The use of well-suited algorithms to a particular language can lead to efficient systems. In this paper we show how static modelling of phonemes, combined with phonetic heuristic knowledge, is a good approach for isolated word speech recognition in Spanish. The system developed is specially indicated in applications where low computational load and easy speaker adaptation is important, while maintaining an acceptable performance.
Bibliographic reference. Leandro, Manuel A. / Villegas, Alvaro / Pardo, Josť M. (1995): "Efficient isolated word recognition in Spanish based on static modeling", In EUROSPEECH-1995, 71-74.