1

CHAPTER 1
INTRODUCTION

1.1 CARDIOVASCULAR DISEASES

Cardiovascular diseases (CVDs) remain the biggest cause of deaths
worldwide. More than 17 million people died from CVDs in 2008. More than 3
million of these deaths occurred before the age of 60 and could have largely been
prevented. The percentage of premature deaths from CVDs ranges from 4% in highincome countries to 42% in low-income countries, leading to growing inequalities in
the occurrence and outcome of CVDs between countries and populations. There are
also new dimensions to this alarming situation. Over the past two decades, deaths
from CVDs have been declining in high-income countries, but have increased at an
astonishingly fast rate in low- and middle-income countries (LMIC).

CVDs are largely preventable. Both population wide measures and improved
access to individual health care interventions can result in a major reduction in the
health and socioeconomic burden caused by these diseases and their risk factors.
These interventions, which are evidence based and cost effective, are described as
best buys in the Global Status Report on Non communicable Diseases (NCDs) 2010.
At present, public health services in developing countries are overstretched by
increasing demands to cope with heart disease, stroke, cancer, diabetes and chronic
respiratory disease.

At the same time, health care systems in many LMIC are let down by a model
based on hospital care focused on the treatment of diseases, often centred around
high-technology hospitals that provide extensive treatment for only a small minority
of citizens. Hospitals consume huge amounts of resources, and health ministries may
spend more than half their budgets on treatment services which depend on hospitals.
As a result, a large proportion of people with high cardiovascular risk remain
undiagnosed, and even those diagnosed have insufficient access to treatment at the

primary health-care level; while evidence suggests two-thirds of premature deaths

due to NCDs including CVDs can be prevented by primary prevention and another
one-third by improving health systems to respond more effectively and equitably to
the healthcare needs of people with NCDs. Two new developments have led to this
report at this juncture.

The first development is the growing international awareness that premature

deaths from CVDs and other NCDs reduce productivity, curtails economic growth,
and pose a significant social challenge in most countries. The second development is
that there is now unequivocal evidence that "best buy" interventions to reduce the toll
of premature deaths due to CVDs and other NCDs are workable solutions and that
they are excellent economic investments -- including in the poorest countries. As the
magnitude of CVDs continue to accelerate globally, the pressing need for increased
awareness and for stronger and more focused international and country responses is
increasingly recognized.

1.2 LEFT VENTRICUAR MYOCARDIUM

In the heart, a ventricle is one of two large chambers that collect and
expel blood received from an atrium towards the peripheral beds within the body and
lungs. Ventricles have thicker walls than atria and generate higher blood pressures.
The physiologic load on the ventricles requiring pumping of blood throughout the
body and lungs is much greater than the pressure generated by the atria to fill the
ventricles. Further, the left ventricle has thicker walls than the right because it needs
to pump blood to most of the body while the right ventricle fills only the lungs. he
left ventricle is shorter and more conical in shape than the right, and on transverse
section its concavity presents an oval or nearly circular outline. It forms a small part
of the sternocostal surface and a considerable part of the diaphragmatic surface of the
heart; it also forms the apex of the heart. The left ventricle is thicker and more
muscular than the right ventricle because it pumps blood at a higher pressure. Non
compaction of the ventricular myocardium is a cardiomyopathy thought to be caused

by arrest of normal embryogenesis of the endocardium and myocardium. This

abnormality is often associated with other congenital cardiac defects, but it is also
seen in the absence of other cardiac anomalies. Clinical manifestations are highly
variable, ranging from no symptoms to disabling congestive heart failure,
arrhythmias, and systemic thrombo emboli. Echocardiography has been the
diagnostic procedure of choice, but the correct diagnosis is often missed or delayed
because of lack of knowledge about this uncommon disease and its similarity to other
diseases of the myocardium and endocardium.

Non compaction of the ventricular myocardium (NVM) is an uncommon finding.

It is thought to be caused by arrest of the normal process of endomyocardial
morphogenesis. NVM was first described in association with other congenital
anomalies, such as obstruction of the right or left ventricular outflow tracts, complex
cyanotic congenital heart disease, and coronary artery anomalies. The abnormal
compaction process in these cases is not fully understood, but pressure overload or
myocardial ischemia preventing regression of the embryonic myocardial sinusoids
has been suggested. This process results in the persistence of deep intertrabecular
recesses in communication with both the ventricular cavity and the coronary
circulation.

Isolated non compaction of the ventricular myocardium (INVM) is characterized

by persistent embryonic myocardial morphology found in the absence of other
cardiac anomalies to explain the abnormal development. In such cases, the resultant
deep recesses communicate only with the ventricular cavity, not the coronary
circulation.

The left ventricle is uniformly affected, but biventricular non compaction has
been reported, with right ventricular non compaction described in less than one-half
of patients. Because of difficulty in distinguishing normal variants in the highly
trabeculated right ventricle from the pathological non compacted ventricle, several
authors dispute the existence of right ventricular non compaction. Furthermore, some
authors have advocated that the term left ventricular hypertrabeculation be used

instead of isolated non compaction, as coexisting cardiac abnormalities have been

described in isolated cases, and the latter term suggests that the pathogenesis is
proven Treatment for non compaction of the ventricular myocardium focuses on the
3 major clinical manifestations: Heart failure, arrhythmias, and systemic embolic
events. Standard medical therapy for systolic and diastolic ventricular dysfunction is
warranted. Cardiac transplantation has been used for those with refractory congestive
heart failure. Only 6 cases of INVM leading to cardiac transplantation have been
published to date. The beneficial effects of the -blocker carvedilol on left
ventricular function, mass, and neurohormonal dysfunction in an infant with INVM
have been described. Because of the frequency of ventricular tachycardia and
significant risk of sudden cardiac death and systemic embolism, assessment for atrial
and ventricular arrhythmias by ambulatory ECG monitoring should be performed
annually. As more information is gathered about NVM and risk of sudden cardiac
death, implantable defibrillator technology may have an expanded role. Biventricular
pacemakers may have a role in the treatment of NVM patients with heart failure,
reduced left ventricular function, and prolonged intraventricular conduction.
Prevention of embolic complications is also an important management issue, and
several authors have recommended long-term prophylactic anticoagulation for all
patients with ventricular non compaction whether or not thrombus has been found.

1.3 MYOCARDIUM SEGMENTATION

Nuclear cardiology, echocardiography, cardiovascular magnetic resonance
(CMR), cardiac computed tomography (CT), positron emission computed
tomography (PET), and coronary angiography are imaging modalities that have been
used to measure myocardial perfusion, left ventricular function, and coronary
anatomy for clinical management and research. Although there are technical
differences between these modalities, all of them image the myocardium and the
adjacent cavity. However, the orientation of the heart, angle selection for cardiac
planes, number of segments, slice display and thickness, nomenclature for segments,
and assignment of segments to coronary arterial territories have evolved
independently within each field. This evolution has been based on the inherent

strengths and weaknesses of the technique and the practical clinical application of
these modalities as they are used for patient management. This independent
evolution has resulted in a lack of standardization and has made accurate intra- and
cross-modality comparisons for clinical patient management and research very
difficult, if not, at times, impossible.
Attempts to standardize these options for all cardiac imaging modalities should
be based on the sound principles that have evolved from cardiac anatomy and clinical
needs. Selection of standardized methods must be based on the following criteria:

(1) Maintain consistency with accepted anatomic and autopsy data.

(2) Utilize, as much as possible, existing and accepted approaches to

myocardial segmentation and nomenclature.

(3) Allow precise localization by using anatomic landmarks whenever

possible.

(4) Provide adequate sampling of the left ventricle and coronary distribution
without exceeding the resolution limits of the imaging modalities or
relevance for clinical and research applications.

(5) Allow linkage of the segments to known coronary arterial topography as

defined by coronary angiography.
An earlier special report from the American Heart Association, American

College of Cardiology, and Society of Nuclear Medicine defined standards for plane
selection and display orientation for serial myocardial slices generated by cardiac 2dimensional (2D) or tomographic imaging. In these standards, recommendations
were not made for echocardiography. The American Society of Echocardiography
and the American Society of Nuclear Cardiology also have made specific
recommendations for their respective modalities. Best-practice guidelines for CMR
have been published. To optimize and facilitate communication between cardiac
imaging

modalities

for

research

and

clinical

applications,

consensus

recommendations will be made for the following: orientation of the heart, names for

cardiac planes, number of myocardial segments, selection and thickness of cardiac

slices for display and analysis, nomenclature and location of segments, and
assignment of segments to coronary arterial territories. These recommendations are
applicable for imaging myocardial perfusion and wall motion.

1.4 COMPUTER TOMOGRAPHY

X-ray computed tomography (x-ray CT) is a technology that uses computerprocessed x-rays to produce tomographic images of specific areas of the scanned
object, allowing the user to see inside without cutting. Digital geometry processing is
used to generate a three-dimensional image of the inside of an object from a large
series of two-dimensional radiographic images taken around a single axis of
rotation. Medical imaging is the most common application of x-ray CT. Its crosssectional images are used for diagnostic and therapeutic purposes in various medical
disciplines. The rest of this article discusses medical-imaging x-ray CT; industrial
applications of x-ray CT are discussed at industrial computed tomography scanning.
As x-ray CT is the most common form of CT in medicine and various other
contexts, the term computed tomography alone (or CT) is often used to refer to x-ray
CT, although other types exist (such as positron emission tomography [PET]
and single-photon emission computed tomography [SPECT]). Older and less
preferred terms that also refer to x-ray CT are computed axial tomography (CAT
scan) and computer-aided/assisted tomography. X-ray CT is a form of radiography,
although the word "radiography" used alone usually refers, by wide convention, to
non-tomographic radiography.
CT produces a volume of data that can be manipulated in order to
demonstrate various bodily structures based on their ability to block the x-ray beam.
Although, historically, the images generated were in the axial or transverse plane,
perpendicular to the long axis of the body, modern scanners allow this volume of
data to be reformatted in various planes or even as volumetric (3D) representations of
structures. Although most common in medicine, CT is also used in other fields, such
as nondestructive materials testing. Another example is archaeological uses such as
imaging the contents of sarcophagi. Individuals responsible for performing CT

exams are calledra diographers or radiologic technologists and are required to be

licensed in most states of the USA.
Usage of CT has increased dramatically over the last two decades in many
countries. An estimated 72 million scans were performed in the United States in
2007. One study estimated that as many as 0.4% of current cancers in the United
States are due to CTs performed in the past and that this may increase to as high as
1.5 to 2% with 2007 rates of CT usage; however, this estimate is disputed., as there
is not a scientific consensus about the existence of damage from low levels of
radiation. Kidney problems following intravenous contrast agents may also be a
concern in some types of studies.
There are several advantages that CT has over traditional 2D medical
radiography. First, CT completely eliminates the superimposition of images of
structures outside the area of interest. Second, because of the inherent high-contrast
resolution of CT, differences between tissues that differ in physical density by less
than 1% can be distinguished. Finally, data from a single CT imaging procedure
consisting of either multiple contiguous or one helical scan can be viewed as images
in the axial, coronal, or sagittal planes, depending on the diagnostic task. This is
referred to as multiplanar reformatted imaging.
CT is regarded as a moderate- to high-radiation diagnostic technique. The
improved resolution of CT has permitted the development of new investigations,
which may have advantages; compared to conventional radiography, for example,
CT angiography avoids the invasive insertion of a catheter. CT colonography (also
known as virtual colonoscopy or VC for short) may be as useful as a barium
enema for detection of tumors, but may use a lower radiation dose. CT VC is
increasingly being used in the UK as a diagnostic test for bowel cancer and can
negate the need for a colonoscopy.
The radiation dose for a particular study depends on multiple factors: volume
scanned, patient build, number and type of scan sequences, and desired resolution
and image quality. In addition, two helical CT scanning parameters that can be
adjusted easily and that have a profound effect on radiation dose are tube current and

pitch. Computed tomography (CT) scan has been shown to be more accurate than
radiographs in evaluating anterior interbody fusion but may still over-read the extent
of fusion.

CHAPTER 2
LITERATURE SURVEY
2.1. EXISTING SYSTEM
A Complete System For Automatic Extraction Of Left Ventricular
Myocardium From CT Images Using Shape Segmentation And
Contour Evolution - Liangjia Zhu, Yi Gao, Vikram Appia, Anthony Yezzi,
Chesnal Arepalli, Tracy Faber

The shape segmentation is used technique for localizing the left ventricle.
Unlike other methods that only use low level information from voxels, our method
captures a global geometric characteristic of the left ventricle that agrees with our
visual perception. Next we use a variational region-growing method to locate the
epicardial surface given the segmentation of the endocardial surface. Then the
localized endocardial and epicardial surfaces are employed as a constraint for the
final segmentation. The shape variability is naturally handled and incorporated into
our system without using training images. Instead of simply imposing a constraint on
the point-wise distance between two contours, the one used in our model is a surfacewise restriction that uses a distance field for guiding contour evolution process. The
overall system is complete in that all the active contour models involved are
initialized automatically and robustly, other than in those systems that active
contours are either used as a single component or initialized manually

A Review Of Segmentation Methods In Short Axis Cardiac MR

Images - C. Petitjean and J. N. Dacher
The complexity of segmenting heart chambers and myocardium mainly relies
on heart anatomy and MRI acquisition specificity. The LV function consists in
pumping the oxygenated blood the aorta and consequently to the systemic circuit.

10

The LV cavity has a shape of ellipsoid and is surrounded by the myocardium. The
RV has a complex crescent shape. The standard imaging plane is perpendicular to the
long (apex-base) axis and called short axis plane. Imaging of the heart in MRI covers
the whole organ with about 810 short-axis slices, distance between two adjacent
slices ranging from 10 to 20 mm. As the heart is a moving organ, images are
acquired throughout the cardiac cycle. In terms of shape, the ventricle varies over
patients, over time and over the long axis. This variability must be accounted in
segmentation algorithms

Active Appearance Models - T. F. Cootes, G. J. Edwards, and C. J. Taylor

Active appearance models are generated by combining a model of shape
variation with a model of texture variation. By texture we mean the pattern of
intensities or colors across an image patch. To build a model, we require a training
set of images where corresponding points have been marked on each sample. For
instance, to build a face model, we require face images marked with points defining
the main features. We apply analysis to align the sets of points and build a statistical
shape model. We then warp each training image so the points match those of the
mean shape. This is raster scanned into a texture vector, which is normalized by
applying a linear transformation. After normalization, eigen-analysis is applied to
build a texture model. Finally, the correlations between shape and texture are learned
to generate a combined appearance model

Mesh Sementation Refinement - L. Kaplansky and A. Tal

This method is used for refining existing mesh segmentations. Given the
segmentation, produced either by an automatic segmentation method or manually.
This algorithm propagates the segment boundaries to more appropriate locations. In
addition, unlike most segmentation algorithms, this method allows the boundaries to
pass through the mesh faces, resulting in smoother curves, particularly visible on
coarse meshes. The method is also capable of changing the number of segments, by

11

enabling splitting and merging of boundary curves during the process. Finally, by
changing the propagation rules, it is possible to segment the mesh by texture-based
segmentations.

Patient Specific Heart Models From High Resolution CT C. Bajaj, S.

Goswami, Z. Yu, Y. Zhang, Y. Bazilevs, and T. Hughes

This paper, provides a solution to the problem of automatically constructing

three dimensional (3D) finite-element mesh models (FEM) of the human heart
directly from high resolution CT. Our overall computational pipeline from 3D
imaging to FEM models has five main steps, (i) discrete voxel segmentation of the
CT (ii) discrete topological noise filtering to remove non-regularized, and small
geometric measure artifacts (iii) a reconstruction of the inner and outer surface
boundaries of the human heart and its chambers (iv) computation of the medial axis
of the heart boundaries and a volumetric decomposition of the heart into tubular,
planar and chunky regions, (v) a flexible match and fit of each of the decomposed
volumetric regions using segmented anatomical volumetric templates obtained from
a 3D model heart. This initial model is not always correct because of missing
information and topological inconsistency so we create a segmented 3D map from
the template heart model and build a correspondence. Further we match and fit the
patient specific model with the segmented 3D map

Active Shape Model Their Training And Application - T. F. Cootes, C.

J. Taylor, D. H. Cooper, and J. Graham

In this paper we present new methods of building and using flexible models
of image structures whose shape can vary. Our technique relies upon each object or
image structure being represented by a set of points. The points can represent the
boundary, internal features, or even external ones, such as the center of a concave
section of boundary.Points are placed in the same way on each of a training set of
examples of the object. The sets of points are aligned automatically to minimize the

12

variance in distance between equivalent points. The statistics of the positions of the
labeled points a Point Distribution Model is derived. The model gives the average
positions of the points, and has a number of parameters which control the main
modes of variation found in the training set. Given a model and an image containing
an example of the object modeled, image interpretation involves choosing values for
each of the parameters so as to find the best fit of the model to the image.

Four Chamber Heart Modelling - Zheng, A. Barbu, B. Georgescu, M.

Scheuering, and D. Comaniciu

Here object detection method based on two elegant techniques, marginal

space learning (MSL) and steerable features are used.The idea of marginal space
learning (MSL) is not to use a classifier directly in the full similarity transformation
space but to incrementally learn classifiers on projected sample distributions. Here
we split the estimation into three problems: position estimation, position-orientation
estimation, and full similarity transformation estimation. To attack the second
challenge, we introduce steerable features, which constitute a very flexible
framework. Basically, we sample a few points from the volume under a sampling
pattern and extract a few local features (e.g., intensity and gradient) for each
sampling point. The efficiency of steerable features comes from the fact that much
fewer points are needed for manipulation, compared to the whole volume. After
similarity transformation estimation, we get an initial estimate of the non rigid shape.
We then use a model based 3D boundary detector to guide the shape deformation in
the active shape model (ASM) framework

Random Forests Classification In Real Time 3D Echocardiography

V. S. Lempitsky, M. Verhoek, J. A. Noble, and A. Blake

3D cardiography is however a challenging modality to work with due to low

signal to- noise ratio, unpredictable speckle patterns, and large variability in shape
and appearance between different subjects. In the paper, we investigate a

13

discriminative approach to this classification task. Thus, our classifier is trained to

model the posterior probability of the class given the data directly, without explicit
modeling of either the statistics of the shape of the heart or the physics of the
imaging process. Furthermore, the non-parametric nature of the classifier we use
means that no restrictive assumptions are made about the form of the posterior. Our
system is thus much simpler than many of the competing approaches. Importantly, as
the output of our method is soft, it could be used as an input for a more complex
algorithm which can incorporate richer geometric information, such as deformable
templates, level sets, or graph cuts. The particular discriminative classifier that we
use is the Random Forest classifier. Random Forests have an excellent track record
for various machine learning problems on a par or even better than boosting method
or max-margin classifiers like SVMs

Localizing Region-Based Active Contours - S. Lankton and A.

Tannenbaum,
In this paper, we propose a natural framework that allows any region-based
segmentation energy to be re-formulated in a local way. We consider local rather
than global image statistics and evolve a contour based on local information.
Localized contours are capable of segmenting objects with heterogeneous feature
profiles that would be difficult to capture correctly using a standard global method.
The presented technique is versatile enough to be used with any global region-based
active contour energy and instill in it the benefits of localization. We describe this
framework and demonstrate the localization of three well-known energies in order to
illustrate how our framework can be applied to any energy. We then compare each
localized energy to its global counterpart to show the improvements that can be
achieved. Next, an in-depth study of the behaviors of these energies in response to
the degree of localization is given. Finally, we show results on challenging images to
illustrate the robust and accurate segmentations that are possible with this new class
of active contour models

14

A Registration-Based Propagation Framework For Automatic

Whole Heart Segmentation Of Cardiac MRI - X. Zhuang, K. S. Rhode,
R. S. Razavi, D. J. Hawkes, and S. Ourselin

Magnetic resonance (MR) imaging has become a routine modality for the
determination of patient cardiac morphology. The extraction of this information can
be important for the development of new clinical applications as well as the planning
and guidance of cardiac interventional procedures. To avoid inter- and intra-observer
variability of manual delineation, it is highly desirable to develop an automatic
technique for whole heart segmentation of cardiac magnetic resonance images.
However, automating this process is complicated by the limited quality of acquired
images and large shape variation of the heart between subjects. In this paper, we
propose a fully automatic whole heart segmentation framework based on two new
image registration algorithms: the locally affine registration method (LARM) and the
free-form deformations with adaptive control point status (ACPS FFDs). LARM
provides the correspondence of anatomical substructures such as the four chambers
and great vessels of the heart, while the registration using ACPS FFDs refines the
local details using a constrained optimization scheme. We validated our proposed
segmentation framework on 37 cardiac MR volumes on the end-diastolic phase,
displaying a wide diversity of morphology and pathology, and achieved a mean
accuracy of 2.14 0.63 mm (rms surface distance) and a maximal error of 4.31 mm.

Multistage

Hybrid

Active

Appearance

Model

Matching:

Segmentation Of Left And Right Ventricles In Cardiac MR Images S. C. Mitchell, B. Lelieveldt, R. J. van der Geest, J. G. Bosch, J. H. C. Reiber,
and M. Sonka,

A fully automated approach to segmentation of the left and right cardiac

ventricles from magnetic resonance (MR) images is reported. A novel multistage
hybrid appearance model methodology is presented in which a hybrid active shape
model/active appearance model (AAM) stage helps avoid local minima of the

15

matching function. This yields an overall more favorable matching result. An

automated initialization method is introduced making the approach fully automated.
Our method was trained in a set of 102 MR images and tested in a separate set of 60
images. In all testing cases, the matching resulted in a visually plausible and accurate
mapping of the model to the image data. Average signed border positioning errors
did not exceed 0.3 mm in any of the three determined contours-left-ventricular (LV)
epicardium, LV and right-ventricular (RV) endocardium. The area measurements
derived from the three contours correlated well with the independent standard (r =
0.96, 0.96, 0.90), with slopes and intercepts of the regression lines close to one and
zero, respectively. Testing the reproducibility of the method demonstrated an
unbiased performance with small range of error as assessed via Bland-Altman
statistic. In direct border positioning error comparison, the multistage method
significantly outperformed the conventional AAM (p < 0.001). The developed
method promises to facilitate fully automated quantitative analysis of LV and RV
morphology and function in clinical setting.

Automatic Model-Based Segmentation Of The Heart In CT Images O. Ecabert, J. Peters, H. Schramm, C. Lorenz, J. von Berg, and M. Walker,.

Automatic image processing methods are a prerequisite to efficiently analyze

the large amount of image data produced by computed tomography (CT) scanners
during cardiac exams. This paper introduces a model-based approach for the fully
automatic segmentation of the whole heart (four chambers, myocardium, and great
vessels) from 3-D CT images. Model adaptation is done by progressively increasing
the degrees-of-freedom of the allowed deformations. This improves convergence as
well as segmentation accuracy. The heart is first localized in the image using a 3-D
implementation of the generalized Hough transform. Pose misalignment is corrected
by matching the model to the image making use of a global similarity transformation.
The complex initialization of the multicompartment mesh is then addressed by
assigning an affine transformation to each anatomical region of the model. Finally, a
deformable adaptation is performed to accurately match the boundaries of the

16

patient's anatomy. A mean surface-to-surface error of 0.82 mm was measured in a

leave-one-out quantitative validation carried out on 28 images. Moreover, the
piecewise affine transformation introduced for mesh initialization and adaptation
shows better interphase and interpatient shape variability characterization than
commonly used principal component analysis.

17

CHAPTER 3
PROPOSED SYSTEM
The existing method is Localized principal component analysis based curve
evolution: A divide and conquer approach. We propose a novel localized principal
component analysis (PCA) based curve evolution approach which evolves the
segmenting curve semi-locally within various target regions (divisions) in an image
and then combines these locally accurate segmentation curves to obtain a global
segmentation. The training data for our approach consists of training shapes and
associated auxiliary (target) masks. The masks indicate the various regions of the
shape exhibiting highly correlated variations locally which may be rather
independent of the variations in the distant parts of the global shape. Thus, in a sense,
we are clustering the variations exhibited in the training data set. We then use a
parametric model to implicitly represent each localized segmentation curve as a
combination of the local shape priors obtained by representing the training shapes
and the masks as a collection of signed distance functions. We also propose a
parametric model to combine the locally evolved segmentation curves into a single
hybrid (global) segmentation. Finally, we combine the evolution of these semi-local
and global parameters to minimize an objective energy function. The resulting
algorithm thus provides a globally accurate solution, which retains the local
variations in shape. We present some results to illustrate how our approach performs
better than the traditional approach with fully global PCA.
The proposed method is automatic extraction of left ventricular from CT
images using shape segmentation and contour evolution The left ventricular
myocardium plays a key role in the entire circulation system and an automatic
delineation of the myocardium is a prerequisite for most of the subsequent functional
analysis. In this paper, we present a complete system for an automatic segmentation
of the left ventricular myocardium from cardiac computed tomography (CT) images
using the shape information from images to be segmented. The system follows a
coarse-to-fine strategy by first localizing the left ventricle and then deforming the

18

myocardial surfaces of the left ventricle to refine the segmentation. In particular, the
blood pool of a CT image is extracted and represented as a triangulated surface.
Then, the left ventricle is localized as a salient component on this surface using
geometric and anatomical characteristics. After that, the myocardial surfaces are
initialized from the localization result and evolved by applying forces from the image
intensities with a constraint based on the initial myocardial surface locations

19

CHAPTER 4
SYSTEM ANALYSIS
4.1. SYSTEM OVERVIEW
The main challenges in extracting the myocardium include large shape
variability within cardiac cycles and between different patients, and weak edges
between epicardium and heart fat or soft tissues. To get an accurate and robust
segmentation, model-based methods have become dominant in this research . Heart
models are commonly used to represent the geometric or intensity features of the
heart, and they are applied either explicitly or implicitly for segmentation. In the first
type of methods, models created off-line are fitted to images for segmentation. For
example, active shape models (ASMs) build a statistical shape model from a set of
aligned shapes by using the principal component analysis (PCA) technique, and have
been used for left ventricle segmentation. Active appearance models (AAMs) extend
this idea by incorporating gray level information and have been used in segmenting
the left and right ventricles from MR images. The deformations allowed in the
parametric models such as ASMs and AAMs are confined to the shape space where
the heart models are embedded.
A more sophisticated way of representing the shape space is by using 3D
diffusion wavelets, which encode shape variations hierarchically. Using deformable
models provides a flexible way to incorporate shape priors that are capable of
adapting to local image content. The local adaptation was achieved by progressively
optimizing the piecewise affine transformations of this model to match image
boundaries. In, a mean shape of the heart was fitted to an image by estimating
similarity transformations, which was then deformed to match image boundaries with
the help of landmark points on the interventricular septum. Instead of deforming a
pre-aligned model, atlas-based methods use shape information implicitly by directly
registering each atlas image to a target image. Then, either the labels from multiple
atlases are fused or one single registered atlas is deformed to extract the heart
region. Model-free methods have also been widely used to explore the characteristics
of heart geometry or intensity distribution from other perspectives. For example, the

20

geometric and intensity features in the myocardial region were learned by using a
random forests method for delineating the myocardium . Active contour models have
been widely used in medical image segmentation because of their flexibility and
robustness.
In these models, energy functionals are commonly defined over image
features such as edge, region statistics, local characteristics, and a combination of
edges and regions , which are optimized by using gradient descent techniques. Prior
information can be incorporated as well to restrict the optimization space. For
example, an active contour model was evolved in the shape space of the left ventricle
obtained by applying the PCA to manually segmented images. Local variations may
be captured by decomposing images into different regions using prior information for
ventricles segmentation or by modeling a shape prior using pixel-wise stochastic
level sets to extract the endocardium . A shape constraint was also employed to
control the search space of the myocardial contours between two consecutive image
slices. Coupled active contours have been proposed with distance constraints
between contours for myocardium extraction, cortex segmentation , and cell tracking
. One important but less studied topic is how to locate the heart initially, especially
for these methods using deformable models, which tend to get stuck in undesirable
local extrema when started without a good initialization. Typically, the geometric
features of the heart are used for localization. In, the endocardium was initialized by
searching for a circular structure in a blood pool mask obtained via thresholding.
Similar empirical rules were used to identify the left ventricle cavity. To capture a
more generic shape of the heart, the generalized Hough transform was utilized for
heart detection.
In , the localization was achieved by searching for a similarity transformation
in a hierarchical way. Atlas based registration has also been used for coarse
initialization. One fact that has been ignored in the literature for the localization is
that the left ventricle is a salient component on the heart surface. This is where the
shape decomposition/ segmentation technique may be utilized to cluster the surface
into meaningful components based on some given criteria as in computer graphics
and geometric modeling .

21

For example, a surface may be hierarchically decomposed into regions of

deep concavities by using fuzzy clustering and graph partition techniques. Prominent
feature points have also been used to cluster a surface into meaningful regions.
Applications of shape segmentation in medical imaging can be found in heart
modeling from images and aneurysm neck detection on vessel surfaces. Active
contour models have been applied as well on surfaces to refine the segmentation or
extract objects of interest. Among the few applications of the shape decomposition
techniques to cardiac image segmentation, the narrowing of vessels around the left
atrium was detected by merging local features based on given criteria to extract the
left atrium. As for the left ventricle localization, the region near the left ventricle is
much more recognizable from the heart surface than from the volumetric data, which
can be identified by a deep concave contour.

4.2. FEASIBILITY STUDY

For successful implementation of every project, feasibility study is to be
performed. This can be created as one the initial steps which are to be undertaken
during the preliminary investigation stage which is a major step in the system
Analysis and an important outcome of the preliminary investigation to determine the
implementation of the personalized digital utility.
Feasibility study is a test of system proposed regarding its functionality,
impact on the organization, ability to meet the needs and effective use of resources.
Feasibility study is made to see if the project on the organization for the amount of
work, effort and the time that is spend on it. Feasibility study lets the developer to
foresee the future of the project and its usefulness.
Feasibility and risk analysis are related in many ways. If project risk is great the
feasibility of producing the quality software is reduced. Feasibility study use three
major yardsticks to measure, or predict a systems success:
1. Technical Feasibility
2. Economical Feasibility

22

3. Operational Feasibility

4.2.1. TECHNICAL FEASIBILITY

It is found that the new system can be implemented and work can be done
with the existing software technology and available man power. All the methods and
equipment used in this system are available in CSI Institute of Technology, Thovalai.
Windows XP is the platform used to develop the system. MATLAB is the software
tool. Thus the system is technically feasible.

4.2.2. ECONOMICAL FEASIBILITY

The cost of conducting the investigation and the development of the system
was not high. Since the company already owns the necessary software and resources
no cost overhead is to be considered. The system can reduce the efforts and time
after it is developed. This evaluation looks at the financial aspects of the project. It
determines whether the investment needed to implement the system will be
recovered. To carry out an economic feasibility study, it is necessary to place actual
money values against any purchases or activities needed to implement the project. It
is also necessary to place money values against any benefits that will accrue from a
new system created by the project. Such calculations are often described as costbenefit analysis. The proposed system is found to be economically feasible, because
as in the testing phase of the system it doesnt need any special tools. So the new
system is economically feasible.

4.2.3. OPERATIONAL FEASIBILITY

Operational feasibility covers two aspects. One is a technical performance

aspect and the other is acceptance within the organization. Technical performance
includes issues such as determining whether the system can provide the right

23

information for the organizations personal, and whether the system can organized so
that it always delivers this information at the right place and on time. Acceptance
revolves around the current system and its personal. Operational feasibility must
determine how the proposed system will fit in with the current operations and what,
if any, job restructuring and retraining may be needed to implement the system. The
scope of the new system fulfills all the salient features of working as desired by
manual. So it is operationally feasible.

24

CHAPTER 5
SYSTEM DESIGN
5.1. SYSTEM STUDY
In this, we present a complete system for automatically extracting the
myocardium from cardiac CT images without using training images. A coarse-to-fine
strategy, consisting of global localization and local deformations, is used for the
myocardium segmentation. Before starting the localization step, the heart surface is
approximated by the blood pool surface. Then, the apex point of the left ventricle on
this surface is detected by using the relative orientation of ventricles with respect to
the physical coordinate system of a CT image. The left ventricle is automatically
detected by examining the distribution of the level sets starting from the apex point,
which is further refined by performing the geometric active contour model on the
blood pool surface. This contour decomposes the surface into two parts, and the one
contains the apex point is chosen as the initial endocardial surface. Once the
endcoardial surface is located, its corresponding mask is obtained via rasterization.
Then, a variational region-growing method is used to extract the initial epicardial
surface based on the endocardium segmentation. Finally, these two surfaces are
refined by employing an active contour model with a shape constraint, and the
myocardium is obtained by extracting voxels between these surfaces. The
contributions of the proposed method are as follows:
1) we utilize the shape segmentation technique for localizing the left
ventricle. Unlike other methods that only use low level information from voxels, our
method captures a global geometric characteristic of the left ventricle that agrees
with our visual perception. Hence, it is not sensitive to such issues as shape
variability and changes of volume coverage. Note that, as an initialization step, the
proposed method can be easily incorporated into other model-based frameworks.
2) we use a variational region-growing method to locate the epicardial surface
given the segmentation of the endocardial surface. Then the localized endocardial
and epicardial surfaces are employed as a constraint for the final segmentation. In
this formulation, the shape variability is naturally handled and incorporated into our

25

system without using training images. In addition, instead of simply imposing a

constraint on the point-wise distance between two contours, the one used in our
model is a surface-wise restriction that uses a distance field for guiding contour
evolution process.
Therefore, the overall system is complete in that all the active contour models
involved are initialized automatically and robustly, other than in those systems that
active contours are either used as a single component or initialized manually.

5.1.1. BLOCK DIAGRAM

.

Fig 5.1 Block Diagram

5.2. LEFT VENTRICLE LOCALIZATION

Assume that the orientation of a CT image is given and intensity contrast
exists between blood pool and myocardium. The localization of the left ventricle is
determined via searching for a deep concave boundary on the blood pool surface as
follows.

26

5.2.1 EXTRACT BLOOD POOL SURFACE

The extraction of the blood pool surface is carried out by a few mature
techniques in the computer vision and graphics: since CT images have calibrated
gray levels, the source image is thresholded to highlight the blood pool region. Then,
the morphological opening operator is applied to remove noisy arteries and cut
spines that may be residing in the same connected component of the heart. After that,
the largest connected component is chosen and triangulated to get the blood pool
surface.

5.2.2 DETECT APEX POINT

Suppose the coordinate system of the source image is Left-Posterior-Inferior
(LPI) as shown in Fig. 5.2. In this system, the XYZ coordinates trace from left to
right, posterior to anterior, and inferior to superior. Even though the long axis of pig
and human heart has different orientations, the directions of left and right are clearly
defined from the inferior view. The apex point is one salient feature that can be used
to locate the left ventricle. Its location is determined as follows:

Fig 5.2 Apex Point Location

27

1) estimate the orientation of ventricles;

2) search for the left ventricle apex, which is the left tip point with respect to the
estimated orientation a reference plane such that the left ventricle points are mainly
above the plane and otherwise for the right ventricle.

5.2.3 IDENTIFY CUT CONTOUR

A two-step segmentation strategy is used to identify the left ventricle by
searching for a cut contour on Mbp.
1) Find an Initial Cut Contour: The initial cut contour, denoted by C0, is determined
based on the distance field starting from papx

Fig 5.3 Identifying the cut contour

The total length of each isocontour increases gradually and then drops slightly as it is
traveling along the left ventricle. After that, it goes up first and then drops rapidly as
it is propagating to the right ventricle and other regions

2) Refine the Cut Contour: The geometric active contour model is utilized to refine
the initial cut contour C0. Suppose a contour on the surface Mbp is represented by
the zero level set of a function

28

5.3 . MYOCARDIUM WALL SEGMENTATION

The endocardial surface indicates the location of the left ventricle, which is
rasterized to get its 3D mask Iendo for refinement. Instead of simply dilating Iendo
for approximating the epicaridal mask Iepi , a variational region-growing model is
used by taking an outward neighborhood of Iendo as the seed region. After that, a
localized region-based active contour model is utilized with a shape constraint
imposed by these initial masks to refine the myocardial segmentation.

5.3.1 INITIALIZE THE ENDOCARDIAL AND EPICARDIAL MASKS

The surface Mendo is closed via triangulating the points along the cut contour
Ccut . Here, Mendo is still used to denote the closed surface. Iendo is created by
rasterizing Mendo with the same origin and resolution as the source image I. To
remove noise and papillary muscles, the convex hull of Mendo is computed and set
as a ROI for performing the morphological closing operation on Iendo.
.
5.3.2. EVOLVE ACTIVE CONTOURS WITH A SHAPE CONSTRAINT

Given a mask image the refinement of the endocardial and epicardial masks
were performed separately. In initializing the epicardial mask, a parameter dw was
used to control the maximum distance allowed in the region-growing process. This
parameter is related to the average thickness of the myocardial wall

5.3.3. EXTRACTING THE MYOCARDIAL WALL

The myocardial wall is defined as the volume between the endocaridal and
epicardial masks. Note that the contour evolution process returns closed masks. To
extract a complete myocardial wall, the voxels inside the blood pool need to be
removed. To this end, the wall is divided into two parts: one in which the

29

myocardium can be completely determined by performing the XOR operation

between the endocardial and epicardial masks, and the other formed by removing the
voxels inside the blood pool from the epicardial mask.

30

CHAPTER 6
SYSTEM REQUIREMENTS
6.1 GENERAL
This chapter deals with the explanation of software we used in this project.
6.1.1 HARDWARE CONFIGURATION
Processor

Intel Core i3

Speed

2.30 GHz

RAM

2 GB

Hard Disk

500GB

General

Key Board, Monitor, Mouse

6.1.2 SOFTWARE CONFIGURATION

Operating System

Windows 7

Software

MATLAB 7.1

6.1.3 SOFTWARE DETAILS

This part explains some introduction of MATLAB.

6.2 INTRODUCTION TO MATLAB

MATLAB (matrix laboratory)
and fourth-generation
MATLAB

programming

is

a numerical

language.

allows matrix manipulations,

computing environment

Developed

plotting

by Math

of functions and

Works,
data,

implementation of algorithms, creation of user interfaces, and interfacing with

programs written in other languages, including C, C++, Java, and Fortran.
Although MATLAB is intended primarily for numerical computing, an
optional toolbox uses the MuPAD symbolic engine, allowing access to symbolic
computing capabilities. An additional package, Simulink, adds graphical multidomain simulation and Model-Based Design for dynamic and embedded systems.
In 2004, MATLAB had around one million users across industry and
academia. MATLAB users come from various backgrounds of engineering, science,

31

and economics. MATLAB is widely used in academic and research institutions as

well as industrial enterprises.
Cleve

Moler,

the

chairman

of

the computer-science department

at

the University of New Mexico, started developing MATLAB in the late 1970s. He
designed it to give his students access to LINPACK and EISPACK without them
having to learn Fortran.
It soon spread to other universities and found a strong audience within
the applied mathematics community. Jack little, an engineer, was exposed to it
during a visit Moler made to Stanford University in 1983.
Recognizing its commercial potential, he joined with Moler and Steve
Bangert. They rewrote MATLAB in C and founded MathWorks in 1984 to continue
its development. These rewritten libraries were known as JACKPAC. In 2000,
MATLAB was rewritten to use a newer set of libraries for matrix
manipulation, LAPACK.
MATLAB was first adopted by researchers and practitioners in control
engineering, Little's specialty, but quickly spread to many other domains. It is now
also used in education, in particular the teaching of linear algebra and numerical
analysis, and is popular amongst scientists involved in image processing.
Structures
MATLAB supports structure data types. Since all variables in MATLAB are
arrays, a more adequate name is "structure array", where each element of the array
has the same field names. In addition, MATLAB supports dynamic field names (field
look-ups by name, field manipulations etc).
Unfortunately, MATLAB JIT does not support MATLAB structures,
therefore just a simple bundling of various variables into a structure will come at a
cost.
6.3 MATLAB
MATLAB is high performance language for technical computing. It integrates
computation, visualization and programming in an easy to use environment,
where problems and solutions are expressed in familiar mathematical. Typical
uses include
Math and computation.

32

Algorithm development.
Data acquisition.
Modeling, simulation and prototyping.
Data analysis, exploration visualization.
Scientific and engineering graphics.
MATLAB is an interactive system whose

basic data element is and

array that does not require dimensioning. This allows us to solve many
technical computing problems especially those

with

matrix

and

vector

formulations.
In a fraction of time it would take to write a program in a scalar
non- interactive language such as C or Fortran.
6.4 MATLAB SYSTEM
The MATLAB system consists of following parts:
Development environment
Mathematical functions library
6.5 DEVELOPMENT ENVIRONMENT
This is a set of tools and facilities that help us to use MATLAB
functions and files. Many of this tools are graphical user interface.
It includes a MATLAB desktop, a
history, an editor debugger and a

command

window,

a command

browser for viewing help, the workspace,

files and the search path.

6.6 MATHEMATICAL FUNCTIONS LIBRARY
This

is a vast collection of computational algorithms ranging from

elementary functions like sum, sine, cosine and complex arithmetic to more
sophisticated functions

like

matrix

functions and Fourier transforms.

inverse, matrix

eigen

values ,

Bessel

33

CHAPTER 7
RESULT AND DISCUSSION
7.1 Input Image

Fig 7.1 Input Image

As we run our program the command window screen shows to ENTER THE
INPUT IMAGE .Then we have to select our input image.

34

7.2 Original Image

Fig 7.2 Original Image

This is the screenshot of the original image. As we enter the input image it is
displayed as the original image in our figure window.

35

7.3 Detect Apex Point

Fig 7.3 Detect Apex Point

The screenshot above shows the message box appears to detect our apex
point. This is given by pausing our program. We must click OK and detect the
apex point

36

7.4 Apex Point Vertex

Fig 7.4 Apex Point Vertex

This is the screenshot of the apex point detected. The vertexes show the
triangulated surface.

37

7.5 Region Based Segmented Image

Fig 7.5 Region based segmented image

The screenshot shows the region based segmented image. The regions are
identified based on the pixel values.

38

7.6 Active Contour Model Set

Fig 7.6 Active Contour Model Set

The screenshot shows the active contour model set . The red boundary shows
the myocardial region obtained by ACM.

39

7.7 Output Image

Fig 7.7 Output Image

The screenshot shows all the output of all the images on the same screen

40

7.8 Cardiac Volumes

Fig 7.8 Cardiac Volumes

The message box arrives after segmenting the epicardium and endocardium showing
their volumes. The epicardial volume is greater than the endocardial volume.

41

7.9 Sensitivity

Fig 7.9 Sensitivity

This screenshot shows the sensitivity of the epicardial volume. The parameter dw
was used to control the maximum distance allowed in the region-growing process.
This parameter is related to the average thickness of the myocardial wall

42

CHAPTER 9
CONCLUSION AND FUTURE ENHANCEMENT
9.1 CONCLUSION
We have presented a complete system for automatically segmenting the
myocardial wall from cardiac CT images. It follows the coarse-to-fine
framework by first detecting the left ventricle, and then refining this result by
employing contour evolution techniques with a shape constraint obtained online. In these tests, our method achieved high accuracy as well as strong
robustness for segmenting the human myocardium. Though the proposed
method was specific for the segmentation of the left ventricular myocardium,
it may be generalized in several possible ways for broader applications in cardi
image segmentations. It is straightforward to apply the shape segmentation
technique to segment the right ventricle from CT images because of the
similarity of ventricle structures. In addition, the proposed method may be
applied to other image modalities as long as a smooth heart surface can be
reconstructed. For example, in MR images, we may manually threshold an
MR image to extract the blood pool and then generate the heart surface.
Moreover, the proposed method can be easily integrated into user interactive
segmentation frameworks, which are widely used in medical image
segmentations. In particular, for ventricles with an arbitrary orientation, a user
can effectively pinpoint the apex point to start the segmentation.

9.2 FUTURE ENHANCEMENT

Therefore, in future work, we plan to extend the current method to segment
ventricles both in CT and MR images and apply segmentation results to
clinical applications such as evaluating the myocardial mass at risk caused by
stenoses.

43

REFERENCE
1. C. Petitjean and J. N. Dacher, A review of segmentation methods in short
axis cardiac MR images, Med. Image Anal., vol. 15, no. 2, pp. 169184,
2011.
2. T. F. Cootes, C. J. Taylor, D. H. Cooper, and J. Graham, Active shape
models-their training and application, Comput. Vis. Image Understand., vol.
61, no. 1, pp. 3859, 1995.
3. H. C. van Assen, M. G. Danilouchkine, M. S. Dirksen, J. H. Reiber, and B. P.
Lelieveldt, A 3-D active shape model driven by fuzzy inference: Application
to cardiac CT and MR, IEEE Trans. Inf. Technol. Biomed., vol. 12, no. 5,
pp. 595605, Sep. 2008.
4. H. C. van Assen, M. G. Danilouchkine, F. F. Frangi, S. Ordas, J. J.
Westenberg, J. H. Reiber, et al., SPASM: A 3D-ASM for segmentation of
sparse and arbitrarily oriented cardiac MRI data, Med. Image Anal., vol. 10,
no. 2, pp. 286303, 2006.
5. T. F. Cootes, G. J. Edwards, and C. J. Taylor, Active appearance models,
IEEE Trans. Pattern Anal. Mach. Intell., vol. 23, no. 6, pp. 681685, Jun.
2001.
6. S. C. Mitchell, B. Lelieveldt, R. J. van der Geest, J. G. Bosch, J. H. C. Reiber,
and M. Sonka, Multistage hybrid active appearance model matching:
Segmentation of left and right ventricles in cardiac MR images, IEEE Trans.
Med. Imag., vol. 20, no. 5, pp. 415423, may 2001.
7. S. Essafi, G. Langs, and N. Paragios, Hierarchical 3D diffusion wavelet
shape priors, in Proc. Int. Conf. Comput. Vis., 2009, pp. 17171724.
8. O. Ecabert, J. Peters, H. Schramm, C. Lorenz, J. von Berg, and M. Walker,
Automatic model-based segmentation of the heart in CT images, IEEE
Trans. Med. Imag., vol. 27, no. 9, pp. 11891201, Sep. 2008.
9. Y. Zheng, A. Barbu, B. Georgescu, M. Scheuering, and D. Comaniciu, Fourchamber heart modeling and automatic segmentation for 3-D cardiac CT
volumes using marginal space learning and steerable features, IEEE Trans.
Med. Imag., vol. 27, no. 11, pp. 16681681, Nov. 2008.

44

10. E. M. van Rikxoort, I. Isgum, Y. Arzhaeva, M. Staring, S. Klein, M. A.

Viergever, et al., Adaptive local multi-atlas segmentation: Application to the
heart and the caudate nucleus, Med. Image Anal., vol. 14, no. 1, pp. 3949,
2010.
11. X. Zhuang, K. S. Rhode, R. S. Razavi, D. J. Hawkes, and S. Ourselin, A
registration-based propagation framework for automatic whole heart
segmentation of cardiac MRI, IEEE Trans. Med. Imag., vol. 29, no. 9, pp.
16121625, Sep. 2010.
12. V. S. Lempitsky, M. Verhoek, J. A. Noble, and A. Blake, Random forest
classification for automatic delineation of myocardium in realtime 3D
echocardiography, in Proc. 5th Int. Conf. Funct. Imag. Model. Heart, 2009,
pp. 447456.
13. S. Kichenassamy, A. Kumar, P. Olver, A. Tannenbaum, and A. Yezzi,
Gradient flows and geometric active contour models, in Proc. 5th Int. Conf.
Comput. Vis., Washington, DC, USA, 1995. pp. 810815.
14. V. Caselles, R. Kimmel, and G. Sapiro, Geodesic active contours, Int. J.
Comput. Vis., vol. 22, no. 1, pp. 6179, 1997.
15. T. F. Chan and L. A. Vese, Active contours without edges, IEEE Trans.
Image Process., vol. 10, no. 2, pp. 20292039, Feb. 2001.
16. X. Wang, D. S. Huang, and H. Xu, An efficient local Chan-Vese model for
image segmentation, Pattern Recognit., vol. 43, no. 3, pp. 603618, 2010. L.
Chen, Y. Zhou, Y. Wang, and J. Yang, Rapid and brief communication:
GACV: Geodesic-aided C-V method, Pattern Recognit., vol. 39, no. 7, pp.
13911395, Jul. 2006