Semantic Relation among Spatial Objects for Situation-Aware Services in Interactive Smart Spaces

In projector-camera based interactive smart spaces, information about spatial relation among users and nearby objects is crucial in order to find the most suitable surfaces on which situation-aware information will be augmented. This paper describes a spatial context management system that continuously tracks semantic relations among spatial objects for situation-aware services in interactive smart spaces. Our contribution is twofold. First, we propose a zone-space model for supporting flexible dynamic collaboration among multiple depth cameras and a seamless handover scheme between the adjacent zones for supporting not-contact people tracking without shadow area. Second, we present three types of spatial operators based on a 3D spatial object model to understand and /or detect various situations in space by querying semantic relations among objects within the space. Experiment results show that our people location tracking scheme shows good performance on both position errors and success rate of handover between adjacent trackers. We also verify the correct functions of the overall system by implementing a proof-of-concept scenario i.e. the smart memo service.

In projector-camera based interactive smart spaces, information about spatial relation among users and nearby objects is crucial in order to find the most suitable surfaces on which situation-aware information will be augmented. This paper describes a spatial context management system that continuously tracks semantic relations among spatial objects for situation-aware services in interactive smart spaces. Our contribution is twofold. First, we propose a zone-space model for supporting flexible dynamic collaboration among multiple depth cameras and a seamless handover scheme between the adjacent zones for supporting not-contact people tracking without shadow area. Second, we present three types of spatial operators based on a 3D spatial object model to understand and /or detect various situations in space by querying semantic relations among objects within the space. Experiment results show that our people location tracking scheme shows good performance on both position errors and success rate of handover between adjacent trackers. We also verify the correct functions of the overall system by implementing a proof-of-concept scenario i.e. the smart memo service.