An Event-Driven, Stochastic, Undirected Narrative (EDSUN) Framework for Interactive Contents

摘要

In this paper, we present an extensible framework for interactive multimodal contents, with emphasis on augmented reality applications. The proposed framework, EDSUN, enables concurrent and variable narrative structures as well as content reusability and dynamic yet natural experience generation. EDSUN's main components include a canonical specification of 5-state lexical syntax and grammar, stochastic state transitions, and extensions for hierarchical grammars to represent complex behavioral and multimodal interactions. The benefits of EDSUN in enabling classical contents to support the affordances of AR environments and in complementing recent published works are also discussed.