Socially-driven computer vision for group behavior analysis

Abstract

The analysis of human activities is one of themost intriguing and important open issues in the video analytics field. Since few years ago, it has been handled following primarily Computer Vision and Pattern Recognition methodologies,where an activity corresponded usually to a temporal sequence of explicit actions (run, stop, sit, walk, etc.).More recently, video analytics has been faced considering a new perspective, that brings in notions and principles from the social, affective, and psychological literature, and that is called Social Signal Processing (SSP). SSP employs primarily nonverbal cues, most of them are outside of conscious awareness, like face expressions and gazing, body posture and gestures, vocal characteristics, relative distances in the space and the like. This paper will discuss recent advancements in video analytics, most of them related to the modelling of group activities. By adopting SSP principles, an age-old problem -what is a group of people?- is effectively faced, and the characterization of human activities in different respects is improved. © 2014 Springer-Verlag Berlin Heidelberg.