Source reduction method using wavelets for the source separation problem

The blind source separation problem is to estimate the number of the unknown source signals and to separate the original source signals from the several observed signals, which are assumed to be linear superpositions of the sources. The coefficients of the superposition are also unknown. In this paper, in order to overcome a difficulty in previously proposed methods, a new approach to solve the problem, named source reduction method, using continuous wavelet transform is proposed.