| L.F. Lamel and J.-L. Gauvain (1992), "Multi-lingual Speech Recognition at LIMSI," Presented at the 1st International Workshop of Speech Translation, Warden, Germany, Oct. 18-20. |
....that stem from the acoustic level. Phone recognition is also useful in determining pronunciation errors in the lexicon and alternate pronunciations that need to be included in the lexicon. Finally, phone recognition is shown to be effective for language identification and for speaker identification[25, 26, 27]. The baseline phone recognizer uses a set of 35 context independent (CI) phone models. Each model is a 3 state left to right HMM with Gaussian mixture observation densities. The covariance matrices of all the Gaussians components are diagonal. The 16 kHz speech was downsampled by 2 and a ....
L.F. Lamel and J.-L. Gauvain (1992), "Multi-lingual Speech Recognition at LIMSI," Presented at the 1st International Workshop of Speech Translation, Warden, Germany, Oct. 18-20.
....parallel the unknown incoming speechby different sets of phone models for each of the languages under consideration, and to choose the language associated with the model set providing the highest likelihood. Experiments have been performed using sets of SI CI phone models for French and for English[13]. For French the set of 35 SI CI models were used. For English a set of 52 SI CI phone models were trained on the training speakers in the TIMIT Corpus[3] Using this approach and processing the entire utterance always gave 100 correct language identification. The identification accuracy as a ....
L.F. Lamel, J.L. Gauvain, "Multi-lingual Speech Recognition at LIMSI," Presented at the 1st Intl. Workshop on Speech Translation, Warden, Germany, Oct. 18-20, 1992.
Online articles have much greater impact More about CiteSeer.IST Add search form to your site Submit documents Feedback
CiteSeer.IST - Copyright Penn State and NEC