Countering counter-forensics: the case of JPEG compression

Authors:
ShiYue Lai;Rainer Böhme
Affiliations:
European Research Center for Information System, University of Münster, Münster, Germany;European Research Center for Information System, University of Münster, Münster, Germany
Venue:
IH'11 Proceedings of the 13th international conference on Information hiding
Year:
2011

Citing 9
Cited 1

Steganalysis of JPEG Images: Breaking the F5 Algorithm

IH '02 Revised Papers from the 5th International Workshop on Information Hiding
Calibration revisited

Proceedings of the 11th ACM workshop on Multimedia and security
Tamper hiding: defeating image forensics

IH'07 Proceedings of the 9th international conference on Information hiding
Feature-Based steganalysis for JPEG images and its implications for future design of steganographic schemes

IH'04 Proceedings of the 6th international conference on Information Hiding
Detecting doctored JPEG images via DCT coefficient analysis

ECCV'06 Proceedings of the 9th European conference on Computer Vision - Volume Part III
Hiding Traces of Resampling in Digital Images

IEEE Transactions on Information Forensics and Security
Detection of Double-Compression in JPEG Images for Applications in Steganography

IEEE Transactions on Information Forensics and Security
A mathematical analysis of the DCT coefficient distributions for images

IEEE Transactions on Image Processing
Identification of bitmap compression history: JPEG detection and quantizer estimation

IEEE Transactions on Image Processing

JPEG anti-forensics using non-parametric DCT quantization noise estimation and natural image statistics

Proceedings of the first ACM workshop on Information hiding and multimedia security

Quantified Score

Hi-index	0.01

Visualization

Abstract

This paper summarizes several iterations in the cat-andmouse game between digital image forensics and counter-forensics related to an image's JPEG compression history. Building on the counterforensics algorithm by Stamm et al. [1], we point out a vulnerability in this scheme when a maximum likelihood estimator has no solution. We construct a targeted detector against it, and present an improved scheme which uses imputation to deal with cases that lack an estimate. While this scheme is secure against our targeted detector, it is detectable by a further improved detector, which borrows from steganalysis and uses a calibrated feature. All claims are backed with experimental results from 2 × 800 never-compressed never-resampled grayscale images.