Advertisement

Weighting Scores to Improve Speaker-Dependent Threshold Estimation in Text-Dependent Speaker Verification

Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 3817)

Abstract

The difficulty of obtaining data from impostors and the scarcity of data are two factors that have a large influence in the estimation of speakerdependent thresholds in text-dependent speaker verification. Furthermore, the inclusion of low quality utterances (background noises, distortion...) makes the process even harder. In such cases, the comparison of these utterances against the model can generate non-representative scores that deteriorate the correct estimations of statistical data from client scores. To mitigate the problem, some methods propose the suppresion of those scores which are far from the estimated scores mean. The tecnique results in a ‘hard decision’ that can produce errors especially when the number of scores is low. We propose here to take a ‘softer decision’ and weight scores according to their distance to the estimated scores mean. The Polycost and the BioTech databases have been used to show the effectiveness of the proposed method.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    Chen, K.: Towards Better Making a Decision in Speaker Verification. Pattern Recognition 36, 329–346 (2003)CrossRefGoogle Scholar
  2. 2.
    Saeta, J.R., Hernando, J.: Automatic Estimation of A Priori Speaker Dependent Thresholds in Speaker Verification. In: Kittler, J., Nixon, M.S. (eds.) AVBPA 2003. LNCS, vol. 2688, pp. 70–77. Springer, Heidelberg (2003)CrossRefGoogle Scholar
  3. 3.
    Saeta, J.R., Hernando, J.: On the Use of Score Pruning in Speaker Verification for Speaker Dependent Threshold Estimation. In: 2004: A Speaker Odyssey, The Speaker Recognition Workshop, pp. 215–218 (2004)Google Scholar
  4. 4.
    Furui, S.: Cepstral Analysis for Automatic Speaker Verification. IEEE Trans. Speech and Audio Proc. 29(2), 254–272 (1981)CrossRefGoogle Scholar
  5. 5.
    Lindberg, J., Koolwaaij, J., Hutter, H.P., Genoud, D., Pierrot, J.B., Blomberg, M., Bimbot, F.: Techniques for A Priori Decision Threshold Estimation in Speaker Verification. In: Proceedings RLA2C, pp. 89–92 (1998)Google Scholar
  6. 6.
    Pierrot, J.B., Lindberg, J., Koolwaaij, J., Hutter, H.P., Genoud, D., Blomberg, M., Bimbot, F.: A Comparison of A Priori Threshold Setting Procedures for Speaker Verification in the CAVE Project. In: Proceedings ICASSP, pp. 125–128 (1998)Google Scholar
  7. 7.
    Zhang, W.D., Yiu, K.K., Mak, M.W., Li, C.K., He, M.X.: A Priori Threshold Determination for Phrase-Prompted Speaker Verification. In: Proceedings Eurospeech, pp. 1203–1206 (1999)Google Scholar
  8. 8.
    Surendran, A.C., Lee, C.H.: A Priori Threshold Selection for Fixed Vocabulary Speaker Verification Systems. In: Proceedings ICSLP, vol. II, pp. 246–249 (2000)Google Scholar
  9. 9.
    Bimbot, F., Genoud, D.: Likelihood Ratio Adjustment for the Compensation of Model Mismatch in Speaker Verification. In: Proceedings 2001: A Speaker Odyssey, The Speaker Recognition Workshop, pp. 73-76 (2001)Google Scholar
  10. 10.
    Gravier, G., Chollet, G.: Comparison of Normalization Techniques for Speaker Verification. In: Proceedings RLA2C, pp. 97–100 (1998)Google Scholar
  11. 11.
    Auckentaler, R., Carey, M., Lloyd-Thomas, H.: Score Normalization for Text-Independent Speaker Verification Systems. Digital Signal Processing 10, 42–54 (2000)CrossRefGoogle Scholar
  12. 12.
    Bimbot, F., Bonastre, F.J., Fredouille, C., Gravier, G., Magrin, I., Meignier, S., Merlin, T., Ortega-García, J., Petrovska, D., Reynolds, D.: A Tutorial on Text-Independent Speaker Verification. In: Proceedings Eusipco, pp. 430–451 (2004)Google Scholar
  13. 13.
    Mirghafori, N., Heck, L.: An Adaptive Speaker Verification System with Speaker Dependent A Priori Decision Thresholds. In: Proceedings ICSLP, pp. 589–592 (2002)Google Scholar
  14. 14.
    Navratil, J., Ramaswamy, G.N.: The Awe and Mystery of T-norm. In: Proceedings Eurospeech, pp. 2009–2012 (2003)Google Scholar
  15. 15.
    Reynolds, D.: The Effect of Handset Variability on Speaker Recognition Performance: Experiments on the Switchboard Corpus. In: Proceedings ICASSP 1996, pp. 113–116 (1996)Google Scholar
  16. 16.
    Reynolds, D.A.: Comparison of Background Normalization Methods for Text-Independent Speaker Verification. In: Proceedings Eurospeech, pp. 963–966 (1997)Google Scholar
  17. 17.
    Heck, L.P., Weintraub, M.: Handset Dependent Background Models for Robust Text-Independent Speaker Recognition. In: Proceedings ICASSP, pp. 1071–1074 (1997)Google Scholar
  18. 18.
    Saeta, J.R., Hernando, J.: New Speaker-Dependent Threshold Estimation Method in Speaker Verification based on Weighting Scores. In: Proceedings of the 3th Internacional Conference on Non-Linear Speech Processing (NoLisp), pp. 34–41 (2005)Google Scholar
  19. 19.
    Li, Q., Juang, B.H., Zhou, Q., Lee, C.H.: Verbal Information Verification. In: Proceedings Eurospeech, pp. 839–842 (1997)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2006

Authors and Affiliations

  1. 1.Biometric Technologies, S.L.BarcelonaSpain
  2. 2.TALP Research CenterUniversitat Politècnica de Catalunya (UPC)BarcelonaSpain

Personalised recommendations