Complex extension of infinite sparse factor analysis for blind speech separation

Kohei Nagira; Toru Takahashi; Tetsuya Ogata; Hiroshi G. Okuno

doi:10.1007/978-3-642-28551-6_48

Complex extension of infinite sparse factor analysis for blind speech separation

Kohei Nagira^*, Toru Takahashi, Tetsuya Ogata, Hiroshi G. Okuno

^*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

3 Citations (Scopus)

Abstract

We present a method of blind source separation (BSS) for speech signals using a complex extension of infinite sparse factor analysis (ISFA) in the frequency domain. Our method is robust against delayed signals that usually occur in real environments, such as reflections, short-time reverberations, and time lags of signals arriving at microphones. ISFA is a conventional non-parametric Bayesian method of BSS, which has only been applied to time domain signals because it can only deal with real signals. Our method uses complex normal distributions to estimate source signals and mixing matrix. Experimental results indicate that our method outperforms the conventional ISFA in the average signal-to-distortion ratio (SDR).

Original language	English
Title of host publication	Latent Variable Analysis and Signal Separation - 10th International Conference, LVA/ICA 2012, Proceedings
Pages	388-396
Number of pages	9
DOIs	https://doi.org/10.1007/978-3-642-28551-6_48
Publication status	Published - 2012
Externally published	Yes
Event	10th International Conference on Latent Variable Analysis and Signal Separation, LVA/ICA 2012 - Tel Aviv, Israel Duration: 2012 Mar 12 → 2012 Mar 15

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	7191 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	10th International Conference on Latent Variable Analysis and Signal Separation, LVA/ICA 2012
Country/Territory	Israel
City	Tel Aviv
Period	12/3/12 → 12/3/15

Keywords

Blind source separation
Infinite sparse factor analysis
Non-parametric Bayes

ASJC Scopus subject areas

Theoretical Computer Science
Computer Science(all)

Access to Document

10.1007/978-3-642-28551-6_48

Cite this

Nagira, K., Takahashi, T., Ogata, T., & Okuno, H. G. (2012). Complex extension of infinite sparse factor analysis for blind speech separation. In Latent Variable Analysis and Signal Separation - 10th International Conference, LVA/ICA 2012, Proceedings (pp. 388-396). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 7191 LNCS). https://doi.org/10.1007/978-3-642-28551-6_48

Complex extension of infinite sparse factor analysis for blind speech separation. / Nagira, Kohei; Takahashi, Toru; Ogata, Tetsuya et al.
Latent Variable Analysis and Signal Separation - 10th International Conference, LVA/ICA 2012, Proceedings. 2012. p. 388-396 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 7191 LNCS).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Nagira, K, Takahashi, T, Ogata, T & Okuno, HG 2012, Complex extension of infinite sparse factor analysis for blind speech separation. in Latent Variable Analysis and Signal Separation - 10th International Conference, LVA/ICA 2012, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 7191 LNCS, pp. 388-396, 10th International Conference on Latent Variable Analysis and Signal Separation, LVA/ICA 2012, Tel Aviv, Israel, 12/3/12. https://doi.org/10.1007/978-3-642-28551-6_48

Nagira K, Takahashi T, Ogata T, Okuno HG. Complex extension of infinite sparse factor analysis for blind speech separation. In Latent Variable Analysis and Signal Separation - 10th International Conference, LVA/ICA 2012, Proceedings. 2012. p. 388-396. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-642-28551-6_48

Nagira, Kohei ; Takahashi, Toru ; Ogata, Tetsuya et al. / Complex extension of infinite sparse factor analysis for blind speech separation. Latent Variable Analysis and Signal Separation - 10th International Conference, LVA/ICA 2012, Proceedings. 2012. pp. 388-396 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{64f46d9fb4474744bbb573c9ff83a596,

title = "Complex extension of infinite sparse factor analysis for blind speech separation",

abstract = "We present a method of blind source separation (BSS) for speech signals using a complex extension of infinite sparse factor analysis (ISFA) in the frequency domain. Our method is robust against delayed signals that usually occur in real environments, such as reflections, short-time reverberations, and time lags of signals arriving at microphones. ISFA is a conventional non-parametric Bayesian method of BSS, which has only been applied to time domain signals because it can only deal with real signals. Our method uses complex normal distributions to estimate source signals and mixing matrix. Experimental results indicate that our method outperforms the conventional ISFA in the average signal-to-distortion ratio (SDR).",

keywords = "Blind source separation, Infinite sparse factor analysis, Non-parametric Bayes",

author = "Kohei Nagira and Toru Takahashi and Tetsuya Ogata and Okuno, {Hiroshi G.}",

year = "2012",

doi = "10.1007/978-3-642-28551-6_48",

language = "English",

isbn = "9783642285509",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

pages = "388--396",

booktitle = "Latent Variable Analysis and Signal Separation - 10th International Conference, LVA/ICA 2012, Proceedings",

note = "10th International Conference on Latent Variable Analysis and Signal Separation, LVA/ICA 2012 ; Conference date: 12-03-2012 Through 15-03-2012",

}

TY - GEN

T1 - Complex extension of infinite sparse factor analysis for blind speech separation

AU - Nagira, Kohei

AU - Takahashi, Toru

AU - Ogata, Tetsuya

AU - Okuno, Hiroshi G.

PY - 2012

Y1 - 2012

N2 - We present a method of blind source separation (BSS) for speech signals using a complex extension of infinite sparse factor analysis (ISFA) in the frequency domain. Our method is robust against delayed signals that usually occur in real environments, such as reflections, short-time reverberations, and time lags of signals arriving at microphones. ISFA is a conventional non-parametric Bayesian method of BSS, which has only been applied to time domain signals because it can only deal with real signals. Our method uses complex normal distributions to estimate source signals and mixing matrix. Experimental results indicate that our method outperforms the conventional ISFA in the average signal-to-distortion ratio (SDR).

AB - We present a method of blind source separation (BSS) for speech signals using a complex extension of infinite sparse factor analysis (ISFA) in the frequency domain. Our method is robust against delayed signals that usually occur in real environments, such as reflections, short-time reverberations, and time lags of signals arriving at microphones. ISFA is a conventional non-parametric Bayesian method of BSS, which has only been applied to time domain signals because it can only deal with real signals. Our method uses complex normal distributions to estimate source signals and mixing matrix. Experimental results indicate that our method outperforms the conventional ISFA in the average signal-to-distortion ratio (SDR).

KW - Blind source separation

KW - Infinite sparse factor analysis

KW - Non-parametric Bayes

UR - http://www.scopus.com/inward/record.url?scp=84863178508&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84863178508&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-28551-6_48

DO - 10.1007/978-3-642-28551-6_48

M3 - Conference contribution

AN - SCOPUS:84863178508

SN - 9783642285509

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 388

EP - 396

BT - Latent Variable Analysis and Signal Separation - 10th International Conference, LVA/ICA 2012, Proceedings

T2 - 10th International Conference on Latent Variable Analysis and Signal Separation, LVA/ICA 2012

Y2 - 12 March 2012 through 15 March 2012

ER -

Complex extension of infinite sparse factor analysis for blind speech separation

Abstract

Publication series

Conference

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this