KERNEL DISCRIMINANT ANALYSIS FOR SPEECH RECOGNITION

Kernel Discriminant Analysis (KDA) has been successfully applied to many pattern recognition problems. KDA transforms the original problem into a space of dimension N whereN is the number of training vectors. For speech recognition,N is usually prohibitively high increasing computational requirements beyond current computational capabilities. In this paper, we provide a formulation of a subspace version of KDA that enables its application to speech recognition, thus conveniently enabling nonlinear feature space transformations that result in discriminatory lower dimensional features.

[1]  Gunnar Rätsch,et al.  An introduction to kernel-based learning algorithms , 2001, IEEE Trans. Neural Networks.

[2]  Hsiao-Wuen Hon,et al.  Speaker-independent phone recognition using hidden Markov models , 1989, IEEE Trans. Acoust. Speech Signal Process..

[3]  G. Baudat,et al.  Generalized Discriminant Analysis Using a Kernel Approach , 2000, Neural Computation.

[4]  Heiga Zen,et al.  On the Use of Kernel PCA for Feature Extraction in Speech Recognition , 2003, IEICE Trans. Inf. Syst..

[5]  Ramesh A. Gopinath,et al.  Maximum likelihood modeling with Gaussian distributions for classification , 1998, Proceedings of the 1998 IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP '98 (Cat. No.98CH36181).

[6]  Sebastian Mika,et al.  Kernel Fisher Discriminants , 2003 .

[7]  B. Scholkopf,et al.  Fisher discriminant analysis with kernels , 1999, Neural Networks for Signal Processing IX: Proceedings of the 1999 IEEE Signal Processing Society Workshop (Cat. No.98TH8468).

[8]  Richard O. Duda,et al.  Pattern classification and scene analysis , 1974, A Wiley-Interscience publication.

[9]  Andreas G. Andreou,et al.  Heteroscedastic discriminant analysis and reduced rank HMMs for improved speech recognition , 1998, Speech Commun..