Automatic nonverbal analysis of social interaction in small groups: A review

摘要

An increasing awareness of the scientific and technological value of the automatic understanding of face-to-face social interaction has motivated in the past few years a surge of interest in the devising of computational techniques for conversational analysis. As an alternative to existing linguistic approaches for the automatic analysis of conversations, a relatively recent domain is using findings in social cognition, social psychology, and communication that have established the key role that nonverbal communication plays in the formation, maintenance, and evolution of a number of fundamental social constructs, which emerge from face-to-face interactions in time scales that range from short glimpses all the way to long-term encounters. Small group conversations are a specific case on which much of this work has been conducted. This paper reviews the existing literature on automatic analysis of small group conversations using nonverbal communication, and aims at bridging the current fragmentation of the work in this domain, currently split among half a dozen technical communities. The review is organized around the main themes studied in the literature and discusses, in a comparative fashion, about 100 works addressing problems related to the computational modeling of interaction management, internal states, personality traits, and social relationships in small group conversations, along with pointers to the relevant literature in social science. Some of the many open challenges and opportunities in this domain are also discussed.