Study On The Embedding Way Of Text Watermark

With the rapid development of computer multi-media and network, more and more digital products apper into the market, and the question of the safy of the products need to be resolved. At the same time, digital watermarking is the way to deal with the question.The digital watermarking can be classified image watermarking, video watermarkiong and text watermarking. Text watermarking has not been given more consideration by many researchers compare to other watermarking, because the Chinese expression is much more abundant than the English, some little modification can change the original meaning of the production. In order to improve the robustness, watermarking capacity, and identify the attacks of deleting and adding characters, this thesis will make use of the idea of zero-watermarking, and proposes a novel text watermarking algorithm based on the feature Chinese characters. The main research work and contributions of this thesis are summarized as follows:1. We study the principle of watermarking algorithm such as a form based approach, a natural language approach, a character feature approach and zero-watermarking, and compare the merit and demerit of these algorithms.2. At the basic of Chinese characters, we will propose a novel text watermarking algorithm based on the stroke of Chinese characters, which takes specified Chinese characters as marks, and gets the feature by dealing with the neighborhood of the marks. The paper will zoom out the embedded level to stroke. Experimental results show that the algorithm has a greater watermarking capacity, stronger robustness and best imperceptibility. At the same time, the algorithm has good robustness against the format attack, and can identify the attacks of deleting and adding characters.3. We propose a novel text watermarking algorithm based on the certification of the content of text. The algorithm take after the Chinese characters and watermarking, making use of the relation between8strokes and the neighborhood of8strokes to achieve the complete certification. Experimental results show that the algorithm can identify and locate the attacks of deleting and adding characters, and show that the alsorithm can satiffy the rule of the text watermarking certification, achieving the better certification.