Biology Reference
In-Depth Information
FIGURE 14.4
DICOM slice
from the scanning boxes. From the
bottom left to right: femur, tibia,
and fibula. From the top left
to
right:
left
humerus,
thoracic
vertebra,
right ulna, and right
humerus.
Image Segmentation
Segmentation is the process of selecting regions or surfaces from three-dimensional
images and then separating the objects into three-dimensional surface models based on
specific grayscale threshold values. DICOM images are the individual transverse 2-D images
created during a CT scan. DICOMs require specialized imaging software to be opened, and
there are shareware computer programs available free on the Internet to simply open and
read the DICOMs.
4
There are other programs designed to create three-dimensional models
from the DICOM. One such computer program is the commercially available program
Amira
, which can segment the high-resolution DICOM images to create 3-D models. See
Figure 14.4
.
There are two types of image segmentation: manual and automatic. For manual segmen-
tation, the researcher opens a series of DICOM images in a 3-D imaging program like Amira.
Starting at one end of the bone, the researcher opens a single DICOM slice. Cortical bone is
dense and appears as voxel values that are very light gray to white, trabecular bone appears
as a range of darker gray, and the background appears black. The grayscale value of the
region of interest (i.e., the cortical bone) is selected, using maximum and minimum threshold
4
Free DICOM readers are available from many Internet sites. Two such sites are OsiriX at
http://www.
osirix-viewer.com/
,
and Cad/Cam Services, Inc. at
http://xrayscan.com/software/free-dicom-viewers.
