Results - Details
Search command : Author="Τσόγκα"
And Author="Χρυσούλα"
Current Record: 7 of 9
|
Identifier |
000398257 |
Title |
Diffusion magnetic resonance imaging techniques. : applications in brain and human body |
Alternative Title |
Εφαρμογές των τεχνικών απεικόνισης διάχυσης μαγνητικού συντονισμού στον εγκέφαλο και το ανθρώπινο σώμα |
Author
|
Ιωαννίδης, Γεώργιος
|
Thesis advisor
|
Καμπάνης, Νίκος
|
Reviewer
|
Μαριάς, Κώστας
Τσόγκα, Χρυσούλα
|
Abstract |
Medical imaging is the process of creating visual representations of the interior of our
body for clinical analysis and medical intervention. Medical imaging reveals the internal
structures hidden by the skin and it is a powerful tool for population screening and
efficient patient management. Thirty years ago, Magnetic Resonance Imaging (MRI)
was introduced in clinical practice and was progressively established among the most
frequently used medical imaging modalities. Apart from conventional MR imaging,
more advanced techniques were arisen such as spectroscopy, diffusion and perfusion
MRI.
The main subject of this work is Diffusion Weighted Imaging (DWI) and is focused on
obtaining diffusion parameters from the Intravoxel Incoherent Motion (IVIM) model
which gives information about water diffusivity and tissue vascularity. In order to
extract the aforementioned parameters from the IVIM model, non-linear least squares
are used. As an iterative process, non-linear least squares need a starting point (initial
guess), a lower and an upper bound. The initial point is of crucial significance for
the convergence of the algorithm. Due to the complexity of the IVIM function (many
local minima), standard non-linear least squares fail to produce clinically reliable results
inside the natural bounds.
To this end, we present a novel mathematical framework which exploits the characteristics
of the Diffusion Weighted (DW) signal curve in order to find the appropriate
initial starting points and bounds for the non-linear least squares. Comparisons between
results obtained by the proposed and those of the standard initialization, for
the non-linear least squares, results in clinically relevant and reliable results with our
method, in contrast to the standard initialization which suffers from reproducibility of
the results.
|
Language |
English |
Issue date |
2015-11-20 |
Collection
|
School/Department--School of Sciences and Engineering--Department of Mathematics and Applied Mathematics--Post-graduate theses
|
|
Type of Work--Post-graduate theses
|
Permanent Link |
https://elocus.lib.uoc.gr//dlib/a/2/9/metadata-dlib-1452668040-595361-10781.tkl
|
Views |
543 |