| 摘要 |
PURPOSE:To facilitate the correction of a KANJI (Chinese character) recognizing result and to improve a recognition processing efficiency by directly displaying recognition result in KANJI and then performing automatically the KANJI/ KANA (Japanese syllabary) conversion through a system when the normal recognition is visually confirmed. CONSTITUTION:An OCR recognizer 2 acquires three bank name candidates and two branch name candidates, for example, as the KANJI recognizing results. A controller 3 edits the image data and the recognizing result received from the recognizer 2 and sends them to a terminal 4 for display on a drawing. Thus the terminal 4 shows the received image data and OCR recognizing results as well as the candidate character strings. When a first candidate of a bank name is normally recognized, an operator visually confirms the candidate and then moves a cursor to the frame of a branch name. If 1st and 2nd candidates of the branch name are incorrect and correct respectively, the operator depresses the ten keys 2 to correct the recognizing results. At the same time, the selected character string is automatically replaced and shown in a display frame. Then the character string undergoes the KANJI/KANA conversion through the controller 3 and then transferred to the next processing. |
