A trainable speech enhancement technique based on mixture models for speech and noise.

Ilyas Potamitis, Nikos Fakotakis, George Kokkinakis

Browse the full Interspeech paper archive.