Spatial Models for Wide-Area Visual Surveillance: Computational Approaches and Spatial Building-Blocks

摘要

Spatial models play a key role when interpreting a dynamic and uncertain world for a wide-area surveillance application. This paper presents two different views to illustrate the range of spatial models. First, we take a top-down look, where we survey various work relevant to the development of spatial models and how they have been used in AI applications. Then we take a more bottom-up look, starting with a promising spatial primitive to identify a useful foundation that can support visual surveillance applications.