摘要 |
PURPOSE:To increase the processing speed of a KANA (Japanese syllabary)/ KANJI (Chinese characters) converter by using the voice information to divide a KANA sentence written solid and therefor decreasing dictionary consulting frequencies in case of KANA/KANJI conversion. CONSTITUTION:Voice inputted from an input part 1 is analyzed by an analyzing part 2 for extraction of a feature value. A KANA sentence written solid is stored in a storing part 3 for KANA sentence written solid and then given to a dividing part 4. Then the KANA sentence is divided based on the feature value obtained by the part 2. Then this KANA sentence divided by the part 4 is given to a KANA/KANJI conversion processing part 5 and KANA/KANJI conversion processing is successively applied to each divided part of said KANA sentence.
|