Shape Matching and Symmetry

Monday, 31 July
10:30 am - 12:15 pm
Hall C
Session Chair: Ioana Boier-Martin, IBM T.J. Watson Research Center

A Planar Reflective Symmetry Transform for 3D Shapes
A symmetry transform that provides a measure of reflective symmetry for all planes. This paper describes efficient construction algorithms and novel applications for alignment, retrieval, segmentation, and viewpoint selection.

Joshua Podolak
Philip Shilane
Aleksey Golovinskiy
Szymon Rusinkiewicz
Thomas A. Funkhouser
Princeton University


Partial and Approximate Symmetry Detection for 3D Geometry
Stochastic clustering of local point-pair symmetry relations in transformation space allows time- and memory-efficient symmetry detection with applications in shape compression, segmentation, symmetry-consistent editing, indexing for retrieval, etc.

Niloy J. Mitra
Leonidas J. Guibas
Stanford University

Mark Pauly
Eidgenössische Technische Hochschule Zürich


Reassembling Fractured Objects by Geometric Matching
Development of algorithms for automatic reassembly of fractured 3D solids. Pair- and multi-piece geometric matching is applied to digital models of the fragments to find the best globally consistent reconstruction.

Qi-Xing Huang
Tsinghua University

Simon Flöry
Technische Universität Wien

Natasha Gelfand
Stanford University

Michael Hofer
Helmut Pottmann
Technische Universität Wien


Perfect Spatial Hashing
Pakcing sparse data into a compact table using a perfect multi-dimensional hash function optimized for efficient SIMD evaluation. Applications include vector images, sprites, 3D painting, simulation, and collision detection.

Sylvain Lefebvre
Hugues Hoppe
Microsoft Research