Generative metadata for the torso and tail of entertainment videos

Presenter(s): Bharath Mohan (Sansara Technologies), Elvis D'Souza (Sensara Technologies)

Presents work in automated metadata creation that takes a hybrid approach. A bottom-up video reasoning system generates several structural artifacts including poster art, trailers, cast information, promotions, brands, title credits, moods, and predict genre automatically. All this is fed into a graph database with global knowledge of TV, and from there on curated by metadata professionals. It's an AI-assisted, metrics-driven process for the creation and maintenance of more complete metadata.