Return to Colloquia & Seminar listing
Don't be perturbed; come to my talk.
Student-Run Research| Speaker: | Matt Herman, UC Davis |
| Location: | 2112 MSB |
| Start time: | Wed, Feb 18 2009, 12:10PM |
Description
The area of signal processing and information theory has recently experienced
a paradigmatic shift due to the theory and techniques of Compressed
Sensing (CS). In the ideal case, a signal is encoded in the form of $y =
Ax$, while in the case of additive noise we have $y = Ax + e$. Special
non-linear methods such as L_1 minimization are often used to recover the
original data x.
In this talk I will review the basics of CS theory, and will present new
results of a general perturbed model of the form: $y = (A+E)x + e$. This
model accounts for multiplicative noise as well as additive noise.
This area of mathematics relies on simple linear algebra, spectral theory,
and harmonic analysis.