Quantifying nonverbal communicative behavior in face-to-face human dialogues

Abstract

The referred study is based on the assumption that understanding how humans use nonverbal behavior in dialogues can be very useful in the design of more natural-looking animated talking heads. The goal of the study is twofold: (1) to explore how people use specific facial expressions and head movements to serve important dialogue functions, and (2) to show evidence that it is possible to measure and quantify the entity of these movements with the Qualisys MacReflex motion tracking system. Naturally elicited dialogues between humans have been analyzed with focus on the attention on those nonverbal behaviors that serve the very relevant functions of regulating the conversational flux (i.e., turn taking) and producing information about the state of communication (i.e., feedback). The results show that eyebrow raising, head nods, and head shakes are typical signals involved during the exchange of speaking turns, as well as in the production and elicitation of feedback. These movements can be easily measured and quantified, and this measure can be implemented in animated talking heads.