Digital watermarks have been proposed in recent literature as the approach for copyright protection of multimedia data. For its invisibili- ty, validity and robustness, digital watermark is an effective Post Ar- chitecture approach for personality identification and information trace. The watermark used for copyright protection is related to the watermark scheme and copyright protection laws.
Multiple claims of the rightful ownership for a digital product, also called the IBM attack, are the threshold for the watermarking usage in the copyright protection. In the normal feature-based algorithms, the feature extractors are decided by the watermarking scheme and open to the public, so the watermarked features are exposed to attackers and the wate- rmarking scheme is vulnerable. At the same time, the inserted watermarks are irrespective for each other and we can not ensure the One-Way charact- eristic of the algorithm, so the IBM attack is possible. In the paper, our feature extractors and inserted watermarks are decided by the water- marking content and transparent to the public, so attackers can not extract the watermarked features from the watermarked multimedia data and do the modification. And the inserted signature is interrelated with the feature extractors, extracted features and original un-watermarked multimedia, so attackers are impossible to fabricate effective watermarks.
To resolve the IBM attack brought by the multiple watermarks in a digital product, we divide the copyright protection system into watermarking algorithm and watermarking deploying strategy two parts in the paper. When they cooperate well, the validity and completeness of the system can be resolved.