ITRW on Non-Linear Speech Processing (NOLISP 03) May 20-23, 2003 Le Croisic, France

Maximization of the Modelisation Error Ratio for Neural Predictive Coding

Mohamed Chetouani, B. Gas, J. L. Zarader

Laboratoire des Instruments et Systèmes d’Ile-de-France, Université Paris VI, France

In this paper, we introduce a model for Discrimant Feature Extraction (DFE): the Neural Predictive Coding (NPC). It is an extension of the Linear Predictive Coding (LPC). The Modelisation Error Ratio (MER), a discriminant criterion adapted for predictive models, is introduced. We propose a theoretical validation of the discriminant properties of the MER. The experimental validation consists on phoneme recognition task. The phonemes are extracted from the Darpa-Timit speech database. The performances are compared with traditional methods: LPC, MFCC, PLP.

Full Paper

Bibliographic reference.  Chetouani, Mohamed / Gas, B. / Zarader, J. L. (2003): "Maximization of the modelisation error ratio for neural predictive coding", In NOLISP-2003, paper 012.