Sparsity and low-rank amplitude based blind Source Separation

Fangchen Feng (1) , Matthieu Kowalski (1)

1 L2S - Laboratoire des signaux et systèmes

This paper presents a new method for blind source separation problem in reverberant environments with more sources than microphones. Based on the sparsity property in the time-frequency domain and the low-rank assumption of the spectrogram of the source, the STRAUSS (SparsiTy and low-Rank AmplitUde based Source Separation) method is developed. Numerical evaluations show that the proposed method outperforms the existing multichannel NMF approaches, while it is exclusively based on amplitude information.

Mots clés: low-rank; multichannel non-negative matrix factorization; Blind source separation; sparsity

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

Fangchen Feng, Matthieu Kowalski. Sparsity and low-rank amplitude based blind Source Separation. The 42nd IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP 2017), Mar 2017, New Orleans, United States. pp.571 - 575, ⟨10.1109/ICASSP.2017.7952220⟩. ⟨hal-01547459⟩