Extraction and Analysis o
Hui-Fuang
Dept. of C
pang@asis.edu.tw
Abstract—Structural characteristics of late
brain medical images from CT scan can
distinguishing between normal brain structu
brain structure. This paper presents im
algorithms for automatic segmentation and
structural features of lateral ventricle in brain
The effectiveness of such features on
discrimination is also examined. Experimen
that structural features of lateral ventricle
images, including Frontal Horn Radius and V
are useful for discriminating between norm
brain structure.
Keywords- brain medical image; feature
ventricle, image processing
I. INTRODUCTION
In clinical medicine, doctors have long b
way to extract and analyze expansion c
ventricle in medical brain images in orde
disease diagnosis, for instances, the diagnos
hydrocephalus and brain atrophy [1]. E
general, patients with brain disease posses
physiological anomalies and less anatomical
is believed that the structural character
ventricle in brain medical images from
tomography (CT) scan can be helpful fo
between normal brain structure and abnorm
such as obstructive hydrocephalus and brain
1 shows the schematic diagrams of imp
features of lateral ventricle, including Ventri
Ventricular Angle (VA), and Frontal Horn
Other features, such as the expansion char
third ventricle, temporal horn, or sulci mig
for the task [2].
However, for a radiologist, to manua
measure such features accurately on CT sca
difficult and time consuming. Image proce
have been successfully used in medical im
Therefore, the main objectives of this study
image processing algorithms for automatic s
measurement of structural features of lat
brain medical images, and to examine th
properties of such features for distinguishing
and abnormal brain structure.
of Structural Features of Lateral Ventr
Medical Images
g Ng, Cheng-Hung Chuang, Chih-Hsueh Hsu
Computer Science and Information Engineering
Asia University
Taichung, Taiwan
w, chchuang@asia.edu.tw, sunny.boy1030@gmail.com
eral ventricle in
n be helpful for
ure and abnormal
mage processing
measurement of
n medical images.
brain structure
ntal results show
in brain medical
Ventricular Index,
mal and abnormal
extraction; lateral
been looking for a
characteristics of
er to assist brain
sis of obstructive
Even though in
ss more apparent
l abnormalities, it
ristics of lateral
m computerized
or distinguishing
mal brain structure
n atrophy [2]. Fig.
portant structural
icular Index (VI),
n Radius (FHR).
racteristics of the
ght also be useful
lly segment and
an images is both
essing techniques
mage analysis [3].
y were to develop
segmentation and
teral ventricle in
he discriminative
g between normal
(a) Original brain CT sc
(b) Ventricular Index (VI) (c)
Figure 1. Structural features of lateral ven
The rest of the paper is organiz
presents the algorithm and proced
measuring structural features of l
medical images. Section 3 contains
concluding remarks are given in sec
II. METHO
The algorithm for segmenting a
features of lateral ventricle consist
First, a segmentation procedure is u
ventricle region from the brain im
ricle in Brain
can image
Ventricular Angle (VA) and
Frontal Horn Radius (FHR)
ntricle in brain medical image.
zed as follows. Section 2
dure for segmenting and
ateral ventricle in brain
experimental results, and
tion 4.
ODS
and measuring structural
s of the following steps.
used to extract the lateral
mage. Next, each of the
2012 Sixth International Conference on Genetic and Evolutionary Computing
978-0-7695-4763-3/12 $26.00 © 2012 IEEE
DOI 10.1109/ICGEC.2012.93
35
2012 Sixth International Conference on Genetic and Evolutionary Computing
978-0-7695-4763-3/12 $26.00 © 2012 IEEE
DOI 10.1109/ICGEC.2012.93
35
2012 Sixth International Conference on Genetic and Evolutionary Computing
978-0-7695-4763-3/12 $26.00 © 2012 IEEE
DOI 10.1109/ICGEC.2012.93
35
2012 Sixth International Conference on Genetic and Evolutionary Computing
978-0-7695-4763-3/12 $26.00 © 2012 IEEE
DOI 10.1109/ICGEC.2012.93
35