Download Advances in Natural and Applied Sciences

Advances in Natural and Applied Sciences, 9(7) June 2015, Pages: 92-99
AENSI Journals
Advances in Natural and Applied Sciences
ISSN:1995-0772 EISSN: 1998-1090
Journal home page: www.aensiweb.com/ANAS
Unsupervised Segmentation based Classification of MICROCALCIFICATION Using
SVM
S. Venkata Lakshmi and Dr. J. Janet
CSE Manonmaniam Sundaranar University, Tirunelveli
ARTICLE INFO
Article history:
Received 12 March 2015
Accepted 28 April 2015
Available online 1 June 2015
Keywords:
breast
Cancer,
Features,
classification,
Particle
swarm
optimization. Fuzzy Enhancement.
ABSTRACT
Advancement of medical image digitalization leads to image processing and computeraided diagnosis systems in numerous clinical applications. These technologies could be
used to automatically diagnose patient or serve as second opinion to pathologists. This
paper briefly reviewed on breast screening techniques, available, so that this can
motivate the researcher to concentrate more on medical related issues which becomes a
small contribution to our country which is very essential. The digital data of the
screening techniques are used as data for the computer screening system, as replaced in
the expert analysis. Four stages of the computer system are enhancement, features
extraction, feature selection, and classification reviewed in detail. Breast cancer takes a
tremendous toll on women, next to skin cancer; it is the most common cancer in women
and the leading cause of mortality in women. Screening program is based on x-ray
examination of the breasts which is the best method for early detection. The
development of the new system is to assist the radiologist and not to replace the
radiologist. The proposed system consists of three stages namely image enhancement,
segmentation
and classification of micro calcification.
Enhancement
of
mammograms is done by normalization, noise removal and pectoral muscle
suppression to improve the quality of the image and in separating the pectoral
muscle from the whole breast area. Detection of suspicious regions is done by
segmentation. In this Particle Swarm Optimization (PSO) has been proposed to
segment the suspicious regions from the enhanced image. Classification is done
in three steps namely feature extraction, feature selection and by classification in
two phases. Features are extracted from energy values which are calculated from
wavelet transform of an image. Support Vector Machine (SVM) is used as a classifier,
which classifies the micro calcification into benign and malignant. SVM improves
classification accuracy. The classifier proved to possess the best recognition ability
due to its ability to deal with nonlinearly separable data sets. Receiver Operating
Characteristic (ROC) analysis is presented to evaluate the classification performances
with Darwinin PSO (DPSO) and FODPSO. Micro calcification images in the
Mammography Image Analysis society (MIAS) database are used to evaluate the
proposed system.
© 2015 AENSI Publisher All rights reserved.
To Cite This Article: S. Venkata Lakshmi and Dr. J. Janet, Unsupervised Segmentation based Classification of MICROCALCIFICATION
Using SVM. Adv. in Nat. Appl. Sci., 9(7): 92-99, 2015
INTRODUCTION
Cancer is the public health problem for men and
women in this century. According to the survey more
than 8% of women are affected by breast cancer
(Jemal, 2008; Imaginis Corporation, 2008) and 29%
of men are affected by prostate cancer (Ferlay, 2007;
Jemal, 2007; Ferlay, 2007; Cheng, 2006; Noble,
2006), 31% of women are affected by cervical
cancer (Suetens, 2009) and 70% of women are
affected by ovarian cancer (Polascik, 2008). Since
the causes of cancer still remain un-known, better
treatment can be provided to detect from the early
stage (Bangma, 2007; Schröder, 2009; Wink, 2007;
Jesneck, 2007; Pelzer, 2005). The most modality for
detecting the diagnosing is mammography (Bangma,
2007). To the low specificity mammography many
biopsy operations are used (Bluth, 2007; Uof
Pittsburgh Medical Centre, 2009; Yang, 2006).
Currently one of the best alternative method is called
ultrasound imaging technique. (Sahiner, 2007;
Huang, 2005; Loch, 2007; Costantini, 2006).
According to the survey showed that more than one
out of every four researches using ultrasound images,
which provides accurate results (Park, 2007).
Ultrasound techniques are more convenient and safer
than mammography (Salcudean, 2006). It is also
cheaper than mammography. Different countries and
continents use ultrasound technique (Cheng, 2010)
which is more sensitive (Kalaivani Narayanan,
2009) and faster method. Hence it is valuable for
people than 35 years of age ACS, 2011).
Corresponding Author: S. Venkata Lakshmi, Research Scholar, CSE Manonmaniam Sundaranar University, Tirunelveli
E-mail: [email protected]
93
S. Venkata Lakshmi and Dr. J. Janet, 2015
Advances in Natural and Applied Sciences, 9(7) June 2015, Pages: 92-99
Elastography is an automatic method for measuring
the elasticity of tissue based on analysis of
ultrasound tissue compression (Kerkar, 2006; Kanter,
2009; Krishna, 2006). This survey focuses different
approaches for breast, prostate (Cohenford, 1997),
cervix (Suetens, 2009) and ovarian (Polascik, 2008)
cancer detection and classification method for
Ultrasound images. Usually this involves four stages.
Our work concentrates on the breast cancer. The
basic steps in image processing steps are carried out
in the processing of the cancer image. While preprocessing, the images can be an ultra sound image,
HDR image or raw image and dicom images and
they need to be pre-processed. Ultrasound images are
affected by noise such as speckle noise (Fung, 1997;
Muhammad, 2009), impulse noise (Eltonsy, 2007),
multiplicative noise (Eltonsy, 2007). To suppress the
noise some filtering techniques like (Eltonsy, 2007;
Eltonsy,2006; Guo, 2005) wavelet domain
techniques and de-speckling methods are used.
Image Segmentation: This method sub categories the
image into number of small portions and differentiate
the object from the background. In Feature extraction
and selection, we extract some features both from
normal tissue and abnormal cancer tissue. So
extracting and selecting some essential features is
important for classification. After the feature
extraction we classify the tissue and identify the
normal and abnormal tissue.
Related Work:
A number of methods have been used to classify
and/or detect abnormalities in medical images, such
as wavelets, fractal theory and statistical methods.
Generally, in these methods use features are
extracted using image-processing techniques. A
CAD system in which features are extracted using
image processing techniques is developed in for
detecting abnormalities. A large variety of techniques
have been applied to the problem of mass detection,
but most follow a two-step scheme. First, one or
more features are computed for each pixel, after
which each pixel is classified and the suspicious
pixels are grouped into a number of suspicious
regions. In the second step, these regions are
classified as normal or abnormal regions, based on
regional features like size, shape or contrast. The
signs can indicate the presence of a lesion: a
radiating pattern of spicules or a central mass. To
detect the whole range from architectural distortions
to circumscribed masses, both signs must be
detected. The central mass is a more or less circular
bright region with a diameter between 5mm and 5
cm. Convolution of the image with a zero mean filter
with a positive centre and a negative surrounding
area was used by a number of research groups to
detect the mass, for example with the Laplacian of
the Gaussian (LoG) or a Difference of Gaussians
filter (DoG). Several works have been done to
develop computer aided breast cancer detection and
diagnosis tools et al. Proposed a technique in which
presence of concentric layers surrounding a focal
area with suspicious morphological characteristics
and low relative incidence in the breast region is used
for malignancy detection. Results were reported with
92, 88 and 81% sensitivity and developed a similar
procedure to detect both masses and architectural
distortion by finding those points which are covered
by concentric layers of image activity. Also used
Hausdorff fractal dimension and an SVM classifier to
differentiate ROIs having architectural distortion.
72.5% accuracy was obtained with a set of 40 ROIs,
of which 21 had normal tissue patterns and 19 had
architectural distortion.
Proposed work:
Thresholding method is one of the most
common methods for the segmentation of images
into two or more clusters It is a simple and popular
method for digital image processing that can be
divided into three different types: global thresholding
method, local thresholding method
and optimal
thresholding method. In the former, global
thresholding methods are used to determinate a
threshold for the entire image. An approach for
breast cancer diagnosis in digital mammogram using
wavelet transform is proposed in .Though there are
different methods with rig let and curvelet our
proposed method gives a better results compared
with other methods. As the pre-processing of the
input images the fuzzy enhancement is carried out.
After the pre processing procedure with the
mammogram image, the
clustering process is
carried out to perform the segmentation .After
segmentation the classification is carried out with
the feature extracted based on the energy with
wavelet. The methodologies are explained in the
below sections .The experimental results are given in
section 4.
.
Fig. 1: Image Decomposition with Wavelet Transformation.
A. Preprocessing:
primary step is preprocessing .There are different pre
In Every image processing technique the
processing steps with different methods but we use
94
S. Venkata Lakshmi and Dr. J. Janet, 2015
Advances in Natural and Applied Sciences, 9(7) June 2015, Pages: 92-99
the fuzzy method to give novelty to the proposed
work. Fuzzy enhancement methods can be used to
enhance masses and micro-calcifications, upon
enhancing mammograms containing masses and
micro calcification and compared between fuzzy
enhancement techniques and histogram equalization
the fig 2 gives the steps involved in preprocessing.
The Fig (4) gives the result of enhancement of the
input image is comparatively good on comparing
with the other preprocessing methods.
Gray scale conversion
preprocessing
Input Image
Histogram Equalization
Fuzzy contrast Enhancement
Fig. 2: Steps involved in Pre processing.
B. Segmentation:
Segmentation is to divide an image into region
corresponding to objects. The fig 3 represents the
various segmentation methods available for the
classification of breast cancer. These objects include
more information than pixels. Indeed, interpretation
of images based on objects gives more illustrative
and meaningful information’s than the information
based on individual pixel interpretation. The
principle aim of image segmentation is to apply a
specific treatment or to interpret the image content.
In this paper, a proposed algorithm for image
segmentation based on PSO is used to automatically
determine the threshold values in multilevel
thresholding problem. The PSO algorithm is based
on swarm behavior of birds where particles (in our
case: pixels), fly through the search space using two
simples equations for velocity and position. This
algorithm has been widely used to solve global and
local optimization problems. This algorithm has
undergone several evolutions, we cite the Darwinian
Particle Swarm Optimization (DPSO) and the
fractional calculus based PSO called FODPSO
(Cohenford, 1997). This work mainly focuses on
using a new variant of PSO.
PSO Algorithm:
Initialization:
1. Initialize the position Xi and the velocity of
each
2.
3.
4.
5. For I from 1 to N
6. Update:
7. If fitness(xi)< fitness(xi) then
Up to date of P best: Pi=Xi
Up to date of g best
8.
For I from 1 to M
9.
Up to date of Vi and Xi
10.
end
11.
end
12.
end
end
Region –Based
Segmentation
Breast –
region
Segmentatio
n
Contour –Based
Techniques
Un Supervised
Segmentation
Mammogram
Segmentation
Techniques
Region of
Interest ROI
Segmentatio
n
Techniques
Cluster
Segmentation
Supervised
Segmentation
Pseudo color
Segmentation
Graph
Segmentation

Fig. 3: Over view of Segmentation methods.
Variant
feature
Transformation
95
S. Venkata Lakshmi and Dr. J. Janet, 2015
Advances in Natural and Applied Sciences, 9(7) June 2015, Pages: 92-99
C. Feature Extraction:
Feature extraction is the key for the
classification of benign and malignant patterns in
digital mammograms. The next step after detection
and segmentation of the lesion is the extraction of
features that would depict the class to which it
belongs. Feature extraction methods examine objects
and images to extract the most important features
that represent various classes of objects. Features are
used as inputs to classifiers that assign them to the
class they represent. The extracted features are then
used by a classifier to extract the different regions. In
the proposed system, the training mammogram
image is decomposed by using 2-level DWT with
various directions.
The fig(1)represents the DWT decomposition of
a single image to one level where the HH, HL, LH,
LL. are obtained. Then the energy of each directional
sub-band is used as feature for the corresponding
training image. This feature extraction method is
applied to all the training samples and the features
are stored in the database called as Feature Database
(FD), and this database is used as one of the input to
the classification phase. The energy of each
directional sub-band of the image I is calculated by
using the equation (1)
Energy (
(1)
Where
band and W, H is width and height of the sub-band
respectively.
D. Classification:
In the classification phase, the Support Vector
Machine (SVM) algorithm is used as a classifier.
SVM is a method for classifying objects based on
closest training examples in the feature space. The
selected features are provided as inputs to the
classifier, with the known values. SVM is used as a
classifier. SVMs are a set of related supervised
learning methods that analyze data and recognize
patterns, used for classification and regression
analysis. The standard SVM is a non probabilistic
binary linear classifier which predicts, the given
input, with the two possible classes the input is a
member of. Classification phase executes two
phases. One is to classify mammograms into normal
and abnormal cases and in the abnormal case it will
represent the tissue contains tumor (micro
calcification). Finally, the abnormal mammogram is
classified into malignant or benign in the second
stage. In this classification stage, SVM classifier in
every phase is trained at specific number of training
set in each category. Receiver Operating
Characteristics (ROC) analysis is presented to
evaluate the classification performance of the
textural features extracted by texture-analysis
method.
is the pixel value of the eth sub-
Fig. 4: (a) Fuzzy Contrasted Image and Histogram Representation(b)Segmentation Algorithm using
FODPSO(c) Wavelet Decomposition.
Exprimental Results:
The above figure shows the results of
enhancement carried out using the proposed
method. Though there are many methods present for
enhancing the input, the fuzzy contrast enhancement
algorithm helps to minimize the computational time.
The first figure represents the fuzzy enhancement
algorithm and the second image represents the
wavelet decomposition of the input with feature
extraction is as shown. The result in the figure5
shows the output with their histogram results. The
extracted features for test data and the trained data
are as shown in the table 1.
96
S. Venkata Lakshmi and Dr. J. Janet, 2015
Advances in Natural and Applied Sciences, 9(7) June 2015, Pages: 92-99
Fig. 5: Experimental results of segmented micro calcification.
The various parameters which is used for the
evaluation is tabulated as in the table 2. The
techniques can be employed to evaluate the
enhancement and segmentation algorithms such as,
sensitivity and specificity, receiver operating
characteristic (ROC) and free-ROC (FROC). The
results of diagnosing a mammogram can be classified
as follows with the example of diagonosation. The
results of the extracted features are tabulated in table
1.
Table I: Extracted Features of Test and Train Dataset
Features
C1
C2
C3
N1
N2
Mean
88.27
41.44
56.01
23.73
65.74
Medium
80
31
72
4
69
Maximum
242
139
132
120
136
Minimum
12
0
0
0
0
Range
230
139
132
120
136
S. D
75.26
31.59
33.85
30.81
29.11
M.A.S.D
68
12
19
4
9
Mn.A..S.D.
66.31
25.69
29.2
26.54
20.63
L1. N
4.384e+06
2.894e+06
3.65e+06
1.47e+06
5.164e+06
L2 .N
2.585e+04
1.377e+04
1.671e+04
9677
2.015e+04
Max norm
242
139
132
120
136
C1, C2,C3 the cancer images N1,N2 are the normal images taken for testing and training .Standard Deviation is abbreviated S.D,
M.A.S.D. is Maximum absolute standard deviation ,Mn.A.S.D. is Minimum absolute standard diviation,L1.N, L2.N N is norm values.
No vertical lines in table. Statements that serve as captions for the entire table do not need footnote letters.
a
Gaussian units are the same as cgs emu for magnetostatics; Mx = maxwell, G = gauss, Oe = oersted; Wb = weber, V = volt, s = second,
T = tesla, m = meter, A = ampere, J = joule, kg = kilogram, H = henry.
Table 2: Segmentation Evaluation Methods.
Parameters
spatial overlap
jaccard coeff
accuracy
sensitivity
specificity
false positive
false negative
ground truth
PSO
0.9649
0.9298
56.0310
93.2918
99.3140
123
2422
36105
The various parameters which is used for the
evaluation is tabulated as in the table 2. The
techniques can be employed to evaluate the
enhancement and segmentation algorithms such as,
sensitivity and specificity, receiver operating
characteristic (ROC) and free-ROC (FROC). The
results of diagnosing a mammogram can be classified
as follows with the example of diagonosation.
DPSO
0.9635
0.9280
56.4077
93.2558
99.0116
179
2435
36105
FODPSO
0.9649
0.9282
57.2386
93.2835
99.0229
180
2425
36105
True positive (TP): sick people correctly diagnosed
sick.
False positive (FP): healthy people incorrectly
diagnosed sick.
True negative (TN): healthy people correctly
diagnosed healthy.
False negative (FN):
sick people incorrectly
diagnosed healthy.
97
S. Venkata Lakshmi and Dr. J. Janet, 2015
Advances in Natural and Applied Sciences, 9(7) June 2015, Pages: 92-99
Where sensitivity (true positive fraction (TPF))
is the fraction of sick people who are correctly
diagnosed as positive and specificity (true negative
fraction (TNF)) is the fraction of healthy people who
are correctly diagnosed as negative. These measures
specify the accuracy of the system for identifying the
actual positive and actual negative patients
respectively. However, the false-positive-fraction
(FPF) and the false-negative fraction represent the
frequencies of incorrect diagnoses, therefore: Falsenegative-fraction = 1 - TPF, and False positive-
1
0.95
0.9
fraction = 1 - TNF. There is an interrelationship
between these measures which makes it important
and meaningful to specify a single pair, either
sensitivity and specificity, or TPF and FPF, but it is
inadequate to use only one measure alone. Equation
2 shows sensitivity measure, while equation 3shows
specificity measure
Sensitivity = number of TP / (number of TP +
number of FN)
(2)
Specificity = number
of TN / (number of TN + number of FN)
(3)
JACCARD …
Fig. 6: The comparative results of various methods.
Conclusion:
This paper presented a comprehensive
classification and evaluation of mammogram
enhancement and segmentation algorithms. The
enhancement techniques were categorized into four
distinct techniques. From the review, it is obvious
that the results produced from the fuzzy enhancement
techniques are best suited for
enhancing both
masses and micro-calcifications. On the other hand,
the mammogram segmentation techniques were
categorized into two distinct categories, and the
regions of interest segmentation techniques were
further categorized into subcategories. under the
unsupervised, single view segmentation techniques
are better suited for the segmentation of masses and
micro-calcifications.
REFERENCES
ACS, What is cervical cancer?” 2011. American
Cancer
Society,
http://www.cancer.org/Cancer/CervicalCancer/Detail
ed- Guide/cervical-cancer-what-is-cervical-cancer.
Bangma, C.H., S. Roemeling and F.H. Schröder,
2007. “Overdiagnosisandovertreatment of early
detected prostate cancer,” World J. Urol., 25: 3–9.
Bluth, E.I. and M.J. Siegel, 2007. Ultrasound: A
Practical Approach toClinical Problems, 2nd ed.
New York, NY: Thieme Medical Publishers.
Cheng, H., X. Shi, R. Mil, L. Hu, X. cai, H. Du,
2006. Approaches for automated detection and
classification of masses in mammogram. Pattern
Recognition, 39(4):646-668.
Cheng, H.D., Juan Shan, 2010. we Ju,
YanhuiGuo, Ling Zhang,” Automated breast cancer
detection and classification using US images: a
survey”, Department of computers cience, Utah state
University, Logan, UT 84322 USA,school of
mathematics and system science, Shandong
university, China, Pattern Recognition, 43: 299–317.
Cohenford, M.A., T.A. Godwin, F. Cahn, P.
Bhandare, T.A. Caputo and B. Rigas, 1997. “Infrared
spectroscopy of normal and abnormal cervical
smears: evaluation by principal componentanalysis,”
Gynecologic Oncology, 66(1): 59–65.
Costantini, M., P. Belli, R. Lombardi, G.
Franceschini, A. Mule, L. Bonomo, 2006.
Characterization of solid breast masses use of the
sonographic breast imaging reporting and data
system lexicon, journal of ultrasound in medicine,
25(5): 649-659.
Eltonsy, N., G. Tourassi, A. Elmaghraby, 2006.
“Investigatingperformance of a morphology-based
CAD scheme in detecting architectural distortion in
screening mammograms.” Proc. 20th Int. Congr.
Exhib. Comput. Assist. Radiol. Surg.
Eltonsy, N., G. Tourassi, A. Elmaghraby, 2007.
A concentric morphology model for the detection of
masses in mammography. IEEE Trans. Med. Imag,
26(6): 880-889.
Ferlay, J., P. Autier, M. Boniol, M. Heanue, M.
Colombet and P. Boyle, 2007. “Estimates of the
cancer incidence andmortality in Europe in 2006,”
Ann. Oncol, 28(3): 581–592.
Ferlay, J., P. Autier, M. Boniol, M. Heanue, M.
Colombet and P. Boyle, 2007. “Estimates of the
cancer incidence and mortality in Europe in 2006
”Ann. Oncol, 18: 581–592.
Fung, M.F.K., M. Senterman, P. Eid, W. Faught,
N.Z. Mikhael and P.T.T. Wong, 1997. “Comparison
of fourier-transform infrared spectroscopic screening
of exfoliated cervical cells with standard
98
S. Venkata Lakshmi and Dr. J. Janet, 2015
Advances in Natural and Applied Sciences, 9(7) June 2015, Pages: 92-99
papanicolaou screening,” Gynecologic Oncology,
66(1): 10–15.
Guo, Q., J. Shao, V. Ruiz, 2005. “Investigation
of support vector machine for the detection of
architectural distortion in mammographic images.” J.
Phys. Conf. Ser., 15: 88–94.
Huang, YL., 2005. Computer aided diagnosis
applied to 3D US of solid breast nodules by using
principal component abnalysis and image retrival, in:
Proceding of the 2005 IEEE engineering in medicine
and biology 27 annual conference, pp: 1802-1805.
Huang, Z., J. Mo, W. Zheng, J. Low, J. Ng and
A. Ilancheran, 2008. “Combining near-infrared
autofluorescence and raman spectroscopy improves
the in vivo detection of cervical precancer,” in
Proceedings of the Conference on Quantum
Electronics and Laser Science Conference on Lasers
and Electro-Optics (CLEO/QELS ’08), San Jose,
Calif, USA, May 2008.
Hwang, KH., JG. Lee, JH. Kim, Hj. Lee, KS.
Om, M. Yoon, W. Choe, Computer aided diagnostic
of breast mass on ultrasonagraphy and seniti
mammography in: proceding of 7th
Imaginis Corporation, 2008. Breast Cancer
Cases/Deaths
per
Year
(U.S.
andWorld);
2008[online]
available:
http://imaginis.com/breasthealth/statistics.asp#1.
INCa. Breast Cancer Incidence in Brazil; 2008
[online]
available:
http://www.inca.gov.br/estimativa/2008/finalversion.
pdf.
Jemal, A., R. Siegel, E. Word, Y. Hao, J. Xu, T.
Murray, MJ. Thun, 2008. Cancer statistics 2008,
CA:A cancer journal for clinicians, 58: 71-96.
Jemal, R., E. Siegel, T. Ward, J. Murray, Xu and
M. Thun, 2007. “Cancerstatistics, 2007,” CA Cancer
J. Clin, 57(1): 43–66.
Jemila Rose R., S. Allwin, 2013. “Computerized
Cancer Detection and Classification Using
Ultrasound Images: A Survey Computerized Cancer
Detection and Classification Using Ultrasound
Images: A Survey International Journal of
Engineering Research and Development e-ISSN:
2278-067X, p-ISSN : 2278-800X, www.ijerd.com,
5(7): 36-47.
Jesneck, J., J. lo, J. Baker, 2007. Breast
masslesion:Computer aided diagnosis models with
mammographic and sonographic descriptors.
Radiology, 244(2):392-398.
Kalaivani Narayanan, S., R.S.D. Wahidabanu,
2009. “A view on despeckling in ultrasound
imaging” , Int.J.signalProc.ImageProcess.Pattern
Recogn, 2(3).
Kanter, E.M., E. Vargis, S. Majumder, 2009.
“Application of raman spectroscopy for cervical
dysplasia diagnosis,” Journal of Biophotonics, 2(12): 81–90.
Kerkar, R.A. and Y.V. Kulkarni, 2006.
“Screening for cervical cancer: an overview,”
Obstetrics and Gynecology of India, 56(2): 115–122.
Krishna, C.M., N.B. Prathima, R. Malini, 2006.
“Raman spectroscopy studies for diagnosis of
cancers in human uterine cervix,” Vibrational
Spectroscopy, 41(1): 136–141.
Linden, R.A., E.J. Trabulsi, F. Forsberg, P.R.
Gittens, L.G. Gomella and E.J. Halpern, 2007.
“Contrast
enhanced
ultrasound
flash
replenishmentmethod for directed prostate biopsies,”
J. Urol., 178:2354–2358.
Loch, T., 2007. “Urologic imaging for localized
prostate cancer in 2007,”World J. Urol., 25: 121–
129.
Muhammad, AJ., H. Ayyaz, MM. Anwar, Ch.
Asmatullah, 2009. “Fuzzy Entropy And Morphology
Based Fully Automated Segmentation Of Lungs
From Ct Scan Images.” Int. J. Innov. Comput., Inf.
Control, 5(12B): 4993-5002.
NCCC, “Cervical cancer,” http://www.nccconline.org/index .php/cervicalcancer.
Noble, J.A., D. Boukerroui, 2006. ”ultrasound
image segmentation: a survey, IEEE trans.Med.Imag,
25(8): 987-1010.
Park, S., S.R. Aglyamov and S.Y. Emelianov,
2007. “Elasticity imaging usingconventional and
high-frame
rate
ultrasound
imaging:
Experimentalstudy,”
IEEE
Trans.
Ultrason.,
Ferroelectr. Freq. Control, 54(11): 2246–2256, Nov.
Pelzer, A., J. Bektic, A.P. Berger, L. Pallwein,
E.J. Halpern, W. Horninger, G. Bartsch and F.
Frauscher, 2005. “Prostate cancer detection inmen
with prostate specific antigen 4 to 10 mg/mL using a
combinedapproach of contrast enhanced color
doppler targeted and systematicbiopsy,” J. Urol.,
173: 1926–1929.
Polascik, T.J. and V. Mouraviev, 2008. “Focal
therapy for prostate cancer,”Curr.Opin.Urol, 18:
269–274.
Sahiner, B., 2007. Malignant and benign breast
masses on 3D US volumetric images:Effect of
computer aided diagnosis on radiologist accuracy.
Radiology, 242(3): 716-724.
Salcudean, S.E., D. French, S. Bachmann, R.
Zahiri-Azar, X. Wen and W. Morris, 2006.
“Viscoelasticity modeling of the prostate region
using vibroelastography,”in Proc. Med. Image
Comput. Comput.-Assist. Intervention (Lecture
Notes in Computer Science), 4190: 389–396.
Schröder, F.H., J. Hugosson, M.J. Roobol, T.L.J.
Tammela, S. Ciatto, V. Nelen, M. Kwiatkowski, M.
Lujan, H. Lilja, M. Zappa, L.J. Denis, F. Recker, A.
Berenguer, L. Määttänen, C.H. Bangma, G. Aus, A.
Villers, X. Rebillard, T. van der Kwast, B.G.
Blijenberg, S.M. Moss, H.J. de Koning and A.
Auvinen, 2009. “Screening and prostate-cancer
mortalityin a randomized European study,” N. Engl.
J. Med., 360(13): 1320–1328.
Seitz, M., A. Shukla-Dave, A. Bjartell, K.
Touijer, A. Sciarra, P.J. Bastian, C. Stief, H. Hricak
and
A.
Graser,
“Functional
magnetic
resonanceimaging in prostate cancer,” Eur. Urol, 55:
99
S. Venkata Lakshmi and Dr. J. Janet, 2015
Advances in Natural and Applied Sciences, 9(7) June 2015, Pages: 92-99
801–814.
Suetens, P., 2009. Fundamentals of Medical
Imaging, 2nd ed. New York,NY: Cambridge
University Press.
Uof Pittsburgh Medical Centre, 2009. Prostate
cancer:Transrectal ultrasound. [Online]. Available:
http://www.upmccancercenters.com/cancer/prostate/
biopsyultrasound.html
Wink, M.H., J.J. M.C.H. de la Rosette, C.A.
Grimbergen and H. Wijkstra, 2007. “Transrectal
contrast enhanced ultrasound for diagnosis
ofprostatecancer,”World J. Urol., 25: 367–373.
Wong, P.T.T., R.K. Wong, T.A. Caputo, T.A.
Godwin and B. Rigas, “Infrared spectroscopy of
exfoliatedhumancervical cells: evidence of extensive
structural
changes
during
carcinogenesis,”
Proceedings of the National Academy of Sciences of
the United States of America, 88(24): 10988–10992.
Yang, F., J. Suri and A. Fenster, 2006.
“Segmentation of prostate from3-D ultrasound
volumes using shape and intensity priors in level
setframework,” in 28th Annual EMBS Int. Conf.,
New York City, USA, Aug, pp: 2341–2344.
Zhi, H., B. Ou, B. Luo, X. Feng, Y. Wen, H.
Yang, 2007. Comparison of ultra sound
elastrography, mammography and sonagraphy in the
diaghonosis of solid breast lesions. Journal of
ultrasound in medicine, 26 (6): 807-815.