Extraction of 3D Feature Descriptor Using the Distribution of Normal Vectors

Authors:
Ami Kim;Oubong Gwun;Juwhan Song
Affiliations:
Div. of Electronics and Information Engineering, Chonbuk National Univ., South Korea;Div. of Electronics and Information Engineering, Chonbuk National Univ., South Korea;School of Liberal Art, University of Jeonju, South Korea
Venue:
FQAS '09 Proceedings of the 8th International Conference on Flexible Query Answering Systems
Year:
2009

Citing 5
Cited 0

Shape distributions

ACM Transactions on Graphics (TOG)
A Developer's Survey of Polygonal Simplification Algorithms

IEEE Computer Graphics and Applications
Mathematics for 3D Game Programming and Computer Graphics, Second Edition

Mathematics for 3D Game Programming and Computer Graphics, Second Edition
The Princeton Shape Benchmark (Figures 1 and 2)

SMI '04 Proceedings of the Shape Modeling International 2004
Feature-based similarity search in 3D object databases

ACM Computing Surveys (CSUR)

Quantified Score

Hi-index	0.00

Visualization

Abstract

We proposed the distribution of mesh normal directions over the surface as a feature descriptor of 3D model. Feature descriptor of 3D model should be invariant to translation, rotation and scale for its model. So this paper normalizes all the model using PCA. The normal is sampled in proportion to each polygon's area so that the information on the surface with less surface area may be less reflected on composing a feature descriptor. Besides it is calculated by weight average method via angles and interpolated. Similarity measurement uses a L1 -norm between histograms, and search efficiency is indicated with precision and recall. The results of mesh simplification show the performance improvement by 17.6% in comparison with original model and the case of normal interpolation show by 22.3%. The experimental results have shown that the performance of retrieval has been improved compared to conventional methods.