Non-Stationary Pattern Recognition
So far, we have discussed pattern recognition for stationary signals. In this chapter, we will discuss pattern recognition for both stationary and nonstationary signals. In speaker authentication, some tasks, such as speaker identification, are treated as stationary pattern recognition while others, such as speaker verification, are treated as non-stationary pattern recognition. We will introduce the stochastic modeling approach for both stationary and nonstationary pattern recognition. We will also introduce the Gaussian mixture model (GMM) and the hidden Markov model (HMM), two popular models that will be used throughout the book.
KeywordsHide Markov Model Speech Signal Gaussian Mixture Model Equal Error Rate Speaker Recognition
Unable to display preview. Download preview PDF.
- 1.Bahl L. R., Brown P. F., de Souza P. V., Mercer R.L.: “Maximum mutual information estimation of hidden Markov model parameters for speech recognition.” in Proc. IEEE Int. Conf. Acoust., Speech, Signal Processing (Tokyo), pp. 49–52, 1986Google Scholar
- 10.Korkmazskiy F., Juang B.-H.: “Discriminative adaptation for speaker verification.” in Proceedings of Int. Conf. on Spoken Language Processing (Philadelphia), pp. 28–31, 1996Google Scholar
- 14.Neyman J., Pearson E. S.: “On the use and interpretation of certain test criteria for purpose of statistical inference.” Biometrika,20A, pp. Pt I, 175–240; Pt II, (1928)Google Scholar
- 16.Rabiner, L. R., Wilpon, J. G., Juang, B.-H.: “A segmental k-means training procedure for connected word recognition”. AT&T Technical Journal 65, 21–31 (1986)Google Scholar
- 17.Rosenberg A. E., Siohan O., Parthasarathy S.: “Speaker verification using minimum verification error training.” in Proc. IEEE Int. Conf. Acoust., Speech, Signal Processing (Seattle), pp. 105–108, May 1998Google Scholar
- 18.Siohan O., Rosenberg A. E., Parthasarathy S.: “Speaker identification using minimum verification error training.” in Proc. IEEE Int. Conf. Acoust., Speech, Signal Processing (Seattle), pp. 109–112, May 1998Google Scholar