A Bayesian approach to attack characterisation using robust watermarks

HD Knowles; DA Winne; CN Canagarajah; DR Bull

doi:10.1117/12.503067

A Bayesian approach to attack characterisation using robust watermarks

HD Knowles, DA Winne, CN Canagarajah, DR Bull

Engineering Faculty Office

Research output: Chapter in Book/Report/Conference proceeding › Conference Contribution (Conference Proceeding)

3 Citations (Scopus)

Abstract

In this paper we propose the use of a Bayesian framework to allow characterisation of image tampering from a library of attacks. We use the double watermarking strategy proposed in our previous work to derive sufficient information to drive the classifier. A non-parametric Bayesian classifier, trained on data derived from Monte Carlo simulations is used. In addition to classification, the effects of varying the input parameters are studied. The results obtained show that the non-parametric Bayesian classifier has a very low misclassification rate for this type of problem. Explanations as to the nature of the results, and some of the practical considerations, are given.

Translated title of the contribution	A Bayesian approach to attack characterisation using robust watermarks
Original language	English
Title of host publication	Visual Communications and Image Processing 2003, Lugano, Switzerland
Publisher	Society of Photo-Optical Instrumentation Engineers (SPIE)
Pages	851 - 861
Number of pages	11
Volume	5150 II
DOIs	https://doi.org/10.1117/12.503067
Publication status	Published - 8 Jul 2003

Access to Document

10.1117/12.503067

Handle.net

Persistent link

Cite this

@inproceedings{f922e6f82b794f4394d8d49e1653b15b,

title = "A Bayesian approach to attack characterisation using robust watermarks",

abstract = "In this paper we propose the use of a Bayesian framework to allow characterisation of image tampering from a library of attacks. We use the double watermarking strategy proposed in our previous work to derive sufficient information to drive the classifier. A non-parametric Bayesian classifier, trained on data derived from Monte Carlo simulations is used. In addition to classification, the effects of varying the input parameters are studied. The results obtained show that the non-parametric Bayesian classifier has a very low misclassification rate for this type of problem. Explanations as to the nature of the results, and some of the practical considerations, are given.",

author = "HD Knowles and DA Winne and CN Canagarajah and DR Bull",

year = "2003",

month = jul,

day = "8",

doi = "10.1117/12.503067",

language = "English",

volume = "5150 II",

pages = "851 -- 861",

booktitle = "Visual Communications and Image Processing 2003, Lugano, Switzerland",

publisher = "Society of Photo-Optical Instrumentation Engineers (SPIE)",

address = "United States",

}

A Bayesian approach to attack characterisation using robust watermarks. / Knowles, HD; Winne, DA; Canagarajah, CN et al.
Visual Communications and Image Processing 2003, Lugano, Switzerland. Vol. 5150 II Society of Photo-Optical Instrumentation Engineers (SPIE), 2003. p. 851 - 861.

Research output: Chapter in Book/Report/Conference proceeding › Conference Contribution (Conference Proceeding)

TY - GEN

T1 - A Bayesian approach to attack characterisation using robust watermarks

AU - Knowles, HD

AU - Winne, DA

AU - Canagarajah, CN

AU - Bull, DR

PY - 2003/7/8

Y1 - 2003/7/8

N2 - In this paper we propose the use of a Bayesian framework to allow characterisation of image tampering from a library of attacks. We use the double watermarking strategy proposed in our previous work to derive sufficient information to drive the classifier. A non-parametric Bayesian classifier, trained on data derived from Monte Carlo simulations is used. In addition to classification, the effects of varying the input parameters are studied. The results obtained show that the non-parametric Bayesian classifier has a very low misclassification rate for this type of problem. Explanations as to the nature of the results, and some of the practical considerations, are given.

AB - In this paper we propose the use of a Bayesian framework to allow characterisation of image tampering from a library of attacks. We use the double watermarking strategy proposed in our previous work to derive sufficient information to drive the classifier. A non-parametric Bayesian classifier, trained on data derived from Monte Carlo simulations is used. In addition to classification, the effects of varying the input parameters are studied. The results obtained show that the non-parametric Bayesian classifier has a very low misclassification rate for this type of problem. Explanations as to the nature of the results, and some of the practical considerations, are given.

U2 - 10.1117/12.503067

DO - 10.1117/12.503067

M3 - Conference Contribution (Conference Proceeding)

VL - 5150 II

SP - 851

EP - 861

BT - Visual Communications and Image Processing 2003, Lugano, Switzerland

PB - Society of Photo-Optical Instrumentation Engineers (SPIE)

ER -

A Bayesian approach to attack characterisation using robust watermarks

Abstract

Access to Document

Handle.net

Fingerprint

Cite this