Quantization-based watermarking performance improvement using host statistics: AWGN attack case

Authors:
Oleksiy Koval;Sviatoslav Voloshynovskiy;Fernando Pérez-González;Frederic Deguillaime;Thierry Pun
Affiliations:
University of Geneva, Geneva, Switzerland;University of Geneva, Geneva, Switzerland;University of Vigo, Vigo, Spain;University of Geneva, Geneva, Switzerland;University of Geneva, Geneva, Switzerland
Venue:
Proceedings of the 2004 workshop on Multimedia and security
Year:
2004

Citing 3
Cited 3

Image Coding based on Mixture Modeling of Wavelet Coefficients and a Fast Estimation-Quantization Framework

DCC '97 Proceedings of the Conference on Data Compression
Guest editorial special issue on signal processing for data hiding in digital media and secure content delivery

IEEE Transactions on Signal Processing
Image coding based on a morphological representation of wavelet data

IEEE Transactions on Image Processing

Quantization-based methods: additive attacks performance analysis

Transactions on data hiding and multimedia security III
How quantization based schemes can be used in image steganographic context

Image Communication
Improved QIM strategies for gaussian watermarking

IWDW'05 Proceedings of the 4th international conference on Digital Watermarking

Quantified Score

Hi-index	0.00

Visualization

Abstract

In this paper we consider the problem of performance improvement of known-host-state (quantization-based) watermarking methods undergo Additive White Gaussian noise (AWGN) attack. The motivation of our research is twofold. The first reason concerns the common belief that any knowledge about the host image taken into account designing quantization-based watermarking algorithms can not improve their performance. The second reason refers to the poor practical performance of this class of methods at low Watermark-to-Noise Ratio (WNR) regime in comparison to the known-host-statistics techniques when AWGN attack is applied. We demonstrate in this paper that bit error probability of Dither Modulation (DM) and Distortion-Compensated Dither Modulation (DC-DM) against AWGN attack can be significantly reduced when the quantizers are designed using the statistics of the host data. For the case when the statistics of the data correspond to i.i.d. Laplacian distribution and using Uniform Deadzone Quantizer (UDQ) we develop close-form analytical models for the analysis of bit error probability of DM and DC-DM. Results of performed experiments demonstrate that significant performance improvement of classical DM and DC-DM with respect to bit error probability can be achieved with the minor increase of design complexity.