Speech production models for ASR in Spanish language

Javier Mikel Olaso, Maria Ines Torres

Abstract: In this paper we undertake the extraction of phonological features applied to Spanish language. Also propose a method to integrate these features into an HMM based speech recognition system using an architecture that uses independent feature streams. In the experimental results we ?nd that higher recognition accuracies and less computational cost can be obtained.

Index Terms: speech recognition, acoustic modeling, phonological features.

Full Paper