Departmental Papers (BE)

Document Type

Conference Paper

Date of this Version

March 2003

Comments

Copyright 2003 IEEE. Reprinted from Proceedings of the 1st International IEEE EMBS Conference on Neural Engineering 2003, pages 16-19.
Publisher URL: http://ieeexplore.ieee.org/xpl/tocresult.jsp?isNumber=26900

This material is posted here with permission of the IEEE. Such permission of the IEEE does not in any way imply IEEE endorsement of any of the University of Pennsylvania's products or services. Internal or personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution must be obtained from the IEEE by writing to pubs-permissions@ieee.org. By choosing to view this document, you agree to all provisions of the copyright laws protecting it.

Abstract

The salience of a contour depends upon several factors, including continuity, closure and curvature consistency. We analyze the statistics of curvature variation using a single image from Shimon Ullman’s [1] original work on contour salience. We develop a measure based on the arc length of a contour segment over which curvature variation remains within a constrained range. Locally, all contours in the image are similar with respect to curvature consistency. However, when the entire contour is considered, the most salient contours are found to have the most consistent curvatures. This finding reinforces Ullman’s point that salience is a global property of the object.

We interpret these results in view of Rosenholtz’s [2] recent model of salience as a statistical measure of outliers from a population. In addition, we speculate on the visual cortical mechanisms in striate and extrastriate cortex required to carry out salience measurements on this class of images.

Keywords

Salience, curvature, cocircularity, visual cortex

Share

COinS

Date Posted: 09 November 2004

This document has been peer reviewed.