Biomedical Engineering Reference
In-Depth Information
tion MRI data will need sophisticated registration techniques to perform data fusion.
In the case of elastic registration, deformations are generally constrained to handle
the large number of degrees of freedom [
38
,
39
]. For internal forces, we may need
physical and statistical parameters [
40
,
41
] for better reconstruction. Such a devel-
opment should take advantage of existing expertise in deformable models [
42
-
45
],
to reconstruct 3D models from volumetric MRI image and image segmentation.
Another important component is 3D visualization which can help to validate the
result of 3D segmentation [
46
]. Different approaches to 3D visualization of MRI data
segmentations exist. Volumetric methods give a good overall picture of the data set.
However, they often appear to be confusing and lacking fine details. Partial occlu-
sion and visual clutter that typically result from the overlay of these traditional 3D
scalar-field visualization techniques make it difficult for users to perceive and recog-
nize visual structures. Surface rendering could be a good alternative but some tissue
surface is usually not directly available in the original 3D volume MRI data. Hence,
an efficient 3D visualization framework suitable for segmentation especially incor-
porating advanced 3D display technologies is a significant research direction and
poses many developmental challenges.
1.4 Building Computational Biomechanics Models for
Deforming Hard and Soft Musculoskeletal Tissues
It is important to note that the reliability of computational approaches strongly
depends on appropriate 3D mathematical descriptions of the material behavior of
soft tissues and their interactions with surrounding structures. Reasonable constitu-
tive models must be designed within the context of associated comprehensive exper-
imental data for tissue, cellular, and molecular structures. Computational models
offer, however, the potential to simulate multi-field coupled processes encountered
in the micro-heterogeneous soft tissues and to realistically predict physiological
functional interactions. Biomechanical simulation has to consider the interaction
between the medical device and biological tissue which is deformable and divisible.
The large deformation and/or puncture of the tissue can be simulated realistically
with a physics-based simulation engine that integrates finite element and multi-body
dynamics codes [
47
]. The simulation engine assumes that the global operating space
is divided intomany sub-regions. Quantification of tissue biomechanics during defor-
mation is often restricted to excised tissue and using conventional mechanical testing
techniques. Despite efforts to recreate in vivo conditions during mechanical testing,
the testing environment is far from physiological, with grossly approximated bound-
ary conditions [
48
]. Image-based quantification will enable deformation analysis to
be carried out in the physiological environment and will be entirely non-invasive,
preserving tissue-tissue interfaces, providing realistic biomechanical responses.
A combination of image processing techniques, together with finite element sim-
