CICAAR - Convolutive ICA with an Auto-Regressive Inverse Model

Mads Dyrholm; Lars Kai Hansen

CICAAR - Convolutive ICA with an Auto-Regressive Inverse Model

Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review

331 Downloads (Pure)

Abstract

We invoke an auto-regressive IIR inverse model for convolutive ICA and derive expressions for the likelihood and its gradient. We argue that optimization will give a stable inverse. When there are more sensors than sources the mixing model parameters are estimated in a second step by least squares estimation. We demonstrate the method on synthetic data and finally separate speech and music in a real room recording.

Original language	English
Title of host publication	Independent Component Analysis and Blind Signal Separation
Publisher	Springer
Publication date	2004
Pages	594-601
Publication status	Published - 2004

Access to Document

Imm3367

http://www.imm.dtu.dk/pubdb/p.php?3367

OpenUrl availability

Full text

Cite this

@inproceedings{898c9d102def417aa9c51c5c5d9c10e0,

title = "CICAAR - Convolutive ICA with an Auto-Regressive Inverse Model",

abstract = "We invoke an auto-regressive IIR inverse model for convolutive ICA and derive expressions for the likelihood and its gradient. We argue that optimization will give a stable inverse. When there are more sensors than sources the mixing model parameters are estimated in a second step by least squares estimation. We demonstrate the method on synthetic data and finally separate speech and music in a real room recording.",

keywords = "Convolutive ICA AR speech deconvolution maximum likelihood",

author = "Mads Dyrholm and Hansen, {Lars Kai}",

year = "2004",

language = "English",

pages = "594--601",

booktitle = "Independent Component Analysis and Blind Signal Separation",

publisher = "Springer",

}

TY - GEN

T1 - CICAAR - Convolutive ICA with an Auto-Regressive Inverse Model

AU - Dyrholm, Mads

AU - Hansen, Lars Kai

PY - 2004

Y1 - 2004

N2 - We invoke an auto-regressive IIR inverse model for convolutive ICA and derive expressions for the likelihood and its gradient. We argue that optimization will give a stable inverse. When there are more sensors than sources the mixing model parameters are estimated in a second step by least squares estimation. We demonstrate the method on synthetic data and finally separate speech and music in a real room recording.

AB - We invoke an auto-regressive IIR inverse model for convolutive ICA and derive expressions for the likelihood and its gradient. We argue that optimization will give a stable inverse. When there are more sensors than sources the mixing model parameters are estimated in a second step by least squares estimation. We demonstrate the method on synthetic data and finally separate speech and music in a real room recording.

KW - Convolutive ICA AR speech deconvolution maximum likelihood

M3 - Article in proceedings

SP - 594

EP - 601

BT - Independent Component Analysis and Blind Signal Separation

PB - Springer

ER -

CICAAR - Convolutive ICA with an Auto-Regressive Inverse Model

Abstract

Access to Document

OpenUrl availability

Fingerprint

Cite this