Informedia-II Conceptual Overview
Notes:
Through the extraction of appropriate metadata from diverse video collections, relevant information can be synthesized and presented driven by the user’s needs. Currently users may visit numerous video collections in search of an answer that reveals itself only in bits at a time, such as an unfolding story of a famous criminal trial, the effects of ocean currents on bird migrations, or shifting political alliances. Video information collages will emphasize dimensions of importance to the user so that the full context can be understood and navigated to narrow the focus to a particular information thread, resulting in only the most useful video pieces then being played.
Video information collages can be built from the integration of speech recognition and image processing technologies to achieve information extraction and summarization in the video domain. There will be numerous templates or organizational schemes for collages, including geo-collages like maps, chrono-collages like timelines, and auto-documentaries in which the collage is not viewed all at once but rather is played like a documentary video.