| 摘要 |
<p><P>PROBLEM TO BE SOLVED: To enable a highly accurate similar document retrieval suppressing deterioration of the retrieval accuracy due to a misrecognized character, even if the misrecognized characters exist in either one of a seed document and a retrieval object document or both of the documents in a document managing system using image data. <P>SOLUTION: In this method, a processing for correcting the misrecognized character existing in a featured character string of the seed document or a registration object document and a processing for allowing the misrecognized character existing in a retrieval object document are individually provided. In the processing for correcting the misrecognized character existing in the featured character string, featured character strings existing in a read document are extracted, the character string including the misrecognized character of the extracted featured character strings is corrected to a proper character string for executing a retrieval, and the featured character string to be used in an actual retrieval is selected. <P>COPYRIGHT: (C)2003,JPO</p> |
