Kernel-based nonlinear independent component analysis for underdetermined blind source separation

S. Miyabe,B. Juang,H. Saruwatari,K. Shikano
DOI: https://doi.org/10.1109/ICASSP.2009.4959915
2009-04-19
Abstract:In this paper we propose a new unsupervised training method for nonlinear spatial filter using a new independent component analysis based on kernel infomax. The nonlinearity of the spatial filter used in this paper is equivalent to the integration of beamforming and spectral subtraction, and the whole structure is optimized by independent component analysis in the reproducing kernel Hilbert space. The optimized filter is shown to be capable of achieving better quality output than the conventional method based on time-frequency binary masking.
What problem does this paper attempt to address?