Abstract
The visual design of virtual agents presents developers with a very large number of choices. We conducted a series of studies using Amazon's Mechanical Turk that demonstrate that there are no design universals for characters, optimal design of character proportion and rendering style depends on the task domain and user characteristics. Specifically, we found these adjustments to an agent's appearance directly effected how users rated it based on whether it was discussing social or medical content. The results of this research aim to help create visual guidelines for the development of domain specific virtual agents. © 2014 Springer International Publishing Switzerland.
Author supplied keywords
Cite
CITATION STYLE
Ring, L., Utami, D., & Bickmore, T. (2014). The right agent for the job? The effects of agent visual appearance on task domain. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 8637 LNAI, pp. 374–384). Springer Verlag. https://doi.org/10.1007/978-3-319-09767-1_49
Register to see more suggestions
Mendeley helps you to discover research relevant for your work.
