Activity Recognition Based on Intra and Extra Manipulation of Everyday Objects

Recognizing activities based on an actor’s interaction with everyday objects is an important research approach within ubiquitous computing. We present a recognition approach which complement objects grabbed or released information with the object’s internal state changes (as an effect of intra manipulation) and the object’s external state changes with reference to other objects (as an effect of extra manipulation). The concept of Intra manipulation is inspired by the fact that many everyday objects change their internal state when manipulated by the human actor, while extra manipulation is motivated by the fact that humans commonly rearrange the spatial relations between everyday objects as part of their activities. A detailed evaluation of our prototype activity recognition system in virtual reality (VR) environment is presented as a “proof of concept”. We have obtained a recognition precision of 92% on the activity-level and 81% on the action-level among 15 everyday home activities. Virtual reality was used as a test-bed in order to speed up the design process of our activity recognition system, allowing us to compensate for the limitations with currently available sensing technologies and to compare the contributions of intra manipulation and extra manipulation for activity