Speech Signal De-Noising Using Negative Entropy Based Independent Component Analysis and Adaptive Wavelet Packet Thresholding

This paper presents an innovative approach to speech signal de-noising in highly Non-Stationary noise Environment. The proposed technique is based on Independent Component Analysis for Blind Source Separation along with efficient and recent popular Adaptive Wavelet packet Thresholding scheme. An Independent Component Analysis is phenomenon which separates various statistically independent components of input speech signal from an observed mixture vector. An Independent component analysis extracts clean speech signal and noise signal from a mixture of individual sources. Negative Entropy measure of non Gaussianity is used as basic principle of independent component analysis. An Adaptive Wavelet packet Thresholding algorithm is applied to eliminate residual noise which remains in extracted individual signals. Prior to this algorithm, an estimation of noise level is carried out. The Time as well as frequency domain Objective Quality Measures such as Log-Likelihood Ratio (LLR), Segmental Signal to Noise Ratio (SNRseg), Weighted Spectral Slope (WSS), Perceptual Evaluation of speech Quality (PESQ), Itakura-Saito (IS) Ratio are then evaluated for resultant Enhanced speech signal with respect to the original desired signal.