A Bayesian Network Approach for Combining Pitch and Reliable Spectral Envelope Features for Robust Speaker Verification

  • Mijail Arcienega
  • Andrzej Drygajlo
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 2688)


In this paper, we provide a new approach in the design of robust speaker verification in noisy environments using some principles based on the missing data theory and Bayesian networks. This approach integrates high-level information concerning the reliability of pitch and spectral envelope features in missing feature compensation process in order to increase the performance of Gaussian mixture models (GMM) of speakers. In this paper, a Bayesian network approach for modeling statistical dependencies between reliable prosodic and spectral envelope features is presented. Within this approach, conditional statistical distributions (represented by GMMs) of the features are simultaneously exploited for increasing the recognition score, particularly in very noisy conditions. Masked by noise data can be discarded and the Bayesian network can be used to infer the likelihood values and compute the recognition scores. The system is tested on a challenging text-independent telephone-quality speaker verification task.


Bayesian Network Gaussian Mixture Model Equal Error Rate Speaker Verification Noisy Speech 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. [1]
    Reynolds, D. A., Rose, R.: Robust text-independent speaker identification using Gaussian mixture models. 3 (1995) 72–83Google Scholar
  2. [2]
    Pearl, J.: Probabilistic reasoning in intelligent systems: networks of plausible inference. Morgan Kaufmann Publishers (1988)Google Scholar
  3. [3]
    Arcienega, M., Drygajlo, A.: A Bayesian network approach for combining pitch and spectral envelope features for speaker verfication. In: COST 275 Workshop, “The Advents of Biometrics over the Internet”, Rome, Italy (2002) 99–102Google Scholar
  4. [4]
    Renevey, P., Drygajlo, A.: Missing feature theory and probabilistic estimation of clean speech componenets for robust speech recognition. Volume 6., Budapest, Hungary (1999) 2627–2630Google Scholar
  5. [5]
    Drygajlo, A., El-Maliki, M.: Integration and imputation methods for unreliable feature compensation in GMM based speaker verification, Crete, Greece (2001) 107–112Google Scholar
  6. [6]
    Arcienega, M., Drygajlo, A.: Robust voiced/unvoiced decision associated to continuous pitch tracking in noisy telephone speech. Volume 4., Denver, Colorado USA (2002) 2433–2436Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2003

Authors and Affiliations

  • Mijail Arcienega
    • 1
  • Andrzej Drygajlo
    • 1
  1. 1.Swiss Federal Institute of Technology Lausanne, Signal Processing InstituteSweden

Personalised recommendations