Project Semantic Multimedia Access (N5)
A filmmaker searching for the missing image fragment to perfect his documentary; a teacher looking for the best animation for his history lesson; a DJ who wants to rock the crowd in a nightclub with that particular music sample...
There's a big chance these people get more hits on their search requests than they need or wish for. Not they, but the current search machines are to blame for this. These are almost one-dimensional, while the user wishes are often much more diverse and specific. Due to the multi-layered information need and its inherent dynamics, the user’s wish can never be made completely specific at search machines like Google. The Semantic Multimedia Access project is therefore developing tools with which multi media archives can be searched on a conceptual level.
Metadata
The project has developed an open source search machine generator named Spiegle. It uses metadata to infer the actual user wish and individual search behavior. In this way, a customized search machine comes into being. There won’t be one search machine for everything and everyone, but every user will have his own personal search machine. The user profile will largely determine the results and improve the quality.
Spiegle un’Google’d
The project team starts from the premise that it’s possible to create a search environment that reduces the inevitable flaws of a general system and compensates for them. This means that the project team is actually working on ‘ungoogling’ the current one-dimensional search machine technology. Spiegle departs from properties of the collection to be searched (formats, domains, language, period, background information), the user (scenario, interaction type, history, preferences), and the platform at use (pc, laptop, telephone).
Participanten
Image and Sound has taken a seat at the worktable as first user. Image and Sound manages about 70% of the Dutch audiovisual heritage. The organization cooperates within MultimediaN with scientists from University of Twente, University of Amsterdam, Center for Math and Computer Science (CWI), and Philips.
Project leader: Prof. dr. Franciska de Jong, University of Twente
E-mail: f.m.g.dejong@ewi.utwente.nl
Learning Features (N1)
Multimodal Interaction (N2)
Ambient Multimedia Databases (N3)
Semantic Multimedia Access (N5)
Professional's Dashboard (N6)
Video At Your Fingertips (N7)
E-Culture (N9C)
PERsonal Information Services (N9MI)