VERFAHREN ZUM ERMITTELN VON DATENBANKEINTRÄGEN DURCH SPRACHEINGABE

摘要

The method involves buffering a speech input and processing it using an automatic speech recognizer (4). A speech recognition result is generated using a first lexicon (6) comprising a first set of entries. A lexicon entry contains at least one acoustic reference. The method then involves searching for data corresponding to the speech recognition result, and if none are found, the first lexicon is adapted so that it contains a second set of different entries and the steps of the method are repeated.