Joe Marks
MERL--Mitsubishi Electric Research Laboratories
Cambridge, Massachusetts, U.S.A.
Home page
Tangible Interaction + Graphical Interpretation:
A New Approach to 3D Modeling
OR
How to Build Virtual Worlds With Blocks and Clay
Computer Science Seminar
Date and Time: Friday, April 19th at 12:00
Venue: 2nd floor Oriel House lecture room, 2.09 , Trinity College (see map).
Artists using standard 3D modeling packages must specify precisely
the geometric and material properties of the models they create,
and therein lies much of the complexity and tedium of using those
tools. By contrast, children playing with construction toys
like Lego and K'nex make simple models easily, and use their
imaginations to fill in the details. We would like to transform
computer-based geometric modeling into that same kind of playful,
tactile experience but without sacrificing the ability to create
the interesting geometric detail and movement that make 3D graphics
and animation compelling. To retain the tactile experience of
model manipulation, we look to tangible-interface technology;
and to create detailed, fully realized models, we use new methods
for graphically interpreting a nascent model by recognizing and
augmenting its salient features.
This combination of tangible interaction and graphical
interpretation is investigated in a pair of case studies.
Tangible modeling can be approached in two ways: either by
directly instrumenting the modeling medium with embedded
computation or by using external sensors to capture the geometry.
Our first system consists of computational building blocks
assembled into physical structures that in their aggregate
determine and communicate their own geometric arrangement.
A rule-based system interprets these structures as buildings,
parses their architectural features, then adds geometric detail
and decorative enhancements automatically. Our second system
uses simple and robust computer vision to capture volumetric
scans of clay models of such common toy-like objects as
people, animals, trees, houses, cars, and boats. A volumetric
matching algorithm allows us to recognize, interpret, and
bring the clay models to (virtual) life.
Joint work with: D. Anderson, J. Frankel, A. Agarwala, P. Beardsley,
J. Hodgins, D. Leigh, K. Ryall, E. Sullivan, and J. Yedidia.
Speaker bio:
Joe Marks grew up in Dublin, Ireland, before coming to the U.S. for
college. He earned a Ph.D. in Computer Science from Harvard University
in 1991. He has worked previously at Bolt Beranek and Newman and
at Digital's Cambridge Research Laboratory. He is currently the
Director of MERL's Cambridge Research Lab. He is also the Chair
of ACM SIGART and an Associate Editor for ACM's Transactions
on Graphics. His areas of interest include computer graphics,
human-computer interaction, and artificial intelligence.
|
