SIGGRAPH 2012

Conference 5–9 August 2012
Exhibition 7–9 August 2012
Los Angeles Convention Center

Sets of Shapes

Technical Papers

Sets of Shapes

Tuesday, 7 August 2:00 PM - 3:30 PM | Los Angeles Convention Center, Room 515AB
Session Chair: Alla Sheffer, The University of British Columbia

Exploring Collections of 3D Models Using Fuzzy Correspondence

This paper introduces an encoding of similarity relationships using fuzzy-point correspondences. The computational framework estimates fuzzy correspondences using only a sparse set of pairwise model alignments. Correspondences are further used to demonstrate applications towards view alignment, smart exploration, and faceted browsing.

Vladimir G. Kim
Princeton University

Wilmot Li
Adobe Systems Incorporated

Niloy J. Mitra
University College London

Stephen DiVerdi
Adobe Systems Incorporated

Thomas Funkhouser
Princeton University

A Probabilistic Model for Component-Based Shape Synthesis

A new generative model of shapes that understands variability in shape structure. The model is trained on a set of compatibly segmented shapes. Then it is used to automatically synthesize new shapes, by identifying new plausible combinations of components from the training shapes.

Evangelos Kalogerakis
Stanford University

Siddhartha Chaudhuri
Stanford University

Daphne Koller
Stanford University

Vladlen Koltun
Stanford University

Synthesizing Open Worlds With Constraints Using Locally Annealed Reversible Jump MCMC

A probabilistic inference algorithm that addresses the challenge of exploring isolated modes in trans-dimensional distributions. The algorithm is a valid MCMC sampler that samples efficiently in the presence of complex constraints. The paper shows how it can synthesize layouts with varying numbers of objects.

Yi-Ting Yeh
Stanford University

Lingfeng Yang
Stanford University

Matthew Watson
Stanford University

Noah D. Goodman
Stanford University

Pat Hanrahan
Stanford University

Fit and Diverse: Set Evolution for Inspiring 3D Shape Galleries

This paper introduces set evolution as a means for creative 3D shape modeling, where an initial population of 3D models is evolved to produce generations of novel shapes.

Kai Xu
Shenzhen VisuCA Key Lab/SIAT and National University of Defense Technology (NUDT)

Hao (Richard) Zhang
Simon Fraser University

Daniel Cohen-Or
Tel Aviv University

Baoquan Chen
Shenzhen VisuCA Key Lab/SIAT