Abstract

Image segmentation is a fundamental task in many computer vision applications. We present a novel unsupervised color image segmentation algorithm named GSEG, which exploits the information obtained from detecting edges in color images. By using a color gradient detection technique, pixels without edges are clustered and labeled individually to identify the image content. Elements that contain higher gradient density are included by a dynamic generation of clusters as the segmentation progresses. By quantizing the colors in the image and extracting texture information from the neighborhood entropy of each pixel, the proposed method obtains accurate models of texture that are highly effective to merge regions that belong to the same object. Experimental results for various image scenarios in comparison with state-of-the-art segmentation techniques demonstrate the performance advantages of the proposed method.

Library of Congress Subject Headings

Computer vision--Mathematical models; Image processing--Digital techniques

Publication Date

8-2007

Document Type

Thesis

Student Type

Graduate

Degree Name

Electrical Engineering (MS)

Department, Program, or Center

Electrical Engineering (KGCOE)

Advisor

Eli Saber

Advisor/Committee Member

Vincent Amuso

Advisor/Committee Member

Sohail Dianat

Comments

Physical copy available from RIT's Wallace Library at TA1634 .G37 2007

Campus

RIT – Main Campus

Download

Share

COinS