DETECTION OF BREAST CANCER

deaths due to cancer. The various factors causes

USING IMAGINGcancer
SENSOR AND
are physical carcinogen are ultraviolet and
ionizing radiation, chemical

MAMMOGRAMcarcinogen
- A REVIEW
are tobacco, smoke, alcohol and
arsenic, biological carcinogen are certain viruses,
bacteria. In 2018, 30% of cancer diagnosed HPV,
Dr. Srigitha S. Nath, Professor Department of Electronics and Communication
hepatitis B, hepatitis C. Both viruses are danger to
Engineering, Saveetha Engineering College,
liver, colon 0000-000
and cervical cancer. Nowadays 30% to
50% of cancer is curable. The cancer burden is
Anjaline Jayapraba.A, Assistant Professor, Department of Electronics and
reduced due to early detection of cancer.
Communication Engineering, Saveetha Engineering College,
HPV tests and pap tests for cervical cancer,
[email protected]
mammogram screening method for breast cancer
patients among 50 to 60 years. Colonoscopy, stool

ABSTRACT
test, sigmoidoscopy to reduce the risk of colorectal
Breast cancer killer of women. Mortality can be
reduced if disease is early detected. Breast cancer cancer screening. CT scan for lung cancer detection
leading the second cancer cause death all over the for 50 to 75 age’s group people is affected. Alpha
world. Conventional method of breast cancer fetoprotein blood test for liver, PSA blood test for
detection is various disadvantage it’s include high cancer screening is not recommended. Transvaginal
cost, heavy radiation and discomfort to patients. The ultrasound imaging test for women ovaries and
alternative method to decrease the death rate is uterus test when women are cause serious means
microwave imaging technique. The main objective of this test occur because they carry harmful gene
paper to provide various techniques in microwave
mutation. A virtual colonoscopy allows the colon
imaging and hand held device for breast cancer
detection.
and rectum are used to tests outside the body.
Microwave imaging means detecting and locating
keywords: Mammogram, Microwave imaging, Lymph object from hidden object using electromagnetic
nodes, imaging algorithm. waves. The two method of microwave imaging are
qualitative and quantitative imaging system.
1 INTRODUCTION Quantitative technique gives electrical and
magnetic property and geometry parameter of an
The most common types of cancer are breast,
image it can solve non linear inverse problem.
colon, lung, and rectum cancer. Among one third of
Qualitative technique gives reflective signal of an
death due to tobacco use, highly alcohol
image. A transmitting antenna send
consumption, low foods intake etc. approximately
electromagnetic waves to sample that signal are
30% of cancer caused by HPV and hepatitis cases
reflected that waves are back to transmitting
is lower middle income countries. But many cancer
antenna, the collected information are send to
are cure when early detection and effective
software for further processing. Microwave
treatment. The most common cancer death in 2020
imaging safer for dense breast patients because x-
are lung (1 million deaths), colon and rectum
ray mammography is not suitable interms of cost
(9,16,000), liver (8,30,000), breast (7,69,000)
and high radiation. This technique has low risk and
nearly 10 million people
low cost

1.1 DIELECTRIC PROPERTY

MEASUREMENT

Dielectric property of tissue range from 1Hz to 30

GHz. Three dispersion are alpha, beta, gamma,
alpha dispersion focus on frequency dependence on
outer cell membrane cover range from 0.1 to 100
KHz. Beta dispersion cover range from 1 to 20
MHz it has high electromagnetic frequency and
enclosed cellular structure with bound water.
Gamma dispersion about 20 GHz it has dipolar
relaxation of water in the tissue. Our blood, body,
muscles and fatty tissue are different dielectric
properties. High water content include blood, brain
and internal organs low water content include fat,
bone, lung and other inner organ. Low water
content tissue is low permittivity. Network analyzer
measures the relative amplitude and phase
difference from the reflected signals. Various
algorithm used to computes the dielectric property,
Joines et.al 1994 in vitro identify the normal versus
malignant tissue it cover 50 to 900 MHz Meaney
et.al 2000 microwave imaging system using breast
cancer imaging contrast between normal and
malignant breast tissue ratio 2:1. Sha et.al dielectric
properties to identify the normal and malignant
tissue at 100 MHz to 1 GHz. Dielectric properties Figure 2: Flow diagram of microwave imaging
are not same for all the tissue based on relative technique
permittivity and conductivity that will
The two common microwave imaging system is
characterized therefore detection of tumor is
microwave tomography and radar based imaging
compared through microwave imaging and
technique. Tomographic method used to solve
dielectric properties of tissue. Breast tissue is
complex inverse problem. Antenna is placed in
affected by disease and temperature that can be
circular arrangement to carried out the information.
identified by variation in dielectric properties.
The microwave signal is passed to object and
scattered signal are collected and process using
algorithm. Radar based detection technique to
detect the tumor. Hologram imaging technique to
detect the tumor using optics. The hybrid method
has higher conductivity in cancer cells rather than
non- cancerous cells. Microwave tomography
requires high data acquisition time. The first
method of MWT imaging for kidney at frequency
3.5 GHz. The transmitter and receiver antenna
move in various direction but this is not suitable for
hardware implementation. To overcome the
disadvantage the second method circular antenna to
Figure 1: Basic diagram of dielectric property
design in mechanical movement. Third method
measurement
called cylindrical antenna to detect the small tumor
1.2 TYPES OF MICROWAVE IMAGING using multi frequency model. Another method
TECHNIQUE three dimension cylindrical antenna to detect small
size tumor.

Radar based imaging technique, the first method

developed by Hagness using monopole antenna to
detect 2mm breast tumor. Using 3D system to
detect the 6mm breast tumor. The major
disadvantage is noise and high frequency
dependent and high penetration. Tissue sensing
adaptive radar protocol to reduce the noise using
this method to detect 4mm diameter breast tumor.
The major disadvantage is high radiation it cause horn antenna bandwidth 1 to 18 GHz and gain 10
patient’s skin. Microwave imaging system to dB . [4] Breast cancer is the most common type of
develop the sensor to attain required bandwidth, cancer mostly occur in females the current
frequency and sensitivity. The main consideration technique are mammogram, x-ray, tomography,
in sensor design it penetrates more into tissue to digital mammogram and PET (Positron emission
obtain low frequency. Various sensor are used the tomography) use low energy to generate images.
first method open coaxial probe the advantage is But the exposure of radiation is harmful to most of
easy to use. To attain high accuracy because of the patients and then in mammogram based on
broad band frequency. Another method tapped slot breast size compression is used so they experienced
antenna it has low cost this method is more pain it is disadvantage of identify the tumor in
popular. Bow tie sensor are easy to manufacture dense breast. Second method ultrasound it uses
and low cost using inverse scattering algorithm to sound waves to create image it frequency greater
reconstruct the image. Microwave imaging system than 20 KHz. Transducer placed on skin to produce
gives more attention to sensor array the various image it is painless method and less radiation
types are spherical, cylindrical, circular and planar exposure. But one serious disadvantage is
shaped microwave imaging array. Circular sensor ultrasound does not predict the tumor whether it is
has more advantage this method used in clinical benign or malignant so the resolution is low. Third
approach and high resolution but major method is MRI (Magnetic resonance imaging) this
disadvantage is it has more cost. Strip sensor used method uses radio waves and magnetic field to
in microwave imaging for detecting tumor it has produce image this method suitable for early
low cost, less expensive and easy to implement in detection of breast cancer. Another method is PET
circuit board. Patch antenna frequency range up to scan use of glucose and glycoprotein to identify the
9 GHz the antenna bandwidth 77% to detect breast cancer cells that cells are consume nutrients so it
tumor. Micro strip antenna were used to detect grow fast. Compare to other method PET scan are
tumor up to 10 GHz frequency range in future 3D identify the cancer in early stage. Microwave
imaging were used. Micro strip slot sensor antenna imaging method is an alternative solution to
achieves good performance up to 8 GHz it can use mammogram newly method is CMOS chip based
for other application. time domain technique. Microwave tomography
uses the inverse scattering to project image the
2 MICROWAVE IMAGING SYSTEM inverse signal map permittivity and conductivity it
take more time for calculation process. Monopole
[1] Ultra wide band MIMO antenna aims to design
antenna operates 300 MHz to 1 GHz frequency
early detection of breast cancer detection. The
range resolve the time problem. The patient’s data
MIMO based antenna design to radiate ultra wide
sets acquire seven different array heights for better
band in the air it covering 2.3 GHz–12.2 GHz.
resultant images.
importantly two elements are used in proposed
MIMO structure low mutual coupling and envelope [6] The modified log periodic array antenna has
correlation. Mutual coupling is the main factor to operating range 2 to 5 GHz. The antenna has placed
achieve high gain and better performance. The circular pattern around phantom totally 16 antenna
proposed antenna 1 has vertically polarized pattern are placed one antenna act as transmitter and
and antenna 2 has horizontally polarized pattern it remaining antenna acts as a receiver. DAS
gives high resolution. [2] In this proposed algorithm used to identify the breast tumor present
hexagonal shaped antenna feed by coupling in the image. [7] Microwave is non ionizing
method. Rectangular shaped antenna very radiation, inexpensive treatment method. Based on
important for volume limited applications. The the electrical properties of breast tissue it will
wide band antenna covering range from 1.5 to 2.5 identify the small tumor. Three methods of
GHz and 4 to 8 GHz. The bandwidth of wideband detecting are active, passive and hybrid, in passive
antenna is nearly 50% is sufficient for breast cancer method based on temperature variation to identify
detection application. [3] For ultra wide band the normal and abnormal cells, the cancer affected
application two antenna has been decide one has cells are higher temperature and different dielectric
spiral slot patch antenna and circular disc properties. Hybrid technique compose of
monopole antenna among this antenna pulse time microwave sensor and transducer to detect the
domain reception compare with standard wide band breast cancer due to higher electrical properties
they emits more acoustic energy than normal tissue. [12] ultra wide band antenna based time domain
[8] Wearable radar based breast cancer detection, microwave imaging system has 16 antenna it has
16 based monopole antenna are surrounded in bra less expensive and low cost more suitable to
surface the antenna does not directly contact with patients. This system used to detect breast cancer
skin surface. The transmitted and receiver antenna up to 1mm diameter.
select the pulse generated form switching network.
Picoscope used to record the received signal. [9]
[10] In the proposed system UWB antenna has
been developed to detect the breast tissue. DAS
algorithm used to reconstruct the image 20mm
diameter cancer cells has been detected. 16 UWB
embedded based antenna using solid state
technology 1cm tumor has been detected.

[11] This structure has multiple sensors to detect

breast tumor. The proposed algorithm to
differentiate the healthy and unhealthy tumor they
respond based on dielectric properties of breast.

Figure 5: Microwave time domain system for

breast cancer detection

[13] The microwave radar system used to detect

tumor the setup has shown in the figure pulse
generator and amplify and send into switching
matrix then collected using transmitter and receiver
antenna. Picoscope used to record the received
signal.

Figure 4: (a) Microwave array sensor schematic

diagram

Figure 6: Wearable microwave antenna array for

(b) Microwave array sensor setup diagram for time domain breast cancer detection
early detection of breast cancer
[14] This system consisting of nine antennas with
adjustable phantom. The antenna adjustable
parameter is ABC using antenna array to
implement the phantom stepper motor to control
the container. VNA generate microwave frequency
transmit to phantom backscattered signal are
collected using receiver.

Figure 9: CMOS based time domain system for

Figure 7: Low cost portable microwave imaging breast cancer detection
system using UWB directional array antenna
Using planar antenna to detect 2mm diameter
[15] Gaussian monocycle pulse transmitter time breast tumor this method improve the performance.
domain system requires high clock speed, fast Another method using Fourier transforms to
sampling rate and low jitter. Time domain approach reconstruct the image. UWB signal are used to
has several advantages less cost and less scan time. detect the breast cancer detection. Bridge method
are used to detect abnormal breast tissue, antenna
are placed indifferent sides and back scattered
signals are reconstructed using beam forming
algorithm. Various algorithm are used to detect the
breast cancer DAS, TR, IFFT, SRR, FDAS, NCCB,
DMAS, TSAA, MFA etc.

Figure 8: Recent advances in microwave imaging Figure 3: Microwave Tomography

for breast cancer detection
3 MAMMOGRAM BASED BREAST CANCER
[16] CMOS based time domain system for breast DETECTION
cancer detection can reduce the scan time within
one second. Breast compression does not require in [17] Breast composed of connective tissue,
microwave imaging system. granular tissue and fatty tissue. In mammogram
white color represent connective and granular
tissue. Black color represents fatty tissue. Breast
contain various parts lobes, nipples, ducts and
areola. Lymphatic consist of lymph vessel and
lymph nodes. Lumph means mass that can find
near breast or arm that tissue are thin and fatty so
this may cause changes in breast size. Most of the
lymph is not cancer another factor nipple discharge
they face inward to breast or some kind of fluid is
not breast milk. Another factor skin changing
itching and redness. Some of the ways to find
breast changes for example clinical breast exam
and mammogram. Mammogram to find small or
large size of tumor but this method is more painful
and discomfort to patient. The best time to take
mammogram is end of your menstrual cycle. [18]
The two types of mammogram are film
mammogram and digital mammogram this both
technique are same but different in storage and
recorder. Film mammogram take the image and
store directly on the film and digital mammogram
store the image on the computer. National cancer
institute (NCI) shower digital mammogram is best
to detect the breast cancer. Mammogram used for
screening and diagnostic both methods are used to
find the breast change. But diagnostic method to
find the tumor in different angle. Another method
Figure 10: Flow diagram of image processing
to diagnostic the tumor are MRI, ultrasound biopsy
is to remove the sample of breast tissue some
methods include fine needle aspiration biopsy, core
biopsy, vacuum assisted biopsy and surgical
biopsy. Breast cancer found in under arm or
inverted nipple, treatment depend on the stages
chemotherapy, radiation therapy, hormonal
therapy, biological therapy and surgery. Small
round called lumps or lobules treatment include
core biopsy or surgical biopsy. Abnormal cells in
breast lobules condition named as ALH treatment Figure 11: Mammogram
include clinical breast exam and mammogram.
Lumps filled with fluid condition called cysts more [20] Double thresholding based segmentation
common in women 30 to 50 years old treatment approach to detect the tumor in mammogram
include ultrasound. Breast cancer in breast duct storage area and processing time will reduce. This
condition called Ductal carcinoma in situ this type method eliminate the image from the background
of cancer treatment includes surgery. [19] New black and grey cell value L = 0.75& U = 0.96. [21]
method for breast cancer classification of benign or Segmentation play a important role in
malignant. Hybrid thresholding to separate ROI mammogram most common segmentation
multiple fractal approach for feature selection using method include thresholding, feature extraction to
ANN to classify the tumor. extract the breast area in the image support vector
machine for classifier. The different types of
mammogram are digital mammogram,
tomosynthesis and full field digital mammogram.
Digital mammogram otherwise called full field
digital mammogram using x-ray film replaced by
solid state detector the electrical signal are used to
reproduce the image and send to computer.
Computer aided detection system highlight the
abnormal area. Tomosynthesis to create 2D and
3D images.
 Figure 15: Direct method
Figure 12: Conventional mammogram
[22] Breast cancer detection using thermal images
give appropriate result when compare to
mammogram. But pattern recognition is difficult
because of different in size, shape and texture
bispectral invariant feature highly concentrate on
extract images from noisy images to depend on
size, shape and features. Using IR camera to
Figure 13: Digital mammogram
capture breast images for better resolution to
differentiate benign or malignant. Fuzzy c-means
clustering used for feature extraction, Ad boost
classifier is used to select the feature. Using this
classifier to achieve 83% accuracy. [23]
Mammogram can cause side effects to most
patients in this proposed method computer based
cancer protection model has been developed.
Support vector machine to classify the image based
on require ANN. [24] using deep learning method
to detect the breast cancer using thermal image.
Segment the images into training set and validation
set 93% accuracy using this DNN method. Feature
mapping done in max pooling layer. [25] Using
diffuse optical tomography to detect the breast
cancer in deep neural network. Optical properties
of breast tissue are collected this method overcome
the limitation such as cost, sensitivity, selectivity
and false rate. This method differentiates benign or
Figure 14(a): Indirect method
malignant more accuracy compare to acoustic
method. Support vector machine used to classify
the image based on scattering, absorption and
refractive index. [26] Optical sensor to detect the
breast cancer depends on transmittance and
absorption parameter. Normal cells are higher
absorption rate when compare to unhealthy cells.
This optical sensor method common for liver, lung
and breast cancer. Confocal microscope use to
checking the normal or cancerous cells this
microscope used to count the number of cells in
Figure 14(b): Indirect method CCD suspension and adjust 10^7 ml cells. After the
suspension the measured cells are collected through Figure 18: Segmented breast cancer using
experimental setup. circular Hough transform

Figure 16: Transmittance and optical sensor

Figure 19: Basic block diagram of deep learning

network

[29] Non contact based breast cancer detection to

Figure 17: Cancer cell detection using optical detect up to 1mm based on reflectance and
sensor transmittance mode they detect using NIROS
detector. CCD camera acts has detector, the target
[27] Histology images to detect breast cancer they for reflectance 1.5 mm and 4mm in transmittance
separate into 4 classes sliding window contain mode the NIROS scanned comprised into hand for
essential information about breast tissue k-means comfortable the reduced risk this method more
clustering algorithm to segment the image. Feature comfortable to patient.
extraction based on structure of tissue and texture.
[28] Thermal images taken in infrared camera to
decide the decision the data given to support vector
machine and deep neural network first remove the
background using Otsu’s thresholding method,
locate the object using hough transform and
segmentation algorithm random walk to segregate
the object. Point with higher score mark cancer
cells.

Figure 20: Hand held device for breast cancer

detection

CONCLUSION

The purpose of this paper to demonstrate various

breast cancer detection technique implement in
image processing tool and various algorithm. The
above mentioned method are recent developed
approach to detect breast cancer it consume less 10. Yadollahpour Ali and Shoghi Hamed Early Breast Cancer
Detection using Mammogram Images: A Review of Image
time and high accuracy. Various segmentation
Processing Techniques BIOSCIENCES BIOTECHNOLOGY
algorithm are compared with previous method in RESEARCH ASIA, March 2015. Vol. 12(Spl. Edn. 1), p. 225-
deep learning algorithm accuracy is high compare 234.
to other method.
11. (IJACSA) International Journal of Advanced Computer
Science and Applications, Vol. 8, No. 10, 2017 117 | P a g e
www.ijacsa.thesai.org Breast Cancer Detection with
Mammogram Segmentation: A Qualitative Study
REFERENCES
12. Y.Ireaneus Anna Rejani et al /International Journal on
1. Liting Wang1 and Bin Huang2 Design of Ultra-Wideband
Computer Science and Engineering Vol.1(3), 2009, 127-130 127
MIMO Antenna for Breast Tumor Detection International
EARLY DETECTION OF BREAST CANCER USING SVM
Journal of Antenna and Propagation Research Article | Open
CLASSIFIER TECHNIQUE
Access Volume 2012 |Article
ID 180158 | https://doi.org/10.1155/2012/180158
13. Sebastien Jean Mambou 1, Petra Maresova 1 Ondrej Krejcar
1, Ali Selamat 1,2 and Kamil Kuca 1 Breast Cancer Detection
2. Abhishek Choubey, Rachna Pal Hexagonal Shaped Ultra
Using Infrared Thermal Imaging and a Deep Learning Model
Wide Band Patch Antenna with a Square Fractal in Ground
Received: 26 May 2018; Accepted: 23 August 2018; Published:
Plane for Breast Cancer Detection International Journal of
25 August 2018 MDPI.
Engineering Research & Technology (IJERT) Vol. 1 Issue 6,
August - 2012 ISSN: 2278-0181
14. Sumitha mishra, aditya prakash, sandip kumar roy “ Breast
cancer detection using thermal images and deep learning”
3. A compact ultra-wideband monopole antenna for breast
cancer detection October 2016
15. Qiwen Xu1,2, Xin Wang1,2 and Huabei Jiang1,2,3*
DOI:10.1109/MWSCAS.2016.7870066 Conference: 2016 IEEE
Convolutional neural network for breast cancer diagnosis using
59th International Midwest Symposium on Circuits and Systems
diffuse optical tomography Xu et al. Visual Computing for
(MWSCAS)
Industry, Biomedicine, and Art (2019) 2:1
https://doi.org/10.1186/s42492-019-0012-y
4. Recent Advances in Microwave Imaging for Breast Cancer
Detection Sollip Kwon and Seungjun Lee Published online
16. Adel najar et al, Label-Free Cancer Cells Detection Using
2016 Dec 21. doi: 10.1155/2016/5054912
Optical Sensors Article in IEEE Access · October 2018 DOI:
10.1109/ACCESS.2018.2872768
5. Microwave Sensors for Breast Cancer Detection Lulu
Wang1,2Published online 2018 Feb 23. doi: 10.3390/s18020655
17. Yuqian li 1, junmin wu1,2, and qisong wu3 Classification of
Breast Cancer Histology Images Using Multi-Size and
6. Zhang, Z.Q.; Liu, Q.H.; Xiao, C.; Ward, E.; Ybarra, G.;
Discriminative Patches Based on Deep Learning SPECIAL
Joines, W.T. Microwave breast imaging: 3-D forward scattering
SECTION ON DEEP LEARNING FOR COMPUTER-AIDED
simulation. IEEE Trans. Biomed. Eng. 2003, 50, 1180–1189.
MEDICAL DIAGNOSIS Received January 6, 2019, accepted
January 28, 2019, date of publication February 7, 2019, date of
Meaney, P.M.; Fanning, M.W.; di Florio-Alexander, R.M.;
current version February 27, 2019. Digital Object Identifier
Kaufman, P.A.; Geimer, S.D.; Zhou, T.; Paulsen, K.D.
10.1109/ACCESS.2019.2898044
Microwave tomography in the context of complex breast cancer
imaging. In Proceedings of the 2010 Annual International
18. Aditya Golatkar(B), Deepak Anand, and Amit Sethi
Conference of the IEEE Engineering in Medicine and Biology
Classification of Breast Cancer Histology Using Deep Learning
Society (EMBC), Buenos Aires, Argentina, 31 August–4
Indian Institute of Technology Bombay, Mumbai, India
September 2010; pp. 3398–3401.
19. Mahmoud al ahmad 1, (senior member, ieee), adel najar2,
7. Kwon, K.C.; Lim, Y.T.; Kim, C.H.; Kim, N.; Park, C.; Yoo,
amine el moutaouakil1, nida nasir1, minas hussein1, shaima
K.H.; Son, S.H.; Jeon, S.I. Microwave tomography analysis
raji1, and ali hilal-alnaqbi Label-Free Cancer Cells Detection
system for breast tumour detection. J. Med. Syst. 2012, 36,
Using Optical Sensors Received August 6, 2018, accepted
1757–1767.
September 12, 2018, date of publication October 1, 2018, date
of current version October 25, 2018. Digital Object Identifier
8. Fear, E.C.; Li, X.; Hagness, S.C.; Stuchly, M.A. Confocal
10.1109/ACCESS.2018.2872768
microwave imaging for breast cancer detection: Localization of
tumours in three dimensions. IEEE Trans. Biomed.
20. Hafiz Talha Iqbal, Bilal Majeed, Uzair Khan and
Eng. 2002, 49, 812–822
Muhammad Awais Bin Altaf Lahore University of Management
Sciences, Lahore, Pakistan [email protected] An
9. Dheyaa Ahmed Ibrahim , Diyar Qader Zeebaree , Mazin
Infrared High classification Accuracy Hand-held Machine
Abed Mohammed , Habibollah Haron , Nechirvan Asaad
Learning based Breast-Cancer Detection System.
Zebari , Robertas Damaševiˇcius , and Rytis Maskeliunas
“Breast Cancer Detection Using Mammogram Images with
Improved Multi-Fractal Dimension Approach and Feature
Fusion”