A new duration intrinsic model for improved speech recognition by HMM techniques is presented. Assuming an exponentially decaying time dependency of the states loop probability, the duration density can be factorized and a path early pruning theorem demonstrated. As a consequence, computational complexity is greatly reduced with respect to explicit models, whereas recognition performances improve considerably.
Continuously Variable Transition Probability HMM for Speech Recognition / Falaschi, Alessandro. - STAMPA. - (1992), pp. 125-130. (Intervento presentato al convegno Proceedings of the NATO Advanced Study Institute tenutosi a Cetraro, Italy nel July 1-13, 1990).
Continuously Variable Transition Probability HMM for Speech Recognition
FALASCHI, Alessandro
1992
Abstract
A new duration intrinsic model for improved speech recognition by HMM techniques is presented. Assuming an exponentially decaying time dependency of the states loop probability, the duration density can be factorized and a path early pruning theorem demonstrated. As a consequence, computational complexity is greatly reduced with respect to explicit models, whereas recognition performances improve considerably.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
