VIENNA, 23 May 2002: Sail Labs Technology launched Media Mining System 2.0, the latest version of the award-winning technology. The Media Mining System 2.0 has the capability to index and summarize Spanish audio data in addition to handling audio data in languages like English, German, and Arabic. The Spanish speech recognition module covers 98% of the words usually used in Spanish news broadcasts. The Spanish version also has the Named Entity Detection feature that uses statistical models to identify elements within the audio data, which belong to categories like Person, Organization, and Location.

The Spanish Media Mining System in its present version is able to identify more than 120 speakers based on a voice print match with existing acoustic models for individual speakers. The system saves the user a lot of time in the search and summarization process by prior identification of topics within the broadcasts based on topic categories. The Spanish version presently has more than 100 topics ranging from Technology to Terrorism. This can be extended to include any number of speakers and topics. Currently the Spanish version of the Media Mining System is specially geared towards the form of Spanish spoken predominantly in North America and Mexico. Users can obtain summarized views of Spanish broadcasts summarized by topic, person, organization, location or speaker.

‘We view the introduction of the Spanish language feature as an important milestone for the broadcasting and media asset management firms operating global networks. The introduction of these languages will enable not only multimedia companies operating in the areas where these languages are the national language but also companies targeting new markets with a dormant need for search and summarization of foreign broadcasts,’ says Dr Thomas Netousek, CEO of Sail Labs Technology.

The Spanish language feature is a feature of the new Media Mining System 2.0 that is based on a distributed CORBA architecture. This enables the distribution of the components of the Media Mining System over several processors for obtaining indexed results in real time. The new Sail Labs Media Mining Indexer 2.0 has a standard application-programming interface that makes its integration into asset management and audio-video archiving products easier and simpler. The Media Mining System uses audio content from various sources like radio and TV channels, press agency news feeds, and archived media streams as input to produce an indexed output in real time. This output can now be searched for relevant information through a user-friendly graphical user interface. Incidentally, the Sail Labs Media Mining System was the recipient of the Merkur Innovation Award for product excellence awarded by the Austrian National Chamber of Commerce and the Austrian Wirtschaftsförderungsinstitut, an acclaimed Austrian institute for economic studies and training.