A regression-based restoration technique for automated watermark removal

  • Authors:
  • Andreas Westfeld

  • Affiliations:
  • Technische Univerität Dresden, Dresden, Germany

  • Venue:
  • Proceedings of the 10th ACM workshop on Multimedia and security
  • Year:
  • 2008

Quantified Score

Hi-index 0.01

Visualization

Abstract

The second Break our Watermarking System (BOWS-2) contest exposed a technology named "Broken Arrows" to attacks from all over the world. For a successful attack participants had to render the watermark unreadable to the online detector in three given images while preserving a minimum quality level of 20 dB PSNR. We applied a generic approach to remove the independent elements of the watermark from dependent media elements. Its core part does not need to interact with the online detector to remove the watermark. In a postprocessing step with ten detector calls the quality was increased by about 1 dB. The first episode of the second Break Our Watermarking System (BOWS-2) contest was aiming to investigate when an image watermarking system can be broken by image processing operations. During this period, from July to October 2007, no information about the watermarking algorithm was provided and only 30 trials per day were allowed. We describe the attack that preserved the highest quality of the content.