A novel method for noise cancellation of speech signals using wavelet packets

Abstract

In this paper a novel method for cancellation of broadband/narrowband noise from speech signals is proposed. Independent component analysis (ICA) and wavelet packet approaches have been combined for blind noise separation from mixtures of speech signals. ICA method is used to estimate matrix A, which defines how the mixture signals have been mixed. Wavelet packets are used for de-correlation of approximation of noise and speech. This algorithm is applied to real data with high SNRs and has been able to completely retrieve speech signal from white background noise successfully