模糊聚类算法在数据库水印中的应用

Application of Fuzzy Clustering Algorithm in Database Watermarking

当前数据库水印技术存在版权证明次数受限与无法抗密钥泄露的问题,一旦参数泄露,水印信息将被完全暴露。为此,提出一种基于FCM模糊聚类的水印方案,利用聚类算法筛选嵌入水印的元组,并通过分类器来完成水印的检测和数据的恢复。研究结果表明,该方案可多次证明版权,即使密钥及水印参数暴露依然不会暴露水印信息,对常见的数据攻击表现出较好的鲁棒性,具有一定商业价值。

The current database watermarking technology has the problems of limited number of copyright certifications and inability to resist key leakage.Once the parameters are leaked,the watermark information will be completely exposed.Aiming at the above problems,a watermarking scheme based on FCM fuzzy clustering is proposed.The tuples embedded in the watermark are filtered through the clustering algorithm,and the watermark detection and data recovery are completed through the classifier.The research results show that this scheme can prove copyright multiple times,even if the key and watermark parameters are exposed,the watermark information will not be exposed,and it can show good robustness against common data attacks and has certain commercial value.