International Journal of Engineering and Advanced Technology (IJEAT)

ISSN: 2249 8958, Volume-2, Issue-3, February 2013

A Review of Thyroid Disorder Detection,

Segmentation and Classification on Medical Images
Sheeja Agustin A, S. Suresh Babu
Another difficulty may arise due to the lack of sufficient
training samples. In the field of Image analysis segmentation
of medical images is a challenging problem due to poor
resolution and weak contrast of medical images. Medical
image segmentation means Segmentation of known
anatomic structures from medical images. Now a days
Medical image analysis has played more and more
important role in many clinical procedures and in detecting
different types of human diseases. Now a days most of the
peoples have thyroid diseases.
For diagnosing thyroid diseases, Ultrasound (US) and
Computerized Tomography (CT) are two of the most
popular imaging modalities. US imaging is inexpensive,
non-invasive and easy to use. US images are often adopted
due to their cost-effectiveness and portability in smaller
hospitals. The thyroid is well suited to ultrasound study
because of its superficial location, size and echogenicity
.Computer-Aided Diagnosis (CAD) of Thyroid Ultrasound
is necessary in order to delineating nodules, classifying
benign/malignant and estimating the volumes of thyroid
tissues to increase reliability and reduce invasive operations
such as biopsy and Fine Needle Aspiration (FNA).

Abstract- Thyroid is a small butterfly shaped gland located in

the front of the neck just below the Adams apple. Thyroid is one
of the endocrine gland, which produces hormones that help the
body to control metabolism. Different thyroid disorders include
Hyperthyroidism, Hypothyroidism, goiter, and thyroid nodules
(benign/malignant). Ultrasound imaging is most commonly used
to detect and classify abnormalities of the thyroid gland. Other
modalities (CT/MRI) are also used. Computer aided diagnosis
(CAD) help radiologists and doctors to increase the diagnosis
accuracy, reduce biopsy ratio and save their time and effort.
Numerous researches have been carried out in thyroid medical
images and that are utilized for the diagnosis process. In this
paper, a broad review on the researches that are developed for
the thyroid diagnosis using medical images is presented along
with the classification. A short description about thyroid, thyroid
diseases and thyroid diagnosis are also presented.
Key terms:-Medical imaging, Thyroid, Thyroid disorders,
Segmentation, Classification.

I.

INTRODUCTION

Image processing is any form of signal processing for

which the input is an image such as photograph or video
frame, the output of image processing may be either an
image or parameters related to the image. Image processing
usually refers to digital image processing. Digital image
processing is the use of computer algorithms to perform
image processing on digital images. Different Imaging
technologies are:-Radiology, Magnetic resonance imaging
(MRI), Nuclear medicine, Photo acoustic imaging, Breast
thermograph, Tomography and ultrasound imaging. Medical
imaging is the technique and process used to create images
of the human body for clinical purpose or medical science
including the study of normal anatomy and physiology.
Image segmentation is the process of partitioning an
image into multiple segment or set of pixels used to locate
object and boundaries. Each of the pixels in a region is
similar with respect to some characteristics such as color,
intensity or texture. One of the common applications of
segmentation is in medical image analysis for clinical
diagnosis that has an important role in terms of quality and
quantity. Medical image segmentation methods generally
have restrictions because medical images have very similar
gray level and texture among the interested objects.
Therefore, significant segmentation error may occur.

Fig1: Normal Thyroid US image

In digital image classification the conventional statistical

approaches for image classification use only the gray values.
Different advanced techniques in image classification like
Artificial Neural Networks (ANN), Support Vector
Machines (SVM), Fuzzy measures, Genetic Algorithms
(GA), Fuzzy support Vector Machines (FSVM) and Genetic
Algorithms with Neural Networks are being developed for
image classification. The use of textural features in ANN
helps to resolve misclassification. SVM was found
competitive with the best available machine learning
algorithms in classifying high-dimensional data sets. ANN
is a parallel distributed processor that has a natural
tendency for storing experiential knowledge. Image
classification using neural networks is done by texture
feature extraction and then applying the back propagation
algorithm.
The rest of this paper is organized as follows. In Section II
describes about different disorders and symptoms, section
III describes review on thyroid disorders, section IV

Manuscript received on February, 2013.

Sheeja Agustin, Research Scholar, Computer science & Engineering,
Noorul Islam Centre for Higher Education, Noorul Islam University,
Kumaracoil,TamilNadu, India.
Dr.S.Suresh Babu, Electronics & Communication Engineering,
Professor, TKM College of Engineering, Kollam, India.

88

A Review of Thyroid Disorder Detection, Segmentation and Classification on Medical Images

describes about Reviewed results. Finally, the conclusions
as well as future directions are summarized in Section V.

cancers. Thyroglobulin (TG) for well differentiated papillary

or follicular adenocarcinoma, and the rare medullary thyroid
cancer has calcitonin as the marker. To differentiate between
different types of hypothyroidism, a specific test may be
used. Thyroid-releasing hormone (TRH) is injected into the
body through a vein. This hormone is naturally secreted by
the hypothalamus and stimulates the pituitary gland. The
pituitary responds by releasing thyroid -stimulating hormone
(TSH ).
Nodules of the thyroid may or may not be cancer. Medical
ultrasonography can help determine their nature because
some of the characteristics of benign and malignant nodules
differ. The main characteristics of a thyroid nodule on high
frequency thyroid ultrasound are as follows[12]

II. DIFFERENT DISORDERS IN THYROID

A. Thyroid Diseases
Thyroid gland produces chemicals (hormones) that help
the body to control metabolism. Thyroid hormone is
normally produced in response to another hormone released
by the pituitary gland. There are four main types of thyroid
disease [11]-hyperthyroidism ( too much thyroid hormone ),
hypothyroidism (too little thyroid hormone ), benign
(noncancerous)
thyroid
disease
and
thyroid
cancer(malignant).
B. Thyroid Disease Symptoms
The signs and symptoms of hypothyroidism includes
fatigue, mental fogginess and forgetfulness, feeling
excessively cold, constipation, dry skin, fluid retention, non
specific aches and stiffness in muscles and joints, excessive
or prolonged menstrual bleeding (menorrhagia), and
depression. Hyperthyroidism is suggested by a number of
signs and symptoms. People with mild hyperthyroidism or
those older than 70 years of age usually experience no
symptoms. Common symptoms of hyperthyroidism includes
excessive sweating, heat intolerance, increased bowel
movements, tremor (usually a fine shake), nervousness;
agitation, rapid heart rate, weight loss, fatigue, decreased
concentration, and irregular and scant menstrual flow.
The two most important thyroid hormones are thyroxine
(T4) and triidothyronine(T3), representing 99.9% and 0.1%
[14] of thyroid hormones respectively. The hormone with
the most biological power is actually T3. Once released
from the thyroid gland into the blood, a large amount of T4
is converted to T3 - the active hormone that affects the
metabolism of cells throughout our body. The thyroid itself
is regulated by another gland located in the brain, called the
pituitary. The pituitary is regulated in part by the thyroid by
another gland called the hypothalamus. The hypothalamus
releases a hormone called thyrotropin releasing hormone
(TRH), which sends a signal to the pituitary to release
thyroid stimulating hormone (TSH). In turn, TSH sends a
signal to the thyroid to release thyroid hormones. If over
activity of any of these three glands occurs, an excessive
amount of thyroid hormones can be produced, thereby
resulting in hyperthyroidism. Similarly, if under activity of
any of these glands occurs, a deficiency of thyroid hormones
can result, causing hypothyroidism. The hormone chain is
represented as Hypothalamus TRH, Pituitary- TSH,
Thyroid - T4 and T3. There is another hormone that is
produced by the thyroid called calcitonin. It is a hormone
that contributes to the regulation of calcium and helps to
lower calcium levels in the blood. Excess calcium in the
blood is referred to as hypercalcemia.
Anatomical problems [12] include goiter, endemic goiter,
diffuse goiter, multinodular goiter, Lingual thyroid and
thyroglossal duct cyst. Tumors include thyroid adenoma,
thyroid cancer-Papillary, Follicular, Medullary, Anaplastic
and lymphomas and metastasis. Defiency in adults causes
Myxedema.

Malignant characteristics
irregular border
hypoechoic (less echogenic
than the surrounding tissue)
microcalcifications
taller than wide shape on
transverse study
significant intranodular blood
flow by power Doppler

Benign characteristics
smooth borders
hyperechoic

comet tail" artifact as sound

waves
bounce
off
intranodular colloid

Table 1: benign/malignant characteristics

Ultrasonography is not always able to separate benign
from malignant nodules with complete certainty. In
suspicious cases, a tissue sample is often obtained by biopsy
for microscopic examination. Thyroid scintigraphy, imaging
of the thyroid with the aid of radioactive iodine, usually
iodine -123, is performed in the nuclear medicine
department of a hospital or clinic. Radioiodine collects in
the thyroid gland before being excreted in the urine. While
in the thyroid the radioactive emissions can be detected by a
camera, producing a rough image of the shape (a radiodine
scan) and tissue activity (a radioiodine uptake) of the thyroid
gland.
A normal radioiodine scan shows even uptake and activity
throughout the gland. Irregularity can reflect an abnormally
shaped or abnormally located gland, or it can indicate that a
portion of the gland is overactive or under active, different
from the rest.. In contrast, finding that a substantial section
of the thyroid is inactive ("cold") may indicate an area of
non-functioning tissue such as thyroid cancer.
The amount of radioactivity can be counted as an
indicator of the metabolic activity of the gland. A normal
quantitation of radioiodine uptake demonstrates that about 8
to 35% of the administered dose can be detected in the
thyroid 24 hours later. Over activity or under activity of the
gland as may occur with hypothyroidism or hyperthyroidism
is usually reflected in decreased or increased radioiodine
uptake. Different patterns may occur with different causes of
hypo- or hyperthyroidism.
A medical biopsy refers to the obtaining of a tissue
sample for examination under the microscope or other
testing, usually to distinguish cancer from noncancerous
conditions. Thyroid tissue may be obtained for biopsy by
fine needle aspiration or by surgery.. The accuracy of the
diagnosis depends on obtaining tissue from all of the
suspicious areas of an abnormal thyroid gland.
Thyroid disease in children and in autism has many
overlapping signs and symptoms. These include but not

C. Diagnosis
The measurement of thyroid hormone levels is often used
by doctors as a screening test. If TSH is abnormal,
decreased levels of thyroid hormones T4 and T3 may be
present. There are two cancer markers for thyroid derived

89

International Journal of Engineering and Advanced Technology (IJEAT)

ISSN: 2249 8958, Volume-2, Issue-3, February 2013
limited to Feeding problems, Prolonged jaundice, Poor Algorithms (GA), Fuzzy support Vector Machines (FSVM)
muscletone, Gastrointestinal abnormalities Constipation, and Genetic Algorithms with Neural Networks are being
Sleep disturbances, Developmental delays, Trouble holding developed for image classification. Artificial neural
up head, Protrusion of belly Hyperactivity, Lethargy, Lack networks can handle non-convex decisions. The use of
of play and interaction with others, Dry skin, Poor Hair textural features in ANN helps to resolve misclassification.
Growth/Bald spots, Pale complexion, Frequent infections, SVM was found competitive with the best available
Cold intolerance, Cold extremities, Weight gain, Difficulty machine learning algorithms in classifying high-dimensional
gaining weight Allergies ,Bed wetting, Poor bone data sets. Fuzzy measures show the detection of textures by
development, Fear, Anxiety depression, decreased ability to analyzing the image by stochastic properties M.Seetha et
concentrate, Speech delay, Fading of the personalitys color al. [21] have proposed Fuzzy support vector machines
and vivacity, Progressive loss of interest and initiative and (FSVM) to overcome the n-class problem in Support Vector
iodine deficiency .
Machines.
Won-Jin Moon et al. [22] have evaluate the diagnostic
accuracy of ultrasonographic (US) criteria for the depiction
III. REVIEW ON THYROID DISORDERS
Preeti Aggarwal et al[15] proposed an automatic of benign and malignant thyroid nodules by using tissue
segmentation method and two tools: Analyze 10.0 and diagnosis as the reference standard. They concluded that
Mazda for segmentation of thyroid US images.They Shape, margin, echogenicity, and presence of calcification
provides a summary of all the results obtained either by are helpful criteria for the discrimination of malignant from
automatic tools as well as by applying specific algorithm benign nodules The diagnostic accuracy of US criteria
(automatic) segmentation on both lung CT as well as on dependent on tumor size.
Rajendra Acharya et al. [23] had presented a Computer
thyroid US .
Aided
Diagnosis (CAD) based technique for automatic
Nikita Singh et al. [16] have proposed classification using
classification
of benign and malignant thyroid lesions in 3D
SVM, KNN and Bayesian and also provide information
contrast-enhanced
ultrasound images. Discrete Wavelet
about segmentation and classification methods that are very
Transform
(DWT)
and
texture based features were extracted
important for medical image processing. They provide
from
the
thyroid
images.
The resulting feature vectors were
features of malignant nodules and done a comparative study.
used
to
train
and
test
three
different classifiers: K-Nearest
The results shows that SVM gives much better accuracy
Neighbor (K-NN), Probabilistic Neural Network (PNN), and
than KNN and Bayesian
Eystratios G. et al.[17] have proposed a computer-aided Decision Tree (DeTr) using tenfold cross validation
diagnosis (CAD) system prototype, named TND (Thyroid technique. The results are shown in table.
Skrgio Okida et al. [24] had presented the SADNT
Nodule Detector), for the detection of nodular tissue in
System
to Support the Diagnosis of Thyroid Nodules (TNs).
ultrasound (US) thyroid images and videos acquired during
TNs
are
common clinical problems, involving a great
thyroid US examinations.
number
of
diseases with different biological behaviors, from
Eleonora Horvath et al. [18] have proposed a study using
a
goiter
to
a cancer. The SADNT is an expert system (ES)
the TIRADS (Thyroid Imaging Reporting and Data System)
integrated
to
a database, that uses production rules, certainty
and The following categories were established:
factors
(CFs)
and qualitative reasoning for the
TIRADS 1: normal thyroid gland.
computational
representation
of knowledge. The SADNT
TIRADS 2: benign conditions (0% malignancy).
integrates
information
referring
to the patient's clinical data
TIRADS 3: probably
benign
nodules
(_5%
(history
and
physical
examination)
and the complementary
malignancy).
exams
(ultrasound,
laboratory
tests,
cytology and thyroid
TIRADS 4: suspicious nodules (580% malignancy
scan)
aiming
to
support
the
clinician
in the diagnosis of
rate). A subdivision into 4a (malignancy
TNs.
This
system
is
composed
by
two
ES: an ES that is
between 5 and 10%) and 4b (malignancy
based
on
clinical
data
which
supplies
the
initial diagnosis in
between 10 and 80%) was optional.
terms
of
the
chance
of
the
nodule
to
be
malign;
a second ES
TIRADS 5: probably malignant nodules (malignancy
integrates
the
results
of
the
complementary
exams
to the
_80%).
diagnostic
provided
by
the
first
ES,
to
give
a
final
diagnosis.
TIRADS 6: category
included
biopsy
proven
The SADNT was evaluated by the expert of the domain and
malignant nodules.
Yutaka Hatakeyama et al. [19] have proposed an the results of the simulations met the expectations.
Nodular thyroid disease is a frequent occurrence in clinical
algorithm for the size of the thyroid gland in
practice
and it is associated with increased risk of thyroid
ultrasonography (US) images on the basis of the position of
cancer
and
hyperfunction. D.E. Maroulis et al. [25] have
the neighboring regions in order to objectively evaluate
proposed
a
novel method for computer-aided detection of
target organs for medical screening and secondary use. The
thyroid
nodules
in ultrasound (US) images. The proposed
proposed algorithm first extracts the target image, with the
method
is
based
on
a level-set image segmentation approach
short-axis view of the thyroid gland and setting information
that
takes
into
account
the inhomogeneity of the US images.
on the scale. Second, it calculates the edge of the carotid
This
novel
method
was
experimentally evaluated using US
region and air passage that are near the target organ. Finally,
images
acquired
from
35
patients. The results show that the
the length is measured on the basis of these edges.
proposed
method
achieves
more accurate delineation of the
Mary C. Frates et al. [20] have focus on which nodules
thyroid
nodules
in
the
US
images and faster convergence
should be subjected to US-guided fine needle aspiration and
than
other
relevant
methods.
which thyroid nodules need not be subjected to fine-needle
Konstantinos K. Delibasis et al. [26] have proposed the
aspiration. They provide US features associated with thyroid
utilization
of a computer-aided diagnosis system based on a
cancer. Different advanced techniques in image
supervised
classification algorithm from the artificial
classification like Artificial Neural Networks (ANN),
immune
systems
to assist the task of thyroid malignancy
Support Vector Machines (SVM), Fuzzy measures, Genetic
diagnosis. The core of the proposed algorithm is the so-

90

A Review of Thyroid Disorder Detection, Segmentation and Classification on Medical Images

called BoxCells, which are defined as parallelepipeds in the
feature space. Properly defined operators act on the
BoxCells in order to convert them into individual,
elementary classifiers. The proposed algorithm is applied on
FNA data from 2016 subjects with verified diagnosis and
has exhibited average specificity higher than 99%, 90%
sensitivity, and 98.5% accuracy.
Elpiniki I. Papageorgiou et al. [27] have prposed a new
methodology using the softcomputing technique of FCMs,
accompanied with datamining and knowledge extraction
techniques. The knowledge extraction techniques with FCM
based decision support system develop a method namely
FCM-TDM. FCM-TDM provides 89.80% accuracy for
thyroid diseases.
Thyroid nodules are solid or cystic lumps formed in the
thyroid gland and may be caused by a variety of thyroid
disorders. Michalis A. Savelonas et al. [28] had presented a
novel active contour model for precise delineation of thyroid
nodules of various shapes according to their echogenicity
and texture, as displayed in ultrasoundd (US) images. The
proposed model, named joint echogenicitytexture (JET), is
based on a modified MumfordShah functional that, in
addition to regional image intensity, incorporates statistical
texture information encoded by feature distributions.The
JET model requires only a rough region of interest within
the thyroid gland as input and automatically proceeds with
precise delineation of the nodules, revealing their shape and
size. The performance of the JET model was validated on a
range of US images displaying hypoechoic and isoechoic
nodules of various shapes. The results shows that the JET
model provides precise delineations of thyroid nodules as
compared to ground truth delineations obtained by experts
and copes with the limitations of the previous thyroid US
delineation approaches as it is capable of delineating thyroid
nodules regardless of their echogenicity or shape.
A fine needle aspiration (FNA) biopsy is the standard
procedure of choice for differentiating between benign and
malignant thyroid nodules. FNA is invasive, costly and
uncomfortable for patients. Jieming Ma et al. [29] have
presented a non-invasive and automatic approach for
differentiating benign and malignant thyroid nodules with
ultrasound elastography based on support vector machines
(SVM) with biased penalties.They used the elastography
data of 98 thyroid nodules (16 malignant and 82 benign)
from 92 subjects previously acquired with a clinical
ultrasound machine, Hitachi Hi Vision 5500. Also they
conducted the leave-one-out cross-validation (LOOCV) in
evaluating the performance of their classification method.
Their goal was to obtain the maximum geometric mean
(MGM) of sensitivity and specificity. The results show that
the method was able to get MGM of 90.1% with the
sensitivity of 93.8% and the specificity of 86.6%.
D.Selvathi et al. [30] have developed an automatic system
that classifies the thyroid images and segments the thyroid
gland using machine learning algorithms. The classifiers
such as SVM, ELM are used. The features such as mean,
variance, Coefficient of Local Variation Feature, Histogram
Feature, NMSID Feature, and Homogeneity are extracted
and these features are used to train the classifiers such as
ELM and SVM. The results are compared with the ground
truth images obtained from the radiologist and the
performance measure such as accuracy is evaluated. It is
observed that the segmentation using ELM is better than
SVM classifier.
Chuan-Yu Chang et al. [1] have presented a progressive
learning vector quantization neural network (PLVQNN)

combined with a reprocessing procedure is proposed for

automatic thyroid segmentation and volume estimation
using computerized tomography (CT) images. The
preprocessing procedure is used to extract the region of
interest (ROI) of thyroid glands and exclude non-thyroid
glands based on thyroid anatomy. The PLVQNN contains
several learning vector quantization neural networks
(LVQNNs), each responsible for segmenting one slice of a
thyroid CT image. The training of the PLVQNN is
conducted starting from the LVQNN of most reliable
(middle) slices, where the thyroid has the largest region. The
training then propagates upwards and downwards to
adjacent LVQNNs using the results of the middle slice as
the initialization values and constraints. Experimental results
show that the proposed method can effectively segment
thyroid glands and estimate thyroid volume.
Jianrui Ding et al. [2] have proposed a new effective,
accurate, and quantitative metric using computer aided
diagnosis (CAD) techniques . The statistical features and
texture features are extracted from the lesion region on the
elastogram. The important and reliable features are selected
by using Minimum-Redundancy-Maximum-Relevance
(mRMR) algorithm. The selected features were input to the
SVM to classify the thyroid nodules. The experiment results
confirm that the method is more accurate and robust than
color score and strain ratio.
Chuan-Yu Chang et al. [3] have proposed the radial basis
function neural network to classify blocks of the thyroid
gland. The integral region is acquired by applying a
specific-region-growing method to potential points of
interest. The parameters for evaluating the thyroid volume
are estimated using a particle swarm optimization algorithm.
The experiment results show that the proposed method can
be used to segment the thyroid gland region and to estimate
thyroid volume directly from US images
Chuan-Yu Chang et al. [4] had presented five support
vector machines (SVM) to select the significant textural
features and to classify the nodular lesions of thyroid.
Experimental results showed the proposed method classifies
the thyroid nodules correctly and efficiently. The
comparison results demonstrated that the capability of
feature selection of the proposed method was similar to the
sequential floating forward selection (SFFS) method.
However, the proposed method is faster than the SFFS
method.
Deepika Koundal1 et al. [5] have provided a summary of
existing automatic tools available to formulate the disease
diagnosis part easier as well efficient. Different performance
evaluation metrics are studied, and the future developments
and trends are also investigated.
Singh1 and Mrs Alka Jindal [6] had uses the groups
Benign (non-cancerous) and Malignant (cancerous) Thyroid
Nodules images . The texture feature method like GLCM
are very useful for classifying texture of images and these
features are used to train the classifiers such as SVM, KNN
and Bayesian. The experimental result shows the
performance of the classifiers and shows the best predictive
value and positively identify the percentage of the noncancerous or cancerous people and shows the best
performance accuracy using the SVM classifier as compare
to the KNN and Bayesian classifier. Use the segmentation
based algorithm localized based active contour (region
based) method is basically to segment the local area of the
images and to segment the nodule which is give the
information of which type nodule exist benign and
malignant. The proper procedure of segmentation algorithm

91

IV. REVIEWED RESULTS

[4]

The segmentation accuracy of different researches are

shown below. Depending on the thyroid diseases the
accuracy is varied.

[5]

[6]

reference
1

Segmentation method
PLVQNN

Accuracy %
98.34

92

ELM

SVM

Decision

PNN

KNN

FCM-TDM

SVM

RBFNN

PLVQNN

accuracy

International Journal of Engineering and Advanced Technology (IJEAT)

ISSN: 2249 8958, Volume-2, Issue-3, February 2013
the thyroid benign image are segmented with Quantitative
3
RBFNN
96.54
measurement of classification accuracy is calculated in term
6
SVM
84.62
of true positive (TP), true negative (TN), false positive (FP),
20
FCM-TDM
89.8
false negative (FN) with respect to the ground truth.
23
KNN
98.9
Edgar Gabriel et al. [7] had presented their experiences
23
PNN
97.8
with two parallel versions of a code used for texture-based
23
Decision tree
96.9
segmentation of thyroid FNAC images, a critical first step in
29
SVM with biased
90.1
realizing a fully automated CAD solution. An MPI version
penalities
of the code is developed to exploit distributed memory
30
ELM
93.56
compute resources such as PC clusters.
M. Savelonas et al. [8] have proposed a novel active
Table 2: comparison of thyroid segmentation accuracies of
contour model named Variable Background Active Contour
different researches
model is proposed and applied for the detection of thyroid
A graphical representation is also shown below.
nodules in ultrasound images. The new model offers edge
independency, no need for smoothing, ability for topological
Accuracy %
changes and it is more accurate when compared to the
Active Contour Without Edges model. improved accuracy
100
is achieved by introducing as background a limited image
subset which appropriately changes shape to reduce the
95
effects of background inhomogeneity.
90
Accuracy %
Nasrul Humaimi Mahmood and Akmal [9]
had
85
presented a simple guide of determine the thyroid lobes in
80
the thyroid ultrasound image using a MATLAB. The image
75
undergoes the contrast enhancement to suppress speckle.
The enhancement image is used for further processing of
segmentation the thyroid region by local region-based active
contour. The thyroid region is segmented into two parts,
which are right and left with the active contour method
1 3 6 20 23 23 23 29 30
separately. This is accordingly to the thyroid have two
classifiers
lobes; right lobe and left lobe. Thyroid ultrasound image of
transverse view is used in this study. Therefore, the
Fig 2: comparison of thyroid segmentation accuracies of
measurements only involve the width, depth and area of the
different researches
thyroid region. The result of thyroid measurement is
successfully calculated in pixel unit. The measurement is
V. CONCLUSION
converted in centimetre (cm) unit. The proposed method is
benefited to enhance the image and segmentation the thyroid
Different researchers have been developed various
lobe. It shows that from five samples, different people have techniques to process thyroid medical images. Such
different size of thyroid, especially in measurement of the techniques helps the radiologists and doctors to efficiently
width, depth and area.
diagnosing the nodules in thyroid images and lead the
Dimitris E. Maroulis [10] had presented a computer- reduction of false diagnosis thyroid diseases. Different
aided approach for nodule delineation in thyroid ultrasound classifiers improve the performance of classifying thyroid
(US) images. The developed algorithm is based on a novel nodules as cancer/non- cancerous. This review work
active contour model, named variable background active provides knowledge about thyroid, thyroid diseases, thyroid
contour (VBAC), and incorporates the advantages of the diagnosis and also gives a summary of thyroid segmentation
level set region-based active contour without edges (ACWE) accuracies of different researches in various thyroid
model, offering noise robustness and the ability to delineate diseases. We think that this work will direct the future
multiple nodules. Unlike the classic active contour models researchers to develop the better methods for the diagnosis
that are sensitive in the presence of intensity of thyroid diseases.
inhomogeneities, the proposed VBACmodel considers
information of variable background regions. VBAC has
REFERENCES
been evaluated on synthetic images, as well as on real
[1] Chuan-Yu Changet al. A Neural Network for Thyroid Segmentation
thyroid US images. From the quantification of the results,
and Volume Estimation in CT Images, November 2011 IEEE
two major impacts have been derived- higher average
COMPUTATIONAL INTELLIGENCE MAGAZINE, pp 43-55
accuracy in the delineation of hypoechoic thyroid nodules, [2] Jianrui Ding et al. A novel quantitative measurement for thyroid
cancer detection based on elastography, 4th International Congress
which exceeds 91% and
faster convergence when
on Image and Signal Processing 2011
compared with the ACWE model.
[3] Chuan-Yu Chang et al. Thyroid Segmentation and Volume
Estimation in Ultrasound Images, IEEE TRANSACTIONS ON
BIOMEDICAL ENGINEERING, VOL. 57, NO. 6, JUNE 2010
Chuan-Yu Chang et al. Classification of the Thyroid Nodules Using
Support Vector Machines, International Joint Conference on Neural
Networks (IJCNN) 2008
Deepika Koundal1 et al., Computer-Aided Diagnosis of Thyroid
Nodule: A Review,International Journal of Computer Science &
Engineering Survey (IJCSES) Vol.3, No.4, August 2012
Singh1 and Mrs Alka Jindal, A segmentation method and
classification of diagnosis for thyroid nodules, . IOSR Journal of
Computer Engineering (IOSRJCE) ISSN : 2278-0661 Volume 1,
Issue 6 , PP 22-27,2012

A Review of Thyroid Disorder Detection, Segmentation and Classification on Medical Images

Edgar Gabriel et al. Towards high performance cell segmentation in
multispectral fine needle aspiration cytology of thyroid lesions,
computer methods and programs in biomedicine, pp 1-10
[8] M. Savelonas et al. Variable Background Active Contour Model for
Automatic Detection of Thyroid Nodules in Ultrasound Images,
Dept. of Informatics and Telecommunication
[9] Nasrul Humaimi Mahmood and Akmal, Segmentation and Area
Measurement for Thyroid Ultrasound Image, International Journal of
Scientific & Engineering Research Volume 2, Issue 12, December
2011
[10] Dimitris E. Maroulis, Variable Background Active Contour Model
for Computer-Aided Delineation of Nodules in Thyroid Ultrasound
Images,
IEEE
TRANSACTIONS
ON
INFORMATION
TECHNOLOGY IN BIOMEDICINE, VOL. 11, NO. 5,
SEPTEMBER 2007
[11] http;//www.nlm.nih.gov/medlinep/us/ency/article/000371.htm
[12] http://www.wikipedia.com
[13] http://thyroid.about.com/library/weekly/qathyroid 101-e.htm
[14] http://ww.medicinenet.com/script/main/at.asp
[15.] Preeti Aggarwal et al. Role of Segmentation in Medical Imaging: A
Comparative Study International Journal of Computer Applications
(0975 8887) Volume 29 No.1, September 2011
[16] Nikita Singh et al. A Segmentation Method and Comparison of
Classification Methods for Thyroid Ultrasound Images International
Journal of Computer Applications (0975 8887) Volume 50 No.11,
July 2012
[17] Eystratios G.et al. ND: A Thyroid Nodule Detection System for
Analysis of Ultrasound Images and Videos Springer
Science+Business Media, LLC 2010
[18] Eleonora Horvath et al. An Ultrasonogram Reporting System for
Thyroid Nodules Stratifying Cancer Risk for Clinical Management, J
Clin Endocrinol Metab 90: 17481751, 2009
[19] Yutaka Hatakeyama et al. Algorithm for Estimation of Thyroid Gland
Size in Ultrasonography Images for Extracting Abnormal Thyroid in
Medical Practice Journal of Advanced Computational Intelligence
and Intelligent Informatics Vol.16 No.1, 2012
[20] Mary C. Frates et al. Management of Thyroid Nodules Detected at
US: Society of Radiologists in Ultrasound Consensus Conference
Statement Management of Thyroid Nodules Detected at US Volume
237 , Number 3
[21] M.Seetha et al. ARTIFICIAL NEURAL NETWORKS AND
OTHER METHODS OF IMAGE CLASSIFICATION, Journal of
Theoretical and Applied Information Technology,2008
[22] Won-Jin Moon et al. Benign and Malignant Thyroid Nodules: US
DifferentiationMulticenter Retrospective Study Radiology:
Volume 247: Number 3June 2008
[23] U Rajendra Acharya et al. Automated Benign & Malignant Thyroid
Lesion Characterization and Classification in 3D Contrast-Enhanced
Ultrasound34th Annual International Conference of the IEEE EMBS
San Diego, California USA, 28 August - 1 September, 2012
[24] Skrgio Okida et al. Expert System to Support the Diagnosis of
Thyroid Nodules, Proceedings of the 22"' Annual EMBS
international Conference, July 23-28,2000, Chicago IL.
[25] D.E. Maroulis et al. Computer-Aided Thyroid Nodule Detection in
Ultrasound Images, Proceedings of the 18th IEEE Symposium on
Computer-Based Medical Systems (CBMS05) 2005 IEEE
[26] Konstantinos K. Delibasis et al. Computer-Aided Diagnosis of
Thyroid Malignancy Using an Artificial Immune System
Classification
Algorithm,
IEEE
TRANSACTIONS
ON
INFORMATION TECHNOLOGY IN BIOMEDICINE, VOL. 13,
NO. 5, SEPTEMBER 2009
[27] Elpiniki I. Papageorgiou et al. Fuzzy cognitive map based decision
support system for thyroid diagnosis management, IEEE
International Conference on Fuzzy Systems(FUZZ ) 2008.
[28] Michalis A. Savelonas et al Active Contours Guided by
Echogenicity and Texture for Delineation of Thyroid Nodulesin
Ultrasound Images, IEEE TRANSACTIONS ON INFORMATION
TECHNOLOGY IN BIOMEDICINE, VOL. 13, NO. 4, JULY 2009
[29] Jieming Ma et al. Differential Diagnosis of Thyroid Nodules with
Ultrasound Elastography based on Support Vector Machines , IEEE
International Ultrasonics Symposium Proceedings 2010
[30] D.Selvathi et al. Thyroid Classification and Segmentation in
Ultrasound Images Using Machine Learning Algorithms
Proceedings of International Conference on Signal Processing,
Communication, Computing and Networking Technologies
(ICSCCN) 2011
[7]

93