Source separation using multiple directivity patterns produced by ICA-based BSS

Takashi Isa, Toshiyuki Sekiya, Tetsuji Ogawa, Tetsunori Kobayashi

    Research output: Chapter in Book/Report/Conference proceedingConference contribution

    Abstract

    In this paper, we propose a multistage source separation method constructed by combining blind source separation (BSS) based on independent component analysis (ICA) and segregation using multiple directivity patterns (SMDP) introduced in our previous paper. We obtain the directivity patterns needed in SMDP by ICAbased BSS. In the SMDP, simultaneous equations of amplitudes of sound sources are generated by using these multiple directivities. The solution of these equations gives good disturbance estimates. We apply spectral subtraction using these disturbance estimates and the speech enhancement of the target source is performed. We conducted experimentation in a real room in the source-number-given condition where there is no priori information about the sound sources and the characteristics of room acoustics. The experimental results of double talk recognition show that the proposed technique is effective in reducing the error rate by 30% compared to frequency domain BSS.

    Original languageEnglish
    Title of host publicationEuropean Signal Processing Conference
    Publication statusPublished - 2006
    Event14th European Signal Processing Conference, EUSIPCO 2006 - Florence, Italy
    Duration: 2006 Sep 42006 Sep 8

    Other

    Other14th European Signal Processing Conference, EUSIPCO 2006
    CountryItaly
    CityFlorence
    Period06/9/406/9/8

    Fingerprint

    Source separation
    Blind source separation
    Independent component analysis
    Acoustic waves
    Speech enhancement
    Acoustics

    ASJC Scopus subject areas

    • Signal Processing
    • Electrical and Electronic Engineering

    Cite this

    Source separation using multiple directivity patterns produced by ICA-based BSS. / Isa, Takashi; Sekiya, Toshiyuki; Ogawa, Tetsuji; Kobayashi, Tetsunori.

    European Signal Processing Conference. 2006.

    Research output: Chapter in Book/Report/Conference proceedingConference contribution

    Isa, T, Sekiya, T, Ogawa, T & Kobayashi, T 2006, Source separation using multiple directivity patterns produced by ICA-based BSS. in European Signal Processing Conference. 14th European Signal Processing Conference, EUSIPCO 2006, Florence, Italy, 06/9/4.
    Isa, Takashi ; Sekiya, Toshiyuki ; Ogawa, Tetsuji ; Kobayashi, Tetsunori. / Source separation using multiple directivity patterns produced by ICA-based BSS. European Signal Processing Conference. 2006.
    @inproceedings{feba56dee1d8406cba14984d587e644c,
    title = "Source separation using multiple directivity patterns produced by ICA-based BSS",
    abstract = "In this paper, we propose a multistage source separation method constructed by combining blind source separation (BSS) based on independent component analysis (ICA) and segregation using multiple directivity patterns (SMDP) introduced in our previous paper. We obtain the directivity patterns needed in SMDP by ICAbased BSS. In the SMDP, simultaneous equations of amplitudes of sound sources are generated by using these multiple directivities. The solution of these equations gives good disturbance estimates. We apply spectral subtraction using these disturbance estimates and the speech enhancement of the target source is performed. We conducted experimentation in a real room in the source-number-given condition where there is no priori information about the sound sources and the characteristics of room acoustics. The experimental results of double talk recognition show that the proposed technique is effective in reducing the error rate by 30{\%} compared to frequency domain BSS.",
    author = "Takashi Isa and Toshiyuki Sekiya and Tetsuji Ogawa and Tetsunori Kobayashi",
    year = "2006",
    language = "English",
    booktitle = "European Signal Processing Conference",

    }

    TY - GEN

    T1 - Source separation using multiple directivity patterns produced by ICA-based BSS

    AU - Isa, Takashi

    AU - Sekiya, Toshiyuki

    AU - Ogawa, Tetsuji

    AU - Kobayashi, Tetsunori

    PY - 2006

    Y1 - 2006

    N2 - In this paper, we propose a multistage source separation method constructed by combining blind source separation (BSS) based on independent component analysis (ICA) and segregation using multiple directivity patterns (SMDP) introduced in our previous paper. We obtain the directivity patterns needed in SMDP by ICAbased BSS. In the SMDP, simultaneous equations of amplitudes of sound sources are generated by using these multiple directivities. The solution of these equations gives good disturbance estimates. We apply spectral subtraction using these disturbance estimates and the speech enhancement of the target source is performed. We conducted experimentation in a real room in the source-number-given condition where there is no priori information about the sound sources and the characteristics of room acoustics. The experimental results of double talk recognition show that the proposed technique is effective in reducing the error rate by 30% compared to frequency domain BSS.

    AB - In this paper, we propose a multistage source separation method constructed by combining blind source separation (BSS) based on independent component analysis (ICA) and segregation using multiple directivity patterns (SMDP) introduced in our previous paper. We obtain the directivity patterns needed in SMDP by ICAbased BSS. In the SMDP, simultaneous equations of amplitudes of sound sources are generated by using these multiple directivities. The solution of these equations gives good disturbance estimates. We apply spectral subtraction using these disturbance estimates and the speech enhancement of the target source is performed. We conducted experimentation in a real room in the source-number-given condition where there is no priori information about the sound sources and the characteristics of room acoustics. The experimental results of double talk recognition show that the proposed technique is effective in reducing the error rate by 30% compared to frequency domain BSS.

    UR - http://www.scopus.com/inward/record.url?scp=84862589822&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=84862589822&partnerID=8YFLogxK

    M3 - Conference contribution

    BT - European Signal Processing Conference

    ER -