Complete blind subspace deconvolution
There is a more recent version of this eprint available. Click here to view it.
In this paper we address the blind subspace deconvolution (BSSD) problem; an extension of both the blind source deconvolution (BSD) and the independent subspace analysis (ISA) tasks. While previous works have been focused on the undercomplete case, here we extend the theory to complete systems. Particularly, we derive a separation technique for the complete BSSD problem: we solve the problem by reducing the estimation task to ISA via linear prediction. Numerical examples illustrate the efficiency of the proposed method.
Available Versions of this Item