Nonparametric dominant point detection

Abstract

ABSTRACT A new method for detecting dominant points is presented. It does not require any input parameter, and the dominant points obtained by this method remain relatively the same even when the object curveis scaled or rotated.Ill this method, for each boundary point, a support region is assigned to the point based on its localproperties. Each point is then smoothed by a Gaussian filter with a width proportional to its determined support region. A significance measure for each point is then computed. Dominant points are finally obtained through nonmaximum suppression.Unlike other dominant point detection algorithms which are sensitive to scaling and rotation of theobject curve, the new method will overcome this difficulty. Furthermore, it is robust in the presence of noise. The proposed new method is compared to a well-known dominant point detection algorithm in termsof the computational complexity and the approximation errors. 1. INTRODUCTION It has been suggested from the viewpoint of the human visual system1 that dominant points along anobject contour are rich in information content and are sufficient to characterize the object contour. Thedominant points are the high curvature points along a digital curve that have important shape attributes.Many algorithms216 have been suggested for detecting dominant points. They fall into two categories