The DARPA 1000-word resource management database for continuous speech recognition

摘要: A database of continuous read speech has been designed and recorded within the DARPA strategic computing recognition program. The data is intended for use in designing evaluating algorithms speaker-independent, speaker-adaptive speaker-dependent recognition. consists sentences appropriate to a naval resource management task built around existing interactive graphics programs. 1000-word vocabulary be logically complete habitable. database, which represents over 21000 utterances from 160 talkers with variety dialects, includes partition training testing purposes. >