This research presents how to use the intersection position of horizontal line, virtually run across text character skeleton line on a text image under the cross ratio applying, to be the marking point of zero watermark; means watermarking without any actual watermark, which consequently increase the remarkable watermark hiding-bit capacity and robustness. One point of intersection across one text character skeleton line is one specific pixel which defined as one hiding-bit of zero watermark; just only marking its pattern, so that the more reference horizontal lines plotting, the more points of intersection would be got and increasing the hiding-bit capacity for watermarking which generally is the most limitation of other text watermarking methods. The high capacity for hiding bits means more chance to make many sets of redundant encrypted watermark data, in case some parts of text document image have been modified or tampered with. In addition, intersection points for virtually embedding watermarks still have a great strength for identifying the point of text which has been modified if all positions of intersection points are collected. Moreover, these intersection positions also has been applied with the cross ratio of four collinear points, to build up its robustness in order to survive from possibly attacking. The testing proved that this watermarking method can be effectively used to hide more redundant data, verify its integrity, be robust and easily apply it to multi-language text images such as English, Thai, Chinese and Arabic languages.
展开▼
