engleski

Cochlear Model based Enhancement of Noisy Speech Signals

Many noise reduction techniques were developed over the past decades and noise reduction is still a major problem in many applications. Although reducing noise, these algorithms typically introduce some distortion to speech signal. Humans are quite capable of detecting speech in background noise, so we propose a method for noise reduction based on the model of cochlear processing of speech signals. Using our model of signal reconstruction from the cochlear output we have achieved improvement in speech quality. Our experiments show that the proposed approach performs better than several other noise reduction methods.

Cochlear model; noise reduction; speech signal; PESQ

nije evidentirano