Temporal structure based adaptive semiblind source separation algorithm
Fasong Wang · Journal of Tsinghua University(Science and Technology) · 2007
An adaptive semiblind signal separation algorithm was developed for the separation of a class of arbitrarily distributed but temporally correlated source signals.The adaptive algorithm is based on only second-order statistical(SOS) information and exploits the assumption that the source signals are not strictly spatially statistically independent and are not identically distributed temporally.To get the ideal separation,the source signals must have different power spectra shapes.The algorithm alleviates the limitation that the classic ICA algorithm can separate at most one Gaussian distributed source.Simulation results for real image recordings show that the algorithm can restore source signals,and its performance is comparable to that of BSS algorithms such as JADE,FPICA,FOBI,and AMUSE.