Information Technology Reference
In-Depth Information
Detection and Removal of Hidden Data in Images
Embedded with Quantization Index Modulation
Kaiwen Zhang, Shuozhong Wang, and Xinpeng Zhang
Communication & Information Engineering, Shanghai University, Shanghai 200072, China
ztszkwzr@sh163.net,shuowang@yc.shu.edu.cn
zhangxinpeng@263.net
Abstract.
Methods of attack on watermarks embedded with quantization index
modulation and block DCT are introduced in this paper. The proposed method
can detect the presence of hidden data by analyzing the histogram or spectrum
of the transform coefficients, and prevent their extraction without further de-
grading the image. It is shown that, in some cases, the embedding-induced dis-
tortion can even be reduced in terms of peak-signal-to-noise ratio. Experimental
results are given to show the effectiveness of the method.
1 Introduction
Digital watermarking has become an active research area over the last decade. Protec-
tion of intellectual property rights and covert communication are typical applications.
It is essential for any data hiding technique that the hidden information causes no
serious degradation to the host signal and, in the meaning time, the embedding must
be robust enough to be able to survive hostile attacks as well as common signal proc-
essing.
Many information embedding techniques are robust in resisting various signal
processing operations such as compression and filtering, but less robust against cer-
tain types of malicious attacks. In this paper, we demonstrate an attack scheme that
can defeat watermarking for still images based upon quantization index modulation
(QIM) [1-3]. It can reveal the presence of hidden information in the absence of the
original cover image, and remove the embedded data without further degrading per-
ceptual quality of the image. The purpose of studying the attack strategies is twofold.
First, successful attacks may provide clues for improving watermarking performance,
and secondly, attack itself is a countermeasure against hostile covert communication.
The latter, referred to as active warden attack, is an important part of the current re-
search topic of information warfare that has attracted much attention form computer
specialists, signal processing and information security professionals, and software
developers [4].
In Section 2 we briefly describe a watermarking technique for images based on
QIM and block DCT. Statistical characteristics of the DCT coefficients of a water-
marked image are studied in Section 3. Section 4 presents a method for detecting and
disabling QIM/DCT-based watermarks. Experimental results are given in Section 5.
Section 6 concludes the paper.
