• DocumentCode
    1998393
  • Title

    Noise classification using Gaussian Mixture Models

  • Author

    Gupta, Hitesh Anand ; Varma, Vinay M.

  • Author_Institution
    Birla Inst. of Technol., Electron. & Commun. Eng., Ranchi, India
  • fYear
    2012
  • fDate
    15-17 March 2012
  • Firstpage
    821
  • Lastpage
    825
  • Abstract
    Gaussian Mixture Models (GMMs) have been proven effective in modeling speech and other acoustic signals. In this study, we have used GMMs to model different noise sources, viz. subway, babble, car and exhibition. Expectation maximization algorithm has been implemented to fit the model. Further, we present the `threshold´ method which uses the energy coefficient of the Mel - Frequency Cepstral Coefficients (MFCC) vector to determine the frames with noise (no speech) data.
  • Keywords
    Gaussian noise; cepstral analysis; expectation-maximisation algorithm; speech processing; GMM; Gaussian mixture models; MFCC vector; acoustic signals; babble; car; energy coefficient; expectation maximization algorithm; mel-frequency cepstral coefficients; noise classification; noise data; noise sources; speech modelling; subway; threshold method; Accuracy; Feature extraction; Hidden Markov models; Signal to noise ratio; Speech; Training; Expectation Maximization; GMM; MFCC; Noise Classification; threshold method;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Recent Advances in Information Technology (RAIT), 2012 1st International Conference on
  • Conference_Location
    Dhanbad
  • Print_ISBN
    978-1-4577-0694-3
  • Type

    conf

  • DOI
    10.1109/RAIT.2012.6194530
  • Filename
    6194530