Title :
Neural network for demixing super-Gaussian signals
Author :
Prieto, B. ; Prieto, A. ; Puntonet, C.G. ; Martin-Smith, P.
Author_Institution :
Dept. de Arquitectura y Tecnologia de Computadores, Granada Univ., Spain
fDate :
8/17/2000 12:00:00 AM
Abstract :
A new method for separating linear mixtures of statistically independent signals with super-Gaussian probability distributions, using a simple neural network, is proposed. The procedure is based on geometric properties, and it is shown that the maxima of the mixed density distribution belong to straight lines, the direction vectors of which, when taken as columns of a matrix, comprise a demixing matrix. The results obtained with synthetic mixtures of real speech signals are shown
Keywords :
signal processing; neural network; signal demixing; speech separation; super-Gaussian probability distribution;
Journal_Title :
Electronics Letters
DOI :
10.1049/el:20001053
