Disney Research at SIGGRAPH Asia 2012

Seven papers from the Disney Research teams were accepted to the Technical Papers and/or the Technical Briefs sessions at the SIGGRAPH Asia 2012 conference in Singapore, November 28th-December 1st. The research areas in which our work was presented include: Physical Models, Dynamics, Video & Image Manipulation, Stereo & Displays, Motion Capture & Synthesis, Global Illumination, and Visual Environment & Interaction.

Additional information can be found below:

Technical Papers

Physical Models

Thursday, 29 November
09:00 – 10:45
Location: Peridot 206

Chopper: Partitioning Models into 3D-Printable Parts

We propose a framework, called Chopper, to print a large 3D object by decomposing it into smaller parts so that each part fits into the printing volume and satisfy a number of objectives including assemblability, having few components, unobtrusiveness of the seams, and structural soundness. This project was developed in collaboration with Linjie Luo (Princeton University), Ilya Baran (Disney Research Zurich), Szymon Rusinkiewicz (Princeton University), Wojciech Matusik (Massachusetts Institute of Technology)

 

Dynamics

Thursday, 29 November
16:15 – 18:15
Location: Peridot 201

 

Speculative Parallel Asynchronous Contact Mechanics

We focus on provably correct simulations and show how to achieve a 200-600x speedup of Asynchronous Contact Mechanics. We replace expensive kinetic data structures with a rollback mechanism for handling collisions. This change alone provides a significant speedup, but it also enables force computation to be easily parallelized. This project was developed in collaboration with Samantha Ainsley (Columbia University), Etienne Vouga (Columbia University), Eitan Grinspun (Columbia University), and Rasmus Tamstorf, (Disney Animation Research Group)

 

Video and Image Manipulation

Friday, 30 November
14:15 – 16:00
Location: Peridot 201

Transfusive Image Manipulation

We present a method for consistent automatic transfer of edits on one image to many other images of the same subject. We introduce novel, content-adaptive weight functions that enhance the Lucas-Kanade non-rigid alignment framework. Our method is robust to changes in view-point, illumination and subject arrangement. This project was developed in collaboration with Kaan Yücer (ETH Zurich), Alec Jacobson (ETH Zurich), Alexander Hornung (Disney Research, Zurich), Olga Sorkine (ETH Zurich)

 

Stereo and Displays

Friday, 30 November
16:15 – 18:15
Location: Peridot 201

The Magic Lens: Refractive Steganography

We construct Magic Lenses, composed of refractive lenslet arrays, to reveal hidden images when placed over potentially unstructured printed/displayed source images. We show how to optimize for the shape of these lenses to ensure manufacturability and demonstrate the technique using a number of physically fabricated lenses. This project was developed in collaboration with Marios Papas (ETH Zurich / Disney Research Zurich), Thomas Houit (ETH Zurich), Derek Nowrouzezahrai (Disney Research Zurich / University of Montreal), Markus Gross (ETH Zurich / Disney Research Zurich), Wojciech Jarosz (Disney Research Zurich)

 

Motion Capture and Synthesis

Saturday, 01 December
09:00 – 10:45
Location: Peridot 206

Data-driven Finger Motion Synthesis for Gesturing Characters

Creating compelling finger motions is a challenging and time-consuming process. Our method automatically adds detailed finger movements to the body motions of gesturing and conversing characters. We locate suitable finger motion segments from a database based on similarity of the arm motions and smoothness of the reconstructed finger motions. This project was developed in collaboration with Sophie Jörg (Carnegie Mellon University, Clemson University), Jessica Hodgins (Carnegie Mellon University, Disney Research, Pittsburgh), Alla Safonova (Disney Research, Pittsburgh)

 

Global Illumination

Saturday, 01 December
11:00 – 13:00
Location: Peridot 206

Practical Hessian-Based Error Control for Irradiance Caching

We introduce a new error metric for irradiance caching based on a new derivation of occlusion-aware irradiance Hessian that significantly outperforms previous approaches such as the split-sphere and adaptive caching. This project was developed in collaboration with Jorge Schwarzhaupt (UC San Diego), Henrik Wann Jensen (UC San Diego), Wojciech Jarosz (Disney Research, Zürich)

Technical Briefs

Virtual Environment & Interaction

Thursday, 29 November
16:15 – 18:00
Location: Garnet 219

Efficient Rasterization for Edge-Based 3D Object Tracking on Mobile Devices

Augmented reality applications on consumer-grade hand-held devices suffer from the limited available processing power. By focusing on key behaviors of edge-based models, we present a sparse depth buffer structure to provide an efficient CPU-based rasterization method. This project was developed in collaboration with Etan Kissling (ETH Zurich), Kenny Mitchell (Disney Research Zurich), Thomas Oskam (Disney Research Zurich), Markus Gross (Disney Research Zurich)