On detecting dominant points

Document Type

Article

Publication Date

1-1-1991

Abstract

Detecting dominant points is a crucial preprocessing step for shape recognition and point-based motion estimation. Polygonal approximation has been a commonly used approach in detecting dominant points. This paper presents two alternatives which detect stable dominant points. In the first method, we find a set of positive maximum and negative minimum curvature points along the Gaussian smoothed boundary, followed by a split-and-merge polygonal approximation algorithm. The resulting break points, vertices of the approximated polygon, are the dominant points. Experimental results show that dominant points obtained by this method are less sensitive to the orientation of the boundary than other polygonal approximation algorithms in the sense that the number and the location of the dominant points along the contour remain relatively unchanged. In the second method, we smooth a boundary by a Gaussian filter using various widths until the extreme curvature points remain relatively unchanged for a range of filter widths. The resulting extreme curvature points which are stable to orientation and a reasonable range of scaling are the dominant points. © 1991.

Identifier

0025920328 (Scopus)

Publication Title

Pattern Recognition

External Full Text Location

https://doi.org/10.1016/0031-3203(91)90057-C

ISSN

00313203

First Page

441

Last Page

451

Issue

5

Volume

24

This document is currently not available here.

Share

COinS