In this paper, in order to improve the quality and reconstruction of the noisy speech signal, an algorithm based on clustering in the spectrographic image texture is proposed and evaluated. In this algorithm, we first ...