Watermarking relational data based on adaptive mechanism

A novel watermarking algorithm is proposed for intellectual property rights (IPR) protection for relational data. An adaptive mechanism is adopted to choose an optimal embedding scheme for each datasheet according to its data characteristics based on predefined several fundamental schemes aiming at making distortion of the watermarked data to be smallest. The error-correcting coding technology and the majority voting principle are employed to improve the robustness of watermarking. Experimental results reveal that the watermarking can meet the invisibility and the blind detection requirements and the proposed algorithm has good resistance to conventional attacks such as subset deleting, adding, updating, and so on.