0085969 20240111140650.020070919235959.9 5 s. CD ROM cav_un_epca*0086003978-83-921340-2-21731-1735PoznanPTETiS2007 Blind source separation multidimensional independent components FCOMBI cav_un_auth*0230093 Gómez-Herrero G. FI 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*0101212 Tichavský Petr UTIA-B Department of Stochastic Informatics Ústav teorie informace a automatizace AV ČR, v. v. i. cav_un_auth*0230094 Egiazarian K. FI textový soubor 367 kB 1M0572 GA MŠk cav_un_auth*0001814 GP102/07/P384 GA ČR CZ cav_un_auth*0227962 GA102/05/0278 GA ČR CZ cav_un_auth*0051169 CEZ:AV0Z10750506 In this article we propose a computationally efficient method (termed FCOMBI) to combine the strengths of non-Gaussianity based blind source separation (BSS) and cross-correlation-based BSS. This is done by fusing the separtion abilities of two well known algorithms: EFICA and WASOBI. The algorithm is suitable for the analysis of very high-dimensional datasets like high-density Electroencephalogram or Magnetoencephalogram recordings. V clanku je navrzen vypocetne nenarocny algoritmus, ktery kombinuje schopnosti slepe separace dvou odlisnych algorimu: EFICA, ktery vyuziva ne-gaussovskosti rozlozeni jednotlivych separovanych zdroju, a WASOBI, ktery vyuziva rozdilnosti frekvencnich spekter jednotlivych zdroju. Algoritmus je vhodny pro analyzu mnohadimensionalnich dat z elektroencefalogramu a magnetoencefaloogramu s vysokym rozlisenim. cav_un_auth*0230095 EUSIPCO 2007: The 15th European Signal Processing Conference Poznan 03.09.2007-07.09.2007 PL 2008 FH http://hdl.handle.net/11104/0148362 2007 textový soubor 367 kB cav_un_epca*0086003 Proccedings of the 15th European Signal Processing Conference. EUSIPCO 2007 978-83-921340-2-2 1731 1735 Poznan PTETiS 2007