Automatic Multiscale Segmentation of MRI
Speaker: Ayelet Akselrod-Ballin , Weizmann Institute of Science
Date: April 11 2008
Time: 2:00PM to 3:00PM
Location: 32-D507
Host: Polina Golland, CSAIL
Contact: Polina Golland, x38005, polina@csail.mit.edu
Relevant URL: Identification and extraction of precise quantitative information for normal
and abnormal anatomical structures in medical images is essential for
improving diagnosis, follow-up, and therapy and for advancing biomedical
research. In this talk I will present an automatic multiscale segmentation
and classification methodology for medical images, derived from algebraic
multigrid.
I will describe two multiscale segmentation frameworks. The first is an
integrated segmentation and classification approach, which combines
multiscale segmentation with a second classification stage based on a
comprehensive set of features describing the segments. The second approach
incorporates probabilistic information into the multiscale segmentation
framework by a Bayesian formulation. The performance of the proposed
algorithms is demonstrated on a variety of medical imaging problems
including segmentation of Multiple Sclerosis lesions and identification of
brain anatomical structures.
See other events that are part of Biomedical Imaging and Analysis 2007/2008
See other events happening in April 2008
