|
With its K-Navigator line of software solutions, Kinomai is applying its proven know-how in real time meta-data generation to the needs of digital video navigation. The solutions are based on a two-level approach: audio and video key properties are computed on the fly from the incoming audiovisual stream and identify meaningful events (change of shot, appearance of a pre-learned image, change of background sound, applause, etc.). Successively, and at any given moment, "broadcast grammars" are used in conjunction with the computed multimedia properties to generate a table of contents of the incoming video. As an example, a television news broadcast will be automatically broken down into segments corresponding to the different news items it contains. Similarly, a tennis match will be broken down into played points and the most dramatic points will be identified. For a talk show, interviews will be segmented and highlighted. A movie feature will be broken down into chapters, offering navigational capabilities similar to those of a DVD.
(http://www.kinomai.com)
|