Mandibular Image Segmentation on the CT-Scan of the head
using the Active Contour Method
Amillia Kartika Sari
1
, Riries Rulaningtyas
2
and Khusnul Ain
2
1
Postgraduate School, Airlangga University, Surabaya, Indonesia
2
Science and Technology Faculty, Airlangga University, Surabaya, Indonesia
Keywords : Image Segmentation, Active Contour Method, Mandibular, CT-Scan
Abstract : Image segmentation is one of the image processing methods with the goal of sharing the image based on
uniformity, one of which is the active contour method. This method is to detect objects on a particular
image by using curve evolution techniques, and can also overcome the deficiencies in the boundary method.
In this study image segmentation was carried out using the active contour method to evaluate the mandible
on the head CT scan. It started with a CT-Scan of the head as input data, and saved with BMP (Bitmap)
format. Then initial contour mandible, and after that the next step is image segmentation with active
contour chan-vese method. From the analysis and evaluation of 108 images of the mandible with *BMP
(Bitmap) format we get to the average accuracy values which were 99.809%, and sensitivity value of
99,806%. The conclusion of this study is that the active contour method gives accurate results of mandibular
bone segmentation on the CT scan of the head.
1 INTRODUCTION
The mandible is the bone that forms the face of a
p
erson, especially the lower third. Like other organs,
the mandible may develop abnormalities such as
tumors, fractures,
or dislocations. Tumor
abnormalities in the mandible may result in bone
defects. Bone defect is a state of partial or complete
loss of bone, which can cause changes in bone
function and anatomy that negatively impact by
psychological weakness and reduced confidence in
social relations
1,2
.
Therefore, mandibular reconstruction surgery is
recommended
immediately. O
ne thing that can be
done to optimize surgical operations is to use a 3D
prototype of the dissected organ. It aims to assess
the severity of bone defects, improve the accuracy of
marginal resection, as an implant pre-contour plate,
and can reduce surgical time
3,4
.
3D prototypes are the result of the printing
technology of 3-dimensional objects from
combining several materials such as plastics,
polymers, ceramics, liquids and living cells. Stages
to obtain 3D prototypes are image acquisition, image
processing, and prototype printing. For image
acquisition, data input is a digital image obtained
from radiology as a CT-scan image. In this study a
CT scan of the head is used.
After obtaining digital image data, image
processing is carried out, namely the segmentation
process. The image segmentation is the process of
dividing an image into a number of parts
5
. Many
methods are used in the image segmentation
processes, one of which is Active Contour. This
method uses evolutionary curve techniques to detect
objects in images
6
. The nature of this method is
finding the boundary or edge of the object becomes
segmented from the influence of internal energy and
external energy. Internal energy regulates continuity
while external energy functions to draw a curve to
the edge of the target
7
.
The Active Contour Method is divided into two
groups: parametric and geometric. Parametric
methods commonly known as deformable can
segment objects with a clear boundary, one of which
is the Snake Active Contour Model. While
geometric method is the method that has the ability
to segment objects with unclear boundaries, one of
which is the Active Contour Level set model. In this
study we used an active contour geometric with
Chan-Vese model.