Campuses:

Multi-dimensional

Wednesday, November 9, 2005 - 2:00pm - 3:00pm
Robert Levis (Temple University)
We propose a new approach to classical detection problem of discrimination
of a true signal from an interferent signal. We show that the detection
performance, as quantified by the receiver operating curve (ROC), can be
substantially improved when the signal is represented by a multi-component
data set that is actively manipulated by a shaped probing pulse. In this
case, the signal sought (agent) and the interfering signal (interferent) are
visualized by vectors in a multi-dimensional detection space. Separation of
Subscribe to RSS - Multi-dimensional