Wideband blind identification and separation of independent sources
Wang Jun · European Signal Processing Conference · 1996
During recent years, there have been much interests focused on blind identification and blind source separation. Many approaches have been proposed mainly for the special and simple case in which the MIMO system is linear memoryless[1][2][4][5]. However, in a variety of applications in which wideband sources are involved, e.g., speech dereverberation, speech enchancement in the presence of background noise and competing speakers separation, these approaches can not be applied. Several solutions have been presented for the two-source two-sensor wideband problem with two channels assumed identity systems[6-8].