High-dimensional pattern analysis and classification: Diagnostic tools based on imaging patterns
Speaker: Christos Davatzikos , University of Pennsylvania
Date: November 4 2010
Time: 3:00PM to 4:00PM
Location: 32-D507
Host: Polina Golland, CSAIL
Contact: Polina Golland, x38005, polina@csail.mit.edu
Relevant URL: This talk will discuss the use of high-dimensional pattern
classification methods for quantifying and detecting brain imaging
patterns. The challenges posed by the high-dimensionality of imaging
data, in conjunction with limited sample sizes typically available will
be discussed, leading to work on feature extraction and selection. The
main focus will be the analysis of patterns of brain atrophy in
Alzheimer's Disease, schizophrenia and normal aging, and results using
standard support-vector machines, as well as transductive SVM's and
hybrid classification/clustering methods will be presented. The ultimate
goal of such methods is to provide more specific and sensitive
diagnostic and prognostic tools.
See other events that are part of Biomedical Imaging and Analysis 2010/2011
See other events happening in November 2010
