Nathan Cahill

Abstract

Optimization problems arise in a wide variety of scientific disciplines. In many practical problems, a global optimum is desired, yet the objective function has multiple local optima. A number of techniques aimed at solving the global optimization problem have emerged in the last 30 years of research. This thesis first reviews techniques for local optimization and then discusses many of the stochastic and deterministic methods for global optimization that are in use today. Finally, this thesis shows how to apply global optimization techniques to two practical problems: the image segmentation problem (from imaging science) and the 3-D registration problem (from computer vision).

Library of Congress Subject Headings

Mathematical optimization; Image processing--Mathematical models; Computer vision--Mathematical models

Publication Date

5-1-2000

Document Type

Thesis

Department, Program, or Center

School of Mathematical Sciences (COS)

Advisor

Bautista, Maurino

Advisor/Committee Member

Kumar, Seshavadhani

Advisor/Committee Member

Wilcox, Theodore

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: QA402.5 .C33 2000

Campus

RIT – Main Campus

Download

Share

COinS