Manifold HLDA and its application to robust speech recognition

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

    Abstract

    A manifold heteroscedastic linear discriminant analysis (MHLDA) which removes environmental information explicitly from the useful information for discrimination is proposed. Usually, a feature parameter used in pattern recognition involves categorical information and also environmental information. A well-known HLDA tries to extract useful information (UI) to represent categorical information from the feature parameter. However, environmental information is still remained in the UI parameters extracted by HLDA, and it causes slight degradation in performance. This is because HLDA does not handle the environmental information explicitly. The proposed MHLDA also tries to extract UI like HLDA, but it handles environmental information explicitly. This handling makes MHLDA-based UI parameter less influenced of environment. However, as compensation, in MHLDA, the categorical information is little bit destroyed. In this paper, we try to combine HLDA-based UI and MHLDA-based UI for pattern recognition, and draw benefit of both parameters. Experimental results show the effectiveness of this combining method.

    Original languageEnglish
    Title of host publicationINTERSPEECH 2006 and 9th International Conference on Spoken Language Processing, INTERSPEECH 2006 - ICSLP
    Pages1551-1554
    Number of pages4
    Volume3
    Publication statusPublished - 2006
    EventINTERSPEECH 2006 and 9th International Conference on Spoken Language Processing, INTERSPEECH 2006 - ICSLP - Pittsburgh, PA
    Duration: 2006 Sep 172006 Sep 21

    Other

    OtherINTERSPEECH 2006 and 9th International Conference on Spoken Language Processing, INTERSPEECH 2006 - ICSLP
    CityPittsburgh, PA
    Period06/9/1706/9/21

    Fingerprint

    Discriminant analysis
    Speech recognition
    Pattern recognition
    Degradation

    Keywords

    • HLDA
    • MHLDA
    • Robust speech recognition

    ASJC Scopus subject areas

    • Computer Science(all)

    Cite this

    Kubo, T., Ogawa, T., & Kobayashi, T. (2006). Manifold HLDA and its application to robust speech recognition. In INTERSPEECH 2006 and 9th International Conference on Spoken Language Processing, INTERSPEECH 2006 - ICSLP (Vol. 3, pp. 1551-1554)

    Manifold HLDA and its application to robust speech recognition. / Kubo, Toshiaki; Ogawa, Tetsuji; Kobayashi, Tetsunori.

    INTERSPEECH 2006 and 9th International Conference on Spoken Language Processing, INTERSPEECH 2006 - ICSLP. Vol. 3 2006. p. 1551-1554.

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

    Kubo, T, Ogawa, T & Kobayashi, T 2006, Manifold HLDA and its application to robust speech recognition. in INTERSPEECH 2006 and 9th International Conference on Spoken Language Processing, INTERSPEECH 2006 - ICSLP. vol. 3, pp. 1551-1554, INTERSPEECH 2006 and 9th International Conference on Spoken Language Processing, INTERSPEECH 2006 - ICSLP, Pittsburgh, PA, 06/9/17.
    Kubo T, Ogawa T, Kobayashi T. Manifold HLDA and its application to robust speech recognition. In INTERSPEECH 2006 and 9th International Conference on Spoken Language Processing, INTERSPEECH 2006 - ICSLP. Vol. 3. 2006. p. 1551-1554
    Kubo, Toshiaki ; Ogawa, Tetsuji ; Kobayashi, Tetsunori. / Manifold HLDA and its application to robust speech recognition. INTERSPEECH 2006 and 9th International Conference on Spoken Language Processing, INTERSPEECH 2006 - ICSLP. Vol. 3 2006. pp. 1551-1554
    @inproceedings{8e4917c26fa349e68fb4f3b49eca2d83,
    title = "Manifold HLDA and its application to robust speech recognition",
    abstract = "A manifold heteroscedastic linear discriminant analysis (MHLDA) which removes environmental information explicitly from the useful information for discrimination is proposed. Usually, a feature parameter used in pattern recognition involves categorical information and also environmental information. A well-known HLDA tries to extract useful information (UI) to represent categorical information from the feature parameter. However, environmental information is still remained in the UI parameters extracted by HLDA, and it causes slight degradation in performance. This is because HLDA does not handle the environmental information explicitly. The proposed MHLDA also tries to extract UI like HLDA, but it handles environmental information explicitly. This handling makes MHLDA-based UI parameter less influenced of environment. However, as compensation, in MHLDA, the categorical information is little bit destroyed. In this paper, we try to combine HLDA-based UI and MHLDA-based UI for pattern recognition, and draw benefit of both parameters. Experimental results show the effectiveness of this combining method.",
    keywords = "HLDA, MHLDA, Robust speech recognition",
    author = "Toshiaki Kubo and Tetsuji Ogawa and Tetsunori Kobayashi",
    year = "2006",
    language = "English",
    isbn = "9781604234497",
    volume = "3",
    pages = "1551--1554",
    booktitle = "INTERSPEECH 2006 and 9th International Conference on Spoken Language Processing, INTERSPEECH 2006 - ICSLP",

    }

    TY - GEN

    T1 - Manifold HLDA and its application to robust speech recognition

    AU - Kubo, Toshiaki

    AU - Ogawa, Tetsuji

    AU - Kobayashi, Tetsunori

    PY - 2006

    Y1 - 2006

    N2 - A manifold heteroscedastic linear discriminant analysis (MHLDA) which removes environmental information explicitly from the useful information for discrimination is proposed. Usually, a feature parameter used in pattern recognition involves categorical information and also environmental information. A well-known HLDA tries to extract useful information (UI) to represent categorical information from the feature parameter. However, environmental information is still remained in the UI parameters extracted by HLDA, and it causes slight degradation in performance. This is because HLDA does not handle the environmental information explicitly. The proposed MHLDA also tries to extract UI like HLDA, but it handles environmental information explicitly. This handling makes MHLDA-based UI parameter less influenced of environment. However, as compensation, in MHLDA, the categorical information is little bit destroyed. In this paper, we try to combine HLDA-based UI and MHLDA-based UI for pattern recognition, and draw benefit of both parameters. Experimental results show the effectiveness of this combining method.

    AB - A manifold heteroscedastic linear discriminant analysis (MHLDA) which removes environmental information explicitly from the useful information for discrimination is proposed. Usually, a feature parameter used in pattern recognition involves categorical information and also environmental information. A well-known HLDA tries to extract useful information (UI) to represent categorical information from the feature parameter. However, environmental information is still remained in the UI parameters extracted by HLDA, and it causes slight degradation in performance. This is because HLDA does not handle the environmental information explicitly. The proposed MHLDA also tries to extract UI like HLDA, but it handles environmental information explicitly. This handling makes MHLDA-based UI parameter less influenced of environment. However, as compensation, in MHLDA, the categorical information is little bit destroyed. In this paper, we try to combine HLDA-based UI and MHLDA-based UI for pattern recognition, and draw benefit of both parameters. Experimental results show the effectiveness of this combining method.

    KW - HLDA

    KW - MHLDA

    KW - Robust speech recognition

    UR - http://www.scopus.com/inward/record.url?scp=44949192117&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=44949192117&partnerID=8YFLogxK

    M3 - Conference contribution

    AN - SCOPUS:44949192117

    SN - 9781604234497

    VL - 3

    SP - 1551

    EP - 1554

    BT - INTERSPEECH 2006 and 9th International Conference on Spoken Language Processing, INTERSPEECH 2006 - ICSLP

    ER -