A new algorithm of Infomax for small numbers of sound signal separation

Abstract

Independent Component Analysis (ICA) is a method of finding unknown source signals from signal mixtures, and it is just one of many solutions to the Blind source separation (BSS )problem. This research focuses on the “Infomax” algorithm, which finds a number of independent source signals from the same number of signal mixtures by maximizing the entropy of the signals. For small numbers of signal mixtures (two to three), the Infomax algorithm is found to be rather efficient.