摘要 |
PURPOSE: To easily and speedily correct errorneous recognition and errorneous conversion at the time of automatic translation. CONSTITUTION: The document of a first language (Japanese, for example) is read as picture data (S1). Then, the first language is character-recognized and it is automatically translated into a second language (English, for example) (S2 and S3). When an error exists in a translated result, the second language after translation is corrected either in errorneous recognition or errorneous translation. Thus, the correct document is generated. |