作者: M. Newman , L. Gillick , Y. Ito , D. McAllaster , B. Peskin
DOI: 10.1109/ICSLP.1996.607297
关键词:
摘要: The authors present a study of speaker verification system for telephone data based on large-vocabulary speech recognition. After describing the recognition engine, they give details algorithm and draw comparisons with other systems. has been tested test set taken from Switchboard corpus conversational speech, results showing how performance varies length utterance, whether or not training transcribed. dominant factor in appears to be channel handset mismatch between testing data.