Attacks on digital watermarks

Two methods for attacking watermarks in images and videos have been proposed. The first attack is to remove robust watermarks by iterative block dropping and concealment. Our study suggests that the block size chosen by a robust watermark embedding system should be large enough to circumventing concealment based attacks. The second attacking method is to forge valid watermarks for authentication systems based on fragile-watermarking. Double watermarking is then proposed as one possible countermeasure.