Two Novel PDF Document Watermarking Algorithms

As more and more electronic documents, including government documents and so on, are spreading in PDF formats, the secure problems occur. Based on profound analysis of the PDF document-structure, this thesis proposes two PDF document watermarking algorithms oriented to document-structure.Firstly, the PDF document watermarking algorithm of special characters is proposed in the thesis. The special characters refer to the logo of the comment objects of a PDF-- "%". The watermarking signals are masqueraded as legitimate comment objects and embedded in the content stream. The watermarking signals can be embedded anywhere in the content stream. We will embed them in the file body and the trailer. What's more, the comment objects will not be shown when the document is opened. Therefore, the algorithm is invisible.Then, the PDF documents watermarking algorithm based on garbage data stream is proposed. The garbage data stream refers to abandoned page objects of a PDF. PDF document viewer does not display the content described by these discarded page objects. Therefore, we masquerade the watermarking signals as legitimate discarded page objects. We embed the watermarking signals in the file trailer. So, we can avoid modifying the cross-reference tables, and the algorithm is invisible.The security and transparency of both algorithms are ensured, and the capacities are huge. Many simulations on the algorithms, including attacking and comparative experiments, are made. The attacking experiments include 6 types attacking of text and page and compression attacking. The results show that both algorithms can resist all of the attacks. The robustness is ensured. In addition, the proposed algorithms are compared with two kinds of PDF document watermarking algorithms oriented to document-structure by experiments separately. The results show that both algorithms are better than the others, such as robustness, capacity and complexity.