Local Neighborhoods for Shape Classification and Normal Estimation

Authors

Cindy Grimm, Washington University in St Louis
William Smart, Washington University in St Louis

Document Type

Technical Report

Department

Computer Science and Engineering

Publication Date

2008-01-01

Filename

wucse-2008-15.pdf

Technical Report Number

WUCSE-2008-15

Abstract

We introduce the concept of local neighborhoods, a generalization of the one-ring on a mesh to unlabeled 3D data points arising from sampling a 2D surface embedded in 3D. The local neighborhood supports both local shape classification and robust normal estimation. In particular, local neighborhoods out-perform traditional approaches in unevenly sampled, curved regions. We show that the local neighborhood can be used in place of a full mesh structure for applications such as smoothing, moving least-squares reconstruction, and parameterization. Longer version of paper submitted to CAGD

Comments

Permanent URL: http://dx.doi.org/10.7936/K78C9TG4

Recommended Citation

Grimm, Cindy and Smart, William, "Local Neighborhoods for Shape Classification and Normal Estimation" Report Number: WUCSE-2008-15 (2008). All Computer Science and Engineering Research.
https://openscholarship.wustl.edu/cse_research/226