Human action recognition for home sensor network

Abstract

Home sensor network, namely smart home has been developing at an incredible speed. In order to improve the intelligence of home sensor network, we committed our efforts to the research on video-based human action recognition to build a home sensor network, which makes it possible for smart home to interact with users through actions and recognize accidents like falling down through surveillance video automatically. In this paper, we present a survey on video-based human action recognition and address this problem from two major aspects: action representation and recognition approach. A comprehensive overview of this field is provided in this paper. We also analyze the characteristics of home sensor network and propose the challenges in front of all researchers for introduction of action recognition into home sensor network at the end of this survey. © 2013 Springer-Verlag.