Social focus of attention as a time function derived from multimodal signals

Abstract

In this paper, we present the results of a study on the social focus of attention as a time function derived from the multisource multimodal signals, recorded by different personal capturing devices during social events. The core of the approach is based on fission and fusion of multichannel audio, video and social modalities to derive the social focus of attention. The results achieved to date on 16+ hours of real-life data prove the feasibility of the approach.