Didn’t see that coming: a survey on non-verbal social human behavior forecasting
Published in Understanding Social Behavior in Dyadic and Small Group Interactions, 2022
Non-verbal social human behavior forecasting has gained growing attention in recent years due to its direct applications in human-robot interaction and socially-aware human motion generation. This survey defines the behavior forecasting problem for multiple interactive agents in a generic way, aiming to unify the traditionally separated fields of social signals prediction and human motion forecasting. It identifies shared challenges such as future stochasticity, context awareness, and history exploitation. The paper also proposes a taxonomy of methods published in the last five years, highlights concerns in the community, and offers a friendly overview of audiovisual datasets featuring non-acted social interactions, alongside common metrics and their issues.
Recommended citation: German Barquero, Johnny Núnez, Sergio Escalera, Zhen Xu, Wei-Wei Tu, Isabelle Guyon, Cristina Palmero. (2022). "Didn’t see that coming: a survey on non-verbal social human behavior forecasting." Understanding Social Behavior in Dyadic and Small Group Interactions. 139-178.
Download Paper | Download Slides