Loquendo-Politecnico di Torino system for the 2009 NIST Language Recognition Evaluation

This paper describes the system submitted by Loquendo and Politecnico di Torino (LPT) for the 2009 NIST Language Recognition Evaluation. The system is a combination of classifiers based on two core acoustic models and on two core phone tokenizers. It exploits several state-of-the-art techniques that have been successfully applied in recent years both in speaker and in language recognition.We illustrate the incremental training procedure that has been devised to deal with broadcast data, we also describe the models, the classification techniques that have been used for this evaluation, and we comment on the performance of the system components alone and in combination.The system obtained using these techniques was among the best participants in this evaluation, obtaining on the 23 languages recognition task an actual DCFx100 of 1.6, 2.8, and 9.2 in the 30, 10 and 3 sec conditions respectively.