The Informedia research group at Carnegie Mellon University has worked with speech recognition, image processing, and language technologies through the years to facilitate machine understanding of video materials. This site, idvl.org, is devoted to aspects of Informedia research especially pertinent to Digital Video Libraries for cultural assets, with NSF funding supporting the application of Informedia technologies for oral histories. Hosted at idvl.org are examples of Informedia Digital Video Library (IDVL) interfaces that build from automatic alignment of transcripts to video, named entity detection, geocoding to support map search, and other multimedia indexing and interface technologies. The IDVL interfaces are built with Adobe Flex, producing a Flash application. Users armed with the Flash plug-in for their web browser can explore the IDVL data sets on their own.
See PUBLICATIONS for papers, slide sets, and more information, including OHA 2012.