Character Animation I

Monday, 6 August
8:30 - 10:15 am
Room 6DE
Session Chair/Discussant: Michiel van de Panne, The University of British Columbia

Active Learning for Real-Time Motion Controllers

An active learning approach to real-time controllable motion. The approach builds the controller in an interactive capture session and optimizes for sample locations.

Seth Cooper
University of Washington

Aaron Hertzmann
University of Toronto

Zoran Popović
University of Washington

Responsive Characters From Motion Fragments

A data-driven animation controller designed for on-line, direct character-control applications that achieves good results by modeling user behavior.

James McCann
Nancy Pollard
Carnegie Mellon University

Optimal Character Animation With Continuous User Control

A real-time character animation system that produces fluid, near-optimal motion under continuously changing multidimensional user control.

Adrien Treuille
Yongjoon Lee
Zoran Popović
University of Washington

Constraint-Based Motion Optimization Using a Statistical Dynamic Model

A method for generating human animation from a variety of spatial-temporal constraints using a low-dimensional, statistical-dynamical model (from motion capture data) as a motion prior in a trajectory optimization framework.

Jinxiang Chai
Texas A&M University

Jessica K. Hodgins
Carnegie Mellon University