Research On Database Watermarking Technology

With the wide application of computer copyright and database technology,more and more attention has been paid to the protection of information and data relations.As an important means of copyright protection,relational database watermarking requires that important information proving the copyright relationship be embedded in the data without damaging the usability of the data itself,and it can still detect the existence of watermarks in a certain proportion after being attacked maliciously,so as to play the role of copyright protection.Therefore,the research of relational database watermarking has high value and significance in practical application,and has become one of the current research hotspots in recent years.Based on the research of the existing domestic and foreign relational database watermarking technology,this paper proposes the following solutions:(1)A database watermarking scheme based on clustering with size constraint is proposed.In this algorithm,a clustering algorithm with size constraint is introduced to strengthen the user's control over the number,strength and data selection of watermark.Experiments show that the scheme has strong robustness and improves the security.(2)In the watermarking scheme,MSB(most significant bit)is applied.That is to say,when the watermark embedding conditions are met,the watermark is selected from the most significant bit of the watermark,and the corresponding watermark information selected by the data owner is embedded into all the tuples of an attribute in turn.The scheme improves the accuracy of watermark extraction and makes the watermark have good robustness and usability.(3)In the scheme,the filtering of the data to be embedded watermark is added.The embedding position of the watermark needs to further filter the relational data according to the lowest bit of the integer part of the relational data,which increases the concealment of the watermark.In addition,the scheme encrypts the watermark information and selects the attributes of the data to be clustered by secret key,which greatly improves the security of the watermark.Aiming at the problem that the selection method of watermark data to be embedded is single in the current relational database watermarking scheme,a new solution is proposed.Combined with the clustering algorithm with size constraint,the data selection can be diversified through data preprocessing.In terms of performance,the proposed clustering algorithm with size constraint is more stable and converges faster and more stably than the traditional clustering algorithm.In addition,aiming at the problem that the functionality of the watermark needs to be strengthened,the original watermark embedding and detection algorithm is improved.The preprocessing of the watermark information,the selection of clustering attributes,the selection of embedded watermark data and other processes are added.The implementation proves that the improved scheme has greatly improved the anti attack ability of the watermark and the concealment of the watermark.