A NN/HMM hybrid for continuous speech recognition with a discriminant nonlinear feature extraction.

Gerhard Rigoll, Daniel Willett

Browse the full ICASSP paper archive.