Abstract

Human skin color is embedded with information about mood, health, and age. While facial color is widely studied in dermatology and cosmetics, there is limited data available to map the subjective experience of skin color to measured data. The ability to derive the color that best represents a person’s skin would be valuable in a wide range of applications, from identifying an appropriate makeup foundation to achieving white balance in a video. When displaying skin tones, it is well-known that preferred renderings are different from accurate renderings. Rendering skin tones becomes more nuanced with complex rendering technologies such as augmented reality. Optical see-through augmented reality (OST-AR) is a technology that superimposes images through a transparent medium, allowing the user to add graphical elements to their view and interact with them. The additive nature of OST-AR makes it so that darker colors appear more transparent in bright environments than lighter colors; this translates to skin tones where people with darker skin often appear more transparent than people with lighter skin, introducing social challenges in OST-AR applications involving renderings of humans. In this dissertation, two topics related to color appearance of skin are studied in OST-AR. The first topic of study is representative color of skin tones on emissive displays and in OST-AR. In this series of studies, memory color and representative color of skin is assessed using eight-alternative forced choice experiments. Results of these studies diverged from previous findings that people prefer more saturated colors when determining what color best represents an object. These results suggest that faces are categorically treated differently than other objects when determining representative color. The second topic of study is preferred skin color rendition of skin on photos and avatars in OST-AR. Two novel contributions are made: assessments of preferred renderings (1) of self-representative stimuli versus zero-acquaintance stimuli and (2) photos versus avatars. In this study, observers adjusted lightness and chroma of faces to their preferred rendition. While observers adjusted photos and avatars with darker skin to have higher lightness in bright environments, this trend diverged when they were assessing their own face. These results emphasize a greater need for self-representative stimuli in color appearance research, especially in OST-AR when skin rendition can be poor.

Publication Date

5-2026

Document Type

Dissertation

Student Type

Graduate

Degree Name

Color Science (Ph.D.)

Department, Program, or Center

Color Science

College

College of Science

Advisor

Susan Farnand

Advisor/Committee Member

Christopher Thorstenson

Advisor/Committee Member

Reynold Bailey

Campus

RIT – Main Campus

Download

Share

COinS