Video Visualization From Speech/Text Semantics – A Theoretical Framework

Abstract

Many a times, words fail to make the impact that visual inputs do. It goes without saying that visual explanation gives better understanding of concepts that speech or written text – both of which require imagination of context for proper understanding. Visualizing speech semantics as video is a novel idea, especially in the context of helping the hearing impaired understand the real-life context of a conversation better. It involves the use of a simple visual device that converts conversation or text into video with fictitious characters, much the same way as the human brain maps language semantics into visual cues. The videos are snippets that decipher sentences into moving visual elements. In this paper, we present the ‘Text or Speech to dynamic Visual System’ – GiVee. Some of the primary and secondary primitives required for designing such a system are presented. Also, the basic algorithm for its implementation and some challenges are also presented.