The reliability of the complete FE analysis package depends on the precise
delineation of objects of interest and their associated properties derived
from multimodal imaging data. The goal is to
create a comprehensive image processing toolbox to provide optimal and
automatic segmentation (some basics about image segmentation can be found
here) and registration (an introduction into
registration techniques
by Prof. Vandermeulen can be found
here) protocols to the clinical end-user.
The imaging data first has to be segmented into the different anatomically
defined tissues, i.e. materials have to be detected and defined in the raw
dataset. Segmentation schemes available within SimBio will range from
thresholding techniques to atlas-based segmentation approaches.
The provided algorithms will
allow to process image data of different modalities such as magnetic
resonance imaging (MRI) and computer tomography (CT).
Image datasets acquired with different imaging techniques or at different
times will be geometrically co-registered first using a voxel-similarity
based rigid registration algorithm . The implemented algorithms will also
be tested and compared with regard to stability, output quality and
performance.