Speech - Non-Speech Separation with GMMs

Type of publication:  Artikel in einem Konferenzbericht
Buchtitel: Acoustic Society of Japan, autumn meeting 2001
Jahr: 2001
Ort: Oita, Japan
Abriss: The separation of speech and non-speech events is an important problem for speech recognition .This paper investigates a separation method based on likelihood ratio using speech and non-speech Gaussioan Mixture models (GMM). The evaluation is done on the SPINE database
Autoren: Binder, Norbert
Markov, K.
Gruhn, R.
Nakamura, S.