Back to Main Conference 2004
LREC 2004main
Transcrigal: A Bilingual System for Automatic Indexing of Broadcast News
Proceedings of the Fourth International Conference on Language Resources and Evaluation (LREC 2004)
Abstract
This paper describes a Broadcast News (BN) database called Transcrigal-DB. The news shows are mainly in Galician language, although around 11% of data is in Spanish. This database has been constructed for automatic speech recognition (ASR) purposes. A BN-ASR reference system is also described and evaluated on the test partition of Transcrigal-DB. The reference system has been designed having in mind that both languages, Spanish and Galician, may be used. Performance of the reference is improved when language adaptation techniques are taken into consideration