RBF dipole surface evolution

Yuntao Jia, Xinlai Ni, Eric Lorimer, Michael Mullan, Ross Whitaker, John C. Hart

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

The level set method can implement a wide variety of shape modeling operations (e.g. offsetting, skeletonization, morphing, blending, smoothing, sharpening, embossing, denoising, sculpting, growing, texturing and fitting) simply by specifying a corresponding speed function that controls the growth of an evolving voxel isosurface. The problem is that the basic level set method is implemented on a fixed resolution grid, which limits the utility of these shape modeling operations. We instead represent surfaces with a collection of radial basis function dipole pairs, and derive the motion of these dipoles to implement a surface propagation similar to the level set method but on a smooth, arbitrary resoluton model. We demonstrate the utility of this approach with new level set methods for surface fitting, blending and center redistribution for RBF dipole models.

Original languageEnglish (US)
Title of host publicationSMI 2010 - International Conference on Shape Modeling and Applications, Proceedings
Pages143-150
Number of pages8
DOIs
StatePublished - 2010
EventInternational Conference on Shape Modeling and Applications - Shape Modeling International Conference, SMI 2010 - Aix-en-Provence, France
Duration: Jun 21 2010Jun 23 2010

Publication series

NameSMI 2010 - International Conference on Shape Modeling and Applications, Proceedings

Other

OtherInternational Conference on Shape Modeling and Applications - Shape Modeling International Conference, SMI 2010
Country/TerritoryFrance
CityAix-en-Provence
Period6/21/106/23/10

Keywords

  • Implicit surface
  • Level set
  • Particle system
  • Radial basis function
  • Surface fitting

ASJC Scopus subject areas

  • Computational Theory and Mathematics
  • Computer Science Applications

Fingerprint

Dive into the research topics of 'RBF dipole surface evolution'. Together they form a unique fingerprint.

Cite this