Abstract

Global illumination rendering algorithms are capable of producing images that are visually realistic. However, this typically comes at a large computational expense. The overarching goal of this research was to compare different rendering solutions in order to understand why some yield better results when applied to rendering synthetic objects into real photographs. As rendered images are ultimately viewed by human observers, it was logical to use psychophysics to investigate these differences. A psychophysical experiment was conducted judging the composite images for accuracy to the original photograph. In addition, iCAM, an image color appearance model, was used to calculate image differences for the same set of images. In general it was determined that any full global illumination is better than direct illumination solutions only. Also, it was discovered that the full rendering with all of its artifacts is not necessarily an indicator of judged accuracy for the final composite image. Finally, initial results show promise in using iCAM to predict a relationship similar to the psychophysics, which could eventually be used in-the-rendering-loop to achieve photo-realism.

Library of Congress Subject Headings

Computer graphics; Image processing--Digital techniques; Visual perception; Psychophysics

Publication Date

4-4-2006

Document Type

Thesis

Department, Program, or Center

School of Photographic Arts and Sciences (CIAS)

Advisor

Fairchild, Mark

Advisor/Committee Member

Salvaggio, Carl

Advisor/Committee Member

Johnson, Garrett

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: T385 H38 2006

Campus

RIT – Main Campus

Download

Share

COinS