A TRAINING METHOD TOLERANT OF EMOTIONAL VARIATION FOR SPEECH RECOGNITION SYSTEM

摘要

PURPOSE: a kind of learning method is improved degree of being quick of perception and is provided to the recognition performance for improving speech recognition system based on learning data in speech recognition system, works as voice input. ;CONSTITUTION: one voice parameter extraction module (202) receives learning data emotion and applies frequency conversion parameter. Speech recognition learning module extracts voice parameter module and passes through one voice parameter of input. Learning data is converted to voice parameter by extracting each conversion parameter. ;The 2011 of copyright KIPO submissions