Avtor/Urednik		Naanaa, Wady; Nuzillard, Jean-Marc
Naslov		A genetic algorithm for blind source separation of positive and partially correlated signals
Tip		članek
Vir		Image Anal Stereol
Vol. in št.		Letnik 20, št. Suppl 1
Leto izdaje		2001
Obseg		str. 281-6
Jezik		eng
Abstrakt		Blind source separation (BSS) consists of processing a set of observed mixed signals to separate them into a set of unobserved components. Various approaches have been employed to solve BSS problems assuming that the source signals are mutually uncorrelated (or orthogonal). However, in many real-life problems, signal orthogonality is not guaranteed. This paper presents a new method for blind source separation that can be applied to positive and partially orthogonal signals. The BSS problem is transformed into a combinatorial optimization problem which is solved by means of a genetic algorithm (GA). The method has been successfully applied to many actual examples from NMR spectroscopy one of which is reported here.
Deskriptorji		MAGNETIC RESONANCE IMAGING IMAGE PROCESSING, COMPUTER-ASSISTED