Beruflich Dokumente
Kultur Dokumente
ABSTRACT
Biomedical image processing has experienced dramatic expansion, and has been an interdisciplinary research
field attracting expertise from applied mathematics, computer sciences, engineering, statistics, physics, biology
and medicine. The advent of computer aided technologies image processing techniques has become increasingly
important in a wide variety of medical applications. The medical image plays an important role in clinical
diagnosis and therapy of doctor and teaching and researching etc. Medical imaging is often thought of as a way
to represent anatomical structures of the body with the help of X-ray computed tomography and magnetic
resonance imaging. But often it is more useful for physiologic function rather than anatomy. With the growth of
computer and image technology medical imaging has greatly influenced medical field. As the quality of medical
imaging affects diagnosis the medical image processing has become a hotspot and the clinical applications
wanting to store and retrieve images for future purpose needs some convenient process to store those images in
details. Intervention between the protection of useful diagnostic information and noise suppression must be
treasured in medical images. Image denoising is an applicable issue found in diverse image processing and
computer vision problems. The principal objectives of this paper are to provide an Introduction to basic concepts
and techniques for medical image processing and to promote interests for further study and research in medical
imaging processing. This paper gives the details about the methods of biomedical image processing and after
that it also describe about medical imaging modalities. Some of the medical imaging modalities are described in
this paper like X-ray imaging, CT, MRI, and ultrasound. The optical modalities like endoscopy, photography
and microscopy are also more important in this field. The following steps of image analysis are explained in this
paper, feature extraction, segmentation, classification, quantitative measurements and interpretation.
Keywords: — Magnetic Resonance Imaging, Pre-processing, Segmentation, Image Segmentation
(SPECT), positron emission tomography (PET)
I. INTRODUCTION and fundus photography. The cost-efficient
health-care and effective diagnosis of medicine
Medical image processing is an area of increasing were revolutionized by the medical images in all
interest. It includes a wide range of methods and major disease areas. Scientists and physicians use
techniques, starting with the acquisition of images less invasive techniques to understand potential
using specialized devices (for example, CT life-saving information from these medical
devices), image enhancement and analysis, to 3D images. The accuracy of these medical images
model reconstruction from 2D images. Thus, the depends on the image acquisition where the multi-
research in this field represents a point of interest pronged functions like, performing reconstruction,
for both doctors and engineers, in their attempt to post-processing of the image data [2], storing the
improve medical techniques, with computer scans and controlling imaging hardware were
assistance, in order to obtain more accurate results assisted by computers. Due to the sufficient
in treating the patients.[1] Medical Image volume of data, the technology driven MIP
Processing (MIP) is a novel research discipline research should also be an application oriented
based on the application of computer vision field. In recent years, effective work in the
methods to data sets acquired via medical imaging medical image processing has been initiated [3].
modalities such as ultrasound, magnetic resonance In-order to assist the physician many applications
imaging (MRI), computed tomography (CT), are being developed that can interpret an image
single photon emission computed tomography for diagnosing possible abnormalities. The recent
research community have received considerable quality images of soft tissues of the human
attention to develop automatic screening systems body. Normally an MRI scanner consists of two
for eye related pathologies. The patients, powerful magnets. Initially the first magnet
especially in rural areas were affected from causes the body’s water molecules to align in
obtaining regular inspection due to the shortage of one direction which are normally scattered. The
ophthalmologists and high examination cost [4]. alignment of the hydrogen atom is then altered
Hence, an automated retinal image processing by turning on and off the second magnetic field.
system that can screen the irregularities at the When turning off the magnetic field these
initial stage is a need of the hour. hydrogen atoms will switch back to its original
state. These changes are detected by the scanner
1.1 Trends in Medical Image Processing: to create a detailed cross-sectional image. This
image helps to visualize internal structures such
There are many types of imaging modalities as joints, muscles, and other structures in an
used for medical diagnosis of various diseases, effective way[7].
and are described in the following subsections. 1.1.4 Fluoroscopy:
1.1.1 Computed Tomography (CT): The body structures are visualized in a real time
CT images are produced using X-ray photons environment using an imaging modality termed
with digital reconstruction technique [4]. The fluoroscopy [8]. These real time images are
X-ray beam is generated by the X-ray tube that generated by continuously emitting and
passes through the patient and is captured by capturing the X-ray beam on the screen. To
the detectors. This captured beam is differentiate various structures, high density
reconstructed to produce a three dimensional contrast agents may be used which helps in the
image. This CT scanner uses different assessment of anatomy and its functions of this
reconstruction algorithm at different angle to structure.
produce an image. The most commonly used 1.2.5 Ophthalmic Imaging
clinical applications in CT studies are CT In addition to other medical imaging modalities,
Brain, Pelvic CT, CT angiography and Cardiac there exist several ophthalmic imaging methods
CT. These are used to locate abnormalities in like fluorescein angiography (FA) and optical
the body such as tumours, abscesses, abnormal coherence tomography (OCT), which plays a
blood vessels etc [5]. significant role in the diagnosis of several eye
1.1.2 Ultrasound: related pathologies[9]. In FA a fluorescent dye
Ultrasound image provides the cross-sectional is infused over the patient’s eye in a systematic
images of the body, which are constructed flow, and the fluorescent properties are
using high-frequency sound waves. In activated by passing light in a specific
ultrasound procedure there is no radiation wavelength. The higher light energy is
exposure and hence it is a very safe process stimulated by the transmitted light and the
with minimum known adverse. The sound fundus photograph captures it when the
waves are emitted by the transducer at some molecules returned to the original state. The
frequency and the returning echoes are captured captured image is termed as angiogram that
at frequencies dependent on the tissues through provides valuable information to the
which the waves traverse [6]. The returned ophthalmologists about the progression of the
sound wave is digitized that appears as dots or disease related to the retinal vasculature.
echoes on the screen. These echoes are used in Sodium fluorescein angiography and indocya-
cardiovascular ultrasound to visualize the nine green angiography are the other types used
peripheral vascular structures of heart, and in diagnosis[10]. In OCT, the ocular tissue
abdominal ultrasound to assess the anatomy of structure is further stretched into a two
the liver and gallbladder. dimensional cross-sectional image, where
1.1.3 Magnetic Resonance Imaging (MRI): propagation direction of the light and the
This imaging technique is commonly used in perpendicular spatial direction were its
radiology which uses magnetic radiation to dimensions. The transparent vascular tissue is
assess detailed internal structures. It is mainly scanned with laser light beam and the reflected
used in medical diagnosis to produce high light is collected by the OCT and the
propagation time delay is measured by the low medical images with higher resolution and
coherence interferometer. Numerous axial scans quality, removing noise in these digital images
and numerous OCT images combines to form remains one of the major challenges in the
an optical coherence tomography image where study of medical imaging. Image de-noising
it is mainly used in the diagnosis of macular still remains a challenge for researchers because
swelling and to analyse its type. Fundus noise removal introduces artifacts and causes
imaging is highly preferred for diagnosing blurring of the images. Image processing has a
retinal diseases due to its low cost, portability wide spectrum of applications and can be
and less risk to the patients. The eye disorders measured into different domains where images
such as glaucoma, diabetic retinopathy (DR) are used [12].
and age-related macular degeneration (ARMD),
can be diagnosed for follow-up management of a) Brain Tumor Detection: Brain is the
patients from the digital fundus images. The most important and vital organ of the
detection of abnormalities such as drusen can human body. The control and
be predicted using effective image processing coordination of all the other vital
algorithms, to detect the severity of the disease. structures is carried out by the brain. A
Appropriate algorithms are needed to make an brain tumor is a very serious-type
effective screening process. It can be among all life threatening diseases
implemented using suitable software in which which is increasing drastically among
the accurate screening of retinal images for the humans. A brain tumor is a mass of
various anatomical and pathological feature tissue formed by an unregulated
extraction and analysis is possible. growth of the abnormal cells in the
brain. CT and MRI are the best
technologies currently being used for
diagnosing brain tumor. It helps the
doctors to classify the tumor from
either least aggressive (benign) or
most aggressive (malignant)[12].
b) Craniofacial Fractures : Craniofacial
Fractures Imaging is one of the most
important tools for orthodontists to
evaluate and record size and form of
craniofacial structures [13].
Craniofacial fractures are faced
frequently. Major causes are vehicle
accidents & sports related injuries.
During the treatment of craniofacial
Fig1.1: Human Eye fractures, 3-D imaging is an imaging
[11] technology ultrasound, CT-scan which
. provides high quality of image to the
1.3 APPLICATIONS OF MEDICAL orthopedics to diagnoses.
IMAGE ANALYSIS: c) Breast Cancer Detection
Breast cancer is most commonly
The arrival of digital medical imaging diagnosed cancer worldwide. Pal et al.
technologies has reformed modern medicine. [14] state that, annually more than a
Extensive use of digital imaging in medicine million women have breast cancer and
today, the quality of digital medical images has 400 thousand of those cases lead to
becomes an important issue. To achieve the death. This fatal disease affects not
best possible diagnosis it is most important that only developing countries but also
medical images should be sharp, clear, noise developed countries. In order to find
free and artifacts . With the advancement in the the cure it is necessary to quickly
technologies continuing for gaining digital diagnose the disease accurately and
applications with great societal value. For In [27] this paper discusses about various
example, urban monitoring, fire detection or image processing techniques and tools which
flood prediction can have a great impact on are available for identification of printing
economical and environmental issues. From a technologies. Printing technology identification
machine learning and signal/image processing and associated problems in document forensics
point of view, all the applications are tackled have been projected as challenges in image
under specific formalisms, such as processing application. Various image
classification and clustering, regression and processing approaches based on textures,
function approximation, image coding, spatial variation, HSV colour space, spatial
restoration and enhancement, source un mixing, correlation, and feature based on histogram and
data fusion or feature selection and extraction. some of the pattern recognition methods, like
In[24] authors presented a hybrid approach for gray level co-occurrence matrix, roughness of
classification of brain tissues, such as white the text, perimeter of edge are highlighted.
matter (WM), gray matter (GM), cerebral spinal In[28],The research depicts a comprehensive
fluid (CSF), background (BKG) and tumour survey of the technical achievements in the
tissues, in MRI based on Artificial neural research area of image retrieval, especially
Networks and Genetic Algorith (GA). In the content-based image retrieval, an area that has
proposed technique after seeded region growing been so active and prosperous in the past few
segmentation and connected component years. Based on the state-of-the-art technology
labelling, two texture and intensity feature sets available now and the demand from real-world
are taken as input. First set consists statistical applications, open research issues are identified
features such as: Entropy, Kurtosis, Skewness, and future promising research directions are
Mean, Energy, Momentum, correlation and suggested.
second feature sets, are derived from wavelet In[29], This research work introduce a family
transformation. of first-order multidimensional ordinary
In [25] this paper worked on states the risk of differential equations (ODE’s) with
having cancer in dense breast on mammogram discontinuous right-hand sides and demonstrate
is higher compared to those who have less. This their applicability in image processing. An
is due to presence of glandular cells in the equation belonging to this family is an inverse
breast parenchyma. Therefore, it is important diffusion everywhere except at local extrema,
for radiologists to pay more attention to denser where some stabilization is introduced.
breasts in order to detect abnormalities. The In[30],The study shows that an original coding
suggested method works based on the statistical scheme is introduced to take advantage of the
parameters including kurtosis, skewness, two-dimensional structural information of
median and mean. The system evaluated 180 images within the genetic algorithm framework.
mammogram images and it was found to be Results are presented showing that this new
92.8% accurate with a strong correlation technique outperforms classical optimization
between the system and radiologists’ estimation methods for the optimization of 32 × 32 and
(K=0.87, p=0.0001). 128 x 128 holograms.
In[26] The study states that Breast-tissue In[31],This research primarily focuses on the
microarrays facilitate the survey of very large predictive technology of identifying the state of
numbers of tumours but their scoring by tumors in the breast tissues. In breast cancer
pathologists is time consuming, typically highly diagnosis, patients are forced to undergo a
quantised and not without error. This paper series of biopsies just to identify and confirm
proposes a computationally efficient approach on the state of tumour, as whether malignant or
that approximates the density of colour and benign. In this research however, an algorithm
local invariant features by clusters in the feature will be developed using MATLAB Image
space, and characterises each spot by a Processing Toolbox to indentify the state of a
frequency histogram of nearest cluster centres. tumor solely based on ultrasound images.
Spots are classified into four main types based Ultrasound images of breast tumors are
on their histograms. imported into MATLAB and are passed
through a set of filters to remove background
noise. Next, the filtered images are run through Based on the values of these features of a
a set of edge detection algorithms which digital mammogram, the authors have made an
identifies and defines the region of interest. The attempt to classify the breast tissue in to four
results are compared with the actual biopsy basic categories like fatty, uncompressed fatty,
results from the IIUM Breast Cancer Research dense and high density.
Institute, Kuantan and all the analyzed results In [37], In this study they state that Breast
matched the biopsy results. cancer produces a high rate of mortality
In[32],authors study shows that Mammography worldwide. Early diagnosis is essential for
is at present one of the available method for treatment, however it is difficult to analyse high
early detection of masses or abnormalities density breast tissues. Computer aided
which is related to breast cancer. The most diagnosis systems have been proposed to
common abnormalities that may indicate breast classify the density of mammograms, having as
cancer are masses and calcifications. Breast a major challenge to define the features that
cancer is diagnosed at advanced stages with the better represent the images to be classified.
help of the digital mammogram image. In this In[38],Authors present a neural networkbased
paper, a method has been developed to make a method for automatic classification of magnetic
supporting tool. resonance images (MRI) of brain under three
In[33],Their study is based on a novel, fully categories of normal, lesion benign, and
automatic, adaptive, robust procedure for brain malignant. The proposed technique consists of
tissue classification from 3D magnetic six subsequent stages; namely, pre-processing,
resonance head images (MRI) is described in seeded region growing segmentation, connected
this paper. Starting from a set of samples component labelling (CCL), feature extraction,
generated from prior tissue probability maps in feature Dimension Reduction, and
a standard, brain-based coordinate system, the classification.
method first reduces the fraction of incorrectly In[39],This paper states that A novel, fully
labelled samples in this set by using a minimum automatic, adaptive, robust procedure for brain
spanning tree graphtheoretic approach. tissue classification from 3D magnetic
In[34],The Authors study suggested, Digital resonance head images (MRI) is described in
mammography systems allow manipulation of this paper. The classification procedure is
fine differences in image contrast by means of robust against variability in the image quality
image processing algorithms. Different display through a non-parametric implementation: no
algorithms have advantages and disadvantages assumptions are made about the tissue intensity
for the specific tasks required in breast distributions.
imaging—diagnosis and screening. In [40], the study on this project is based on
In[35], Their study suggested, Mammography machine learning techniques to perform tumour
is the primary tool for the early detection of vs. normal tissue classification using gene
breast cancer. The depiction of fine micro- expression microarray data, which was proven
calcifications and subtle soft-tissue masses on to be useful for early-stage cancer diagnosis and
mammograms is key to the detection of early cancer subtype identification. We compare the
breast cancer. results of both supervised learning (k-
In[36],In this paper authors have made an nearestneighbours, SVMs, boosting) and
attempt to classify the breast tissue based on the unsupervised learning (k-means clustering,
intensity level of histogram of a mammogram, hierarchical clustering) routines on three
Statistical features of a mammogram are datasets.
extracted using simple image processing In[41],The research is based on analysis of
techniques. The proposed scheme uses texture composition of concrete mixture digital image.
models to capture the mammographic The concrete mixture is combination of various
appearance within the breast. The statistical Cement, Air-voids and Aggregates. To analyze
features extracted are the mean, standard the compositions of the concrete mixture, the
deviation, smoothness, third moment, X-ray CT images are used. Digital image
uniformity and entropy which signify the processing algorithm is applied to analyze the
important texture features of breast tissue. obtained image. Using this Digital image
processing algorithm the obtained image is (CRT) display is a nonlinear function (the
processed and filtered. The resultant image is gamma function) of the input video signal
compared with the X-ray CT image and the voltage. In most analogy video systems,
measured and predicted mixture proportions are compensation for this nonlinear transfer
compared to analyze the absolute errors. The function is implemented in the camera
threshold range T1 and T2 were found for amplifiers.
aggregates, cement materials and air-voids. In[47], They discussed on Kilo- to Terabyte
In[42],authors stateed that an automated digital challenges regarding (i) medical image
image processing (DIP) algorithm called management and image data mining, (ii) bio
Volumetric based Global Minima (VGM) imaging, (iii) virtual reality in medical
thresholding algorithm for processing asphalt visualizations and (iv) neuro-imaging. Due to
concrete (AC) X-ray computed tomography the increasing amount of data, image processing
(CT) images. It utilizes known volumetric and visualization algorithms have to be
properties of AC mixtures as the main criterion adjusted. Scalable algorithms and advanced
for establishing the air-mastic and mastic- parallelization techniques using graphical
aggregate gray scale boundary thresholds. processing units have been developed. They are
Several DIP techniques were utilized to summarized in this paper.
characterize the AC microstructure. In [48], They described some central
In[43], authors state that many problems in mathematical problems in medical imaging.
imaging are actually inverse problems. One The subject has been undergoing rapid changes
reason for this is that conditions and parameters driven by better hardware and software. Much
of the physical processes underlying the actual of the software is based on novel methods
image acquisition are usually not known. utilizing geometric partial differential equations
Material parameters in geological structures as in conjunction with standard signal/image
unknown parameters for the simulation of processing techniques as well as computer
seismic wave propagation with sparse graphics
measurement on the surface, or temporal facilitating man/machine interactions.
changes in movie sequences given by intensity In [49], Their study describes CT scanner,
changes or moving image edges and resulting Ultrasound and Magnetic Resonance Imaging
from deformation, growth and transport took over x-ray imaging by making the doctors
processes with unknown fluxes. to look at the body's elusive third dimension.
In[44],This paper concludes that Image With the CT Scanner, body's interior can be
registration is the fundamental task used to bared with ease and the diseased areas can be
match two or more partially overlapping images identified without causing either discomfort or
taken. It is a fundamental image processing pain to the patient.
technique and is very useful in integrating In [50], They studied on quantitatively assess
information from different sensors, finding caries changes of teeth by using digital analysis.
changes in images taken at different times, They studied on the digital images of stained
inferring three-dimensional information from sections of crowns of teeth were acquired with
stereo images, and recognizing model-based a computer-assisted light microscope. In these
objects. images, they found spots representing the main
In[45],authors found that In digital image and total demineralization of enamel were
forensics, it is generally accepted that segmented to determine their area. The area of
intentional manipulations of the image content total demineralization was significantly
are most critical and hence numerous forensic different between premolars with sealed
methods focus on the detection of such fissures and unprotected premolars as indicated
`malicious' post-processing. The researcher by the Mann–Whitney test.
presented a simple yet effective technique to In [51], Their study is on Classification of
detect median filtering in digital images a dental caries is important for the diagnosis and
widely used de-noising and smoothing operator. treatment planning of the dental disease, which
In [46], This paper concludes that the has been affecting a very large population
luminance emitted from a cathode ray rube throughout the globe. It is also helpful for
conducting detailed study and investigations multiscale analysis functions have been used
about the nature of the dental disease. Dental for compact signal and image representations in
caries are, clearly visible in the x-ray changes denoising, compression and feature detection
and it can be detected from the caries lesion processing problems for about twenty years.
present in the radiographs. Numerous research works have proven that
In [52], They state in her study on medical space-frequency and space-scale expansions
image researches for brain tumor detection are with this family of analysis functions provided
attaining more curiosity since the augmented a very efficient framework for signal or image
need for efficient and objective evaluation of data. In [57], Authors studied on
large amounts of data. Medically, tumours are “Morphological Image sequence processing.”
also known as neoplasms, which are an The study presented morphological multi-scale
abnormal mass of tissue resulting from method for image sequence processing, which
uncontrolled proliferation or division of cells results in a truly coupled spatio-temporal
happening in the human body. anisotropic diffusion. The aim of the method is
In [53], authors carried out their study on “x- not to smooth the level-sets of single frames but
ray clinical medical image”. The study suggests to de-noise the whole sequence while retaining
that the use of LoG filter for contrast geometric features such as spatial edges and
improvement process in place of methods highly accelerated motions. This is obtained by
discussed in. An adaptive technique is an anisotropic spatio-temporal level-set
suggested to improve the contrast quality of evolution, where the additional artificial time
dental Xray image using the Laplacian-of-a- variable serves as the multi-scale parameter.
Gaussian (LoG) Filter. Biologically, LoG Filter In [58], Authors studied on “Real time Image
has a Processing Algorithms for the Detection of
similar profile to the response of the receptive Road and Environmental Conditions” The study
fields in the Human Visual System (HVS). depicts image processing algorithms for the
In [54] studied on “Image processing and recognition of environmental and road
cryptography”. The study focus on conditions from real-time camera images. This
Improvement of pictorial information for study addresses various implementation
betterment of human perception like de- techniques for and considerations of the
blurring, de-noising in several fields such as implementation of algorithms to extrapolate
satellite imaging, medical imaging etc are various features from images taken by
renewed research thrust. Specifically we would stationary traffic cameras.
like to elaborate our experience on the In [59]authors show in their study on
significance of computer vision as one of the “Reconstruction of mechanically Recorded
domains where hardware implemented Sound by Image Processing”. The study shows
algorithms performs far better than those Audio information stored in the undulations of
implemented through software. grooves in a medium such as a phonograph
In [55] authors studied on “One dimensional record may be reconstructed, with no or
image processing for eye tracking using minimal contact, by measuring the groove
derivative dynamic time warping.” The study shape using precision metrology methods and
shows that one dimensional image processing digital image processing. The effects of
for eye tracking approach is presented using a damage, wear, and contamination may be
low cost webcam. It was found that the compensated, in many cases, through image
algorithm is resistant to illumination variation processing and analysis methods.
and is able to track eyes of users with and In [60]authors study depicts “Real Time Image
without glasses in multiple orientations. The Processing based on Reconfigurable Hardware
size and location of the region of interest (ROI), Acceleration”. The research is concerned with a
which contains both eyes, is adaptive. substantial speed
In [56],authors studied on “Wavelets in up of image processing methods on 2D and 3D
Medical Image Processing: De-noising, images making use of modern FPGA (Field
Segmentation, and Registration.” The study programmable Gate Array) technology. The
shows that Wavelet transforms and other applications of this class of methods ranges
from 2D and 3D image de-noising and [1] Himadri Nath Moulick et al,” Wavelets in
restoration, segmentation, morphological shape Medical Image Processing On Hip
recovery and matching to vector field Arthroplasty and De-Noising,
visualization and simulation. Segmentation” , 2278-0661, p- ISSN:
In [61],authors studied on “Paper-based 2278-8727Volume 12, Issue 6 (Jul. - Aug.
Watermark Extraction with Image Processing”. 2013), PP 20-3.
The study presents frameworks for the [2]Cohen, L. D., 1991. On Active Contour
digitisation, localisation, Extraction and Models and Balloons. CVGIP: Image
graphical representation of paper-based Understanding, 53:211-8. Cootes, T. F.,
watermark designs embedded in paper texture. Edwards, G. J., Taylor, C. J., 2001. Active
These operations determine a suitable Appearance Models. IEEE Trans. Pattern
configuration of parameters to allow optimal Analysis and Machine Intelligence, 23:
content processing, in addition to the detection 681-85.
and extraction of chain lines. The second [3]Abramoff, M., Garvin, M. and Sonka, M.:
approach uses a model of the back-lighting 2010, Retinal imaging and image analysis,
effect to locate a watermark in pages of archaic Biomedical Engineering, IEEE Reviews
documents. It removes recto information, and in 3, 169–208.
highlights remaining ‘hidden’ data, and then [4]McInerney, T., Terzopoulos, D., 1995. A
presents a statistical approach to locate Dynamic Finite Element Surface Model
watermarks from a known lexicon. for Segmentation and Tracking in
Multidimensional Medical Images with
IV. CONCLUSION Application to Cardiac 4D Image
Analysis. Computerized Medical Imaging
In this paper, we sketched some of the and Graphics, 19:69-83.
fundamental concepts of medical image [5]Mohamed, N. A., Ahmed, M. N., Farag
processing. It is important to emphasize that A., 1998. Modified Fuzzy C-mean in
none of these problem areas has been Medical Image Segmentation.
satisfactorily solved, and all of the algorithms Proceedings of the 20th Annual
we have described are open to considerable International Conference of the IEEE,
improvement. A relative study is made on 3:1377-80.
various techniques. This paper also deals with [6]. R.A. Gonzalez, R.E. Woods, S.L.
the different methods in image classification as Eddins, ‘DigitalImage Processing Using
i) Image Pre-processing and Segmentation ii) Matlab’. PrenticeHall 2004.
Feature Reduction and iii) Classification. Many [7]. J.K. Udupa, G.T. Herman, ‘3D Imaging
algorithms have been proposed in the literature in Medicine’ CRC Press 2000.
for each image processing stage. After [8]. Dougherty D. ‘Digital Image Processing
evaluation, it is clearly shown the various for Medical Applications’ Cambridge:
methods which can detect the medical disease Cambridge University Press; 2009.
efficiently and provide accurate result.. From [9]. Guy C, Ffytche D. ‘Introduction to the
the previous review, we classify the current Principles of Medical Imaging’. London:
methods and summarize their features. Also Imperial College Press; 2005.
each method has its suitable application fields, [10]. Kim Y, Horii SC (eds). ‘Handbook of
and researchers should combine the application Medical Imaging’. Vol. 3: Display and
background and practical requirements to PACS. Bellingham: SPIE Press; 2000.
design proper algorithms. Accuracy, [11]. Rangayyan RM. ‘Biomedical Image
complexity, efficiency and interactivity of a Analysis’. New York: CRC Press; 2005.
segmentation method should all be the [12] NAVNISH GOEL et al,” Medical Image
considered factors. Processing: A Review” 978-1-4673-9080-
4/16.
REFERENCES [2] Gregor Miller, Sidney Fels and Steve
Oldridge, “A Conceptual Structure for
Computer Vision” IEEE-2011 Canadian
[37] Sheshadri Holalu Seenappa and [50] Elzbieta Kaczmarek et al.,(2003) studied
Arumugam Kandaswamy (2006) studied on “Digital Image Analysis in Dental
on “Breast Tissue classification Using Research Applied for Treatment of
Statistical Feature Extraction of Fissures on Occlusal Surfaces of
Mammogram”. Premolars”.
[38] Silva W. R. and D. Menotti (2008) [51] Reddy M.V.Bramhananda et al., (2012)
studied on “Classification of studied on “Dental X-Ray Image Analysis
Mammograms by the Breast by Using Image Processing Techniques.”
Composition”. [52] Agrawal Richa(2013), states in her study
[39] Mehadi Jafari and Shohresh Kasaei on"A Comparative Study of Various
(2011) worked on “Automatic Brain Brain Tumor Detection Algorithms",
Tissue Detection in MRI Images using [53] Tiwari R.B. (2011) carried out their
seeded Region Growing segmentation and study on “x-ray clinical medical image”.
Neural network classification.”(ISSN [54] Sangeet Saha et al.(2008) studied on
1991-8178). “Image processing and cryptography”.
[40] Chris A. Cocosco et al.,(2003) on their [55] N. Mokhtar et al.(2010) studied on “One
study on “A fully automatic and robust dimensional image processing for eye
brain MRI tissue classification method” tracking using derivative dynamic time
(ISSN:513-527) warping.”
[41] Pei-Chun Chen et al.,(2008) worked on a [56] Yinpeng Jin et al. (2006) studied on
project based on “Cancerous Tissue “Wavelets in Medical Image Processing:
Classification using Microarray Gene De-noising, Segmentation, and
Expression”. Registration.”
[42] Joshi Snehal K.(2014) studied on “On [57] Karol Mikula et al. (2008) studied on
Application of Image Processing: Study “Morphological Image sequence
of Digital Image Processing Techniques processing.”
for Concrete Mixture Images and Its [58] James G. Haran et al. (2004) studied on
Composition”. “Real time Image Processing Algorithms
[43] H.M. Zelelew et al., (2008) in their work for the Detection of Road and
on “Application of Digital Image Environmental Conditions”
Processing Techniques for Asphalt [59] Vitaliy Fadeyev and Carl Haber (2003)
Concrete Mixture Images” shows in their study on “Reconstruction
[44] Michail Kulesh et. al. (2009) in their of mechanically Recorded Sound by
study on “Inverse Problems and Image Processing”.
Parameter Identification” [60] Steffen Klupsch et al. (2002) study
[45] Prabhsharan Kaur et al.,(2013) on their depicts “Real Time Image Processing
study on “Image Registration in Digital based on Reconfigurable Hardware
Image Processing” Acceleration”.
[46] Eli Pell (1999) in the study of “Display [61] Hazem Ali Abd Al Faleh Al Hiary
nonlinearity in digital image processing (2008) studied on “Paper-based
for visual communications” Watermark Extraction with Image
[47] Ingrid Scholl, Till Ach et al.(2011) Processing”.
studied on Challenges of Medical Image
Processing.
[48] Sigurd Angenent et al.,(2006) studied
“Mathematical methods in Medical Image
processing”.
[49] Rao K.M.M. and Rao V.D.P. (2001)
studied on “Medical Image processing”.