Recent years have seen a great proliferation of papers on watermarking of digital data. These have usually started from a very generalised view of the nature of the data and concentrated on the quality of the security algorithm.
This presentation considers watermarking from the opposite viewpoint. It takes as its starting point the knowledge of image presentation acquired in the commercial world including familiarity with the variety of common transformations such as screening and compression. The question is raised as to how effectively proposed methods survive such transformations and how to devise embedded signals that are robust and not too harmful to image quality.