PCA based single channel speech enhancement method for highly noisy environment

In this paper, we proposed speech enhancement method using principal component analysis (PCA) for noisy signal. This algorithm is based on the PCA which is subspace approach. Subspace method separates the signal and noise subspace using eigenvalue analysis. Improved signal can be reconstructed by removing the noise subspace and retaining the signal subspace by selecting suitably the number of principal components. In this paper the experimental results shows the good noise reduction with minimum signal distortion.