This paper presents a research framework for understanding communicative emotions aroused between people while interacting in conversation. Our advance is to consider how these emotions are perceived by other people, rather than what the target's internal state really is. Because such perception is subjective, we introduce the concept of using a collection of subjective external observations to objectively identify a fact. By treating the difference in perceived state as a probability distribution, we propose a computational model that describes the relationship between the perceived emotion and participants' key nonverbal behaviors, i.e. gaze and facial expressions. We also propose an evaluation method to assess the model by comparing the distributions estimated by using it with those of observers'. This paper describes initial experiments and discusses its potential.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.