Natural Language Digital Watermark Method on Paragraph Level
Guojiang Xin · Jisuanji gongcheng · 2010
This paper proposes a natural language digital watermark method on paragraph level,by introducing spread spectrum technique to encode the watermark.All the named entities are picked out to create a vector space.A sub-space is selected to embed watermark according to the secret key.In order to embed watermark,the number of named entities is modified by employing the anaphora technique.Whether the watermark is embedded in text is determined by comparing the extracted vector with the vector generated by the original watermark.Experimental results show that the method is robust,and can resist some common active attacks.