A Spatial Masking Technique for Optimal Frequency-Based Image Watermarking

Authors

Abstract

This paper describes an adaptive spatial masking technique
for frequency-based image watermarking schemes. The
strategy consists of defining minimum and maximum
allowable pixel perturbations. These minimum pixel
perturbations, defined as constraints of an optimal watermark
embedding problem, are obtained from measuring
the effects in terms of inaccuracy of the most damaging
interpolation methods over images. By including these
perturbations, the robustness of watermarking schemes is
improved: the watermark strength in images regions which
are less vulnerable to distortions is increased.
The effectiveness of the proposed scheme is illustrated
with experiments made on two watermarking algorithms
where the described masking technique was applied.

Keywords

Watermarking, spatial masking, minimum pixel perturbations

Subject

Watermarking

Conference

Int. Conf. on Signal and Image Processing, August 2004