Feature sensitive multiscale editing on surfaces

Ulrich Clarenz

doi:10.1007/s00371-004-0245-3

Feature sensitive multiscale editing on surfaces

2004-07-01

33

PRE

AI

Marc Alexander Schweitzer

A

Alexandru Telea

DOI：10.1007/s00371-004-0245-3

Original

Original request for help

Save

Abstract

En 中文

A novel editing method for large triangular meshes is presented. We detect surface features, such as edge and corners, by computing local zero and first surface moments, using a robust and noise resistant method. The feature detection is encoded in a finite element matrix, passed to an algebraic multigrid (AMG) algorithm. The AMG algorithm generates a matrix hierarchy ranging from fine to coarse representations of the initial fine grid matrix. This hierarchy comes along with a corresponding multiscale of basis functions, which reflect the surface features on all hierarchy levels. We consider either these basis functions or distinct sets from an induced multiscale domain decomposition as handles for surface manipulation. We present a multiscale editor which enables Boolean operations on this domain decomposition and simply algebraic operations on the basis functions. Users can interactively design their favorite surface handles by simple grouping operations on the multiscale of domains. Several applications on large meshes underline the effectiveness and flexibility of the presented tool.

Keywords:

surface processing

algebraic multigrid

multiscale feature detection

Feature sensitive multiscale editing on surfaces

Abstract

Journal

Organization