Sail Labs unveils Chinese language Media Mining Indexer
April 27, 2003
Sail Labs Technology demonstrated the Chinese language version of its speech technology based Media Mining Indexer, which enables users to search and summarize Chinese news broadcasts in real time.
Users can view transcriptions of audio clips belonging to specific speakers through the Speaker Identification feature. Reading and Analysis of news becomes easier with the Named Entity Detection feature, highlighting keywords belonging to specific categories within the text output. Knowledge of Chinese is not a barrier for information retrieval as the English translation appears next to specific keywords. The introduction is a move to answer the need for a high accuracy audio indexing tool in Mandarin Chinese speaking regions. Sail Labs already offers Media Mining Indexer in accented English version for these regions.
Speech is currently a difficult medium to browse and search. The Media Mining Indexer takes audio input from various sources to produce a searchable indexed output in real-time. It is currently available in English, German, French, Spanish and Arabic versions. Chinese is a language option offered with the Media Mining Indexer 2.3. In the creation of the Chinese version a new mode of signal extraction was implemented to handle the special tonal qualities of the language for higher speech recognition accuracy. Another addition in version 2.3 has been the incorporation of the pay-per-use feature enabling application service providers and media observation companies to initiate operations and start generating revenue with minimal capital expenditure.
While there are numerous Chinese sub-languages and dialects, the current company focus is on speech recognition of natural, unconstrained speech from broadcasts in Mandarin Chinese. Sail Labs' primary intent is to create core enabling, audio indexing platforms suitable for a wide range of advanced applications.
For more information visit
Back to the top
comments powered by Disqus.