| 摘要 |
<P>PROBLEM TO BE SOLVED: To accurately extract embedded information from a document image using an interlinear watermark. <P>SOLUTION: This image processor comprises: an input means for inputting the document image as image data; an image reduction means for, when the number of horizontal pixels in the inputted image data is W0 and that of vertical pixels is H0, for generating reduction image data in which at least the number of horizontal pixels is reduced to W1 (1<W1<W0); a detection means for scanning the reduction image data vertically to detect an average character height; a reduction image correction means in which, when the detected average character height is within a preset allowable range, the reduction image data is outputted as watermark extraction target image data, and when the detected average character height is outside of the preset allowable range, data obtained by a variable power of the reduction image data vertically based on the average character height and allowable range is outputted as the watermark extraction target image data; and watermark information extraction means for scanning the watermark extraction target image data to extract watermark information. <P>COPYRIGHT: (C)2009,JPO&INPIT |
