Applied Math Seminar
Fall Quarter 2008
Friday September 26, 3:15p.m.Deanna Needell
Department of Mathematics
University of California at Davis
Greedy Signal Recovery in Compressed Sensing
Abstract: When acquiring a signal with the conventional methodology, the entire signal is first obtained, and only after this costly step is it compressed. This scheme spends a high amount of computational energy acquiring a signal only to throw most of it away immediately. This seemingly wasteful approach to signal acquisition has sparked the field of Compressed Sensing, which has provided many algorithms to improve upon the conventional method. Two distinct approaches to the problem arose, both providing their own advantages and disadvantages. I discuss these methods as well as two new algorithms, ROMP and CoSaMP, which bridge the gap between the two major approaches in Compressed Sensing.
Seminar website: http://math.stanford.edu/~applmath/