An improved algorithm to watermark numeric relational data

  • Authors:
  • Fei Guo;Jianmin Wang;Zhihao Zhang;Xiaojun Ye;Deyi Li

  • Affiliations:
  • School of Software, Tsinghua University, Beijing, China;School of Software, Tsinghua University, Beijing, China;School of Software, Tsinghua University, Beijing, China;School of Software, Tsinghua University, Beijing, China;School of Software, Tsinghua University, Beijing, China

  • Venue:
  • WISA'05 Proceedings of the 6th international conference on Information Security Applications
  • Year:
  • 2005

Quantified Score

Hi-index 0.00

Visualization

Abstract

This paper studies an improved algorithm to watermark numeric attributes in relational databases for copyright protection. It reviews related researches and presents an improved insertion algorithm, a detection algorithm and a recover algorithm. We introduce a varied-size grouping method in our insertion algorithm to insert a meaningful watermark. We also introduce a new mechanism to insert watermarks using the mark itself to decide marked positions. This insertion mechanism can be validated in our detection algorithm to decide whether a watermark exists or not. A badly destroyed marked relation or only a small part of it could still be detected successfully. Our recover algorithm introduces a competing mechanism to help recover the exact meaningful watermark after the detection result confirms the existence of the watermark. The experiments show it's robust to various attacks.