Abstract
Following the seminal work of Stone [Independent Component Analysis, The MIT Press, Cambridge, 2004], this paper presents a new metric for blind source separation (BSS). It is proved that the metric value of any linear combination of source signals is less than the largest one of sources under a loose condition. Further, the global optimization of this new metric is achieved by formulating it as a generalized eigenvalue (GE) problem. Subsequently, we give out a fast BSS algorithm. Moreover, we analyze the solution properties of ill-posed BSS, and further show that the proposed algorithm is applicable to such a case as well. The numerical simulations demonstrate the efficacy of our algorithm.
Original language | English |
---|---|
Pages (from-to) | 973-982 |
Number of pages | 10 |
Journal | Neurocomputing |
Volume | 71 |
Issue number | 4-6 |
DOIs | |
Publication status | Published - Jan 2008 |
User-Defined Keywords
- Blind source separation
- Generalized eigenvalue problem
- Ill-posed ICA
- Independent component analysis
- Optimal solution