Research On Digital Text Watermarking Algorithms Resistive To Print-Scan Process

Since the21st century, the Internet began to be tightly integrated into the main aspects of public life, becoming an important way for people to obtain information. Digital works have been widely used and spread rapidly through the network. By contrast, violation of the copyright of the digital work is an increasingly serious problem, a serious violation of the legitimate rights and interests of the authors.A main route to pirate digital text is making scanned images of paper text, and illegally spread these scanned images. In order to track the source of piracy or proof of ownership of the digital text, text digital watermarking technology could be used. In order to resist the text print-scan attack, you need to use text digital watermarking algorithm which can be robust to print scan attack. For this purpose, this paper researched digital text watermarking technology and proposed a new digital text watermarking algorithm resistive to Print-Scan Process.First, the researching background and current study of digital watermarking technology is introduced, then the current researching of digital text watermarking technology is introduced in detail. Then the advantages and disadvantages of typical digital text watermarking algorithms which is resistive to print-scan process are analyzed. Based on an analysis of existing text digital watermarking techniques, a digital watermarking algorithms for binary image of Chinese text is proposed, which has a high resistance of print-scan process. The algorithm is based on an approximate invariant of print-scan process. A new concept’virtual average dots number’is proposed to achieve a capacity of one watermark bit per Chinese character. Meanwhile, a simplified strategy for flipping pixels in binary image is used in the algorithm. The strategy is much simplified but has nearly the same effect with much complex strategies when used in binary Chinese text image. By matching the characteristics of Chinese characters with the watermark, the algorithm reinforced its resistance to print-scan process. Experimental results show that the algorithm has a higher capacity than similar algorithms, and its resistance of print-scan process is less affected by characteristics of the text for watermarking.