Robust approach for blind separation of noisy mixtures of independent and dependent sources

In this paper, a new Blind Source Separation (BSS) method that handles mixtures of noisy independent/dependent sources is introduced. We achieve that by minimizing a criterion that fuses a separating part, based on Kullback–Leibler divergence for either dependent or independent sources, with a regularization part that employs the bilateral total variation (BTV) for the purpose of denoising the observations. The proposed algorithm utilizes a primal-dual algorithm to remove the noise, while a gradient descent method is implemented to retrieve the signal sources. Our algorithm has shown its effectiveness and efficiency and also surpassed the standard existing BSS algorithms.