Differential Geometric Algorithms for Vision and Image Processing - Talk by Frank C. Park (Seoul National University)

Host:  Dinesh Pai

Location:  X836 Board Room

Abstract:
Differential geometric methods have now entered the mainstream
of computer vision and image analysis research, but for most
practical applications their advantages vis-a-vis existing methods
have only been demonstrated in limited, narrowly defined settings.
In this talk we present a collection of geometrically motivated
vision and image processing algorithms that, taken together, show
that for a wide range of applications, geometric methods are more
robust and accurate (and sometimes more efficient) than existing methods.
We present geometric optimization algorithms on the rotation and
Euclidean groups that solve a wide range of calibration and registration
problems, from robotic hand-eye calibration and simultaneous
visual-inertial sensor calibration to multimodal image registration.
We show that visual tracking can be formulated as a particle
filtering problem on the affine group. We introduce a
Riemannian distance metric on the space of multivariate normal
distributions and show that it can be effectively used for
DT-MRI image segmentation, and classification problems in a more
general machine learning context. The common thread among these
various approaches is the use of differential geometric methods
to formulate algorithms that are coordinate-invariant; with only
a modest increase in computation, robustness and performance can
often be enhanced considerably.

Bio:
Frank Chongwoo Park received his B.S. in Electrical Engineering
from MIT in 1985, and Ph.D. in Applied Mathematics from Harvard
University in 1991. From 1991 to 1995 he was assistant professor
of mechanical and aerospace engineering at the University of
California, Irvine. Since 1995 he has been professor of mechanical
and aerospace engineering at Seoul National University. From
2009-2012 was an adjunct professor in the Department of Interactive
Computing at the Georgia Institute of Technology, and in 2001-2002
was a visiting professor at the Courant Institute of Mathematical
Sciences at New York University. His research interests are in
robot mechanics, planning and control, visual tracking, and related
areas of applied mathematics. In 2007-2008 he was an IEEE Robotics
and Automation Society (RAS) Distinguished Lecturer, and has served
as secretary of RAS from 2009-2010 and 2012-2013. He has served
as senior editor of the IEEE Transactions on Robotics, an area editor
of the Springer Handbook of Robotics and Advanced Tracts in Robotics
(STAR), and as an associate editor of the ASME Journal of Mechanisms
and Robotics. He is a fellow of the IEEE, and incoming editor-in-chief
of the IEEE Transactions on Robotics.