Diffuse noise robust multiple source localization based on matrix completion via trace norm minimization

Nobutaka Ito (1, 2) , Emmanuel Vincent (1) , Nobutaka Ono (2) , Rémi Gribonval (1) , Shigeki Sagayama (2)

1 METISS - Speech and sound data modeling and processing
2 UTokyo - The University of Tokyo

We aim to improve the robustness of sound localization to diffuse noise coming from various directions. Using a subspace model of the noise covariance matrix, we estimate the signal covariance matrix from the observed covariance matrix and subsequently apply MUSIC (MUltiple SIgnal Classification). In [2], we proposed a maximum likelihood method given the rank of the signal covariance matrix. In this paper, we propose an alternative method based on minimizing the trace norm.

Domaines: Traitement du signal et de l'image [eess.SP]; Traitement du signal et de l'image [eess.SP]

Nobutaka Ito, Emmanuel Vincent, Nobutaka Ono, Rémi Gribonval, Shigeki Sagayama. Diffuse noise robust multiple source localization based on matrix completion via trace norm minimization. ASJ Spring Meeting, Mar 2011, Tokyo, Japan. ⟨inria-00596138⟩