Technical Papers

Data-Driven Animation

Tuesday, 23 July 3:45 PM - 5:35 PM
Session Chair: Jinxiang Chai, Texas A&M University

Non-Polynomial Galerkin Projection on Deforming Meshes

This paper presents an efficient extension of Galerkin projection to a large class of non-polynomial functions and demonstrates the broad applicability of the approach by applying it to two strikingly different problems: fluid simulation and radiosity rendering.

Matt Stanton
Carnegie Mellon University

Yu Sheng
Carnegie Mellon University

Martin Wicke
Otherlab

Federico Perazzi
Carnegie Mellon University

Amos Yuen
Carnegie Mellon University

Srinivasa Narasimhan
Carnegie Mellon University

Adrien Treuille
Carnegie Mellon University

Near-Exhaustive Precomputation of Secondary Cloth Effects

Using over several thousand hours to perform a massive exploration of the space of secondary clothing effects on a character animated through a large motion graph, this method successfully samples the complex dynamical space to low visual error and compresses the resulting dataset to enable real-time animation.

Doyub Kim
Carnegie Mellon University

Woojong Koh
University of California, Berkeley

Rahul Narain
University of California, Berkeley

Kayvon Fatahalian
Carnegie Mellon University

Adrien Treuille
Carnegie Mellon University

James O'Brien
University of California, Berkeley

Modeling Friction and Air Effects between Cloth and Deformable Bodies

Data-driven models to improve the simulation of cloth-body interactions from three perspectives: collision, friction, and air pressure. The resulting system allows efficient and accurate simulation of deformable objects with a three-layer structure such as pillows, comforters, down jackets, and stuffed toys.

Zhili Chen
The Ohio State University

Huamin Wang
The Ohio State University

Renguo Feng
The Ohio State University

Flow Reconstruction for Data-Driven Traffic Animation

A novel algorithm to reconstruct and display continuous virtual traffic flows based on discrete spatio-temporal traffic-sensor data for a dynamic virtual environment.

David Wilkie
University of North Carolina at Chapel Hill

Jason Sewall
Intel Corporation

Ming Lin
University of North Carolina at Chapel Hill

Dynamic-Element Textures

Dynamic-element textures describe phenomena that are repetitive in both shapes and motions, such as particles, threads, and sheets. They are often hard to control during authoring. This method facilitates controls at coarse scales through spatial-temporal constraints and fine scales through input exemplars. It combines constrained optimization with data-driven computation.

Chongyang Ma
The University of British Columbia, Tsinghua University

Li-Yi Wei
The University of Hong Kong, Microsoft Research

Sylvain Lefebvre
INRIA

Xin Tong
Microsoft Research Asia