SYSTEM AND METHOD OF MULTI MODEL ADAPTIVE AND VOICE RECOGNITION CAPABLE OF MAXIMIZING THE VOICE ADAPTIVE EFFECT

摘要

PURPOSE: A system and method of multi model adaptive and voice recognition is provided to use different independent model by each user or group.;CONSTITUTION: A model number selection unit(110) selects a model for designating a voice adaptation. A feature extraction unit(120) extracts the feature vector from the voice of a speaker. An adaptation processing unit(130) stores the voice of the speaker into a designated model after the voice adaptation process. The voice is adapted according to the designation of the speaker by character in a multi adaptation model(160).;COPYRIGHT KIPO 2012