Third European Conference on Speech Communication and Technology

Berlin, Germany
September 22-25, 1993


Low Cost Speaker Dependent Isolated Word Speech Preselection System Using Static Phoneme Pattern Recognition

Manuel A. Leandro, Jose M. Pardo

Grupo de Tecnologia del Habla, Departamento de Ingenieria Electronica, Universidad Politecnica de Madrid, Spain E.T.S.I. Telecomunicaciones, Ciudad Universitaria, Madrid, Spain

We present here a system that builds up a phonetic string from speech at its input. The front-end system has been designed in terms of low cost computation for easy real time implementation, easy and automatic speaker enrollment and low use of linguistic knowledge. It is compared with Olivetti's phonetic string build up system adapted for Spanish in ESPRIT Polyglot project, showing that it is competitive in terms of computation and recognition performance.

Full Paper

Bibliographic reference.  Leandro, Manuel A. / Pardo, Jose M. (1993): "Low cost speaker dependent isolated word speech preselection system using static phoneme pattern recognition", In EUROSPEECH'93, 117-120.