基于中英文字符型属性的数据库水印算法

提出了一种基于中英文字符型属性的数据库数字水印算法。通过元组哈希值和字符属性值长度计算出预测嵌入位置,根据预测嵌入位置所在的字节序号确定水印嵌入位置。以ASCII字符集的不可见字符作为水印信号,将版权图像嵌入宿主数据库。嵌入的水印信号具有不可见性,并且保持数据库的原始可用性,可实现水印的盲提取。该算法对添加元组、删除元组、修改属性值、添加与删除属性以及元组与属性重排等常见数据库攻击具有较强的鲁棒性。

Database Watermarking Algorithm Based on Chinese and English Character Attribute

A database watermarking algorithm based on Chinese and English character attribute is proposed in this paper. The predicting embedding position can be calculated using tuple hash value and the length of character data. The watermark embedding location is determined according to the sequence number of the predicting embedding position. The algorithm uses invisible character in ASCII as watermark signal, embeds copyright image into the host database. The embedded watermark is invisible, does not affect the availability of the database, and enables the blind extraction of the watermark. The proposed algorithm can achieve strong robustness to the common database attacks such as tuple addition, tuple deletion, attribute value alteration, attribute addition, attribute deletion, tuple sorting and attribute sorting.