Research On Content-based English Text Watermarking

With the development of Internet and digital technologies, the media and information transmission mode have been deeply changed. Since varieties of digital production are conveyed and published through the digital networks, it is urgent to solve the issues of security control and copyright protection. Digital watermarking, which can overcome some uncontrollability issues for decrypting encrypted messages, has become a new information security area under this circumstance.Currently, Watermarking researches are mostly focused on image and video instead of text. Some acquired algorithm is mainly implemented by altering the space between lines, words and the characteristic of character slightly. Through deep analysis on the existent text watermarking technologies, it is found that watermarking capacity and robustness need to be improved greatly. The main reason lead to this result is because most of watermarks are not truly being embedded into the content of the text.Based on analysis of disadvantages of the present algorithms and the special demands for the text, this paper presents a new content-based text-watermarking algorithm. According to the similarity between the characteristics of English characters and Greek characters, it is feasible to embed watermark into text by replacing English characters with Greek characters on the premise of not damaging the usability of the text and keeping disguise. Further more, by binding the watermark with the error control code, this new algorithm implements the function of blind-detection, extraction and recoverability, it also attains the goal of security control and tamper recovery.In the experiment the watermarking software was programmed in Java, mainly emphasis on the factors of capacity, invisibility, robustness and security. The attained data of experiment are about embeddable characters, detecting performance and recoverable proportion. The results show that the watermark can be detected blindly, and its integrality and recoverability can be realized easily, the watermarking capacity can be effectively enlarged and the robustness can be successfully increased.The text-watermarking algorithm in this paper breaks through the difficulty that the watermark can't be embedded truly into the content of the text. It's a beneficial attempt to resolve the problem of difficulties to control the document's security issue in the Internet, and this algorithm has an important signification for researches and developments of text watermarking.