Friday, October 31, 2014
Have you ever asked yourself where a scene in your favorite sitcom was filmed? Or whether the movie you are watching on TV is based on a book? Swoozy, an interactive television system from Saarbrücken, Germany, enables viewers to search for this type of information on the web and semantic web while watching TV. When a video is played, additional information such as related pictures, facts, videos, or shopping recommendations can be retrieved via content from well-known web services like YouTube™, DBpedia, Freebase, or Flickr™.
Swoozy not only analyzes the content of each video frame, but it also processes contextual information from sources like electronic program guides (EPGs) or HbbTV®. When a famous actor appears in a scene, for example, the actor's name is automatically extracted from contextual information during the semantic analysis process and displayed as a term - or so-called "Grabbable" - semi-transparently on the screen. These terms are automatically generated and linked to short descriptions of the featured objects, actors, or even locations.
"Grabbables" are also key elements for the gesture-based interactions used to control Swoozy's user interface. By dragging and dropping "Grabbables," users can begin a semantic query. Swoozy supports several input devices for gesture interactions, such as depth cameras (e.g. Microsoft Kinect), LeapMotion controllers, and gyroscope-based remote controllers. The system also supports cross-media interaction, enabling individuals to utilize a second-screen app to simultaneously work with content displayed on TV on another monitor like a tablet. In addition, Swoozy can be used for other video-based systems, including interactive e-learning systems, business TV channels, and in-room hotel solutions.
Swoozy received the 2013 CeBIT Innovation Award and is a spin-off product from the German Research Center for Artificial Intelligence (DFKI).
For more information, visit http://swoozy.de/.
Source & Image: © Swoozy.de