Low Rank and Sparsity Analysis Applied to Speech Enhancement via Online Estimated Dictionary

TL;DR

This paper introduces an online dictionary-based speech enhancement method that leverages low rank and sparsity constraints for improved noise suppression in single-channel audio, utilizing local dictionary estimation via probabilistic modeling.

Contribution

The novel contribution is the integration of low rank and sparse matrix decomposition with an online estimated dictionary for speech enhancement, employing a probabilistic approach for dictionary estimation.

Findings

Significant noise reduction compared to four baseline algorithms.

Effective local dictionary estimation improves speech quality.

The method demonstrates robustness in various noisy environments.

Abstract

We propose an online estimated dictionary based single channel speech enhancement algorithm, which focuses on low rank and sparse matrix decomposition. In this proposed algorithm, a noisy speech spectral matrix is considered as the summation of low rank background noise components and an activation of the online speech dictionary, on which both low rank and sparsity constraints are imposed. This decomposition takes the advantage of local estimated dictionary high expressiveness on speech components. The local dictionary can be obtained through estimating the speech presence probability by applying Expectation Maximal algorithm, in which a generalized Gamma prior for speech magnitude spectrum is used. The evaluation results show that the proposed algorithm achieves significant improvements when compared to four other speech enhancement algorithms.