SenSee Framework for Personalized Access to TV Content

In this paper we present our framework, SenSee, designed to provide context-aware and personalized search that can be used to support multimedia applications in making content recommendations. Our primary demonstration application built upon the framework targets the growing digital television domain where we foresee an increasing need for user-adaptive functionality to counter the looming content explosion which will make current television zapping inadequate. Via our AJAX-based interface we show how different user contexts, such as location, time and audience, in combination with a standard user profile can improve the multimedia (TV content) consumption experience. In modeling the user (profile and context) SenSee exploits ontologies to express the semantics involved. We illustrate via experiments the influence of the user’s current context and demonstrate the difference when we dis/enable the inclusion of different ontologies describing time, geographical, lexical and TV domain knowledge.