Date of Award
9-2019
Document Type
Thesis (Master's)
Department or Program
Computer Science
First Advisor
Lorie Loeb
Second Advisor
Xia Zhou
Third Advisor
Temiloluwa O. Prioleau
Abstract
This research is focused on the Uncanny Valley Effect of humanoid avatars, in a Virtual Reality (VR) /Augmented Reality (AR) environment. We designed and conducted a user study to observe people's reaction and behavior when interacting with virtual avatars of different levels of human likeness in VR and AR. Subjective surveys and objective measurement including EEG (Electroencephalogram), GSR (Galvanic Skin Response), Pupil Dilation sensors are used for quantification. Our result shows that, overall, in AR, people tend to like more human-like avatars, and users are more sensitive to facial expressions compared to body resemblance, especially on highly human likeness avatars. Compared to traditional Uncanny Valley researches, we defined an additional factor in determining users' experience, Avatar-Environment Contrast, the contrast between cartoonic characters and the real background. It plays a more significant role over other factors, when avatars are far enough from users. In VR, contrary to AR, people show a more negative feeling towards more human-like avatars. We think the imperfections of the animations and character design are exaggerated when characters are more human-like, and when they are up close to users. In addition, we discussed the limitations of this study, and how the results from our results and other researchers could be applied on virtual character design, and finally, from a psychological perspective, how human perceive virtual characters with humanoid mind attributions.
Recommended Citation
Yan, Ting, "Exploring the Uncanny Valley Effect in VR/AR" (2019). Dartmouth College Master’s Theses. 314.
https://digitalcommons.dartmouth.edu/masters_theses/314
