Augmented reality (AR), as a key component of the future metaverse, has leaped from the research labs to the consumer and enterprise markets. AR optical see-through (OST) devices utilize transparent optical combiners to provide visibility of the real environment as well as superimpose virtual content on top of it. OST displays distinct from existing media because of their optical additivity, meaning the light reaching the eyes is composed of both virtual content and real background. The composition results in the intended virtual colors being distorted and perceived transparent. When the luminance of the virtual content decreases, the perceived lightness and brightness decrease, and the perceived transparency increases. Lightness, brightness, and transparency are modulated by one physical dimension (luminance), and all interact with the background and each other. In this research, we aim to identify and quantify the three perceptual dimensions, as well as build mathematical models to predict them. In the first part of the study, we focused on the perceived brightness and lightness with two experiments: a brightness partition scaling experiment to build brightness scales, and a diffuse white adjustment experiment to determine the absolute luminance level required for diffuse white appearances on 2D and 3D AR stimuli. The second part of the research targeted at the perceived transparency in the AR environment with three experiments. The transparency was modulated by the background Michelson contrast reduction in either average luminance or peak-to-peak luminance difference to investigate, and later illustrated, the fundamental mechanism evoking transparency perception. The first experiment measured the transparency detection thresholds and confirmed that contrast sensitivity functions with contrast adaptation could model the thresholds. Subsequently, the transparency perception was investigated through direct anchored scaling experiment by building perceived transparency scales from the virtual content contrast ratio to the background. A contrast-ratio-based model was proposed predicting the perceived transparency scales. Finally, the transparency equivalency experiment between the two types of contrast modulation confirmed the mechanism difference and validated the proposed model.

Publication Date


Document Type


Student Type


Degree Name

Color Science (Ph.D.)


Gabriel J. Diaz

Advisor/Committee Member

Michael J. Murdoch

Advisor/Committee Member

Mark D. Fairchild


RIT – Main Campus