Language Model Support for Continuous Speech Recognition in Czech Language

D. Nejedlova and J. Nouza (Czech Republic)


Speech Processing, Speech Recognition, Language Models, Czech Language


The article describes the system for continuous speech recognition of the Czech language. It gives the basic information about the creation of statistical n-gram language models and their importance for the improvement of speech recognition. The pivot information provided here is the role, behavior, and optimal sets of parameters that influence the speech recognition system as well as its best accuracy attained and the time requirements on its operation.

Important Links:

Go Back