Biomedical Engineering Reference
In-Depth Information
In practice, this rezoning procedure has proved to be highly efficient and
effective for large three-dimensional Warping problems. It has allowed for the
registration of image data sets that otherwise could not be successfully regis-
tered using Hyperelastic Warping. In the first example found below, rezoning
allowed for the successful intersubject registration of mouse brain micro-MRI
images. Analysis of these image data sets without rezoning led to incomplete
registration of the internal structure of the brain as well as incomplete external
registration of the cerebellum.
12.3
Applications
The following examples illustrate the broad range of problems that have been
analyzed using Hyperelastic Warping. The first example is an image registration
problem in which MRI images of two normal mouse brains were registered. The
second example illustrates how the results of a registration analysis of micro-CT
images of the gerbil middle ear may be used to provide the boundary conditions
for a second, traditional, FE analysis of the malleus bone. The remaining exam-
ples illustrate applications in cardiovascular mechanics.
12.3.1
Quantification of Changes in Mouse Brain
Morphology
Quantification of time-dependent changes in three-dimensional morphology of
brain structures and neural pathways is a fundamental challenge in anatomi-
cal studies of neurodevelopment and in tracking brain remodeling and/or pro-
gression of certain neurological diseases. The morphometric problem can be
approached using
in vivo
gross-scale (submillimeter) magnetic resonance med-
ical imaging (MRI) of the brain. Tracking anatomical changes
in vivo
has been
a major motivation for the development of higher resolution CT, MRI and ra-
diographic imaging systems. While it is currently routine in clinical MRI of hu-
mans to obtain 1
×
1
×
2 mm resolution, micro-MRI images of small animals
have been obtained with isotropic resolution on the order of 40
µ
m resolution
sometimes termed
magnetic resonance microscopy
(MRM). This type of MRI
data is sufficient to resolve the neuroanatomical structures of interest but it re-
mains difficult to extract quantitative structure-specific morphological measures
