| 摘要 |
PURPOSE:To easily specify a cursor, which is required at the time of editing sentences inputted by voice, by automatically retrieving a position where sentences are corrected, inserted or deleted and moving the cursor to the target position. CONSTITUTION:Based on position information on the cursor, which is obtained by collation processing, and control information, a cursor control part 110 moves the cursor to the top character in the part most similar to an input voice. Moreover the part 110 can also move the cursor to the top of a paragraph including the part retrieved by specification through a keyboard. Since the data structure of an input text memory part 104 and that of a voice recognition result memory part 102 are the same, collation processing never develops trouble. However, if a collation target is a Japanese text, one vs one collation cannot be made because KANJI (Chinese character) code is stored. Accordingly the number of syllables of KANJI is added as additional information, and stored in a Japanese text memory part 107, whereby correspondence to a syllable series can be taken.
|
