ISCA Archive Eurospeech 1997
ISCA Archive Eurospeech 1997

Nonlinear discriminant analysis for improved speech recognition

Vincent Fontaine, Christophe Ris, Jean-Marc Boite

Linear Discriminant Analysis (LDA) has been widely applied to speech recognition resulting in improved recognition performance and improved robustness. LDA designs a linear transformation that projects a m-dimensional space on a m-dimensional space (m < n) such that the class separability is maximum. This paper presents new results related to our previous work [6] on nonlinear discriminant analysis (NLDA) based on the discriminant properties of Artificial Neural Networks (ANN) and more particularly MLP. Experiments performed on the isolated word large vocabulary Phone- book database show that NLDA provides a method for designing discriminant features particularly efficient as well for continuous densities HMM as for hybrid HMM/ANN recognizers.


doi: 10.21437/Eurospeech.1997-548

Cite as: Fontaine, V., Ris, C., Boite, J.-M. (1997) Nonlinear discriminant analysis for improved speech recognition. Proc. 5th European Conference on Speech Communication and Technology (Eurospeech 1997), 2071-2074, doi: 10.21437/Eurospeech.1997-548

@inproceedings{fontaine97_eurospeech,
  author={Vincent Fontaine and Christophe Ris and Jean-Marc Boite},
  title={{Nonlinear discriminant analysis for improved speech recognition}},
  year=1997,
  booktitle={Proc. 5th European Conference on Speech Communication and Technology (Eurospeech 1997)},
  pages={2071--2074},
  doi={10.21437/Eurospeech.1997-548},
  issn={1018-4074}
}