Modern Computer Vision, Augmented Reality, and Interactive Modeling

                    Chris Brown
             University of Rochester

In an augmented reality display, computer graphics are appropriately
mixed with live video or superimposed on a live view.  One of the
crucial problems is how to register the graphics with the view of the
real world.  Most popular are approaches that rely on calibration and
subsequent accurate measurement of geometry of interest -- for
instance the position and orientation of the live video camera.  This
approach uses the familiar Euclidean frame, which has many conceptual
and computational advantages.  However, in some situations calibration
is a very difficult problem.

Two techniques of current interest in the computer vision community
are non-Euclidian (affine or projective) object representations and
real-time feature tracking.  Together, they allow the registration
problem to be solved without prior calibration, and without a known
calibration object in the scene.

The theory is illustrated with current work involving augmented
reality overlays, animations, and visual input of solid models.

----------------------------------

Christopher Brown (BA Oberlin 1967, PhD U. Chicago 1972) is Professor
of Computer Science at the University of Rochester, his home since
finishing a postdoctoral fellowship at the School of Artificial
Intelligence at the University of Edinburgh in 1974.  He is coauthor
of COMPUTER VISION with his Rochester colleague Dana Ballard.  He
spent two years heading the software development of the PADL-2 solid
modeling package from the Production Automation Project at Rochester.
His current research interests are computer vision and robotics,
especially the interaction of visual and cognitive capabilities and
motor behavior.  This interest leads to involvement in real-time
operating and control systems, real-time vision algorithms, and in the
basic connections between planning, learning, and control.  Many of
the same real-time vision issues can be used in augmented reality
applications.