Speaker verification robust to talking style variation using multiple kernel learning based on conditional entropy minimization

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

    1 Citation (Scopus)

    Abstract

    We developed a new speaker verification system that is robust to intra-speaker variation. There is a strong likelihood that intra-speaker variations will occur due to changes in talking styles, the periods when an individual speaks, and so on. It is well known that such variation generally degrades the performance of speaker verification systems. To solve this problem, we applied multiple kernel learning (MKL) based on conditional entropy minimization, which impose the data to be compactly aggregated for each speaker class and ensure that the different speaker classes were far apart from each other. Experimental results showed that the proposed speaker verification system achieved a robust performance to intra-speaker variation derived from changes in the talking styles compared to the conventional maximum margin-based system.

    Original languageEnglish
    Title of host publicationProceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH
    Pages2741-2744
    Number of pages4
    Publication statusPublished - 2011
    Event12th Annual Conference of the International Speech Communication Association, INTERSPEECH 2011 - Florence, Italy
    Duration: 2011 Aug 272011 Aug 31

    Other

    Other12th Annual Conference of the International Speech Communication Association, INTERSPEECH 2011
    CountryItaly
    CityFlorence
    Period11/8/2711/8/31

    Fingerprint

    Speaker Verification
    Conditional Entropy
    Entropy
    kernel
    Robust Performance
    Margin
    Likelihood
    Learning
    Style
    Kernel
    Experimental Results
    Class

    Keywords

    • Intra-speaker variation
    • MCEM
    • Multiple kernel learning
    • Speaker verification

    ASJC Scopus subject areas

    • Language and Linguistics
    • Human-Computer Interaction
    • Signal Processing
    • Software
    • Modelling and Simulation

    Cite this

    Ogawa, T., Hino, H., Murata, N., & Kobayashi, T. (2011). Speaker verification robust to talking style variation using multiple kernel learning based on conditional entropy minimization. In Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH (pp. 2741-2744)

    Speaker verification robust to talking style variation using multiple kernel learning based on conditional entropy minimization. / Ogawa, Tetsuji; Hino, Hideitsu; Murata, Noboru; Kobayashi, Tetsunori.

    Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH. 2011. p. 2741-2744.

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

    Ogawa, T, Hino, H, Murata, N & Kobayashi, T 2011, Speaker verification robust to talking style variation using multiple kernel learning based on conditional entropy minimization. in Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH. pp. 2741-2744, 12th Annual Conference of the International Speech Communication Association, INTERSPEECH 2011, Florence, Italy, 11/8/27.
    Ogawa T, Hino H, Murata N, Kobayashi T. Speaker verification robust to talking style variation using multiple kernel learning based on conditional entropy minimization. In Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH. 2011. p. 2741-2744
    Ogawa, Tetsuji ; Hino, Hideitsu ; Murata, Noboru ; Kobayashi, Tetsunori. / Speaker verification robust to talking style variation using multiple kernel learning based on conditional entropy minimization. Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH. 2011. pp. 2741-2744
    @inproceedings{6ca9d4d5d63b4860a1941a5184a6a067,
    title = "Speaker verification robust to talking style variation using multiple kernel learning based on conditional entropy minimization",
    abstract = "We developed a new speaker verification system that is robust to intra-speaker variation. There is a strong likelihood that intra-speaker variations will occur due to changes in talking styles, the periods when an individual speaks, and so on. It is well known that such variation generally degrades the performance of speaker verification systems. To solve this problem, we applied multiple kernel learning (MKL) based on conditional entropy minimization, which impose the data to be compactly aggregated for each speaker class and ensure that the different speaker classes were far apart from each other. Experimental results showed that the proposed speaker verification system achieved a robust performance to intra-speaker variation derived from changes in the talking styles compared to the conventional maximum margin-based system.",
    keywords = "Intra-speaker variation, MCEM, Multiple kernel learning, Speaker verification",
    author = "Tetsuji Ogawa and Hideitsu Hino and Noboru Murata and Tetsunori Kobayashi",
    year = "2011",
    language = "English",
    pages = "2741--2744",
    booktitle = "Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH",

    }

    TY - GEN

    T1 - Speaker verification robust to talking style variation using multiple kernel learning based on conditional entropy minimization

    AU - Ogawa, Tetsuji

    AU - Hino, Hideitsu

    AU - Murata, Noboru

    AU - Kobayashi, Tetsunori

    PY - 2011

    Y1 - 2011

    N2 - We developed a new speaker verification system that is robust to intra-speaker variation. There is a strong likelihood that intra-speaker variations will occur due to changes in talking styles, the periods when an individual speaks, and so on. It is well known that such variation generally degrades the performance of speaker verification systems. To solve this problem, we applied multiple kernel learning (MKL) based on conditional entropy minimization, which impose the data to be compactly aggregated for each speaker class and ensure that the different speaker classes were far apart from each other. Experimental results showed that the proposed speaker verification system achieved a robust performance to intra-speaker variation derived from changes in the talking styles compared to the conventional maximum margin-based system.

    AB - We developed a new speaker verification system that is robust to intra-speaker variation. There is a strong likelihood that intra-speaker variations will occur due to changes in talking styles, the periods when an individual speaks, and so on. It is well known that such variation generally degrades the performance of speaker verification systems. To solve this problem, we applied multiple kernel learning (MKL) based on conditional entropy minimization, which impose the data to be compactly aggregated for each speaker class and ensure that the different speaker classes were far apart from each other. Experimental results showed that the proposed speaker verification system achieved a robust performance to intra-speaker variation derived from changes in the talking styles compared to the conventional maximum margin-based system.

    KW - Intra-speaker variation

    KW - MCEM

    KW - Multiple kernel learning

    KW - Speaker verification

    UR - http://www.scopus.com/inward/record.url?scp=84865770271&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=84865770271&partnerID=8YFLogxK

    M3 - Conference contribution

    AN - SCOPUS:84865770271

    SP - 2741

    EP - 2744

    BT - Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH

    ER -