0332915 20240111140729.520091202235959.9 10.1016/j.sigpro.2009.04.021 15 s. www cav_un_epca*02550760165-16848912 (2009)2570-2584Elsevier Independent component analysis blind source separation Cramer-Rao lower bound cav_un_auth*0108100 Koldovský Zbyněk UTIA-B Department of Stochastic Informatics Ústav teorie informace a automatizace AV ČR, v. v. i. cav_un_auth*0050739 Málek J. CZ 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*0238913 Deville Y. FR cav_un_auth*0238914 Hosseini S. FR pdf http://library.utia.cas.cz/separaty/2009/SI/tichavsky-blind separationofpiecewisestationarynon-gaussiansources.pdf 1M0572 GA MŠk cav_un_auth*0001814 GA102/09/1278 GA ČR cav_un_auth*0253174 GA102/07/P384 GA ČR CZ CEZ:AV0Z10750506 We address Independent Component Analysis (ICA) of piecewise stationary and nonGaussian signals and propose a novel ICA algorithm called Block EFICA that is based on this generalized model of signals. The method is a further extension of the popular nonGaussianity-based FastICA algorithm and of its recently optimized variant called EFICA. In contrast to these methods, Block EFICA is developed to effectively exploit varying distribution of signals, thus, also their varying variance in time (nonstationarity) or, more precisely, in time-intervals (piecewise stationarity). In theory, the accuracy of the method asymptotically approaches Cramer-Rao lower bound (CRLB) under common assumptions when variance of the signals is constant. On the other hand, the performance is practically close to the CLRB even when variance of the signals is changing. V práci je navržen algoritmus pro slepou separaci lineární okamžité směsi nezávislých po částech stacionárních negaussovských zdrojů. Algoritmus je nazván "bloková EFICA" a je rozšířením a zobecněním dřívějších algorimů FastICA a EFICA. V případě, že separované signálz mají konstantní varianci, dosahuje přesnost separace asymptoticky příslušné Rao-Cramerovy meze. Přínos algoritmu je ukázán na separaci okamžité směsi zvukových signálů. 2010 BB http://hdl.handle.net/11104/0178031 1.283 0.261 7.7 0.0127 0.528 3980 261 68 0.991 25.02 1.77 2009 pdf cav_un_epca*0255076 Signal Processing 0165-1684 1872-7557 Roč. 89 č. 12 2009 2570 2584 Elsevier