| 摘要 |
<p><P>PROBLEM TO BE SOLVED: To provide a new mechanism for further accurately recognizing a character written in a document. <P>SOLUTION: A plurality of field-by-field term dictionary data bases 11a, 11b and 11c storing terms or characters classified by field are prepared, and a field to which the content described in the document belongs is determined. A field-by-field term dictionary data base related to the determined field is selected from the plurality of field-by-field term dictionary data bases 11a, 11b and 11c, and the character recognition is performed using, as a candidate, a term or character stored in the field-by-field term dictionary data base. In this method, since the field to which the content of the document belongs is determined, an appropriate field-by-field term dictionary data base according to the field is selected, and the character recognition is performed using it, the recognition accuracy can be expected to be improved. <P>COPYRIGHT: (C)2006,JPO&NCIPI</p> |
