Blind Source Separation using Independent Component Analysis |
Blind Source Separation (BSS) is a process of estimating a set of signals by using their mixed representations. In this separation process, the user does not have a prior knowledge of the original signals nor their mixing process. This is also called the Cocktail Party Problem. In this paper, a fixed point algorithm of Independent Component Analysis (ICA) was used to separate linearly mixed stationary and non-stationary mixed audio signals. Also, an attempt was made to solve the permutation problem of linearly mixed non-stationary signals by using the sliding window ICA and Minimum Distance Classifier (MDC).
Kilerci CK, Ahmed MZ