Watermarking security incorporating natural scene statistics

Jiangqun Ni, Rongyue Zhang, Chen Fang, Jiwu Huang, Chuntao Wang, Hyoung Joong Kim

Research output: Chapter in Book/Report/Conference proceedingConference contribution

3 Citations (Scopus)


Watermarking security has emerged as the domain of extensive research in recent years. This paper presents both information theoretic analysis and practical attack algorithm for spread-spectrum based watermarking security incorporating natural scene statistics (NSS) model. Firstly, the Gaussian scale mixture (GSM) is introduced as the NSS model. The security is quantified by the mutual information between the observed watermarked signals and the secret carriers. The new security measures are then derived based on the GSM model, which allows for more accurate evaluation of watermarking security. Finally, the practical attack algorithm is developed in the framework of variational Bayesian ICA, which is shown to increase the performance and flexibility by allowing incorporation of prior knowledge of host signal. Extensive simulations are carried out to demonstrate the feasibility and effectiveness of the proposed algorithm.

Original languageEnglish
Title of host publicationInformation Hiding - 10th International Workshop, IH 2008, Revised Selected Papers
PublisherSpringer Verlag
Number of pages15
ISBN (Print)3540889604, 9783540889601
Publication statusPublished - 2008
Event10th International Workshop on Information Hiding, IH 2008 - Santa Barbara, CA, United States
Duration: 2008 May 192008 May 21

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume5284 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349


Other10th International Workshop on Information Hiding, IH 2008
Country/TerritoryUnited States
CitySanta Barbara, CA

ASJC Scopus subject areas

  • Theoretical Computer Science
  • General Computer Science


Dive into the research topics of 'Watermarking security incorporating natural scene statistics'. Together they form a unique fingerprint.

Cite this