Spoken Language Processing: A Guide to Theory, Algorithm, and System Development
Spoken Language Processing: A Guide to Theory, Algorithm, and System Development
Tree-based state tying for high accuracy acoustic modelling
HLT '94 Proceedings of the workshop on Human Language Technology
Speaker normalization using efficient frequency warping procedures
ICASSP '96 Proceedings of the Acoustics, Speech, and Signal Processing, 1996. on Conference Proceedings., 1996 IEEE International Conference - Volume 01
Hi-index | 0.00 |
The article presents a limited-vocabulary speaker independentcontinuous Estonian speech recognition system based on hiddenMarkov models. The system is trained using an annotated Estonianspeech database of 60 speakers, approximately 4 hours in duration.Words are modelled using clustered triphones with multiple Gaussianmixture components. The system is evaluated using a numberrecognition task and a simple medium-vocabulary recognition task.The system performance is explored by employing acoustic models ofincreasing complexity. The number recognizer achieves an accuracyof 97%. The medium-vocabulary system recognizes 82.9% wordscorrectly if operating in real time. The correctness increases to90.6% if real-time requirement is discarded.