Research On Software Watermarking With The Improved Dynamic Graph

Currently, there are more and more attention attached to copyrightprotection of software products, software watermarking technology hasbecome particularly important. Although the software watermarkingtechnology has been actively explored domestically and internationally,and some software watermarking algorithms have been presented, thereare still some distance from meeting the actual comprehensive needs ofdata rate, concealment, robustness and so on. For these problems, wepropose a dual software watermarking method which is both static anddynamic. This technology allows software developers to use the softwarewatermarking method in their software products by adding copyrightinformation in the programs and authorizing the purchaser, which providestrong evidence in the event of copyright disputes, as well as maintainingits intellectual property rights to prevent software piracy.Dynamic graph watermarking technology is an important branch ofsoftware watermarking software which can protect the intellectualproperty of the software owners effectively. This thesis focuses on thedynamic graph watermarking feature and model as well as the commoncoding scheme to analyze the watermarking embedding and extractionmethods. Also, we introduce common watermarking attack methods andtheir corresponding protection strategy. For the existing deficiencies lyingin the data rate, robustness and recoverability of software watermarkingalgorithms, we propose a combination of dynamic and static dualwatermarking technology by using Dex files on Android platform. Thistechnology chooses to combine the dummy function and the dynamicgraph algorithms, the former one has great data embedding rate while thelatter one owns strong resistance to deformation attacks. Moreover, wepropose a dual watermarking scheme with the repair function, and designand implement a software watermarking copyright protection system usingDex files. After doing experiments, we come up with an objective analysisof the experimental data, and finally verify the feasibility andeffectiveness of the system of the program. Thereby we solve the existingdeficiency of the algorithms, that it can’t take into account the data rate, concealment and the robustness. At the same time, we settle the problemof watermarking recovery. The specific innovations are embodies in theseaspects are as follow:We propose a software watermarking method combining static anddynamic algorithms which combines the dummy function algorithm withthe dynamic graph algorithm. It uses the advantage of dummyfunction——the high data rate, and of the dynamic graphalgorithm——robustness, to achieve double protection of copyrightinformation.After analyzing the structure of dex files, this thesis finds a suitablewatermarking algorithm for software protection. Double watermarks areembedded in Dex file, whose extraction process are relatively independent.However, if any one of them is destroyed, the recovery of doublewatermarks is interrelated, thus it greatly improves the security of thewatermark.This thesis implements a combination of dynamic and static dualwatermarking scheme. The experiment shows that the algorithm dataembedding rates have been improved significantly, and has goodconcealment and anti-aggressive ability. When the watermark is subject tosome degree of damage, it can be repaired.