Multimed Tools Appl

DOI 10.1007/s11042-016-4032-8

An efficient reversible data hiding method for AMBTC

compressed images

Wien Hong 1,2 & Yuan-Bo Ma 2 & Hung-Che Wu &

3

Tung-Shou Chen 4

Received: 12 February 2016 / Revised: 10 September 2016 / Accepted: 4 October 2016

# Springer Science+Business Media New York 2016

Abstract Analyzing multimedia data in mobile devices is often constrained by limited

computing capacity and power storage. Therefore, more and more studies are trying to
investigate methods with algorithm efficiency. Sun et al. proposed a low computing cost
reversible data hiding method for absolute moment block truncation coding (AMBTC) images
with excellent embedding performance. Their method predicts quantization values and uses
encrypted data bits, division information, and prediction errors to construct the stego codes.
This method successfully embeds data while providing a comparable bit-rate; however, it does
not fully exploit the correlation of neighboring pixels and division of prediction error for better
embedment. Therefore, the payload and bit-rate are penalized because the embedding perfor-
mance directly depends on the prediction accuracy and division efficiency. In this paper, we
use median edge detection predictor to better predict the quantization values. We also employ
an alternative prediction technique to increase the prediction accuracy by narrowing the range
of prediction values. Besides, an efficient centralized error diversion technique is proposed to
further decrease the bit-rate. The experimental results show that the proposed method offers
8 % higher payload with 5 % lower bit-rate on average if compared to Sun et al.’s method and
has better embedding performance than prior related works.

Keywords Reversible data hiding . AMBTC . Prediction . Low computing cost

* Yuan-Bo Ma
[email protected]

1
School of Computer and Software, Nanjing University of Information Science and Technology,
Nanjing, China
2
Department of Electronic Communication and Software Engineering, Nanfang College of Sun
Yat-Sen University, Guangzhou, China
3
Department of Business Administration, Nanfang College of Sun Yat-Sen University, Guangzhou,
China
4
Department of Computer Science and Information Engineering, National Taichung University of
Science and Technology, Taichung, Taiwan
 Multimed Tools Appl

1 Introduction

Information security is a relatively new research area and has been widely investigated in multiple
fields such as cloud service [6, 31], networking [9, 23], and multimedia applications [17, 29, 30].
As one of the emerging research area, data hiding in social multimedia is in a constant progress
toward a new era. Nowadays, data hiding technology is able to embed data into the vast majority of
digital media, such as images, audios and videos. Among them, images are widely used as cover
images since they are easily available. After embedding, the image containing secret data is called
the stego image. According to whether the original image can be reconstructed, data hiding
methods are classified into lossy [10, 11, 15, 20, 22] and reversible [13, 19, 21, 24, 27]. For lossy
methods, the cover image is permanently destroyed after embedding so that the original image is
unable to recover. However, in some special applications such as medical and military usages,
people need to restore the cover image while hiding the secret data. Therefore, the reversible data
hiding (RDH) methods have been widely investigated and fast developed.
RDH methods can be applied to images in spatial or compressed domain. Spatial domain
RDH methods directly modify image pixels in the spatial domain to insert data. Methods of
this type are easily implemented and offer a relatively high payload and quality. Therefore,
researchers have widely paid attention to investigating RDH methods of this type. In 2003,
Tian [27] proposed an efficient RDH method based on difference expansion for digital images.
Later in 2006, Ni et al. [21] advanced a histogram based RDH method. Data is embedded into
peak points by shifting the pixels of histogram. Many other RDH methods [13, 19, 24] have
been proposed to extend Ni et al.’s and Tian’s works and have better embedding performance.
On the other hand, compressed domain RDH methods embed data into an image’s compressed
format. As the compressed image is the most common form in mobile social networks,
investigating RDH methods for compressed image have become an important issue. As a
result, the RDH methods for compressed images, such as joint photographic experts group
(JPEG) [1], vector quantization (VQ) [2], and block truncation coding (BTC) [3, 12, 18, 25,
33], are proposed recently to fulfill the requirements in this field. The RDH methods for JPEG
compressed images embed data by modifying the coefficients in the frequency domain, while
RDH methods for VQ compressed images embed data by modifying the index table or
rearranging the corresponding codebook. RDH methods for compressed images of these two
types often require significant amount of computing cost.
Similar to JPEG and VQ compressed images, BTC is also a block-based compression
i ; qi ; Bi ), where qi
H
method of which the compressed code of block i is represented by a trio (qH L

and qLi are a pair of quantization values, and Bi is a bitmap. This method was first presented by
Delp and Mitchell [5] in 1979 and later Lema and Mitchell [16] proposed the absolute mean
block truncation coding (AMBTC) method to improve the performance of BTC. Unlike VQ
methods using machine learning techniques [4, 7, 8, 28, 32, 34] to obtain the codebook,
AMBTC requires insignificant computing cost and can be easily implemented. Therefore, it is
suitable for low bandwidth communication channels and low power image processing appli-
cations, such as personal digital assistant (PDA), field-programmable gate array (FPGA) and
portable image signal processor. Due to the vast demands for embedding data into low
computing cost compressed images, a number of lossy and reversible data hiding methods
for AMBTC codes have been proposed.
For lossy data hiding method, AMBTC-compressed code is usually classified as smooth or
complex blocks by a predefined threshold. In methods proposed by Ou et al. [22] in 2015,
more secret data bits are embedded into smooth blocks and two quantization levels are
Multimed Tools Appl

recalculated to minimize the distortion. In 2016, Huang et al. [15] proposed a hybrid method to
embed more secrets into the complex blocks with large variation of the blocks. At the same
year, Malik et al. [20] modified the AMBTC technique by providing two bit plane with four
quantization levels to improve image quality and payload. Although the aforementioned
methods achieve good visual quality and payload, they cause permanent distortion to the
original AMBTC codes. To fulfill the requirement of reversibility of the AMBTC codes, some
RDH methods for AMBTC are proposed [3, 12, 18, 25, 33] for diverse applications.
According to the format of AMBTC stego codes, RDH methods for AMBTC are grouped
into two types. In type I method, the AMBTC stego codes that mimic the traditional AMBTC
codes are created as output. That is, the stego codes can be decoded via a standard AMBTC
decoder and transformed into a stego image directly. Chen et al. [3] proposed a simple but
efficient RDH for AMBTC codes of this type. Their method embeds data bits by exchanging
the quantization value qH L
i and qi to achieve the reversibility of data hiding. Inspired by Chen
et al.’s method, some advanced methods with enhanced capacity and image quality are
proposed [12, 18]. In 2015, Lin et al. [18] proposed a type I method which secret data is
embedded into selected AMBTC block using different combinations of the mean value and the
standard deviation. Although methods of this type create formal AMBTC stego codes, they
suffer from low embedding capacity and high bit-rate.
The type II RDH methods for AMBTC codes represent the AMBTC stego codes with
specific coding structures [25, 33]. Besides bitmap Bi, additional information such as
encrypted data bits, division information, and prediction errors are added to a binary code
stream to form the stego codes. Methods of this type cannot be directly decoded by a standard
AMBTC decoder but only with the decoders designed for the corresponding methods.
Nevertheless, with the help of the special structure of stego codes, type II methods often offer
higher payload and smaller stego codes than those of type I methods. This means that more
secret data can be embedded while saving the storage space.
In 2013, Sun et al. [25] proposed a type II RDH for AMBTC method which combines the
joint neighboring coding (JNC) technique for better embedding performance. JNC was firstly
proposed by Chang et al. [2] to improve the embedding capacity. In [2], a key-selected
neighbor is chosen to predict each to-be-encoded element, and the error between the selected
neighbor and the to-be-encoded element is calculated. Once the error is obtained, it is
represented by different numbers of bits according to a pre-defined classification rule. Since
the to-be-encoded element is highly correlated to its neighbors, the prediction error tends to be
small. As a result, the codes can be represented with fewer bits and they thus provide extra
rooms for data insertion. Sun et al. exploit the JNC technique to propose a RDH method for
AMBTC codes. In their method, secret data are embedded into pairs of quantization values. To
embed data, the key-selected neighbor of the to-be-encoded quantization value is chosen to
predict this quantization value, and later the prediction error is calculated. The prediction error
is then encoded based on a pre-defined error division. Finally, the encrypted data bits, division
information and encoded prediction error are concatenated to generate final compressed stego
codes.
In type II method, a stego code of low bit-rate is crucial. The bit-rate can be reduced either
by using a better predictor to reduce the prediction errors, or by employing a better classifi-
cation rule to efficiently encode the prediction errors. However, the JNC technique adopted in
Sun et al.’s method unnecessarily increases the bit-rate because the selection of prediction
value is equivalent to a random selection. A random selected prediction value often causes
larger prediction errors, and subsequently increases the bit-rate. Moreover, the prediction error
 Multimed Tools Appl

classification rule used in Sun et al.’s method may cause more bits to encode because they use
m bits to record the prediction errors falling in the range [−(2m − 1), 2m − 1]. However, the most
frequently occurred prediction error is zero, and thus a prediction error of zero should be
represented by using fewer bits than other less occurred prediction errors.
To address the aforementioned problems, this paper proposes three techniques to perform
Type II RDH for AMBTC codes to improve Sun et al.’s method. The main contribution of this
paper is to effectively reduce the size of the stego code, as summarized as follows:

1) We adopt the median edge detection (MED) predictor instead of JNC to predict the
quantization values. Since the neighboring quantization values are highly correlated and
MED always chooses either the best or the second-best prediction values from three
neighboring candidates, MED provides a better prediction result than that of JNC used in
Sun et al.’s method.
2) We observe that prediction value of the higher quantization value cannot be smaller than its
lower counterpart and vice versa. Therefore, we propose an alternative prediction (AP) technique
to narrow the range of prediction values and thus the prediction errors can be further reduced.
3) Because zero is the most frequent occurred prediction error, we further propose a
centralization error division (CED) technique to modify the classification rule to catego-
rize zero prediction errors into a new division and require no bits to record.

The rest of this paper is organized as follows. Section 2 introduces the AMBTC coding
method, Sun et al.’s method, and MED predictor. Section 3 illustrates the proposed method and
discusses the basis of how it works. In Section 4, we conduct several experiments to validate the
applicability of the proposed method. The concluding remarks are given in Section 5.

2 Related work

In this section, AMBTC coding method is briefly introduced, followed by Sun et al.’s RDH
method for AMBTC codes and MED predictor.

2.1 AMBTC coding method

AMBTC is a block-based lossy compression technique. The concept of AMBTC is to preserve

the mean and first absolute central moment of each block in images. In the encoding
procedure, an image I is partitioned into a set of non-overlapping blocks {Φi}Ni = 1 of size
v × v, where N is the total number of blocks. Let {Φi,j}vj =×1v be the j-th pixel in block Φi, and Φi
be the block mean. Two quantization values, i.e., higher quantization value qH i and lower
quantization value qLi of block Φi, can be calculated as
1 X
i ¼
qH Φi; j ; ð1Þ
yi
Φi; j ≥ Φi

1 X
qLi ¼ Φi; j ; ð2Þ
v  v−yi
Φi; j < Φi
Multimed Tools Appl

where yi denotes the number of pixels having a value larger than or equal to Φi . The
compressed code for block Φi is then represented by the trio (qH i ; qi ; Bi ), where Bi = {Bi,j}j = 1
L v×
v
∈ (0, 1) is the bitmap for recording the quantized result. The j-th bit in Bi can be obtained by
using the following equation:
(
Bi; j ¼ 1 Φi; j ≥Φi ; ð3Þ
0 otherwise:

Each block is processed in the same manner, and the AMBTC compressed codes
 N  N
qi ; qLi ; Bi i¼1 are obtained. To decode compressed codes qH
H
i ; q i ; Bi
L
i¼1
, each trio
(qH
i ; q L
i ; B i ) is investigated. Let ~
Φ i be the de-compressed block of the i-th trio, and Φ~ i; j be
~ ~ ~
the j-th pixel in Φi . If Bi,j = 0, we set Φi; j ¼ qi . On the other hand, if Bi,j = 1, we set Φi; j ¼ qH
L
i .
 N
After decoding all the compressed codes, the de-compressed image I ¼ Φi i¼1 can be ~ ~
obtained.

2.2 Sun et al.’s method

In 2013, Sun et al. [25] proposed a type II RDH method for AMBTC codes using JNC. This
method selects a prediction value according to the to-be-embedded bits and obtains the
prediction errors for data embedment. To embed data, Sun et al.’s method classifies the
quantization values into referential and predictable groups. The quantization values in the first
row, the first column, and the last column are chosen as the referential quantization values for
recovering the predictable ones. Other quantization values are predictable, where two bits
secret data can be carried by each quantization value. Since the embedment in higher and
lower quantization values share the same procedures, we only take higher ones for illustration.
Let {qH L N H
i , qi , Bi}i = 1 be the AMBTC compressed codes and qi be the higher quantization
value to be encoded. One of the four neighboring quantization values of qH i is used to predict
the value of qHi . The locations of these four neighboring quantization values are represented by
binary indices (00)2, (01)2, (10)2 and (11)2, respectively, as shown in Fig. 1. To select which
neighbor should be used to predict qH i , two secret bits si,1si,2 are extracted from secret data S

Fig. 1 The Neighbors of

(01)2 (10)2 (11)2

(00)2 qiH
 Multimed Tools Appl

and are xor-ed with two key-generated random bits ri,1ri,2 to obtain the xor-ed results
0 0
si;1 si;2 ¼ si;1 si;2 ⊕ri;1 ri;2 , where ⊕ is the xor operator. The selected quantization value for
predicting qH H
i , denoted by pi , is the neighboring quantization value whose binary indices
′ ′
are si,1si,2.
Secret bits si,1si,2 are embedded by inserting s′i,1s′i,2 into the codes of qH H
i . To encode qi , the
prediction error ei = qi − pi is calculated. ei is then represented by 8-bit or m-bit binary codes
H H H H

according to the equations given below:

8 H 
>
> ei ; m 2 ; 0≤eH i ≤2 −1;
m
>  
< −eH ; m ; −ð2m −1Þ ≤eH < 0;
 H
ei 2 ¼  H i  2 i
ð4Þ
>
> ei ; 8 2 ; eH > 2m −1;
>
:   i
−eHi ; 8 2; i < −ð2 −1Þ;
eH m

where (eH H
i , x)2 denotes the x-bit binary code of ei . Eq. (4) partitions the range of ei into four
H

divisions and uses two bits di,1di,2 to record the division information. The assignment of the
two-bit division information is given in Fig. 2. The xor-ed secret bits s′i,1s′i,2, division
information di,1di,2, and the prediction error (eH i )2, are concatenated. Therefore, the final codes
′ ′
CHi = s s ||d d
i,1 i,2 i,1 i,2 ||(eH
)
i 2 for qH
i can be obtained, where || is the concatenation operator. The
block size and the length of secret data are used as the side information to extract the secret
data and recover the AMBTC code.
To extract the embedded secret bits and recover qH H
i from Ci , the key-generated random bits
ri,1ri,2 are obtained. Extract the first two bits si,1si,2 from Ci , and xor s′i,1s′i,2 with ri,1ri,2, the
′ ′ H

secret bits si,1si,2 are obtained. According to s′i,1s′i,2, the prediction value pH i that is used to
predict qH i can also be determined from one of the four neighboring quantization values of qH
i
which have been recovered previously. Extract the next two bits di,1di,2, we have the division
information and then extract (eH H H
i )2 from Ci . Convert (ei )2 into decimal form, qi can be
H
H H H
recovered by using the equation qi = ei + pi .
Here we use a simple example to illustrate the encoding and decoding procedures of Sun
et al.’s method. Let qH i
¼ 176 be the higher quantization value to be encoded, and m = 4.
Figure 3 shows the arrangement of qH i
and its four neighbors with pre-defined location (00)2,
(01)2, (10)2 and (11)2, respectively. Let the two key-generated random bits ri,1ri,2 be 012 and
two secret bits si,1si,2 be 102. The xor-ed bits can be calculated as s′i,1s′i,2 = 102 ⊕ 012 = 112.
Therefore, the location of prediction value pH i
is s′i,1s′i,2 = 112. According to the location defined
in Fig. 1, the upper-right neighboring quantization value located at (11)2 is selected as the
prediction value pH i
¼ 189. The prediction error can then be calculated by eH i = qi − pi = 176
H H

− 189 = − 13. Because − (2 − 1) ≤ ei < 0, the division information is di,1di,2 = 102. Using
m H
′ ′
Eq. (4), the 4-bit binary prediction error is (eH i )2 = 11012. Concatenate si,1si,2, di,1di,2, and
H H ′ ′ H
(ei )2, the final codes Ci = si,1si,2||di,1di,2||(ei )2 = 11 10 11012 are obtained.
To recover qH i
and extract the embedded bits from CH i = 111011012, the first two bits
′ ′
si,1si,2 = 112 are read. The embedded secret bits si,1si,2 = 102 can be obtained by taking the

002 102 112 012

0
Fig. 2 Assignment of the division information
Multimed Tools Appl

Secret bits

102 si ,1si,2 piH

192 191 189
si ,1si,2 CiH
112 177 176 111011012

qiH
012 ri ,1ri ,2

Key-generated
random bits
Fig. 3 An example of Sun et al.’s method

xor operation on s′i,1s′i,2 = 112 and the key-generated random bits ri,1ri,2 = 012. Since s′i,1s′i,2 =
112, the prediction value pH i = 189 can be located. By reading the next two bits di,1di,2 = 102
from CHi , the division information is known and the prediction error is recorded in the next four
bits. Therefore, (eH )
i 2 = 1101 2 and the prediction error is negative. The quantization value qH i

can then be recovered by qi = ei + pH i = − 13 + 189 = 176.

H H

2.3 Median Edge Detection (MED) predictor

The MED predictor [14] is an edge detection method which is initially used in image
compression and computer vision. The essence of the MED predictor is to detect the
most change in intensity by extracting the boundary between the visited pixel and its
neighbors in images. The MED predictor provides a more accurate result with insignif-
icant computing cost. As a result, the proposed method adopts the MED predictor to
predict quantization values. Let qi be the visited higher or lower quantization value. The
three neighboring quantization values of qi, denoted by qi,N, qi,NW, and qi,W, respectively,
are showed in Fig. 4.

Fig. 4 The context of the MED

predictor
 Multimed Tools Appl

The MED predictor predicts qi by using its neighboring quantization values to obtain the
prediction value pi, as given in the following equation:
8    
< min qi;N ; qi;W  ; qi;N W ≥max qi;N ; qi;W ;
 
pi ¼ max qi;N ; qi;W ; qi;N W ≤ min qi;N ; qi;W ; ð5Þ
:
qi;N þ qi;W −qi;N W ; otherwise:

3 Proposed method

In contrast to type I method, Sun et al.’s method belongs to type II and offers a significant
higher embedding capacity with lower bit-rate. However, there are some issues which can be
re-considered to improve the embedding performance. Firstly, the prediction value in Sun
et al.’s method is randomly selected from four neighboring quantization values. The random
selection comes out a relatively inaccurate prediction value, which directly leads to a larger
error and thus it requires more compressed codes. Secondly, the prediction error is encoded by
assigning different bits based on divisions shown in Fig. 2. Note that Sun et al.’s method uses
m bits to record the prediction errors within the range [−(2m − 1), 2m − 1]. However, a
prediction error of zero is the most frequently occurred value and is unnecessary to be
represented by m bits. A better division of errors should take this property into account in
order to minimize the bit-rate.
We propose a novel type II RDH method for AMBTC codes which advances new
techniques to improve Sun et al.’s method. First of all, in the proposed method, the
prediction value is chosen by median edge detection (MED) predictor instead of random
selection to achieve a better result. Secondly, we propose an alternative prediction (AP)
technique to further narrow the range of prediction errors. As for an AMBTC code
i ; qi ; Bi ), the equation qi ≥qi always holds. Therefore, the minimal prediction value
(qH L H L

pi of qi should be qi , and the maximal prediction value pLi of qLi should be qH

H H L
i . By

restricting the minimum value of pH i or the maximum value of pLi , the AP technique
successfully narrows the prediction errors so as to reduce the bit-rate. Thirdly, the
centralized error division (CED) technique is used to better classify the errors into
divisions. The CED technique classifies the prediction error of zero into a new division
and requires none of the bits to record in this division. Therefore, the bit-rate can be
further decreased.

3.1 Alternative prediction technique

The AMBTC codes for each block have a pair of quantization values with the
numerical order qH i ≥ qi . Therefore, the minimal prediction value pi
L H
of qH i can be
restricted to qi . As a result, if pi < qi , we may simply replace pi by an alternative
L H L H

prediction value ^ pH
i ¼ qLi to narrow the range of prediction errors. Similarly, if pLi > qHi ,

an alternative prediction value ^pi ¼ qi can be set. We term the replacement of

L H

prediction values by alternative values as the alternative prediction (AP) technique.

However, the prediction of qH i using the AP technique requires the information of qLi
and vice versa. Therefore, the AP technique can only be used to predict either qH i or qLi .
For the security consideration, we use a key to randomly select which quantization
Multimed Tools Appl

value should be predicted using the AP technique. If qH

i is selected to be alternatively
predicted, the prediction values of qH
i and qL
i can be recalculated by
8
< H  
^pi ¼ max qi L ; pH
i ; ð6Þ
: L
^pi ¼ pi L :

On the other hand, if qLi is selected to be predicted, we use the following equation to
recalculate the prediction values of qH
i and qLi :
8 H
< ^p ¼ pH ;
i
L
i   ð7Þ
: ^p ¼ min qH ; pL :
i i i;

Note that, when only the MED predictor is applied, the range of prediction error ei for a
given 8-bit quantization value is [−255, 255]. With the proposed AP technique, the range of
prediction errors can be reduced to |ei| ≤ qH
i − qi and thus the bit-rate can be further reduced
L

with only a small amount of calculation.

In the decoding procedure, the prediction values piH and piL are obtained using the MED
predictor. According to the given key, the decoder knows whether qH i or qLi is selected for
prediction at the encoding stage, and thus Eq. (6) or (7) can be applied to obtain the prediction
pi H and ^
value ^ pi L .

3.2 Centralized Error Division (CED) technique

In type II RDH method for AMBTC codes, the division arrangement of prediction errors
affects the bit-rate significantly. A well-designed error division should consider the occurrence
of individual prediction errors, and use as fewer bits as possible to represents the most occurred
error. Let ei ¼ qi −^pi be the prediction error between the to-be-encoded quantization value qi
and the prediction value ^pi obtained by the AP technique. In the proposed method, we divide
the range of prediction error ei into four divisions and use 0, m or 8 bits to encode ei. If ei = 0,
none of the bits are required to encode ei. If − (2m − 1) ≤ ei < 0 or 0 < ei ≤ 2m − 1, | ei | is encoded
with m bits. However, if |ei| > 2m − 1, since ei itself needs 9 bits to represent but qi only requires
8 bits, we thus encode qi using 8-bit binary representation of ei. The rules of encoding the
prediction error ei are given in the following equation:
8
> empty; ei ¼ 0;
>   
>
>
< ei ; m ; −ð2m −1Þ≤ei < 0;
ðei Þ2 ¼
ðei ; mÞ2 ;
2
0<
ð8Þ
 e i ≤2 −1;
m
>
>
>
>  
: ðqi ; 8Þ ; ei  > 2m −1:
2

To record the division information which is required in the decoding stage, two bits di,1di,2
are assigned to represent the divisions used in CED, as shown in Fig. 5.

102 012 002 112 102

0
Fig. 5 Assignment of division information di,1di,2 for CED
 Multimed Tools Appl

In the decoding procedure, 2-bit division information di,1di,2 is extracted and x-bit predic-
tion error (ei)2, where x = 0, m, or 8 can be determined. The original quantization value qi
can then be recovered by applying the following equation:
8
>
> ^pi ; d i;1 d i;2 ¼ 002 :
>  
>
< ^pi −ei ; d i;1 d i;2 ¼ 012 :
qi ¼   ð9Þ
>
>  
>
> ^p þ ei ; d i;1 d i;2 ¼ 112 :
: i
ei ; d i;1 d i;2 ¼ 102 :

3.3 Embedding procedures

To embed secret data S, image I is partitioned into N = NR × NC blocks of size v × v, where NR

and NC are the number of rows and columns of the partitioned blocks, respectively. Let
{qH L N
i , qi , Bi}i = 1 be the AMBTC compressed codes of I and C be the final AMBTC stego codes.
All the quantization values in the first column and the first row are pre-saved with eight bits as
referential quantization values, and they do not insert any data. The rest quantization values are
predictable and two secret bits can be inserted into each quantization value. The detailed
embedding procedures are listed as follows.
Input: AMBTC compressed codes {qH L N
i , qi , Bi}i = 1, secret data S, keys k1, k2 and a pre-
defined integer m.
Output: The AMBTC stego codes C.

Step 1 Scan the quantization values {qH N L N

i }i = 1 and {qi }i = 1 using the raster scan order. Use
eight bits each to encode referential quantization values. Let R be the encoded
results.
Step 2 Scan the predictable quantization values and use the MED predictor to predict qH i and
qLi . Use the key k1 to randomly select whether qH i or qL
i should be predicted using the
AP technique, as described in Section 3.1. Let ^ pH i and ^
pL
i be the predicted results.
Step 3 Generate four random bits ri,1ri,2ri,3ri,4 using the key k2, and read four secret bits
si,1si,2si,3si,4 from S. Perform the bitwise xor operation on ri,1ri,2ri,3ri,4 and si,1si,2si,3si,4
to obtain the xor-ed bits s′i,1s′i,2s′i,3s′i,4.
Step 4 Calculate the prediction errors eH i ¼ qi −^ i and ei ¼ qi −^
pH pLi . According to eH
H L L
i
L
and ei , the division information di,1di,2 and di,3di,4 is obtained, respectively.
The x-bit prediction errors (eH L
i )2 and (ei )2 can then be calculated, as described
in Section 3.2.
Step 5 Concatenate the xor-ed secret bits s′i,1s′i,2, division information di,1di,2, and prediction
′ ′
error (eH H H H
i )2, we obtain the codes Ci = si,1si,2||di,1di,2||(ei )2 for qi . Similarly, the codes
L ′ ′ L L
Ci = si,3si,4||di,3di,4||(ei )2 for qi can be obtained.
Step 6 Repeat Steps 2-5, until all the secret data are embedded and the quantization values
are processed. Concatenate R, the codes of bitmap {Bi}Ni = 1, and {CH N L N
i }i = 1, {Ci }i = 1,
the final AMBTC stego codes C are obtained.
To extract the embedded secret data and recover the original AMBTC com-
pressed code, we have to know NR, NC and v so that R and {Bi}Ni = 1 can be
accurately recovered. As a result, in addition to k1 and k2, the parameters NR, NC
and v are also taken as the keys and have to be transferred to the receiver side
via a secret channel.
Multimed Tools Appl

3.4 Decoding and extraction procedures

Once the receiver has the AMBTC stego codes C and the required keys, the embedded secret
data S can be extracted, and the original AMBTC codes {qH L N
i , qi , Bi}i = 1 can be covered. The
detailed procedures are listed as follows.
Inputs: AMBTC stego codes C, keys k1, k2, NR, NC, v, and integer m.
Outputs: AMBTC codes {qH L N
i , qi , Bi}i = 1 and secret data S.

Step 1 Read the first 2 × (NR × NC − 1) × 8 bits from C and reconstruct all the referential
quantization values R, which should be located on the first row and the first column
of the quantization array. Read the next {Bi}Ni = 1 × 16 bits to reconstruct the bitmap
{Bi}Ni = 1, and then read the remaining to obtain {CH N L N
i }i = 1 and {Ci }i = 1, respectively.
′ ′ H ′ ′ L
Step 2 Extract two bits si,1si,2 from Ci and two bits si,3si,4 from Ci . Use the key k2 to
generate four random bits ri,1ri,2ri,3ri,4. The embedded four secret bits si,1si,2si,3si,4 can
be obtained by performing the xor operation on s′i,1s′i,2s′i,3s′i,4 and ri,1ri,2ri,3ri,4.
Step 3 Calculate prediction values pH L
i and pi using the MED predictor. According to the key
k1, the alternative prediction values ^pH i and ^
pLi are obtained.
Step 4 Read the next two bits di,1di,2 from Ci . Based on di,1di,2, read next x bits from CH
H
i to obtain
(eH )
i 2 , where x could be 0, m, or 8. Convert (e H
)
i 2 into its decimal form to obtain |eH
i |.
L L
Similarly, read the next two bits di,3di,4 from Ci , and |ei | can then be obtained with similar
calculation.
Step 5 Use Eq. (9) to recover the original predictable quantization values qH L
i and qi with the
aids of di,1di,2 and di,3di,4.
Step 6 Repeat Steps 2-5 until all the secret data S are extracted and all the original
quantization values {qH N L N
i }i = 1 and {qi }i = 1 are recovered.
Once all the codes {qi }i = 1, {qi }i = 1, and {Bi}Ni = 1 are recovered, the AMBTC
H N L N

compressed image Ĩ can be constructed, as described in Section 2.1.

3.5 A simple example

In this sub-section, we use a pair of predictable quantization values as an example to illustrate the
proposed method. For simplicity, the processing of referential quantization values is not illustrat-
ed in this example. Let qH i
¼ 176 and qLi ¼ 172 be the to-be-encoded predictable quantization
values, as shown in Fig. 6. Suppose si,1si,2si,3si,4 = 10002, ri,1ri,2ri,3ri,4 = 01012 and m = 4. The four
xor-ed secret bits s′i,1s′i,2s′i,3s′i,4 can be calculated by si,1si,2si,3si,4 ⊕ ri,1ri,2ri,3ri,4 = 11012. According
to Eq. (5), the prediction values pH L
i = 176 and pi = 177 are calculated. Suppose the key k1 selects
L
the AP technique to be applied on qi , and thus we use Eq. (7) to obtain the final prediction values
 
^pHi ¼ pi ¼ 176 and ^
H
pLi ¼ min qH i ; pi;
L
¼ min(176, 177) = 176. The prediction errors eH i

¼ qHi −^
pHi ¼ 176 − 176 = 0 and ei ¼ qi −^
L L
pLi ¼ 172 − 176 = − 4 are obtained. To encode qH i ,
H
since prediction error is ei = 0, we have the division information di,1di,2 = 002. According to
Eq. (8), none of the bits are required to encode eH H H
i . The codes Ci = 11002 for qi can be obtained
′ ′ H L
by concatenating si,1si,2, di,1di,2 and (ei )2. To encode qi , the division information is di,3di,4 = 012
since − (2m − 1) ≤ eLi < 0. Therefore, |eLi | will be encoded by four bits (eLi )2 = 01002. Finally, the
code for qLi is CLi = s′i,3s′i,4||di,3di,4||(eLi )2 = 0101 01002. The detailed encoding procedures are
shown in Fig. 6.
 Multimed Tools Appl

192 191 qiH

176 176
H

MED predictor AP technique

75 182

L 70 172

qiL

Fig. 6 The encoding example of the proposed method

To extract the embedded secret data and recover qH i

and qLi , read the first two bits
s′i,1s′i2 = 112 from Ci and two bits si,3si,4 = 012 from Ci . The bits s′i,1s′i,2s′i,3s′i,4 are xor-ed with
H ′ ′ L

ri,1ri,2ri,3ri,4 = 01012, and the secret data si,1si,2si,3si,4 = 10002 are extracted. The prediction
values pH L
i = 176 and pi = 177 can then be calculated by using the MED predictor. The
alternative prediction values ^pH i ¼ 176 and ^ pLi ¼ 176 are recalculated according to Eq. (7)
under the guidance of the key k1. By reading the next two bits di,1di,2 = 002 from CH i , we know
the prediction error is 0. Therefore, according to Eq. (9), qH i ¼ ^
pH
i ¼ 176. Read the next two
bits di,3di,4 = 012 from CLi , we know the next four bits are prediction error (eLi )2 = 01002, and its
decimal value is four. Using Eq. (9), the quantization value qLi = 176 − 4 = 172 can be
recovered. The detailed decoding procedures are shown in Fig. 7.

192 191 192 191

H
176 176 176
6

si ,1si ,2 si ,3si ,4 qiH

MED predictor AP technique

75 182 75 182

70 70 172
L

qiL
Fig. 7 The decoding example of the proposed method
Multimed Tools Appl

4 Experimental results

In this section, we perform several experiments to demonstrate the performance of the

proposed method and compare the results with other related works. Six standard 8-bit
grayscale images of size 512 × 512, including Lena, Jet, Peppers, Baboon, Tiffany, and House,
are used as cover images, as shown in Fig. 8. All images are obtained from the SIPI image
database [26]. In the experiment, a block size of 4 × 4 is used to generate the AMBTC codes.
In the experiment, the peak signal-to-noise ratio (PSNR) is used to measure the image
quality of the AMBTC compressed image. In general, the higher the PSNR, the closer the
image is to its original. We also compare the embedding performance of the proposed method,
including payload, bit-rate and embedding efficiency (EF) with other related methods. The
payload is used to evaluate the total bits embedded into the AMBTC codes, while the bit-rate is
applied to measure the number of bits required to record a pixel (bpp) of a compressed image.
A low bit-rate compressed image requires less storage space, which is an important issue
regarding applications such as memory limited storage devices or streaming data on the
internet. The embedding efficiency, as its name implies, measures the performance of data
embedment and is defined by
Payload
EF ¼ ð10Þ
Size of AMBTC stego codes
An embedding method with higher embedding efficiency indicates that the method offers
larger payload under the same size of stego codes.

(a) Lena (b) Jet (c) Peppers

(d) Baboon (e) Tiffany (f) House

Fig. 8 Six test images used in the experiment
 Multimed Tools Appl

4.1 Performance of the proposed method

To see how the integer parameter m affects the embedding performance of the proposed method,
we plot a graph of bit-rate (in bpp) versus m for the six test images in Fig. 9. Since m = 8 is the worst
condition which cannot reflect any improvements, we only plot the range 1 ≤ m ≤ 7. In the
experiments, all the images are fully embedded with the same size of secret data (64,316 bits).
Figure 9 reveals that for all the images, m = 3 or 4 has the smallest bit-rate. This implies
that proper assignment of the parameter m results in a smaller size of compressed codes. If m is
too small, more prediction errors will be represented with eight bits and thus the bit-rate is
increased. If m is too big, a considerable number of small prediction errors are forced to be
encoded by m bits unnecessarily, and therefore the bit-rate is also increased. It should be noted
that the bit-rate for a standard AMBTC compressed image of a block sized 4 × 4 is two bpp.
However, the proposed method for smoother images such as Jet, Lena, Peppers and Tiffany
offers a bit-rate which is lower than two bpp, even additional two bits are inserted into each
quantization value (which is equivalent to the embedment of four bits into each AMBTC
block). The results indicate that the proposed method not only embeds data into the AMBTC
codes but offers a more comparable bit-rate than that of the original AMBTC codes. Besides,
the bit-rate of the Baboon image is the largest due to the inaccurate prediction because the
Baboon image contains the richest texture.

4.2 Bit-rate comparisons of the proposed and Sun et al.’s methods

To examine how the proposed three techniques (MED, AP, and CED) affect the bit-rate, we
apply these techniques to Sun et al.’s method [25] to see how each technique influences the
performance individually. When these three techniques are all applied together, we name the
proposed method MAC, which is the acronym of MED, AP, and CED. The results are shown
in Fig. 10a–f, respectively. In the legend, B[25] + MED^ denotes Sun et al.’s prediction method

Fig. 9 The bit-rate comparison for six test images

Multimed Tools Appl

(a) Lena (b) Jet

(c) Peppers (d) Baboon

(e) Tiffany (f) House

Fig. 10 Bit-rate comparisons using different techniques

is replaced by the proposed MED technique, while B[25] + AP^ implies the AP technique is
added with Sun et al.’s prediction method. B[25] + CED^ represents Sun et al.’s error division
technique which is replaced by the proposed CED technique.
 Multimed Tools Appl

Figure 10a–f show the contribution of each technique to the reduction of bit-rate when
they are applied to Sun et al.’s method independently. We take the Lena image as an
example. In Sun et al.’s method, the bit-rate (2.10 bpp) is the lowest when m = 3. When
the AP technique is applied to Sun et al.’s method, the bit-rate is slightly reduced to 2.09
bpp. This is because some of the larger prediction errors are confined to smaller ones and
thus fewer bits are required to record these errors. When the division technique used in
Sun et al.’s method is replaced by the CED technique, bit-rate is further reduced to 2.07
bpp. The result implies that the proposed CED technique efficiently classifies the errors
into divisions and more prediction errors are recorded using fewer bits. When the MED
predictor is employed instead of Sun et al.’s prediction method, the bit-rate is reduced to
2.03 bpp. The results show that the MED predictor improves the prediction accuracy to a
large extent. Finally, when all the MED, AP, and CED techniques are applied, the bit-rate
is significantly reduced from 2.10 bpp to 1.98 bpp.
It should be noticed that though the above analyses focus on m = 3, the proposed MAC
method with other m values also performs better than Sun et al.’s method. Moreover, the tests
on other images also show similar results. The results indicate that the proposed MAC method
indeed successfully reduces the bit-rate while providing a slightly higher payload than that of
Sun et al.’s method, as will be discussed in the next sub-section.

4.3 Performance comparison with related works

In this section, we compare the proposed method with two related type II works, including
Zhang et al.’s [33] and Sun et al.’s methods [25]. We also intentionally compare a type I
method - Chen et al.’s method [3] for reference. In Chen et al.’s method, the embedment of
additional secret bits in bitmap is implemented to achieve the highest embedding capacity, as
described in [3]. In Zhang et al.’s method, four neighboring quantization values are employed
as an embedding unit to achieve the best embedding result. In Sun et al.’s and the proposed
MAC methods, the parameter m is properly selected such that the bit-rate is the smallest. We
compare the four benchmarks, including PSNR, payload, bit-rate, and embedding efficiency
EF for each method. The results are shown in Table 1.
Table 1 shows that the PSNR between the original image and AMBTC compressed
image varies from one image to another. In general, a complex image has a lower PSNR
while a smoother one is higher. However, for each test image, all the RDH methods for
the AMBTC codes offer the same PSNR as the AMBTC compressed image since the
original AMBCT codes can be completely recovered. Chen et al.’s method has signifi-
cantly lower payload and bit-rate since this method belongs to type I. Zhang et al.’s
method provides lower bit-rate than other methods, but the payload of the MAC and Sun
et al.’s methods is two times more than that of Zhang et al.’s method. The increased
payload is attributed to the fact that the MAC and Sun et al.’s methods carry 2-bit data in
each quantization value while Zhang et al.’s method carries only one bit. Apparently, in
contrast to Sun et al.’s method, the MAC method provides better embedding performance
because the bit-rate is successfully reduced while keeping the payload slightly higher. In
addition, the MAC method has a great improvement in EF. Since Chen et al.’s method
only embeds one-bit secret data into a pair of quantization values, the EF is the lowest.
When comparing with other two type II methods, the MAC method also has remarkable
improvements. The EF is approximately two times higher than that of Zhang et al.’s
method, and is around 8 % higher on average than that of Sun et al.’s method.
Multimed Tools Appl

Table 1 Performance comparisons of the proposed method and other related works

Performance Methods Lena Jet Peppers Baboon Tiffany House

PSNR (dB) AMBTC 33.2 31.95 33.39 26.91 30.36 30.87

Chen et al. 33.2 31.95 33.39 26.91 30.36 30.87
Zhang et al. 33.2 31.95 33.39 26.91 30.36 30.87
Sun et al. 33.2 31.95 33.39 26.91 30.36 30.87
MAC 33.2 31.95 33.39 26.90 30.35 30.87
Bit-rate (bpp) AMBTC 2.00 2.00 2.00 2.00 2.00 2.00
Chen et al. 2.00 2.00 2.00 2.00 2.00 2.00
Zhang et al. 1.91 1.89 1.91 2.03 1.89 1.90
Sun et al. 2.10 2.10 2.10 2.18 2.07 2.14
MAC 2.00 1.99 2.00 2.13 1.97 2.00
Payload (bits) Chen et al. 16,384 16,384 16,384 16,384 16,384 16,384
Zhang et al. 32,768 32,768 32,768 32,768 32,768 32,768
Sun et al. 64,008 64,008 64,008 64,008 64,008 64,008
MAC 64,516 64,516 64,516 64,516 64,516 64,516
EF Chen et al. 0.0417 0.0417 0.0417 0.0417 0.0417 0.0417
Zhang et al. 0.065 0.066 0.065 0.064 0.067 0.065
Sun et al. 0.116 0.116 0.116 0.112 0.118 0.114
MAC 0.123 0.124 0.123 0.116 0.125 0.123

5 Conclusions

In this paper, we propose a type II RDH method for the AMBTC compressed code. In general,
the stego codes of type I methods can be decoded by a standard AMBTC decoder, but the
embedding capacity is low. Sun et al.’s type II method has good embedding performance;
however, some issues can be further reconsidered. We propose three techniques to improve
Sun et al.’s method by using better prediction techniques and more efficient error divisions.
The proposed MED and AP predictors better predict the quantization values and thus the
accuracy of prediction is significantly improved. The proposed CED technique divides
prediction errors into divisions more efficiently. Therefore, fewer bits are required to encode
the prediction errors. The experiment results show that the proposed MAC method offers a
comparable embedding capacity while providing a lower bit-rate and higher efficiency than
Sun et al.’s method and other related works.

References

1. Chang CC, Chen TS, Chung LZ (2002) A steganographic method based upon JPEG and quantization table
modification. Inform Sci 141(1-2):123–128
2. Chang CC, Kieu TD, Wu WC (2009) A lossless data embedding technique by joint neighboring coding.
Pattern Recogn 42(7):1597–1630
3. Chen J, Hong W, Chen TS, Shiu CW (2010) Steganography for BTC compressed images using no distortion
technique. Imag Sci J 58(4):177–185
 Multimed Tools Appl

4. Chen B, Shu H, Coatrieux G, Chen SX, Coatrieux JL (2015) Color image analysis by quaternion-type
moments. J Math Imag Vision 51(1):124–144
5. Delp EJ, Mitchell OR (1979) Image compression using block truncation coding. IEEE Trans Commun 27(9):
1335–1342
6. Fu Z, Sun X, Liu Q, Zhou L, Shu J (2015) Achieving efficient cloud search services: multi-keyword ranked
search over encrypted cloud data supporting parallel computing. IEICE Trans Commun E98-B(1):190–200
7. Gu B, Sheng VS, Tay KY, Romano W, Li S (2015) Incremental support vector learning for ordinal
regression. IEEE Trans Neural Netw Learn Syst 26(7):1403–1416
8. Gu B, Sheng VS, Wang Z, Ho D, Osman S, Li S (2015) Incremental learning for v-support vector regression.
Neural Netw 67:140–150
9. Guo P, Wang J, Xue HG, Chang SK, Kim JU (2014) A variable threshold-value authentication architecture
for wireless mesh networks. J Internet Technol 15(6):929–936
10. Hong W (2013) Adaptive image data hiding in edges using patched reference table and pair-wise embedding
technique. Inform Sci 221(1):473–489
11. Hong W, Chen TS (2012) A novel data embedding method using adaptive pixel pair matching. IEEE Trans
Inform Forensics Sec 7(1):176–184
12. Hong W, Chen TS, Shiu CW, Wu MC (2011) Lossless data embedding in BTC codes based on prediction
and histogram shifting. Appl Mech Mater 65:182–185
13. Hong W, Chen TS, Chen J (2015) Reversible data hiding using Delaunay triangulation and selective
embedment. Inf Sci 308(1):140–154
14. Hong W, Chen TS, Shiu CW (2009) Reversible data hiding for high quality images using modification of
prediction errors. J Syst Software 82:1833–1842
15. Huang YH, Chang CC, Chen YH (2016) Hybrid secret hiding schemes based on absolute moment block
truncation coding. Multimed Tools Appl. doi:10.1007/s11042-015-3208-y
16. Lema M, Mitchell OR (1984) Absolute moment block truncation coding and its application to color images.
IEEE Trans Commun 32(10):1148–1157
17. Li J, Li X, Yang B, Sun X (2015) Segmentation-based image copy-move forgery detection scheme. IEEE
Trans Inform Forensics Sec 10(3):507–518
18. Lin CC, Liu XL, Tai WL, Yuan SM (2015) A novel reversible data hiding scheme based on AMBTC
compression technique. Multimed Tools Appl 74(11):3823–3842
19. Lou DC, Hu CH (2012) LSB steganographic method based on reversible histogram transformation function
for resisting statistical steganalysis. Inf Sci 188(4):346–358
20. Malik A, Sikka G, Verma HK (2016) A high payload data hiding scheme based on modified AMBTC
technique. Doi: 10.1007/s11042-016-3815-2
21. Ni ZC, Shi YQ, Ansari N, Su W (2006) Reversible data hiding. IEEE Trans Circ Syst Video Technol 16(3):
354–362
22. Ou D, Sun W (2015) High payload image steganography with minimum distortion based on absolute
moment block truncation coding. Multimed Tools Appl 74(21):9117–9139
23. Shen J, Tan H, Wang J, Wang J, Lee S (2015) A novel routing protocol providing good transmission
reliability in underwater sensor networks. J Internet Technol 16(1):171–178
24. Shiu CW, Chen YC, Hong W (2015) Encrypted image-based reversible data hiding with public key
cryptography from difference expansion. Signal Process Image Commun 39:226–233
25. Sun W, Lu ZM, Wen YC, Yu FX, Shen RJ (2013) High performance reversible data hiding for block
truncation coding compressed images. SIViP 7(2):297–306
26. The USC-SIPI Image Database. http://sipi.usc.edu/database
27. Tian J (2003) Reversible data embedding using a difference expansion. IEEE Trans Circ Syst Video Technol
13(8):890–896
28. Wen X, Shao L, Xue Y, Fang W (2015) A rapid learning algorithm for vehicle classification. Inf Sci 295(1):
395–406
29. Xia Z, Wang X, Sun X, Liu Q, Xiong N (2014) Steganalysis of LSB matching using differences between
nonadjacent pixels. Multimed Tools Appl. doi:10.1007/s11042-014-2381-8
30. Xia Z, Wang X, Sun X, Wang B (2013) Steganalysis of least significant bit matching using multi-order
differences. Security Commun Netw 7(8):1283–1291
31. Xia Z, Wang X, Sun X, Wang Q (2016) A secure and dynamic multi-keyword ranked search scheme over
encrypted cloud data. IEEE Trans Parallel Distrib Syst 27(2):340–352
Multimed Tools Appl

32. Yuan C, Sun X, Rui L (2016) Fingerprint liveness detection based on multi-scale LPQ and PCA. Chin
Commun 13(7):60–65
33. Zhang Y, Guo SZ, Lu ZM, Luo H (2013) Reversible data hiding for BTC-compressed images based on
lossless coding of mean tables. IEICE Trans Commun E96.B(2):624–631
34. Zheng Y, Jeon B, Xu D, Wu QMJ, Zhang H (2015) Image segmentation by generalized hierarchical fuzzy C-
means algorithm. J Intell Fuzzy Syst 28(2):961–973

Wien Hong received his M.S. and Ph.D. degree from the State University of New York at Buffalo, USA in 1994
and 1997, respectively. He is currently a professor in the Department of Electronic Communication and Software
Engineering at Nanfang College of Sun Yat-Sen University. He is also a distinguished researcher at Nanjing
University of Information Science and Technology. His research interests include steganography, watermarking
and image compression.

Yuan-Bo Ma received her M.S. degree in IC Design from Hong Kong University of Science and Technology,
Hong Kong in 2011, and B.S. degree in Electronic Engineering from Sun Yat-Sen University at Guangdong,
China in 2010. She is currently a full-time teacher in the Department of Electronic Communication and Software
Engineering at Nanfang College of Sun Yat-Sen University, Guangdong, China. Her research interests include
information security, and digital signal processing.
 Multimed Tools Appl

Hung-Che Wu received his B.A. from Western Michigan University, USA, in 1999, M.S. from the University of
Wisconsin-Stout, USA, in 2001, and Ph.D. from Lincoln University, New Zealand, in 2010. He is currently an
associate professor in the Department of Business Administration at Nanfang College of Sun Yat-Sen University,
China. His research interests involve hospitality security and travel database technology.

Tung-Shou Chen received his B.S. and Ph.D. degrees from National Chiao Tung University in 1986 and 1992,
respectively, both in Computer Science and Information Engineering. Currently, he is a professor and head of the
Department of Computer Science and Information Engineering at National Taichung University of Science and
Technology, Taichung, Taiwan. His current research interests include data mining, image cryptosystems,
steganography, image compression and bioinformatics.