Abstract

Focus of Attention plays an important role in perception of the visual environment. Certain objects stand out in the scene irrespective of observers' goals. This form of attention capture, in which stimulus feature saliency captures our attention, is of a bottom-up nature. Often prior knowledge about objects and scenes can influence our attention. This form of attention capture, which is influenced by higher level knowledge about the objects, is called top-down attention. Top-down attention acts as a feedback mechanism for the feed-forward bottom-up attention. Visual search is a result of a combined effort of the top-down (cognitive cue) system and bottom-up (low level feature saliency) system. In my thesis I investigate the process of goal directed visual search based on color cue, which is a process of searching for objects of a certain color. The computational model generates saliency maps that predict the locations of interest during a visual search. Comparison between the model-generated saliency maps and the results of psychophysical human eye -tracking experiments was conducted. The analysis provides a measure of how well the human eye movements correspond with the predicted locations of the saliency maps. Eye tracking equipment in the Visual Perceptual Laboratory in the Center for Imaging Science was used to conduct the experiments.

Library of Congress Subject Headings

Color vision--Computer simulation; Visual perception--Computer simulation; Eye--Movements; Computer vision

Publication Date

2004

Document Type

Thesis

Student Type

Graduate

Degree Name

Computer Science (MS)

Department, Program, or Center

Computer Science (GCCIS)

Advisor

Roger Gaborski

Advisor/Committee Member

Roxanne Canosa

Advisor/Committee Member

Ankur Teredesai

Comments

Physical copy available from RIT's Wallace Library at QP483 .V35 2004

Campus

RIT – Main Campus

Download

Share

COinS