Experiments towards a multi-language LVCSR interface

This paper describes experiments towards a multilanguage human-computer speech interface. Our interface is designed for large vocabulary continuous speech input. For this purpose a multilingual dictation database has been collected under GlobalPhone, which is a project at the Interactive Systems Labs. This project investigates LVCSR systems in 15 languages of the world, namely Arabic, Chinese, Croatian, English, French, German, Italian, Japanese, Korean, Portuguese, Russian, Spanish, Swedish, Tamil, and Turkish. Based on a global phoneme set we build di erent multilingual speech recognizer and present several performance results in language independent and language adaptive setups.

[1]  A. Constantinescu,et al.  On cross-language experiments and data-driven units for ALISP (Automatic Language Independent Speech Processing) , 1997, 1997 IEEE Workshop on Automatic Speech Recognition and Understanding Proceedings.

[2]  Chalapathy Neti,et al.  Towards a universal speech recognizer for multiple languages , 1997, 1997 IEEE Workshop on Automatic Speech Recognition and Understanding Proceedings.

[3]  Tanja Schultz,et al.  Fast bootstrapping of LVCSR systems with multilingual phoneme sets , 1997, EUROSPEECH.

[4]  Paul Dalsgaard,et al.  Data-driven identification of poly- and mono-phonemes for four european languages , 1993, EUROSPEECH.

[5]  Joachim Köhler.  Language adaptation of multilingual phone models for vocabulary independent speech recognition tasks , 1998, ICASSP.

[6]  Patrizia Bonaventura,et al.  Multilingual speech recognition for flexible vocabularies , 1997, EUROSPEECH.

[7]  Kazuhiro Kondo,et al.  An evaluation of cross-language adaptation for rapid HMM development in a new language , 1994, Proceedings of ICASSP '94. IEEE International Conference on Acoustics, Speech and Signal Processing.