Fast, robust, and faithful methods for detecting crest lines on meshes

Shin Yoshizawa, Alexander Belyaev, Hideo Yokota, Hans Peter Seidel

Research output: Contribution to journalArticlepeer-review

57 Citations (Scopus)

Abstract

The crest lines, salient subsets of the extrema of the principal curvatures over their corresponding curvature lines, are powerful shape descriptors which are widely used for shape matching, interrogation, and visualization purposes. In this paper, we develop fast, accurate, and reliable methods for detecting the crest lines on surfaces approximated by dense triangle meshes. The methods exploit intrinsic geometric properties of the curvature extrema and provide with an inherent level-of-detail control of the detected crest lines. As an immediate application, we use of the crest lines for adaptive mesh simplification purposes. © 2008 Elsevier B.V. All rights reserved.

Original languageEnglish
Pages (from-to)545-560
Number of pages16
JournalComputer Aided Geometric Design
Volume25
Issue number8
DOIs
Publication statusPublished - Nov 2008

Keywords

  • Crest lines
  • Focal surfaces
  • Möbius-invariant energies
  • Ridges
  • Surface Laplacian

Fingerprint

Dive into the research topics of 'Fast, robust, and faithful methods for detecting crest lines on meshes'. Together they form a unique fingerprint.

Cite this