In this paper, we describe some attack strategies we have applied to the three protected images proposed within the particular context of BOWS (break our watermarking system). The technique we have first used exploits the notion self-similarities included in images. This technique alone was powerful enough for the first image whereas for the two other images the watermark has been removed but the use of some additional processing steps was necessary to preserve the minimum required PSNR. Additional tools used in this context were composed of basic signal processing techniques (e.g. filtering). In our article, we also propose a possible use of BOWS as a teaching tool for master students. A homework (derived from BOWS) was then proposed. 30 students were asked to break the watermark of the three images while keeping the highest possible PSNR. They had to return their results as well as a report within a delay of two weeks. This homework was an attractive training for students to better understand the difficulty of breaking a watermark and to develop a critical point of view about watermarking algorithm; in short, to increase their global understanding and knowledge on digital watermarking.
This paper is published in Security, Steganography, and Watermarking of Multimedia Contents IX part of Electronic Imaging, 28 January-1 February 2007, San Jose, USA | SPIE proceedings Volume 6505 and is made available as an electronic preprint with permission of SPIE. One print or electronic copy may be made for personal use only. Systematic or multiple reproduction, distribution to multiple locations via electronic or other means, duplication of any material in this paper for a fee or for commercial purposes, or modification of the content of the paper are prohibited.
