Atlas-based Segmentation in Brain Magnetic Resonance Images

PhD Qualifying Examination


Title: "Atlas-based Segmentation in Brain Magnetic Resonance Images"

by

Miss Siqi BAO


Abstract:

Segmentation of brain Magnetic Resonance (MR) images plays a significant 
role in disease diagnosis, surgery planning and therapy assessment. 
However, manual labeling is time-consuming and can suffer from inter- and 
intra- labeler inconsistencies. A variety of approaches has been proposed 
to obtain the segmentation result semi-automatically or fully 
automatically and atlas-based methods get popular as a result of the 
relatively good performance.

In this survey, we present a comprehensive review about the atlas-based 
segmentation for brain MR images. Three specific issues are discussed in 
the pair-wise segmentation, consisting of rigid and non-rigid 
registration, intensity and shape priors, general framework for target 
image segmentation. The pair-wise registration can be utilized for 
multi-atlas based segmentation, assisted with atlas selection and label 
fusion procedures. Recent works on group-wise segmentation have also been 
investigated intensively and extensively.

For performance evaluation, we present a brief introduction of the 
commonly used data sets and evaluation metrics. Three state-of-the-art 
atlas-based segmentation methods have been selected and experiments on two 
data sets have been carried out to evaluate their performances.


Date:			Friday, 9 May 2014

Time:                  	10:00am - 12:00noon

Venue:                  Room 5501
                         Lifts 25/26

Committee Members:	Dr. Albert Chung (Supervisor)
  			Prof. Ting-Chuen Pong (Chairperson)
  			Dr. Huamin Qu
  			Prof. Chi-Keung Tang


**** ALL are Welcome ****