There is an increasing interest in the concept of intelligent environments where a closed or delimited public space (shopping mall, station, museum, hospital etc) is endowed with some automatic ability to interpret human behavior. Intelligent environments interact with their users, aiding, serving and pre-emptying them. In a not too distant future, this paradigm - in Europe called ambient intelligence - will soon include robotic platforms. Both intelligent environment and robotic platforms will collaborate to better inform the inhabiting or visiting user. This paper presents some steps towards that direction, describing a study on some scene descriptors, which can be employed to provide an automatic interpretation of the clutter and dynamics of a complex scene.
Beibei Zhan, Ndedi D. Monekosso, Paolo Remagnino,