Some solutions for coping with the problem of recognition of spontaneous speech dialogues are presented. Starting from a baseline system, developed for dictation tasks, some modifications are introduced at the acoustic level and in the language model. Acoustic model parameters are modified for addressing speaking rate variations, and specific models of extra-linguistic phenomena are defined and included into the language model. Different acoustic model sets are arranged in a single recognizer though their integration into a multi-model search space. Experiments and evaluations were carried out on a corpus of spontaneous dialogues collected in our laboratory
Automatic Speech Recognition of Spontaneous Dialogues
Cettolo, Mauro;Falavigna, Giuseppe Daniele
1997-01-01
Abstract
Some solutions for coping with the problem of recognition of spontaneous speech dialogues are presented. Starting from a baseline system, developed for dictation tasks, some modifications are introduced at the acoustic level and in the language model. Acoustic model parameters are modified for addressing speaking rate variations, and specific models of extra-linguistic phenomena are defined and included into the language model. Different acoustic model sets are arranged in a single recognizer though their integration into a multi-model search space. Experiments and evaluations were carried out on a corpus of spontaneous dialogues collected in our laboratoryI documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
