|
Paul Reitsma Graduate Student Graphics Lab Computer Science Department School of Computer Science Carnegie Mellon University Pittsburgh, PA 15213, USA
Office: 3719 Wean Hall
|
|
My research focuses on ways to measure how well these two
criteria have been fulfilled, with an eye towards using that information
to create better animation systems; once we understand the tradeoffs,
we can intelligently exploit them. Ultimately, one goal is to
understand the factors involved in motion generation so well
that we can automate most of it away, creating a system simple enough
to allow a child to author animated stories, but powerful enough to drive
the animated characters in high-end games and movies.
My projects in this area have also led to work in related areas, such as
perceptual psychology and graph algorithms. The latter, in particular,
has occupied significant amounts of my time of late, and has led to some
interesting discussions with colleagues.
Research
An automated animation-generation system such as a motion graph has to
fulfill two simple requirements: it needs to be able to create
all the animations we want, and everything it creates must be up
to our standards of quality. In practice, these requirements are
not so simple; in fact, they are often conflicting;
the canonical example is motion editing, where allowing greater changes
to a motion allows greater flexibility in which animations can be
created, but typically risks inflicting correspondingly greater errors
and visual artifacts on the results.
Selected Projects
Evaluating Motion Graphs for Character Animation
Paul Reitsma, Nancy Pollard.
To appear in ACM Transactions on Graphics.
Evaluating Motion Graphs for Character Navigation
Paul Reitsma, Nancy Pollard.
In SCA 2004.
Perceptual Metrics for Character Animation
Paul Reitsma, Nancy Pollard.
In SIGGRAPH 2003.