|
|
Peter Ennis
ennisp02@csse.uwa.edu.au
Entry year: 2004
Enrolment status: confirmed
Degree: BCM
Degree status: complete semester 2, 2004
Project: Segmentation of the Hippocampus from MR Images
Supervisor(s): Mohammed Bennamoun
Project status: complete semester 2, 2004
|
|
The interest and research into medical imaging has grown dramatically in the past few years largely due to the increase in computer power and the subsequent development of application environments that allow developers and medical experts to automate the analysis of the multidimensional medical images.
An important part of research in medical imaging is the classification or segmentation of structures in the brain for medical diagnosis.
One of the structures of interest in the brain is the hippocampus, an internal part of the brain located inside the temporal lobe.
The hippocampus plays an important part in the brain and its volume appears to be affected by some chronic diseases such as epilepsy, schizophrenia, and Alzheimers.
In this project, we present different techniques used to segment the hippocampus.
SNAP, a powerful, geodesic level-set segmentation tool, has been found to be a useful tool to be used for hippocampus segmentation.
This tool has been in daily use by clinicians to segment the local structure such as the ventricle and amygdala in the human brain.
This tool has been found to be a reliable tool to segment the hippocampus, except in an area around the hippocampus called HATA.
Several possible solutions have been proposed to solve HATA problem.
A recent article on normalized convolution, which makes use of the signal/certainty principles, is one of the proposed techniques to solve this problem.
|
 Proposal
Dissertation
|
Last update: Mon Mar 7 21:24:34 2005
|
|
