HPI developed a platform that allows video content to be found not only based on a title or keywords, but thanks to the audiovisual content itself.
"Conventional search engines are reaching their limits here, as it is particularly important that the information content of the individual sequences is captured and understood in its meaning"
explains Harald Sack, senior researcher in HPI's Internet Technologies and Systems department. The solution provides a combination of novel video analysis methods with semantic web technologies.
In the Semantic Web, information expressed in natural language is supplemented by an unambiguous description of its meaning (semantics), which can also be correctly interpreted and thus understood by computers.
By automatically enriching the movie data with semantic information, HPI researchers enable users to identify and browse content links between the web video content.
This allows users to make use of visual tables of contents or navigate through the video offering with the help of visualized audio text provided by automatic speech recognition.
This optimizes access to and use of movies, computer animations, lecture and conference recordings, among other things.
Download as PDF only for members. Please create an account Here
