BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:DIMENSIONALITY REDUCTION and FEATURE SELECTION IN HYPERSPECTRAL IM
 AGE CLASSIFICATION - Joel Trussell\, North Carolina State University
DTSTART:20070809T120000Z
DTEND:20070809T133000Z
UID:TALK7784@talks.cam.ac.uk
CONTACT:Taylan Cemgil
DESCRIPTION:Hyperspectral images provide a vast amount of information abou
 t a scene.\nHowever\, much of that information is redundant as the bands a
 re highly correlated. For computational and data compression reasons\, it 
 is desired to reduce the dimensionality of the data set while maintaining 
 good performance in image analysis tasks. We present a method of dimension
 ality reduction based on neural networks that uses a novel penalty functio
 n to successfully reduce the number of active neurons\, which corresponds 
 to the dimensionality of the data for the task of interest. This method ca
 n be extended to select the best features from an arbitrary feature set\, 
 where "best" is defined in terms of reduction of the number of features wh
 ile maintaining a desired level of classification performance. \n\n
LOCATION:LR6\, Engineering\, Department of
END:VEVENT
END:VCALENDAR