Speech Enhancement Based on Sparse Theory under Noisy Environment

Recently, the sparse algorithm for sparse enhancement is more and more popular issues. In this paper, we classify the process of the sparse theory to enhance speech signal into two parts, one is for dictionary training part and the other is signal reconstruction part. We focus on the White Gaussian Noise. Clean speech dictionary D is trained by K-SVD algorithm. The orthogonal matching pursuit(OMP) algorithm is used to obtain the sparse coefficients X of clean speech dictionary D. Denoising performance of the experiments shows that our proposed method is superior than other methods in SNR, LLR, SNRseg and PESQ. Keywords-Speech enhancement, sparse representations, K-SVD, discrete cosine transform (DCT), orthogonal matching pursuit (OMP).