Communications of the ACM
Customize Your Favorite Tv Show
Credit: USC
With TVs utilizing gestural and voice controls, and analytics to examine your backlog of DVRed programs and suggest new shows, a team of French developers thought something was still missing.
"We think it's time users had search and browsing tools to use with these digital video collections," says Herve Bredin of the Computer Sciences Laboratory for Mechanics and Engineering Sciences in Orsay. Bredin and a team from the Toulouse Institute for Computer Science Research have developed StoryVisualizer (StoViz), a PC-based service that can successfully parse a TV episode based on characters, narrative, and setting. StoViz does this by deinterlacing the visuals and dialogue of a scene into individual threads, seeking specific details such as an actor's face, background scenery, and specific words and phrases of dialogue. This allows the program to construct a collection of scenes from across an episode or an entire series based on any of these threads on the fly.
Bredin says his team has successfully tested StoViz on three different series so far. He hopes StoViz will eventually become a feature in many DVRs.
From NewScientist.com
View Full Article
Abstracts Copyright © 2012 Information Inc., Bethesda, Maryland, USA
No entries found