|
Simon Baker
Robotics Institute
Carnegie Mellon University
Face Recognition Across Pose: A Database, an Evaluation, an Algorithm, and Some Theory
Face recognition across pose is the task of recognizing someone when the
probe and gallery images have different poses. For example, the gallery
(i.e. training) images might be frontal mug-shots, and the probe (i.e.
testing) images might be 3/4 views. More generally, the gallery and
probes might each consist of several images with disjoint poses. For
example, the gallery might consist of a frontal and a profile view, like
in police photographs, and the probe might consist of a couple of
intermediate shots. How do we combine information from multiple distinct
views to improve face recognition?
In this talk I will first describe a database we collected to investigate
the task of face recognition across pose. Next, I will present empirical
results evaluating several face recognition algorithms on this database,
including one of the most successful commerical systems. I will then
describe an algorithm for performing face recognition across pose using
a concept known as "eigen light-fields." Finally, I will present a
theoretical analysis of the information content of the light-field of a
face.
Bio: Simon Baker is a Research Scientist in the Robotics Institute at
Carnegie Mellon University, where he conducts research in Computer
Vision. Before joining the Robotics Institute in September 1998, he
was a Graduate Research Assistant at Columbia University, where he
obtained his Ph.D. in the Department of Computer Science. He also
spent a summer visiting the Vision Technology Group at Microsoft
Research. His current research focuses on a wide range of Computer
Vision problems from stereo reconstruction and the estimation of 3D
scene motion, to illumination modeling and sensor design. His work has
appeared in a number of international Computer Vision conferences and
journals. He received his B.A. in Mathematics from the University of
Cambridge in June 1991, his M.Sc. in Computer Science from the
University of Edinburgh in November 1992, and his M.A. in Mathematics
from the University of Cambridge in February 1995.
|
