of ultra-wideband antennas operating
in a biological medium, IEEE Trans. Antennas Propag.,
55(4), 2007, 1184-1188.
[17] S. Gabriel, R.W. Lau, & C. Gabriel, The dielectric
properties of biological tissues: III. parametric models for
406
View publication stats
the dielectric spectrum of tissues, Phys. Med. Bio., 41(11),
2007, 2271-2293.
[18] F.G. Boese, On impulse distortion in dispersive
media, Proc. Appl. Math. Mech., 6(1), 2006, 603-604.
[19] K.E. Oughstun, Pulse propagation in a linear,
causally dispersive medium, Proc. IEEE, 79(10), 1991,
1379-1390.
[20] Z. Šipuš, S. Škokić, & N. Burum, Performance
analysis of spherical stacked-patch antennas, 18th Int.
Conference on Applied Electromagnetics and
Communications, Dubrovnik, Croatia, 2005, 1-4.
[21] R.C. Conceição, M. O’Halloran, M. Glavin, & E.
Jones, Comparison of planar and circular antenna
configurations for breast cancer detection using
microwave imaging, Progress In Electromagnetics
Research, PIER 99, 2009, 1-20.
[22] X. Yun, R.H. Johnston, & E.C. Fear, Radar-based
microwave imaging for breast cancer detection: Tumor
sensing with cross-polarized reflections, Antennas and
Propagation Society Int. Symposium, California, USA,
2004, 2432-2435.
[23] C. Canbay, Anten ve propagasyon I (İstanbul,
Yeditepe University Press, 1997).
[24] D. Lamensdorf & L. Susman, Baseband-pulse-
antenna techniques, IEEE Antennas Propag. Magazine,
36(1), 1994, 20-30.
[25] A. Abbosh & M. Bialkowski, Compact directional
antenna for ultra wideband microwave imaging systems,
APSURSI, USA, 2009, 1-4.
[26] İ. Ünal, S.B. Malla, & C. Canbay, Determining of
interaction mechanism between scattering
electromagnetic fields and breast cancer tumor using
theoretical human body models, Istanbul Conference on
Mathematical Methods and Modeling in Life Sciences and
Biomedicine 2011 (ICMMM-LSBM), İstanbul, Turkey,
2011.
[27] M. Lazebnik, M. Okoniewski, J.H. Booske, & S.C.
Hagness, Highly accurate debye models for normal and
malignant breast tissue dielectric properties at microwave
frequencies, IEEE Microwave and Wireless Components
Letters, 17(12), 2007, 822-824.
[28] D.W. Winters, J.D. Shea, Panagiotis Kosmas, B.D.
Van Veen, & S.C. Hagness, Three-dimensional
microwave breast imaging: Dispersive dielectric
properties estimation using patient-specific basis
functions, IEEE Trans. Medical Imag., 28(7), 2009, 969-
981.