METHOD FOR RECOGNIZING KOREAN NUMERIC VOICE USING SIMULTANEOUS ASSEMBLY SOUND MODEL

摘要

PURPOSE: A method for recognizing Korean numeric voice using a simultaneous assembly sound model is provided to recognize continuous numeric sound and to test detected continuous numeric voice with all voice information of Korean numeric voice. CONSTITUTION: A method for recognizing Korean numeric voice using a simultaneous assembly sound model includes a first and a second steps. At the first step(502), a simultaneous assembly sound for two continuous numeric voices is detected. At the second step, the two numeric candidates having corresponding simultaneous assembly sound are arranged(505), and the previous simultaneous assembly sound candidates are compared with next simultaneous assembly sound candidates in order to detect the input continuous numeric sounds(507).