JPEG quantization tables forensics: a statistical approach

Authors:
Babak Mahdian;Stanislav Saic;Radim Nedbal
Affiliations:
Institute of Information Theory and Automation of the ASCR, Czech Republic;Institute of Information Theory and Automation of the ASCR, Czech Republic;Institute of Computer Science of the ASCR, Czech Republic
Venue:
IWCF'10 Proceedings of the 4th international conference on Computational forensics
Year:
2010

Citing 3
Cited 0

JPEG Still Image Data Compression Standard

JPEG Still Image Data Compression Standard
A bibliography on blind methods for identifying image forgery

Image Communication
Detection of Double-Compression in JPEG Images for Applications in Steganography

IEEE Transactions on Information Forensics and Security

Quantified Score

Hi-index	0.00

Visualization

Abstract

Many digital image forensics techniques using various fingerprints which identify the image source are dependent on data on digital images from an unknown environment. As often software modifications leave no appropriate traces in image metadata, critical miscalculations of fingerprints arise. This is the problem addressed in this paper. Modeling information noise, we introduce a statistical approach for noise-removal in databases consisted of "unguaranteed" images. In this paper, employed fingerprints are based on JPEG quantization tables.