Speaker:
Assistant Professor Wotao Yin
Institution:
Rice University
Time:
Monday, March 3, 2008 - 4:00pm
Location:
MSTB 254
This talk will introduce compressed sensing through examples
in MATLAB and medical imaging. The compressed sensing
technique allows one to acquire a signal from a much fewer number of
measurements than what is usually necessary.
A main step in compressed sensing is solving a nonsmooth optimization
problem for a sparse solution. Although the solution is expected to be
sparse, the data isoften of an extremely large scale. We outline the
numerical difficulties and introduce new L1-based algorithms.