T. Kristensen and B. Treck (Norway)
Backpropagation, Sampa for Norwegian, Phoneme recognition.
This paper describes a Backpropagation program that learns to convert Norwegian text to phonemes. The original database consists of about 60000 Norwegian words and their transcriptions. The phoneme database is developed by the Norwegian Telecom and is further developed to be used in this program. The Sampa notation for Norwegian is used for the phoneme transcriptions. The experiments have been performed on a PC-platform and on only a sample of words from this database.
Important Links:
Go Back
