On convolutive Blind Source Separation in a noisy context and a total variation regularization

T.Z. Z Boulmezaoud (1) , M. El Rhabi (2) , H. Fenniri (3) , E. Moreau (4)

1 LMV - Laboratoire de Mathématiques de Versailles
2 IMI - Département Ingénierie Mathématique et Informatique, École des Ponts ParisTech
3 CRESTIC - Centre de Recherche en Sciences et Technologies de l'Information et de la Communication - EA 3804
4 LIS - Laboratoire d'Informatique et des Systèmes (LIS) (Marseille, Toulon)

We propose a new strategy for improving classical Blind Source Separation (BSS) methods. This strategy consists in denois-ing both the observed signal and the estimated source signal , and is based on the minimization of regularized criterion which takes into account the Total Variation of the signal. We prove by the way that the method leads to a projection problem which is solved by means of projected gradient algorithm. The effectiveness and the robustness of the proposed separating process are shown on numerical examples.

Domaines: Analyse numérique [math.NA]; Statistiques [math.ST]; Traitement du signal et de l'image [eess.SP]; Théorie de l'information [cs.IT]

T.Z. Z Boulmezaoud, M. El Rhabi, H. Fenniri, E. Moreau. On convolutive Blind Source Separation in a noisy context and a total variation regularization. 2010 IEEE 11th International Workshop on Signal Processing Advances in Wireless Communications (SPAWC 2010), Jun 2010, Marrakech, Morocco. ⟨10.1109/SPAWC.2010.5671015⟩. ⟨hal-01812533⟩