Speaker modeling with various speech representations

    Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review


    Although numerous speech representations have been reported to be useful in speaker recognition, there is much less agreement on which speech representation provides a perfect representation of speaker-specific information. In this paper, we charaterize a speaker's identity through the simultaneous use of various speech representations of his/her voice. We present a parametric statistical model, generalized Gaussian mixture model, and develop an EM algorithm for parameter estimation. Our approach has been applied to speaker recognition and comparative results on KING corpus demonstrate its effectiveness. © Springer-Verlag Berlin Heidelberg 2004.
    Original languageEnglish
    Title of host publicationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)|Lect. Notes Comput. Sci.
    PublisherSpringer Nature
    Number of pages7
    Publication statusPublished - 2004
    EventInternational Conference on Biometric Authentication - Hong Kong
    Duration: 1 Jan 1824 → …


    ConferenceInternational Conference on Biometric Authentication
    CityHong Kong
    Period1/01/24 → …


    Dive into the research topics of 'Speaker modeling with various speech representations'. Together they form a unique fingerprint.

    Cite this