Michael L. Gleicher

Assistant Professor

Computer Sciences Department
University of Wisconsin
1210 W. Dayton St.
Madison, WI 53706-1685

telephone: (608) 262-1204
fax: (608) 262-9777
email: gleicher@cs.wisc.edu
http://www.cs.wisc.edu/~gleicher/

Research Summary

The overall goal of my research is to create tools that make it easier to create pictures, video, animation, and virtual environments; and to make these visual artifacts more interesting, entertaining, and informative. My recent work has explored editing motions for animation, creating animation from video examples, and using image analysis techniques to automate the creation of visual effects in video. Methodologically, I like to build interactive systems and devise methods based on a mathematical core. While this work is predominantly in the area of computer graphics and animation, it often crosses the boundaries into computer vision and user interfaces.

One theme in my work is that it may be easier to adapt some existing animation or video than to create one from scratch. This will require new tools that are better tuned to editing, rather than creating. For example, most techniques used in computer animation results in motions that are single purpose: they apply to a specific character, performing a specific action in a specific environment. I have been working on methods for altering motions so they can be re-used: adapting the walk of a tall character so it applies to a shorter character, or adapting the motion of a character picking up an object to a different object location.

Recently, I have been addressing these Animation editing tasks with Spacetime Constraint methods. These methods consider entire animated motions in a single numerical problem, in contrast to most previous methods that consider each frame independently. By considering more information, Spacetime Constraint methods allow us to edit motions while preserving the characteristics of the original motion, and to construct consisted motions based on incomplete observations. While the Spacetime Constraint approach requires solving large nonlinear constrained optimization problems, care in how these problems are posed and solved can make the methods practical for real problems.

Sample Recent Publications

Projective registration with difference decomposition, Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), San Juan, June 1997.

Motion editing with spacetime constraints, Symposium on Interactive 3D Graphics, Providence, April 1997.