Technical Papers

Geometry Algorithms & Sampling

Tuesday, 27 July | 3:45 PM - 5:30 PM | Room 502 B
Session Chair: Mark Pauly,École Polytechnique Fédérale de Lausanne
Improving Chen & Han's Algorithm on the Discrete Geodesic Problem

This paper discusses the “single source, any vertex" shortest-path problem. The main aim of this research is to show how to filter out useless windows using current estimates of the distances to the vertices and to explore another technique, involving maintaining a priority queue.

Guo-jin Wang
Zhejiang University

Shi-Qing Xin
Zhejiang University

Feature-Preserving Triangular Geometry Images for Level-of-Detail Representation of Static and Skinned Meshes

This paper introduces triangular-chart geometry images as the GPU-side representation for view-dependent level-of-detail rendering of skinned dynamic meshes, with support for feature preservation. It results in a 10-fold improvement in fidelity compared to quad-chart geometry images.

Wei-Wen Feng
University of Illinois at Urbana-Champaign

Byung-Uck Kim
University of Illinois at Urbana-Champaign

Yizhou Yu
University of Illinois at Urbana-Champaign

Liang Peng
Intel Corporation

John Hart
University of Illinois at Urbana-Champaign

Controllable Conformal Maps for Shape Deformation and Interpolation

A novel 2D shape-deformation system that generates conformal maps, yet provides the user with a large degree of control over the result. For example, it allows singularities at user-specified boundary points, so true "bends” can be introduced into the deformation.

Ofir Weber
Technion - Israel Institute of Technology

Craig Gotsman
Technion - Israel Institute of Technology

Accurate Multidimensional Poisson-Disk Sampling

This paper presents an accurate and efficient sampling method in n-dimensional space based on a Poisson-disk distribution. The method is accurate because it obeys the statistical properties of the dart-throwing algorithm. It is efficient because it employs a spatial subdivision structure to signal regions of empty space.

Manuel Gamito
Lightwork Design Ltd

Steve Maddock
The University of Sheffield

Multi-Class Blue Noise Sampling

An extension of traditional single-class algorithms in which each class and their unions exhibit blue-noise properties. The paper presents two flavors of methods that offer explicit controls for sample count or spacing, based on the notion of hard/soft disks.

Li-Yi Wei
Microsoft Research

In the News
SIGGRAPH 2010 Video