Share Email Print

Proceedings Paper

Watermarking capacity in the presence of multiple watermarks and a partially known channel
Format Member Price Non-Member Price
PDF $14.40 $18.00
cover GOOD NEWS! Your organization subscribes to the SPIE Digital Library. You may be able to download this paper for free. Check Access

Paper Abstract

The effect of a partially known watermarking channel, additive noise, and multiple watermarks on the watermarking capacity region is studied. A channel can be partially known because of the following reasons : (a) randomly time-varying channel characteristics, (b) unknown attacks by an adversary, (c) uncertainty due to estimation errors such as in oblivious watermarking techniques. A mathematical model for this scenario is introduced. No assumptions are made regarding the probability distribution of the channel. Lower and upper bounds on the feasible watermarking rate region are derived. It is shown that, in terms of watermarking capacity, it is better to cancel the effect of an interfering watermark than to treat is as noise. As a special case, it is also observed that the capacity estimates based on the popular additive Gaussian noise model tend to either over or under estimate the capacity for a single watermark channel. Numerical results are also presented. Finally, we observe that the proposed mathematical model can be applied to real-life applications such as image/video watermarking. Image processing operations such as scaling, geometrical transformations etc. that distort the image (not just add noise) fall under the proposed mathematical model.

Paper Details

Date Published: 12 November 2001
PDF: 6 pages
Proc. SPIE 4518, Multimedia Systems and Applications IV, (12 November 2001); doi: 10.1117/12.448205
Show Author Affiliations
Rajarathnam Chandramouli, Stevens Institute of Technology (United States)

Published in SPIE Proceedings Vol. 4518:
Multimedia Systems and Applications IV
Andrew G. Tescher; Bhaskaran Vasudev; V. Michael Bove Jr., Editor(s)

© SPIE. Terms of Use
Back to Top