Paramétrisation de la Parole basée sur une Modélisation des Filtres Cochléaires: Application au RAP
Zied Hajaiej, Kaïs Ouni, Noureddine Ellouze.

Signal processing front end for extracting the feature set is an important stage in any speech recognition system. The optimum feature set is still not yet decided. There are many types of features, which are derived differently and have good impact on the recognition rate. This paper presents one more successful technique to extract the feature set from a speech signal, which can be used in speech recognition systems. Our technique based on the human auditory system characteristics and relies on the gammachirp filterbank to emulate asymmetric frequency response and level dependent frequency response. For evaluation a comparative study was operated with standard MFCC and PLP.