CPS Seminar Speaker David Lowe SEA 4.244

Contact Name: 
Jenna Whitney
Date: 
Apr 28, 2008 12:00pm - 1:00pm

Speaker/Affiliation: David Lowe Ph.D. Professor
Computer Science Department University of British Columbia

When/Lo

cation: 4/28/2008 12:00 PM SEA 4.244

Title of Talk: ''Object Recog

nition from Invariant Local Features''

Reception with Refreshments a

t 11:30 AM

Abstract: Human vision is so powerful that we seldom give
a second thought to our ability to immediately identify the objects in our
surroundings. However the problem of object recognition has proven to be

very challenging for computer vision. A major source of the difficulty is t

he large range of variations in appearance that may occur due to factors s

uch as changes in 3D viewpoint varying illumination partial visibility a

nd background clutter. Fortunately there has been rapid progress on this p

roblem within the past few years through the use of invariant local feature
matching. Thousands of these local features can be extracted from an image
to describe small overlapping regions and each feature is designed to be

invariant to a range of image transformations such as changes in scale or

ientation brightness and local deformations. Furthermore the features are
designed to be very distinctive so that a single feature can be used to s

elect a correct match from a large database. Fast methods for nearest-neigh

bor access in high dimensional spaces allow large databases of features to

be matched in real time. This talk will present an overview of the invarian

t feature approach as well as some recent applications such as location re

cognition and automated stitching of digital images into panoramas.