|
VGTU talpykla >
Doktorantūros skyrius / Department for Doctoral Studies >
Technologijos mokslų daktaro disertacijos ir jų santraukos >
Please use this identifier to cite or link to this item:
http://dspace.vgtu.lt/handle/1/1545
|
| Title: | Computer Aided Analysis of Brain in Magnetic Resonance Images |
| Other Titles: | Galvos smegenų magnetinio rezonanso vaizdų kompiuterinė analizė |
| Authors: | Rokicki, Jaroslav |
| Issue Date: | 2012 |
| Publisher: | VGTU leidykla „TECHNIKA“ |
| Citation: | Rokicki, J. 2012. Computer Aided Analysis of Brain in Magnetic Resonance Images: doctoral dissertation. Vilnius: Technika. 144 p. |
| Description: | In this work magnetic resonance imaging (MRI) based image processing
methods for the automatic detection and prevention of dementia are developed.
MRI a powerful, non-invasive hardware technology used for number of
tasks, one of them is to screen and visualize the brain of patients with diagnosed
dementia. Currently, there is no widely accepted automatic signal or
image processing methods for the Alzheimer’s disease detection or description
of the blood vessel quality in medical routines. Development of reliable
and automatic MRI digital image analysis methods can reinforce the disease
diagnosis based on cognitive scores and help to start patient treatment earlier.
The goal of this work is to develop digital image processing methods suitable
for the early detection of dementia from the MRI scans of the brain.
Thesis consists of introduction followed by 4 chapters: review of skeletonization
and Alzheimer’s disease detection methods, blood vessel skeletonization,
materials and methods for Alzheimer’s disease and results of automatic
classification. In the Introduction the research object, main goal and tasks, scientific
novelty, research object, practical significance and defended statements
are presented.
Chapter 1 reviews the most recent literature about skeletonization and
automatic Alzheimer’s disease detection methods. Chapter is concluded by
defining the problems, which are being solved in this PhD thesis. Chapter 2
deals with the blood vessel segmentation methods. Two new skeleton extraction
methods were proposed. First one, called step-wise, is based on the stepwise
traversing along the longest ray through the blood vessel tree. Second one,
called kernel-based was expanded in this work to 3D space. Algorithms quality
was discussed, qualified and compared against manually extracted skeleton for
the blood vessel models. Moreover, all the algorithms were compared against
iterative thinning method. Chapter 3 describes used Alzheimer’s subjects data
together with its analysis methods. Second part of this chapter describes the
cross sectional and longitudinal changes in the brain among different groups of
subjects. Chapter 4 discusses the robustness of the MRI based disease markers.
10 cortical and subcortical MRI based volumetric markers were investigated.
MRI based markers were compared against the cognitive scores (MMSE test),
also influence of subject’s age was investigated. In the second part of the chapter
automatic classification results were discussed. Work is summarized by
general conclusions chapter.
There are 3 appendixes attached to the thesis. Thesis consists of 124 pages,
with 34 figures, 27 tables and 117 references. |
| URI: | http://dspace1.vgtu.lt/handle/1/1545 |
| ISBN: | 978-609-457-391-0 |
| Appears in Collections: | Technologijos mokslų daktaro disertacijos ir jų santraukos
|
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.
|
