The Research On Chinese Text Zero-Watermarking Based On Chinese Mathematical Expressions

Digital watermarking refers to a distinguished digital signal or mode, which is embedded in other data (host data) to prove their copyright. It can be applied to various multimedia data including text document, image, audio, and video etc. Generally speaking, image, audio and video watermarking are comparatively similar in algorithm realization, with their redundancy and the trait of HVS (Human Video System) or HAS (Human Audio System), in which we can embed watermark. But other than the aforementioned, there is no redundancy to transfer secret info in text document. Nowadays, to embed watermark in text documents are limited to methods such as shifting the line and word, amending the characters' traits and disposing in the level of semantics. All these algorithms are not robust or lack of concealment, generally not serving the turn of Chinese text documents rifely. After studying the trait of text documents and deficiency of presently text Digital watermarking algorithms, a novel Chinese text zero-watermarking method is brought forward in this paper based on Mathematical Expression of Chinese Characters. With the statistic result of the frequency of each component in the text document, we embed the watermark by distilling the content which satisfied its demand before sending it to the Register Center. While the copyright was oppugned, we compared the watermark which detected from the dubious text document with that registered in the Register Center, if they are detected from the same mode, we cognize there is original watermark in the text document and the copyright is belongs to the author, or vice versa. Different from common watermarking method, there are no any modification to the text document during the construction of zero-watermark and it do not rely on the format of text document or characters' characteristic. It shows according to experiments that the watermark cannot be wiped off with the format adjusting, characteristic amending or contents' deleting. Moreover, the zero-watermark is more robust and concealment than those based on the format of text or trait of characters; compared with the watermarking that based on semantics, it suits more to Chinese text document than those based on its language structure. It is consequently known to us that the method brought forward in this paper has certain value of theory and application concerned with the request of robust text watermarking, copyright protection and the experimental result, under the wish to bring some novel thoughts for development of Chinese text document watermarking.