Presentation | 2001/12/14 Blind Separation Using Adaptive Nonlinear Functions Controlled by Kurtosis Takayuki Sakai, Kenji Nakayama, Akihiro Hirano, |
---|---|
PDF Download Page | PDF download Page Link |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | Convergence and separation performances are highly dependent on a relation between probability density functions(pdf)of signal sources and nonlinear functions used in updating coefficients of a separation block. This relation was analyzed based on kurtosis κ_4. It was suggested that than y and y^3, where y is the output, are useful nonlinear functions for super-Gaussian(κ_4>0)and sub-Gaussian(κ_4<0), respectively. In this paper, an adaptive nonlinear function is proposed. It has a form of f(y)=a tanh y+(1-a)y^3/4, where a is controlled by kurtosis. It is assumed that the pdf p(y)of the output signal y satisfies the stability condition f(y)=(dp(y)/dy)/p(y). Based on this assumption, the parameter a and the kurtosis is related. First, the kurtosis is calculated for given a, which takes value in 0≤a≤1. Next, this numerical relation is approximated by a function a=q(κ_4). In a learning process, κ_4(n)of the output signals is calculated at each sample n, and a is determined by a(n)=q(κ_4(n)). Then, the nonlinear function f(y) is adjusted. Blind separation of music signals of 2~5 channels were simulated. The proposed method is superior to a method, which switches tanh y and y^3 based on polarity of κ_4(n). |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | Blind separation / Nonlinear functions / Kurtosis / Stabilization / Learning |
Paper # | NC2001-79 |
Date of Issue |
Conference Information | |
Committee | NC |
---|---|
Conference Date | 2001/12/14(1days) |
Place (in Japanese) | (See Japanese page) |
Place (in English) | |
Topics (in Japanese) | (See Japanese page) |
Topics (in English) | |
Chair | |
Vice Chair | |
Secretary | |
Assistant |
Paper Information | |
Registration To | Neurocomputing (NC) |
---|---|
Language | ENG |
Title (in Japanese) | (See Japanese page) |
Sub Title (in Japanese) | (See Japanese page) |
Title (in English) | Blind Separation Using Adaptive Nonlinear Functions Controlled by Kurtosis |
Sub Title (in English) | |
Keyword(1) | Blind separation |
Keyword(2) | Nonlinear functions |
Keyword(3) | Kurtosis |
Keyword(4) | Stabilization |
Keyword(5) | Learning |
1st Author's Name | Takayuki Sakai |
1st Author's Affiliation | Division of Electronics and Computer Science, Graduate School of Natural Science and Technology Dept.of Information and Systems Eng.Faculty of Eng.Kanazawa University() |
2nd Author's Name | Kenji Nakayama |
2nd Author's Affiliation | Dept.of Information and Systems Eng.Faculty of Eng.Kanazawa University |
3rd Author's Name | Akihiro Hirano |
3rd Author's Affiliation | Dept.of Information and Systems Eng.Faculty of Eng.Kanazawa University |
Date | 2001/12/14 |
Paper # | NC2001-79 |
Volume (vol) | vol.101 |
Number (no) | 534 |
Page | pp.pp.- |
#Pages | 8 |
Date of Issue |