Technical Papers

Full-Body Character Animation

Friday, 30 November 09:00 - 10:45 |  Peridot 206

Generative Statistical Models for Semantic Motion Analysis and Synthesis - Picture

Generative Statistical Models for Semantic Motion Analysis and Synthesis

We introduce a new generative statistical model for semantic motion analysis and synthesis. We have demonstrated its power by exploring a wide variety of applications, ranging from automatic motion segmentation, recognition, and annotation, online/offline motion synthesis in both kinematics and behavior level, performance-based animation to semantic motion editing.


Jianyuan Min, Texas A&M University
Jinxiang Chai, Texas A&M University
Terrain Runner: Control, Parameterization, Composition, and Planning for Highly Dynamic Motions - Picture

Terrain Runner: Control, Parameterization, Composition, and Planning for Highly Dynamic Motions

We present methods for the control, parameterization, composition, and planning for highly dynamic motions. More specifically, we learn the skills required by real-time physics-based avatars to perform parkour-style fast terrain crossing using a mix of running, jumping, speed-vaulting, and drop-rolling.


Libin Liu, Tsinghua University
KangKang Yin, National University of Singapore
Michiel van de Panne, University of British Columbia
Baining Guo, Microsoft Research Asia
Falling and Landing Motion Control for Character Animation - Picture

Falling and Landing Motion Control for Character Animation

Landing safely is one of the most fundamental skills in many high dynamic activities such as martial arts, acrobatics and free running. We develop a physics-based landing controller that allows the character to fall from a wide range of initial conditions without inducing a large amount of joint stress.


Sehoon Ha, Georgia Institute of Technology
Yuting Ye, Industrial Light & Magic
C. Karen Liu, Georgia Institute of Technology
Synthesis of Concurrent Object Manipulation Tasks - Picture

Synthesis of Concurrent Object Manipulation Tasks

We introduce a physics-based technique to synthesize human activities involving concurrent full-body manipulation of multiple objects. Given an environment map and manipulation graphs, our algorithm generates a continuous animation of the character manipulating multiple objects and environment features concurrently at various locations in a constrained environment.


Yunfei Bai, Georgia Institute of Technology
Kristin Siu, Georgia Institute of Technology
Karen Liu, Georgia Institute of Technology