An extended model for speech segregation

Speech segregation is an important task of auditory scene analysis (ASA), in which the speech of a certain speaker is separated from other interfering signals. Wang and Brown (1999) proposed a multistage neural model for speech segregation, the core of which is a two-layer oscillator network. We extend their model by adding further processes based on psychoacoustic evidence to improve the performance. These processes include estimation of the pitch of target speech and refined generation of a target speech stream with the estimated pitch. Our model is systematically evaluated and compared with the Wang-Brown model, and it yields significantly better performance