Reconnaissance de parole non native fondée sur l'utilisation de confusion phonétique et de contraintes graphèmiques

This paper presents a fully automated approach for the recognition of non native speech based on acoustic model modication. For a native language (LM) and a spoken language (LP), pronunciation variants of the phones of LP are automatically extracted from an existing non native database. These variants are stored in a confusion matrix between phones of LP and sequences of phones of LM. This confusion concept deals with the problem of non existence of match between some LM and LP phones. The confusion matrix is then used to modify the acoustic models (HMMs) of LP phones by integrating corresponding LM phone models as alternative HMM paths. We introduce graphemic contraints in the confusion extraction process. We claim that prononciation errors may depend on the graphemes related to each phone. The modied ASR system achieved a signicant improvement varying between 20.3% and 43.2% (relative) in isentence error ratei and between 26.6% and 50.0% (relative) in iword error ratei. The introduction of graphemic contraints in the phonetic confusion allowed improvements while using the wordloop grammar.