Speaker identification method for use during speech recognition in infotainment system in car, involves assigning user model to associated entry, extracting characteristics from linguistic expression of user and selecting one entry

摘要

The method involves extracting user specific characteristics required for later speaker identification during training of speaker-dependent entries and storing characteristics as an user model. The model is assigned to the associated entry. The characteristics are extracted from a linguistic expression of an user during later speaker identification. An entry is selected from multiple entries trained by different users, so that only speaker-dependent entries suitable with predetermined confidence are supplied to a vocabulary for speech recognition, isolated or selected in past-processing.