摘要 |
PROBLEM TO BE SOLVED: To improve recognition rate of input words. SOLUTION: The edited word are converted to speech signals and the agreement of the converted speech signals and the words in a dictionary is computed. If the words of the highest agreement are different from the edited words, decision is made that the edited words are inadequate, and the result of decision is outputted As a result, a user is eventually urged of the reediting of the words which are decided to be inadequate, and the recognition rate is improved. |