Yue Zhang

Abstract

Much recent research in computer vision has been aimed at the recognition of objects from scenes. Perceptual grouping has demonstrated its importance in creating descriptions of objects. Elementary edge descriptors are grouped into more complex structures based on relations such as proximity, parallelism, symmetry, and junction. However, exact determinations of these relations are impossible, because the relation is naturally uncertain and the output of low-level processing is not perfect. In this paper, an edge finding scheme based on fuzzy perceptual grouping is introduced. The geometrical relations are considered as fuzzy relations. The grouping operations that are based upon the fuzzy relations are discussed. The results of this edge-finder and their comparison with those of human beings are illustrated. The software implementation is also described.

Library of Congress Subject Headings

Image processing--Digital techniques--Mathematics; Fuzzy systems; Pattern recognition systems; Computer vision

Publication Date

7-1-1994

Document Type

Thesis

Department, Program, or Center

Chester F. Carlson Center for Imaging Science (COS)

Advisor

Rhody, Harvey

Advisor/Committee Member

Arney, Jonathan

Advisor/Committee Member

Anderson, Peter

Comments

Note: imported from RIT’s Digital Media Library running on DSpace to RIT Scholar Works. Physical copy available through RIT's The Wallace Library at: TA1632.Z427 1994

Campus

RIT – Main Campus

Download

Share

COinS