Rival Penalized Competitive Learning Based Separator on Binary Sources Separation
Yiu‐ming Cheung, Lei Xu · 1998
This paper 1 presents an approach named Rival Penalized Competitive Learning based Binary Source Separator (RPCLBSS) , which has two major advantages: (1) fast in implementation, (2) able to automatically determine the number of binary sources, and (3) able to reduce the noise effects. Experiments have shown that RPCL-BSS algorithm can not only find out the correct number of sources quickly, but also be insensitive to Gaussian noise such that the source signals can be exactly recovered under the low-level noise interference. KEYWORDS: Rival Penalized Competitive Learning, Binary Source Separation, Number of Sources, Noise Interference 1. Introduction In recent years, many approaches have been proposed to solve Blind Source Separation (BSS) problems due to its attracting application on wireless communication [1], image processing [2] and so on. Typical examples include Informationmaximization (INFORMAX) [3] , Minimum Mutual Information (MMI) [4] and Learned Parametric Mixture (LPM...