Probabilistic blind deconvolution of nonstationary sources 
Rasmus Kongsgaard Olsson, Lars Kai Hansen

Abstract  We solve a class of blind signal separation problems using a constrained linear Gaussian model. The observed signal is modelled by a convolutive mixture of colored noise signals with additive white noise. We derive a timedomain EM algorithm `KaBSS' which estimates the source signals, the associated secondorder statistics, the mixing filters and the observation noise
covariance matrix. KaBSS invokes the Kalman smoother in the Estep to infer the posterior probability of the sources, and onestep lower bound optimization of the mixing filters and noise covariance in the Mstep. In line with (Parra and Spence, 2000) the source signals are assumed time variant in order to constrain the solution sufficiently. Experimental results are shown for mixtures of speech signals. 
Keywords  Blind source separation, independent component analysis, speech processing 
Type  Conference paper [With referee] 
Conference  12th European Signal Processing Conference 
Year  2004 Month September pp. 16971700 
Electronic version(s)  [pdf] 
BibTeX data  [bibtex] 
IMM Group(s)  Intelligent Signal Processing 