摘要 |
An apparatus and method for extending a pronunciation dictionary for speech recognition are provided. The apparatus and the method may segment speech information of an input utterance into at least one phoneme, collect segmentation information of the at least one segmented phoneme, analyze a pronunciation variation of the at least one segmented phoneme based on the collected segmentation information, and select a substitutable phoneme group for the at least one phoneme where the pronunciation variation occurs, and extend the pronunciation dictionary. |