A fast view of real life video segmentation and a slower<br/><br/>view of learning dictionaries for efficient representations

Tuesday, October 6, 2009 - 9:45am - 10:15am
Lind 305
Guillermo Sapiro (University of Minnesota, Twin Cities)
After spending about 5 minutes showing recent results on
video segmentation (joint work with Adobe), I will describe some
recent works in my group in the area of dictionary learning and sparse coding.
In particular I will present new models derived from information theory,
new models dedicated to go beyond standard sparse coding applications and
into unsupervised clustering, and
new models related to compressed sensing.
MSC Code: