Browse Prior Art Database

Hierarchical feature edge extraction based on local shape Disclosure Number: IPCOM000013084D
Original Publication Date: 2000-May-01
Included in the Prior Art Database: 2003-Jun-12

Publishing Venue




This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 27% of the total text.

Page 1 of 12

Hierarchical feature edge extraction based on local shape


This document describes a new method for
extracting feature edge from polyhedral data.
The method extracts feature edges with
corresponding priorities, and can be used in
a variety of applications such as 3D-shape
similarity searching or automatic ordering
the vertices of polyhedral data for LOD
(Level of Detail) data generation or
progressive data transfer.

The key characteristics of the method are a
new definition of feature edges -a scale
oriented feature edge definition- and a
method for estimating local shape of
polyhedral data by using two kinds of error

1. 1. Introduction

This document describes a new definition of
feature edges of polyhedral data and a method
for extracting feature edges from given
polyhedral data, which enables feature edges
to be extracted with their corresponding

In the area of simplification methods of
polyhedral data, objective of the application
is to reduce numbers of the polygon by
removing smaller features. The policy of most
simplification methods is to remove vertices
while keeping with in a certain margin of
error from the given polyhedral data.
However, as most of these methods do only
this, there are no definitions of features
and methods for estimating features. Thus,
what these methods can extract is not the
characteristics of a given shape, but one
version of the simplified shape. Several
methods have used the idea of feature edges.
For example [Sch92][Hop94] define feature
edges as "keen edges" or "sharp edges", and
[Sch92] defines feature edge as edges for
which the gaussian angle of bounded faces is
longer than a certain angle. But we cannot
distinguish whether such edges are elements
of smaller features or larger ones.

In our feature edge definition, we introduce
scale as a means of prioritizing the feature
edges. Each feature edge has a corresponding
priority that indicates the scale of the
feature (how large the feature is).

This allows to conceive of the following
application. In the case of 3D-shape
similarity searching, we can find the corners
of feature edges and the priority of each
corner, and can use these corners as the


Page 2 of 12

corresponding points of two sets of
polyhedral data.

In the case of polyhedral data
simplification, the rule for vertex
elimination differs according to the local
shape of the neighboring area. For example,
the vertices of a smooth area can be
eliminated in any way, but the vertices of a
crease area should treated carefully so as
not to remove features. In addition, the
determination of a local shape changes
according to the scale. Since our feature
edge definition takes account of these
distinctions, it is possible to create a
variety of levels of detailed data by
changing the weights of feature edges.

In this document, we first define our feature
edge, then describe our local shape
estimation method, and finally describe a
feature edge extraction method.

The principal characteristics of our approach
are scale-orien...