UTCS Colloquium/AI - Vittorio Ferrari/Swiss Federal Institute of Technology Zurich "Visual Learning with Generic Knowledge", ACES 2.302

Contact Name: 
Jenna Whitney
Date: 
Dec 3, 2010 11:00am - 12:00pm

There is a sign-up schedule for this event that can be found at

http://www.cs.utexas.edu/department/webevent/utcs/events/cgi/list_event

s.cgi

Type of Talk: UTCS Colloquium/AI

Speaker/Affiliation: Vitto

rio Ferrari/Swiss Federal Institute of Technology Zurich

Date/Time: Fr

iday, December 3, 2010, 11:00 a.m.

Location: ACES 2.302

Host:

Kristen Grauman

Talk Title: Visual Learning with Generic Knowledge

Talk Abstract:
The dream of computer vision a machine capable of interp

reting images of complex scenes. Central to this goal is the ability to rec

ognize objects as belonging to classes and to localize them in the images.

In the traditional paradigm, each class is learned starting from scratch,
typically from training images where the location of objects is manually a

nnotated (fully supervised setting). In this talk, I will present a scenar

io where knowledge generic over classes is first learned from images of var

ious classes with given object locations and then employed to support learn

ing any new class without location annotation. Generic knowledge provides a
strong basis which facilitates learning in this weakly supervised setting.
This strategy enables learning from challenging images containing extensiv

e clutter and large scale and appearance variations between object instance

s. In turn, this opens the door to learning a large number of classes with
little manual labelling effort.

Speaker Bio:
Vittorio Ferrari is an
Assistant Professor at the Swiss Federal Institute of Technology Zurich (E

THZ). After receiving his PhD from ETHZ in 2004, he was a post-doctoral re

searcher at INRIA Grenoble and the University of Oxford. His research inter

ests are in visual learning, with emphasis on reducing the amount of manua

l effort needed to learn visual concepts.