Impact of Varying Vocabularies on Controlling Motion of a Virtual Actor
An ideal verbally controlled virtual actor would allow the same interaction as instructing a real actor with a few words. Our goal is to create virtual actors that can be controlled with natural language instead of a predefined set of commands. In this paper, we present results related to a questionnaire where people described videos of human locomotion using verbs and modifiers. The verbs were used almost unanimously for many motions, while modifiers had more variation. The descriptions from only one person were found to cover less than half of the vocabulary of other participants. Further analysis of the vocabularies against the numerical descriptors calculated from the captured motions shows that verbs appeared in closed areas while modifiers could be scattered to disconnected clusters. Based on these findings, we propose modeling verbs with a hierarchical vocabulary and modifiers as transitions in the space defined by the numerical qualities of motions.
Keywordsmotion capture natural language virtual actors
Unable to display preview. Download preview PDF.