| 摘要 |
PROBLEM TO BE SOLVED: To carry out training of new commands in a time-saving manner, in a speaker-dependent voice recognition method and a voice recognition system for the method. SOLUTION: The voice recognition system which is in a voice recognition mode records user's utterance, and a first voice pattern is created from this utterance. When the command corresponding to the first voice pattern is found in the voice recognition system, the command is executed. When a similar voice pattern is not found, or the voice pattern with a significant similarity degree is not found, in the voice recognition system, and when the utterance is not recognized by the voice recognition system, the voice recognition system provides the user with change to a training mode to immediately assign the voice utterance to a new command. COPYRIGHT: (C)2009,JPO&INPIT
|
