Biomedical Engineering Reference
In-Depth Information
Figure 4.
Block diagram of the proposed approach. See attached CD for color version.
4.1. Brain Extraction
Brain extraction, also known as
skull stripping
, consists in removing all non-
brain tissues (such as skull, eyes, and fat) from brain MRI scans. For this purpose
the MRICro program (free software available at www.sph.sc.edu/comd/rorden
/mricro.html) has been used to extract brain tissues and exclude all other arti-
facts. Figure 5 shows the results of applying this program to an MRI slice in a
normal case. Figure 6 shows the results of applying MRICro on a more compli-
cated MRI slice that contains larger parts of the skull and eyes, so as to show the
reliability of the program when applied on different brain slices.
4.2. Segmentation
In this section we present the framework used to extract the white matter
from the dataset described earlier. Recall that the datasets at hand consist of
brain grayscale MR images. The following sections describe the adaptive multi
modal image segmentation approach using level sets [9]. The curve evolution
representation is given in addition to the partial differential equation that is related
to the region information.
4.2.1. Introduction to level set segmentation
3D segmentation of anatomical structures is very important for various med-
ical applications. Due to image noise and inhomogeneities as well as the com-
plexity of anatomical structures, the segmentation process remains a tedious and
challenging one. Therefore, this process cannot rely only on image information,
but has to involve prior knowledge of the shapes and other properties of the objects
of interest. In many applications, 3D segmentation is performed using deformable
