0041069 20240111140639.220060913235959.9 5 s. CD-ROM cav_un_epca*00767641-5FlorenceEURASIP2006 independent component analysis blind source separation cav_un_auth*0101212 Tichavský Petr UTIA-B Department of Stochastic Informatics Ústav teorie informace a automatizace AV ČR, v. v. i. cav_un_auth*0213972 Doron E. IL cav_un_auth*0213973 Yeredor A. IL cav_un_auth*0213020 Nielsen Jan UTIA-B Ústav teorie informace a automatizace AV ČR, v. v. i. datovy soubor 212kB 12B 1M0572 GA MŠk cav_un_auth*0001814 CEZ:AV0Z10750506 The second-order blind identification (SOBI) algorithm for separation of stationary sources was proved to be useful in many biomedical applications. This paper revisits the so called weights-adjusted variant of SOBI, known as WASOBI, which is asymptotically optimal (in separating Gaussian parametric processes), yet prohibitively computationally demanding for more than 2-3 sources. A computationally feasible implementation of the algorithm is proposed, which has a complexity of the same order as SOBI. Excluding the estimation of the correlation matrices, the post-processing complexity of SOBI is $O(d^4M)$, where $d$ is the number of the signal components and $M$ is the number of covariance matrices involved. The additional complexity of our proposed implementation of WASOBI is $O(d^6+d^3M^3)$ operations. However, for WASOBI, the number $M$ of the matrices can be significantly lower than that of SOBI without compromising performance. WASOBI is shown to significantly outperform SOBI in simulation, and can be applied, e.g., in the processing of low density EEG signals. Algoritmus SOBI je popularnim algortmem pro slepou separaci signalu, pouzivanym v biomedicine. Tento clanek se zabyva implementaci asymptoticky optimalni varianty tohoto algoritmu zname pod akronymem WASOBI. Je navrzena varianta tohoto algorimu ktera ma radove stejnou vypocetni narocnost jako puvodni algoritmus SOBI a umoznuje separaci 20 nezavislych zdroju v casovem horizontu minut. Je ukazana zlepsena presnost separace v porovnani s algoritmem SOBI. cav_un_auth*0216569 European Signal Processing Conference. EUSIPCO /14./ Florence 04.09.2006-08.09.2006 IT 2007 BB http://hdl.handle.net/11104/0134655 2006 datovy soubor 212kB cav_un_epca*0076764 Proceedings of 14th European Signal Processing Conference. EUSIPCO 2006 1 5 Florence EURASIP 2006