Research On Software Copyright Authentication Technology Based On Watermark And Feature

With the continuous popularization of network services,the problems of software infringement,piracy,arbitrary tampering and malicious attacks are increasing,which seriously hinders the healthy and sustainable development of the software industry.Software protection technology has been valued by governments,entrepreneurs and scholars of various countries.The software protection technologies such as copyright management,anti-piracy,and software identification have also become key and hot research issues.At present,there are many software protection technologies,and there are protection methods for hardware,such as dongle,etc.There are software-based protection methods,such as watermark,birthmark feature,and encryption technology.Software watermarking is a type of digital watermarking.It is more concealed than other technologies such as dongle,and it is more secure than encryption technology.It can provide legal proof.Software feature recognition technology is a new idea of software copyright protection.By extracting representative features and performing feature matching to determine whether one of the software infringes the copyright of another software.Many software watermarking and software birthmark algorithms have been proposed,but they cannot meet the application requirements from the aspects of security,robustness and credibility.And the existing software watermarking and feature recognition algorithms basically use the program source code or Java bytecode as the research object,but the research using the PE file(executable files)as the carrier is still relatively small.Based on the analysis of existing software watermarking technology,this paper analyzes the format characteristics of PE files and focuses on the feasibility of PE files as carriers.In view of the different infringements against software,a new Algorithm based on PE files.Aiming at software copyright disputes,according to the instruction encoding principle,watermark information is hidden in the code segment,and then important information such as the watermark length,key,and random position are hidden in the index image such as the resource segment icon using the algorithm of the copy palette.The algorithm combines the cryptographic algorithm to improve the security of the algorithm,and combines the dual measures of Hamming coding and threshold secret sharing scheme to resist tampering;In view of the problems of information loss,duplication,and tampering in software transmission,this algorithm extracts structural features such as file length,number of segments,and length of each segment,and combines bundling technology and character segment matching algorithm for integrity check;For illegal reuse of software modules,extract software special instruction features to identify the core code section of the PE file and preprocessed watermark information Perform XOR operation and store it as zero watermark authentication information in the zero watermark information database for verification.The experimental results show that the software watermarking algorithm based on instruction tamper-proof coding and resources proposed in this paper can resist about 5%random tampering attacks,and improve the performance indicators of concealment,security and robustness.For the structure-based feature extraction algorithm proposed in this paper,the PE file is destroyed by changing the PE file program entry,adding useless instructions,adding and deleting section tables,modifying the optional header size of the PE file,modifying the base address of the code,and initializing the data.Integrity,the software exits and displays the software copyright;for the special instruction feature extraction algorithm proposed in this paper,through the chi-square test of the features extracted by this experimental software and variant software,it proves that this extraction algorithm is robust,and by this experimental software Perform chi-square test with the features extracted by different software to prove that the extraction algorithm is credible,and thus set the threshold to 5.5.When the similarity between the suspected object and the genuine software is checked,the chi-square value is less than this threshold.Consider the suspected person infringing copyright.