| 摘要 |
<p><P>PROBLEM TO BE SOLVED: To solve the problem that inter-line watermark information may be erroneously extracted because post-script or noise may be frequently added to a document image generally, the document image may include a part with reduced characters within one line and if a line interval is acquired by obtaining a character string rectangle for such a document image, a line interval suitable for extracting inter-line watermark information may not be always obtained. <P>SOLUTION: In a document processing apparatus, a height and a line interval value of a character string are acquired as character string information in a document image (S504), and variances are calculated, respectively, as variations in the height and line interval value of the character string (S505). If a value of the calculated variance is lower than or equal to a threshold (S506), it is regarded suitable to extract an inter-line watermark for such character string information, and inter-line watermark information is extracted from the character string information (S507). <P>COPYRIGHT: (C)2010,JPO&INPIT</p> |
