Reversible Data Hiding for Block Truncation Coding Compressed Images Based on
Prediction-Error Expansion
Kan Wang and Yongjian Hu
School of Electronic and Information Engineering
South China University of Technology
Guangzhou, P.R. China
w.kan@mail.scut.edu.cn, eeyjhu@scut.edu.cn
Zhe-Ming Lu*
School of Aeronautics and Astronautics
Zhejiang University
Hangzhou, P.R. China
*Corresponding author: zheminglu@zju.edu.cn
Abstract—In this paper, we propose a new reversible data
hiding scheme for images compressed by Block Truncation
Coding (BTC). The original image is divided into n × n sized
non-overlapping blocks and encoded by BTC to obtain the
BTC-compressed bitstream. Each block is presented by a low
mean value, a high mean value and a binary bitmap. The
proposed scheme exploits the correlation among adjacent
blocks and embeds data into low mean value and high mean
value by prediction-error expansion. Meanwhile, no additional
overflow/underflow location map is required. For those blocks
with the same low mean value and high mean value, data are
embedded into the bitmaps. Experimental results show that the
proposed scheme achieves high embedding capacity and great
image quality.
Keywords-data hiding; reversible data hiding; BTC; block
truncation coding; prediction-error expansion
I. INTRODUCTION
The rapid development of information technology
improves people's living style and enlarges communication
methods. All kinds of resources are sharing across the
Internet, but the problems of network security subsequently
come. Data transmitted over the Internet could be easily
intercepted and forged for its openness and sharing. Data
hiding is one way to help solve the security problem. Data
hiding techniques embed secret data into cover media
without visual distortion. Conventional data hiding
techniques cause irreversible distortion, while reversible data
hiding techniques can retrieve the hidden data and the
original cover media perfectly. Reversibility is a very
important property for several special fields like medical,
military and forensic fields.
Nowadays, most digital images are stored in compressed
formats such as JPEG and JPEG2000. Block Truncation
Coding (BTC) is another efficient lossy compression
technique that significantly reduces the image file size and
keep an acceptable visual quality. The first work of data
hiding for BTC-compressed images was proposed by Lu et al.
[1] in 2002. The watermark was embedded by modifying the
VQ-BTC encoding process according to the watermark bits.
In 2004, Lin et al. [2] proposed a data hiding scheme for
BTC-compressed images by performing LSB substitution
operators on BTC high and low means, meanwhile
performing the minimum distortion algorithm on BTC
bitmap. In 2006, Chang et al. [3] proposed a BTC scheme to
embed data into the smooth regions of the image. However,
above works are not reversible. In 2008, Hong et al. [4]
proposed a lossless data hiding scheme for BTC-compressed
images for the first time. Secret data were embedded by
swapping low mean value and high mean value meanwhile
flipping bitmaps. However, when low mean value and high
mean value were equal, there was a problem in extracting the
secret data. Later in 2010, Chen et al. [5] overcame this
problem and proposed an improved lossless data hiding
scheme. The difference between the high mean value and the
low mean value of each block was used to determine
whether only one bit of secret data was to be hidden for the
block or to toggle bits in the bitmap to hide more bits. In
2011, Li et al. [6] presented an efficient reversible data
hiding algorithm for BTC-compressed images. Their scheme
introduced the histogram shifting technique to BTC-
compressed mean tables to further improve the hiding
capacity while maintaining the image quality.
Most of existing data hiding techniques for BTC-
compressed are irreversible, only a few dealt with reversible
data hiding. There is a huge margin for improvement in
capacity and image quality. In this paper, we propose a new
reversible data hiding scheme for BTC-compressed images
by introducing the prediction-error expansion technique.
Experimental results confirm the superiority of the proposed
approach for reversible data hiding.
The rest of this paper is organized as follows. Sec. II
briefly reviews BTC compression technique and prediction-
error expansion. Sec. III describes the proposed data hiding
scheme in detail. Experimental results are presented and
explained in Sec. IV, followed by the conclusions in Sec. V.
II. R
ELATED WORKS
In this section, a brief introduction on BTC compression
technique is given and then followed by the concept of
prediction-error expansion.
A. Block Truncation Coding
Block Truncation Coding (BTC) is a lossy image
compression technique proposed by Delp and Mitchell [7]. It
divides the original images into blocks and then uses a
quantizer to reduce the number of gray levels and maintains
the same mean and standard deviation. Later Lema and
Mitchell [8] proposed another variation Absolute Moment
Block Truncation Coding (AMBTC). It preserves the first
2012 Eighth International Conference on Intelligent Information Hiding and Multimedia Signal Processing
978-0-7695-4712-1/12 $26.00 © 2012 IEEE
DOI 10.1109/IIH-MSP.2012.83
317