S\'eparation de sources doublement non stationnaire

TL;DR

This paper introduces a novel algorithm for blind source separation in doubly nonstationary environments, where both the mixing matrix and sources vary over time, using wavelet transform approximations.

Contribution

It proposes a new joint BSS and deformation estimation algorithm tailored for nonstationary signals with time-varying mixing matrices and source spectra.

Findings

Algorithm outperforms existing nonstationary BSS methods in simulations.

Effective estimation of time-dependent mixing matrices and source deformations.

Demonstrates robustness to nonstationarity in source signals.

Abstract

Blind source separation (BSS) techniques aims at joint estimation of source signals and a mixing matrix from observations of mixtures. This paper addresses a doubly nonstationary BSS problem, where the mixing matrix is time dependent and sources are nonstationary, more precisely deformed stationary signals, following the model of [1]. An algorithm for joint BSS and estimation of stationarity-breaking deformations and spectra is introduced, that exploits suitable approximations for the behavior of the wavelet transform of such nonstationary signals. The performance of the approach is evaluated on numerical simulations, and compared with other nonstationary BSS algorithms.