Weakness analysis of singular value based watermarking

Singular value (SV)-based image watermarking has recently been gaining attention. The advantage of this watermarking approach is its robustness to distortion attacks. In this paper, we mathematically prove that SV-based watermarking algorithms are robust against distortion attacks as long as the attacks are not severe. The same mathematical proof, however, also confirms that such watermarking schemes are unfortunately susceptible to false watermark detection. Thus a false watermark (not the embedded one) can be detected from a watermarked image that is not severely distorted or even from an image that is not watermarked. We then propose a method that allows an attacker to claim the ownership of any watermarked image whether or not it is severely distorted. SV-based watermarking algorithms cannot thus be used for protecting the ownership of an image.