ArticlePDF Available

Abstract and Figures

In this paper, a wavelet-based scheme for digital image watermarking is presented. This proposed scheme is blind, which means that it requires neither the original image nor any side information in watermark recovery. It is based on inserting the watermark bits into the coarsest scale wavelet coefficients. Three-level wavelet decomposition and a watermark equal in size to the detail sub-bands in the coarsest scale are used. Only, perceptually significant wavelet coefficients are used to embed the watermark bits. The proposed scheme differs from the traditional wavelet-based schemes in the use of quantization and non-additive watermark embedding. It produces watermarked images with less degradation than the traditional wavelet-based schemes.
Content may be subject to copyright.
International Journal of Signal Processing, Image Processing and Patter
International Journal of Signal Processing, Image Processing and PatterInternational Journal of Signal Processing, Image Processing and Patter
International Journal of Signal Processing, Image Processing and Pattern Recognition
n Recognitionn Recognition
n Recognition
Vol. 4, No. 1,
Vol. 4, No. 1, Vol. 4, No. 1,
Vol. 4, No. 1, March 2011
March 2011March 2011
March 2011
15
Blind Wavelet-Based Image Watermarking
Hanaa A. Abdallah*, Mohiy M. Hadhoud
#
,
Abdalhameed A. Shaalan
*
and Fathi E. Abd El-samie**
*Faculty of Engineering, Zagazig university, Zagazig, Egypt.
#
Faculty of Computers and Information, Menoufia University, Shebin Elkom, Egypt.
**Department of Electronics and Electrical Communications,
Faculty of Electronic Engineering
Menoufia University, Menouf, 32952, Egypt.
E-mails: flower002a@yahoo.com, mmhadhoud@yahoo.com,
dr_shaalan2005@yahoo.com, fathi_sayed@yahoo.com
Abstract
In this paper, a wavelet-based scheme for digital image watermarking is presented.
This proposed scheme is blind, which means that it requires neither the original image nor
any side information in watermark recovery. It is based on inserting the watermark bits into
the coarsest scale wavelet coefficients. Three-level wavelet decomposition and a watermark
equal in size to the detail sub-bands in the coarsest scale are used. Only, perceptually
significant wavelet coefficients are used to embed the watermark bits. The proposed scheme
differs from the traditional wavelet-based schemes in the use of quantization and non-additive
watermark embedding. It produces watermarked images with less degradation than the
traditional wavelet-based schemes.
Keywords: Image watermarking, Wavelet transform, quantization.
1. Introduction
Digital image watermarking has attracted the attention of several researchers in the
last decades. The motivation behind the work in this area is the desire to achieve information
security, information hiding, authentication, and fingerprinting. Several approaches have been
proposed for digital image watermarking. One of such approaches is the discrete wavelet
transform (DWT) approach. The DWT finds a great popularity in the field of watermarking as
it is able to decompose the available images into sub-bands , in which watermarks can be
embedded, selectively [1,2].
Taking the cue from spread spectrum communication, binary watermark data can be
embedded in the wavelet coefficients chosen in a random order. For extraction of the hidden
data, the random sequence must be made available to the extractor. Cox et al. were the first to
apply the spread spectrum technique to data hiding [3]. Were the first to apply the spread
spectrum technique to data hiding Transform domain used DCT and DWT has been used in
[4]. Use of DWT has advantages of speed and robustness against wavelet based compression
[5].
Dugad et al. presented a blind additive watermarking scheme operating in the wavelet
domain [1]. A three-levels wavelet decomposition with Daubechies 8-tap filters was used. No
watermark was inserted into the low-pass sub-band. Unlike some non-blind watermarking
schemes [6,7], this scheme allows a watermark to be detected without access to the original
International Journal of Signal Processing, Image Processing and Pattern Recognition
International Journal of Signal Processing, Image Processing and Pattern RecognitionInternational Journal of Signal Processing, Image Processing and Pattern Recognition
International Journal of Signal Processing, Image Processing and Pattern Recognition
Vol. 4, No. 1, March 2011
Vol. 4, No. 1, March 2011Vol. 4, No. 1, March 2011
Vol. 4, No. 1, March 2011
16
image.It performs an implicit visual masking
as only wavelet coefficients with large magnitude are selected for watermark insertion. These
coefficients correspond to regions of texture and edges in an image. This scheme makes it
difficult for a human viewer to perceive any degradation in the watermarked image. Also,
because wavelet coefficients of large magnitude are perceptually significant, it is difficult to
remove the watermark without severely distorting the watermarked image. The most novel
aspect of this scheme was the introduction of a watermark consisting of pseudorandom real
numbers. Since watermark detection typically consists of a process of correlation estimation,
in which the watermark coefficients are placed in the image, changes in the location of the
watermarked coefficients are unacceptable. The watermarking scheme proposed by Dugad et
al. is based on adding the watermark in selected coefficients with significant energy in the
transform domain in order to ensure the non-erasability of the watermark. This scheme has
overcome the problem of “order sensitivity”.
Unfortunately, this scheme has also some disadvantages. It embeds the watermark in
an additive fashion. It is known that blind detectors for additive watermarking schemes must
correlate the possibly watermarked image coefficients with the known watermark in order to
determine if the image has or has not been marked. Thus, the image itself must be treated as
noise, which makes the detection of the watermark exceedingly difficult [8]. In order to
overcome this problem, it is necessary to correlate a very large number of coefficients, which
in turn requires the watermark to be embedded into several image coefficients at the insertion
stage. As a result, the degradation in the watermarked image increases. Another drawback is
that the detector can only tell if the watermark is present or not. It cannot recover the actual
watermark.
The scheme in [9] is another example of wavelet-based watermarking schemes. A
noise-like Gaussian sequence is used as a watermark. To embed the watermark robustly and
imperceptibly, watermark components are added to the significant coefficients of each
selected sub-band by considering the human visual system (HVS) characteristics. Some small
modifications are performed to improve the HVS model. The host image is needed in the
watermark extraction procedure.
In this paper, we present a new scheme to avoid these drawbacks. It is possible to use
the advantages of the watermarking scheme presented by Dugad, whilst avoiding its
disadvantages. This can be accomplished by using a binary watermark equal in size to the
detail sub-bands in the coarsest wavelet scale in conjunction with an adapted version of the
scalar quantization insertion/detection technique. The proposed watermarking scheme is
blind. Only, perceptually significant coefficients are used to embed the watermark bits. The
proposed scheme is expected to produce watermarked images with less degradation than the
Dugad’s scheme.
This paper is organized as follows. Sections 2 and 3 introduce two traditional
wavelet-based watermarking schemes. Section 4 introduces the proposed watermarking
scheme. Section 5 introduces the perceptual quality metrics that will be used for the
assessment of watermarking schemes. Section 6 introduces the experimental results. Finally,
section 7 gives the concluding remarks.
International Journal of Signal Processing, Image Processing and Patter
International Journal of Signal Processing, Image Processing and PatterInternational Journal of Signal Processing, Image Processing and Patter
International Journal of Signal Processing, Image Processing and Pattern Recognition
n Recognitionn Recognition
n Recognition
Vol. 4, No. 1,
Vol. 4, No. 1, Vol. 4, No. 1,
Vol. 4, No. 1, March 2011
March 2011March 2011
March 2011
17
2. Dugad’s scheme
Dugad et al. presented an additive watermarking scheme operating in the wavelet
domain [1]. The steps of watermark embedding and detection in this scheme are summarized
in the following subsections
.
2.1. Embedding algorithm
The steps of watermark embedding in Dugad’s scheme can be summarized as follows:
1. Wavelet decomposition is performed on the original image. After this decomposition,
we get four components; the approximation (LL
1
) component, the horizontal details (HL
1
)
component, the vertical details (LH
1
) component, and the diagonal details (HH
1
)
component.
2. A random watermark matrix of zero mean and unit variance, which is equal in size to
the detail components of the input image, is generated with a known seed value.
3. All wavelet coefficients in the HL
1
and LH
1
components with magnitude greater than
a threshold t
1
are selected. This ensures that only perceptually significant coefficients are
used.
4. The watermarking is performed on the wavelet coefficients selected in step 3 as
follows[1]:
ˆ
ij ij ij ij
w w w x
α
= +
(1)
where w
ij
is a selected wavelet coefficient at indices (i,j),
α
is a scaling parameter, x
ij
is a
watermark value, and
ˆ
ij
w
is the watermarked wavelet coefficient.
2.2 Detection algorithm
1- The watermark is regenerated using the known seed value.
2- Wavelet decomposition is performed on the possibly corrupted watermarked image.
3- All wavelet coefficients, from all components barring the LL
1
, of magnitude greater than t
2
are selected. Note that by setting t
2
> t
1
, the robustness of the algorithm is increased, as the
magnitude of some wavelet coefficients, which were originally below t
1
, may become greater
than t
1
due to image manipulations.
4- The selected coefficients are correlated with the watermark values at the same locations.
After this correlation process, a yes or no answer will be given for the presence of the
watermark.
3. Miyazaki’s scheme
Two watermarking schemes were presented by Miyazaki et al. in [2]. Both schemes
are implemented in the wavelet domain, but each targets a different set of coefficients for
insertion. The first scheme operates upon insignificant coefficients, whereas the second
scheme operates upon significant coefficients. Thus, both insertion schemes could be applied
to the same image at the same time. However, the reported results indicate that the insertion
technique utilizing the significant coefficients is more robust than the insertion technique
operating utilizing the insignificant coefficients. For this reason, only the insertion technique
utilizing the significant coefficients will be considered in this paper.
International Journal of Signal Processing, Image Processing and Pattern Recognition
International Journal of Signal Processing, Image Processing and Pattern RecognitionInternational Journal of Signal Processing, Image Processing and Pattern Recognition
International Journal of Signal Processing, Image Processing and Pattern Recognition
Vol. 4, No. 1, March 2011
Vol. 4, No. 1, March 2011Vol. 4, No. 1, March 2011
Vol. 4, No. 1, March 2011
18
This scheme depends on a three levels wavelet decomposition of the image to be
watermarked and inserts the watermark into the detail coefficients at the coarsest scale. It is a
quantization based scheme, which aims at modifying the wavelet coefficients of high
magnitude, and thus embedding the watermark into the edge and texture regions of the image.
It is a semi-blind scheme as it requires a file containing the locations, where the watermark
was embedded in order for the detector to work.
4. Proposed watermarking scheme
The proposed watermarking scheme is a blind quantization based scheme. A block
diagram detailing its steps is shown in Figure 1.
Figure 1. The proposed image watermarking scheme.
4.1 Watermark Embedding
The steps of watermark embedding can be summarized as follows:
1. The host image is transformed into the wavelet domain; three levels Daubechies wavelet
with filters of length 4 is used.
2. The coefficients in the third wavelet level (excluding the LL
3
and HH
3
sub-bands) with
magnitude greater than t
1
and less than t
2
are selected. Let
max
f
be the wavelet coefficient
DWT
(3 levels)
N
×
N
Input image
HL
3
LH
Owner seed
Binary watermark
01010001110………
….
Embedding via
quantization IDWT
Watermarked
image
HL
1
HH
1
LH
1
HH
2
HL
2
LH
2
HH
3
LL
3
HL
3
LH
3
International Journal of Signal Processing, Image Processing and Patter
International Journal of Signal Processing, Image Processing and PatterInternational Journal of Signal Processing, Image Processing and Patter
International Journal of Signal Processing, Image Processing and Pattern Recognition
n Recognitionn Recognition
n Recognition
Vol. 4, No. 1,
Vol. 4, No. 1, Vol. 4, No. 1,
Vol. 4, No. 1, March 2011
March 2011March 2011
March 2011
19
with maximum absolute in both HL
3
and LH
3
sub-bands. A threshold t= α.
max
f
is
selected, where
0.01 <α <0.1 and t
2
> t
1
>t. (2)
3. A binary watermark of the same size as the two sub-bands of interest is created using a
secret key, which is a seed of a random number generator.
4. Each
s
ij
w
of the selected wavelet coefficients is quantized. The quantization process can
be summarized as follows:
If
ij
x = 1 and
s
ij
w
> 0, then
s
ij
w
'
= t
2
– x
1
,
If
ij
x = 0 and
s
ij
w
> 0, then
s
ij
w
'
= t
1
+ x
1
,
If
ij
x= 1 and
s
ij
w
< 0, then
s
ij
w
'
= -t
2
+ x
1
,
If
ij
x= 0 and
s
ij
w
< 0, then
s
ij
w
'
= -t
1
- x
1
, (3)
where
ij
x is the watermark bit corresponding to
s
ij
w
, and
s
ij
w
'
is the watermarked wavelet
coefficient. The parameter x
1
narrows the range between the two quantization levels t
1
and t
2
in order to perform a robust oblivious detection. Figure (2) shows the watermark embedding
in a positive wavelet coefficient.
5. After all the selected coefficients are quantized, the inverse discrete wavelet transform
(IDWT) is applied and the watermarked image is obtained.
Figure 2. Watermark embedding for positive wavelet coefficients in the
proposed scheme.
4.2. Watermark Detection
1. The possibly corrupted watermarked image is transformed into the wavelet domain using
the same wavelet transform as in the embedding process.
Lowest value
of Wavelet
coefficients
Highest value
of Wavelet
coefficients
Out of
range Out of
range
Within range
t
1
t
2
x
1
x
1
t
1
+ x
1
t
2
- x
1
WM= WM=
s
ij
w
International Journal of Signal Processing, Image Processing and Pattern Recognition
International Journal of Signal Processing, Image Processing and Pattern RecognitionInternational Journal of Signal Processing, Image Processing and Pattern Recognition
International Journal of Signal Processing, Image Processing and Pattern Recognition
Vol. 4, No. 1, March 2011
Vol. 4, No. 1, March 2011Vol. 4, No. 1, March 2011
Vol. 4, No. 1, March 2011
20
2. The extraction is performed on the coefficients in the third wavelet level (excluding the
LL
3
and HH
3
sub-bands).
3. All the wavelet coefficients of magnitude greater than or equal to t
1
+ x
2
and less than or
equal to t
2
x
2
are selected; these are denoted
s
ij
w
'
. Note that x
2
should be less than x
1
.
This helps to ensure that all the marked coefficients are recovered and
dequantized after being attacked. Unmarked coefficients are unlikely to drift into the
range of selected coefficients after an attack. The introduction of the parameters x
1
and x
2
to the watermarking algorithm gives a degree of tolerance to the system against attacks,
i.e., they collaborate to give a noise margin. The watermark bits are extracted from each
of the selected wavelet coefficients with Eq. (4). Figure (3) illustrates the watermark
detection process.
Figure 3. Watermark detection in the proposed scheme.
If
s
ij
w
'
< (t
1
+ t
2
)/2, then the recovered watermark bit is a 0.
If
s
ij
w
'
≥ (t
1
+ t
2
)/2, then the recovered watermark bit is a 1 (4)
4. The recovered watermark is then correlated with the original watermark in the watermark
file, obtained via the secret key, only in the locations of the selected coefficients. This
allows a confidence measure to be ascertained for the presence or absence of a watermark
in an image.
5. Perceptual quality metrics
Two metrics for ascertaining the quality of a watermarked image are highlighted in
this section. These metrics are the Mean Square Error (MSE), and the Peak Signal to Noise
Ratio (PSNR). The MSE measures the average pixel-by-pixel difference between the original
image (I) and the watermarked image (
I
ˆ
) [9].
2
,
,
,
)
ˆ
(
1
nm
nm
nm
II
MN
MSE =
(5)
MSE
I
dBPSNR
peak
2
10
log10)( =
(6)
where I
peak
is the peak intensity level in the original image (most commonly 255 for an 8-bit
grayscale image), M and N are the dimensions of the image.
3 levels
DWT
Watermarked
image
Extraction of
s
ij
w
'
From HL
3
and LH
3
If
s
ij
w
'
< (t
1
+
t
)/2
If
s
ij
w
'
≥ (t
1
+
WM=0
WM=1
International Journal of Signal Processing, Image Processing and Patter
International Journal of Signal Processing, Image Processing and PatterInternational Journal of Signal Processing, Image Processing and Patter
International Journal of Signal Processing, Image Processing and Pattern Recognition
n Recognitionn Recognition
n Recognition
Vol. 4, No. 1,
Vol. 4, No. 1, Vol. 4, No. 1,
Vol. 4, No. 1, March 2011
March 2011March 2011
March 2011
21
The limitations of pixel based image quality metrics lead to other quality metrics that
are based on the HVS. Two of such metrics were presented by Lambrecht et al. [10] and
Watson [11]. The Lambrecht metric was described by Kutter et al. as a fair and viable method
for determining the amount of degradation suffered by a watermarked image. It makes use of
coarse image segmentation to examine contrast sensitivity as well as the masking phenomena
of the HVS. This metric then returns an overall measure of the distortion of the watermarked
image compared to the original image. The Watson metric was incorporated into the
Checkmark package [12]. It operates in the DCT domain and utilizes contrast sensitivity,
luminance masking and contrast masking in order to calculate a Total Perceptual Error (TPE)
value between the watermarked and original images.
The original and recovered messages or watermarks can be compared by computing the
Normalized Correlation (NC)[9]:
mm
mm
NC .
.
*
*
=
(7)
where m is the original message and
*
m is the recovered message. For unipolar vectors, m
{0, 1}, and for bipolar vectors, m
{−1, 1}.
6. Simulation Results
This section presents experimental results to compare between the Dugad’s scheme,
Miyazaki’s scheme, and the proposed scheme for image watermarking. Images are
watermarked using the three watermarking schemes and subjected to attacks. In order to
measure the degradation suffered by host images after watermark insertion, the PSNR and the
TPE are used. The higher the TPE value, the more degraded an image would appear to a
human viewer. The Checkmark package is used to determine the TPE value.
For all the tests in this paper, MATLAB is used. All tests are performed upon the 8-
bit grayscale 256 × 256 Mandrill and Hat images. To simulate the watermarking schemes on
the Mandrill image, we set t
1
= 115, t
2
= 200. These thresholds are obtained from Figs. (4-a)
and (4-b) to make a trade-off between the required high PSNR of the watermarked image and
high NC of the extracted watermark in the presence of a resizing attack. Resizing is
performed from size 256 × 256 to 128 × 128 and back to 256 × 256. The thresholds used for
the Hat image watermarking are obtained from Figs. (4-c) and (4-d) as t
1
= 90, t
2
= 200. To
simulate the proposed watermarking scheme, we find f
max
and set α = 0.1 to obtain the value
of T=0.1f
max
. We also take x
1
=10 and x
2
=5. Results of all schemes for the Mandrill and Hat
images are shown in Figs.(5) and (6), respectively. The numerical evaluation metrics for all
schemes in the absence and presence of attacks are tabulated in Tables (1) to (6). From Tables
(1) and (4), we notice that the proposed watermarking scheme achieves the lowest distortion
in the watermarked images in the absence of attacks. From Tables (2) and (5), we notice that
the proposed blind watermarking scheme has a better performance than Miyazaki’s scheme,
which is also blind, for most of the attacks. The Dugad’s scheme gives a better performance
than the both the proposed scheme and Miyazaki’s scheme because it is a non-blind scheme.
In fact, the need to blind watermarking schemes is more urgent than that for non-blind
schemes. From Tables (3) and (6), we notice also that a percentage of around 50% of the
input watermark bits can be extracted in the proposed scheme with most of the attacks.
International Journal of Signal Processing, Image Processing and Pattern Recognition
International Journal of Signal Processing, Image Processing and Pattern RecognitionInternational Journal of Signal Processing, Image Processing and Pattern Recognition
International Journal of Signal Processing, Image Processing and Pattern Recognition
Vol. 4, No. 1, March 2011
Vol. 4, No. 1, March 2011Vol. 4, No. 1, March 2011
Vol. 4, No. 1, March 2011
22
(a) (b)
(c) (d)
Figure 4. (a) Variation of the PSNR of the host image with thresholds t
1
and t
2
for the Mandrill image. (b) Variation of the NC between the original watermark
and the extracted watermark with thresholds t
1
and t
2
for the Mandrill image in
the presence of a resizing attack. (c) Variation of the PSNR of the host image
with thresholds t
1
and t
2
for the Hat image. (d) Variation of the NC between the
original watermark and the extracted watermark with thresholds t
1
and t
2
for
the Hat image in the presence of a resizing attack.
(a)
International Journal of Signal Processing, Image Processing and Patter
International Journal of Signal Processing, Image Processing and PatterInternational Journal of Signal Processing, Image Processing and Patter
International Journal of Signal Processing, Image Processing and Pattern Recognition
n Recognitionn Recognition
n Recognition
Vol. 4, No. 1,
Vol. 4, No. 1, Vol. 4, No. 1,
Vol. 4, No. 1, March 2011
March 2011March 2011
March 2011
23
(b) (c) (d)
Figure 5. (a) Original Mandrill image. (b) Mandrill image marked with Dugad’s
scheme in the absence of attacks. (c) Mandrill image marked with Miyazaki’s
scheme in the absence of attacks. (d) Mandrill image marked with the
proposed scheme in the absence of attacks.
(a)
(b) (c) (d)
Figure 6. (a) Original Hat image. (b) Hat image marked with Dugad’s scheme in
the absence of attacks. (c) Hat image marked with Miyazaki’s scheme in the
absence of attacks. (d) Hat image marked with the proposed scheme in the
absence of attacks.
International Journal of Signal Processing, Image Processing and Pattern Recognition
International Journal of Signal Processing, Image Processing and Pattern RecognitionInternational Journal of Signal Processing, Image Processing and Pattern Recognition
International Journal of Signal Processing, Image Processing and Pattern Recognition
Vol. 4, No. 1, March 2011
Vol. 4, No. 1, March 2011Vol. 4, No. 1, March 2011
Vol. 4, No. 1, March 2011
24
Table 1. Evaluation metrics values for all schemes for the Mandrill image.
Scheme PSNR (dB) TPE
Dugad’s scheme (Blind) 42.48 0.01
Miyazaki’s scheme
(Non-blind) 44.65 0.0079
Proposed scheme
(Blind) 46.60 0.007
Table 2. The NC of the extracted watermarks for all schemes for the Mandrill
image.
Dugad’s
scheme
Miyazaki’
s
scheme
Propose
d
Scheme
No
attacks 0.57 1 1
JPEG
Q5 0.21 0.75 0.14
JPEG
Q10 0.22 1 0.48
JPEG
Q15 0.52 1 0.85
Gaussia
n noise 0.53 0.87 0.54
Impulsiv
e noise 0.58 0.95 0.79
Croppin
g 0.11 0.35 0.48
Resizing 0.23 0.75 0.39
International Journal of Signal Processing, Image Processing and Patter
International Journal of Signal Processing, Image Processing and PatterInternational Journal of Signal Processing, Image Processing and Patter
International Journal of Signal Processing, Image Processing and Pattern Recognition
n Recognitionn Recognition
n Recognition
Vol. 4, No. 1,
Vol. 4, No. 1, Vol. 4, No. 1,
Vol. 4, No. 1, March 2011
March 2011March 2011
March 2011
25
Table 3. The extracted watermark length in the proposed scheme for the
Mandrill image. The input watermark length is 102 bits.
Table 4. Evaluation metrics values for all schemes for the Hat image.
Scheme PSNR TPE
Dugad’s scheme (Blind) 40.09 0.021
Miyazaki’s scheme (Non-
blind) 44.62 0.013
Proposed scheme (Blind) 45.36 0.012
Type of
attack
Extracted
watermark
length
No attacks 102
JPEG Q5 53
JPEG Q10 77
JPEG Q15 79
Gaussian
noise 54
Impulsive
noise 79
Cropping 38
Resizing 48
International Journal of Signal Processing, Image Processing and Pattern Recognition
International Journal of Signal Processing, Image Processing and Pattern RecognitionInternational Journal of Signal Processing, Image Processing and Pattern Recognition
International Journal of Signal Processing, Image Processing and Pattern Recognition
Vol. 4, No. 1, March 2011
Vol. 4, No. 1, March 2011Vol. 4, No. 1, March 2011
Vol. 4, No. 1, March 2011
26
Table 5. The NC of the extracted watermarks for all schemes for the Hat image.
Table 6. The extracted watermark length in the proposed scheme for the Hat
image. The input watermark length is 367 bits.
7. Conclusions
This paper presented a blind wavelet-based image watermarking scheme. This
scheme depends on the quantization of certain wavelet coefficients within certain amplitude
ranges in a binary manner to embed meaningful information in the image. Experimental
results have shown the superiority of the proposed scheme from the host image quality point
of view and the blindness point of view.
Dugad’s
scheme
Miyazaki’s
scheme
Proposed
scheme
No attacks 0.45 1
1
JPEG Q5 0.27 0.44 0.28
JPEG Q10 0.38 0.66 0.46
JPEG Q15 0.45 1 0.88
Gaussian
noise 0.37 0.75 0.57
Impulsive
noise 0.42 0.79 0.45
Cropping 0.20 0.32 0.39
Resizing 0.36 0.5 0.49
Type of
attack
Extracted
watermark
length
No attacks 367
JPEG Q5 203
JPEG Q10 271
JPEG Q15 319
Gaussian
noise 250
Impulsive
noise 293
Cropping 78
Resizing 222
International Journal of Signal Processing, Image Processing and Patter
International Journal of Signal Processing, Image Processing and PatterInternational Journal of Signal Processing, Image Processing and Patter
International Journal of Signal Processing, Image Processing and Pattern Recognition
n Recognitionn Recognition
n Recognition
Vol. 4, No. 1,
Vol. 4, No. 1, Vol. 4, No. 1,
Vol. 4, No. 1, March 2011
March 2011March 2011
March 2011
27
References
[1] R. Dugad, K. Ratakonda and N. Ahuja, A new wavelet-based scheme for watermarking images,Proc. IEEE
Intl. Conf. on Image Processing, ICIP’98,Chicago, IL, USA, Oct. 1998, 419-423.
[2] A. Miyazaki, A. Yamamoto and T. Katsura,A digital watermarking technique based on the wavelet transform
and its robustness on image compression and transformation, IEICE Trans., Special Section on Cryptography
and Information Security,E82-A, No. 1, Jan. 1999, 2-10.
[3] I. J. Cox, F. T. Leighton and T. Shamoon, “Secure spread spectrum watermarking for multimedia,” IEEE
Trans. on Image Processing, Vol. 6, Dec. 1997, 1673-1678.
[4] M. S. Raval and P. P. Rege, "Discrete Wavelet Transform Based Covert Communication Technique," Journal
of the Computer Society of India, vol.34, No.1, pp.69-75, Jan-Mar.2004
[5] D. Salomon, Data Privacy and Security, Springer-Verlag, NewYork, 2003.
[6] M. Corvi and G. Nicchiotti, “Wavelet-based image watermarking for copyright protection, Scandinavian
Conference on Image Analysis,” SCIA ’97, Lappeenranta, Finland, June 1997, 157-163.
[7] P. Meerwald, Digital image watermarking in the wavelet transform domain, Master thesis, Department of
Scientific Computing, University of Salzburg, Austria, 2001.
http://www.cosy.sbg.ac.at/˜pmeerw/Watermarking/
[8] A. Zolghadrasli, S. Rezazadeh, “Evaluation of Spread Spectrum Watermarking Schemes in the Wavelet
Domain Using HVS Characteristics,” international journal of information science&technology, volume 5,
number2, July-December, 2007
[9] S. Voloshynovskiy, S. Pereira, V. Iquise, and T. Pun. “Attack modeling: Towards a second generation
watermarking benchmark” Journal of Signal Processing,80 (6) , May 2001.
[10] Checkmark benchmarking project [online]. Available from World Wide Web (date accessed: December,
2004): http://watermarking.unige.ch/Checkmark/.
[11] A. B. Watson, “DCT quantization matrices visually optimized for individual images, Human Vision,” Visual
Processing and Digital Display IV, Proc. SPIE, Vol.1913, San Jose, CA, USA, Feb. 1993, 202-216.
[12] A. Mayache, T. Eude and H. Cherefi, “A comparison of image quality models and metrics based on human
visual sensitivity,” Proc. IEEE Intl. Conf. on Image Processing,ICIP’98, Chicago, IL, USA, Oct. 1998, 409-
413.
Authors
Hanaa A. Abdallah received the BSc and MSc. degrees from the
faculty of Engineering from zagazig University, Egypt in 1998 and
2002, respectively. She is currently an Assistant Lecturer in the Dept.
of Electronics and Communications engineering, Faculty of
Engineering, zagazig University.She is currently working towards
the Ph.D. degree in Communications Engineering from the zagazig
University. Her areas of interests are image processing, image
enhancement image compression, data hiding,
steganography,watermarking.
Mohiy M. Hadhoud
received the BSc and MSc degrees in Electrical
Engineering from Menoufia University in Egypt in 1976 and 1981
respectively. He received the PhD degree from Southampoton
University in 1987. He is currently the dean of the Faculty of
Computers and Information, Menoufia University. His areas of
interests are signal processing, Image Processing and Digital
Communications
International Journal of Signal Processing, Image Processing and Pattern Recognition
International Journal of Signal Processing, Image Processing and Pattern RecognitionInternational Journal of Signal Processing, Image Processing and Pattern Recognition
International Journal of Signal Processing, Image Processing and Pattern Recognition
Vol. 4, No. 1, March 2011
Vol. 4, No. 1, March 2011Vol. 4, No. 1, March 2011
Vol. 4, No. 1, March 2011
28
Abdelhamid A. Shaalan received his MSc in microwave
engineering from Faculty of Engineering, Cairo University, Egypt in
1991. He received his PhD in Microwave Engineering from
Faculty of Engineering, Cairo University in 1996. He is an associate
Professor in communication engineering at Faculty of engineering,
Zagazig University, Egypt. His research interests include antenna
engineering and its applications.
Fathi E. Abd El-Samie received the B.Sc. (Honors), M.Sc., and
PhD. from the Faculty of Electronic Engineering, Menoufia
University, Menouf, Egypt, in 1998, 2001, and 2005, respectively.
He joined the teaching staff of the Department of Electronics and
Electrical Communications, Faculty of Electronic Engineering,
Menoufia University, Menouf, Egypt, in 2005. He is a co-author of
about 130 papers in national and international conference
proceedings and journals. He has received the most cited paper
award from Digital Signal Processing journal for 2008. His current
research areas of interest include image enhancement, image restoration, image interpolation,
superresolution reconstruction of images, data hiding, multimedia communications, medical
image processing, optical signal processing, and digital communications.
... These algorithms are simple, fast, and have a high embedded capacity [27]. These techniques are susceptible to changes that can damage the watermark [28,29]. Spatial-based techniques' disadvantages are more than their advantages that cannot withstand many attacks, such as added noise and compression methods [30]. ...
... An important issue in this area is to choose the best place for watermarking to avoid image distortion. The transform field is resistant to change, attack and the watermark can be invisibly embedded in the image [28,29]. Techniques of this method include DCT [34], Discrete Fourier Transform (DFT) [35], and DWT [6]. ...
Article
Digital watermarking is a method for data hiding that ensures the security of multimedia data. The watermark can be a digital image or data stored within digital content. The Shearlet transform, a multi-resolution and multi-directional conversion, can be used for watermarking in digital images. Due to its superior features, this conversion can increase the efficiency of applications such as watermarking of images. In this paper, Shearlet and the Singular Value Decomposition (SVD) transforms are used with the Whale optimization algorithm to obtain the most appropriate scaling factor in the watermark extraction step after applying different types of image processing attacks. Shearlet transform has more transparency than traditional converts. The SVD transform also increases the robustness of watermarking operations. The results of different experiments show that the new method presented in this paper, in terms of robustness and imperceptibility compared to the evaluated methods, performs better than most image processing attacks. © 2021 Materials and Energy Research Center. All rights reserved.
... It has become a significant technique in image processing and watermarking thanks to its excellent energy compaction properties. [23,26,27]. ...
... Finally, singular value decomposition follows by algebraic properties and the small changes in the singular values would not perceptually influence on an image, so the image would approximately remain the same even after inserting the watermark. Each singular value in a SVD matrix specifies the luminance of an image layer while the subsequent pair of singular vectors identify the geometry of the image layer [27][28]. Table 2 refers to the pros and cons of each transform technique mentioned above. ...
... DWT is a multipurpose mathematical transform which splits an image into four frequency subbands based on small waves with varying frequencies and limited periods. For each level of decomposition in DWT, a lower resolution of approximation component (LL) and three other corresponding detail components like horizontal (HL), vertical (LH) and diagonal (HH) are represented [1,21,22,45]. ...
Article
Full-text available
In this paper, an analysis on hybrid Discrete Wavelet Transform (DWT) and Singular Value Decomposition (SVD) for image watermarking is carried out to investigate the effect of a deeper level of the SVD on imperceptibility and robustness to resist common signal processing and geometric attacks. For this purpose, we have designed two hybrid watermarking schemes, the first one with DWT and first level of SVD, whereas, in the second scheme, the same design is employed with a second level of SVD. In this experiment, a comprehensive analysis is performed on the two designed schemes and the effect of robustness and imperceptibility is compared in the first and second levels of SVD in each DWT sub-band. Having analyzed more than 100 medical and non-medical images in standard datasets and real medical samples of patients, the experimental outcomes show a remarkable increase in both imperceptibility and robustness in the second level of SVD, in comparison to the first level. In addition, the achieved result shows that the SVD2 scheme offers the highest imperceptibility in the LL sub-band (more than 60 dB on average PSNR), with satisfactory robustness against noise attacks, but less persistence in some geometric attacks such as cropping. For the HH sub-band, strong robustness against all types of tested of attacks is obtained, though its imperceptibility is slightly lower than the achieved PSNR in the LL sub-band. In HH sub-band, an average growth of 5 dB in PSNR and 2% in NC can be observed from the second level of SVD in comparison to the first level. These results make SVD2 a good candidate for content protection, especially for medical images.
... Watermarking by embedding data related to its use in the main data, provides purposes such as copyright protection, data authentication, copy control, data tracking, etc. various methods have been suggested to meet these goals. Methods of watermarking in frequency domain are one of the best image watermarking methods that are used [1,2]. For example, it can be included the watermarking methods by using Discrete Cosine Transform (DCT) [[3] - [5]], and Discrete Wavelet Transform (DWT) [ [6] - [9]]. ...
Article
Full-text available
One of the ways to enhance the security and concealment of data used today is image watermarking. In image watermarking operation, we try to hide image inside another image without letting others know about the hidden image. In this paper, Contourlet Transform and SVD Transform are used to embedded watermark in the host image. The PSO Optimization Algorithm is also used in the watermark extraction step to find the best scale factor. The results of the proposed algorithm in this article show an improvement over the comparative methods.
... Abdallah et al. [16] proposed a blind watermarking technique under wavelet transform, by which an identity mark is fetched into the coarsest scale wavelet coefficients. The image undergoes the three-level decomposition and the copyright mark is installed into the coarse scale. ...
Article
Full-text available
Watermarking in multimedia content has secured attractive approaches in research society since nowadays the transmission of digital content in wireless medium has been enormous. Along with the distribution of digital content, it is important to claim ownership. A non-blind detection technique is proposed in this paper. First, an original image undergoes the wavelet decomposition such as LL1, LH1, HL1, and HH2. Afterward, the medium level sub-band coefficients of the image are subject to the comfortable shares for merging. Next, one of the shares from the LH1 band is merging with another one of the shares in HL1 band. In this stage, the copyright mark is fetched into the merged sub-band coefficients. Finally, an inverse wavelet transform is applied to receive the watermarked image. To prove the authentication, the original image and the watermarked image undergo the same operation and acquire copyright information. Experimental results achieve that the proposed approach can withstand various image processing attacks.
... Hanaa A. Abdallah et al 2011 [11] presented a blind wavelet-based image watermarking scheme. It was based on inserting the watermark bits into the coarsest scale wavelet coefficients of the third-level wavelet decomposition. ...
Article
Full-text available
Protection of digital multimedia content has become an increasingly important issue for content owners and service providers. Watermarking is identified as a major means to achieve copyright protection. The algorithm proposed in this work, is to use a blind watermarking scheme based on the Discrete Wavelet Transform (DWT). Watermark components are embedded in the LL subband of the 4th DWT level of the host image by quantizing coefficients of the LL subband to improve the watermark robustness. The Genetic Algorithm (GA) is used for optimizing the quantization step size parameter, and the strength of factors. The host image used is a 512×512 gray scale image and the watermark image is a 32×32 binary logo. The proposed scheme was tested against mostly known threats and it proves to give good robustness. Also, it still gives a high quality watermarked image. MATLAB Program was used to perform the watermarking task.
... This useful information is represented as images. These domains include Fourier domain, wavelet domain empirical mode decomposition (EMD) domain, and discrete coine transform (DCT) domain [3]. ...
Article
Full-text available
This paper presents an audio watermarking technique based on singular value decomposition (SVD) and fractional Fourier transform (FRT). The basic idea of this technique is to implement SVD watermarking on the audio signals in the FRT domain due to its recommended degree of security resulting from using a rotation angle in addition to the frequency-domain transformation. The SVD has an invariance to changes in the signal after watermark embedding. Hence, the proposed technique has a large degree of security and resistance to attacks. This technique is based on embedding an image watermark in either the audio signal or a transformed version of this signal. Experimental results show that watermark embedding in the FRT of an audio signal achieves less distortion of the audio signal in the absence of attacks. In the presence of attacks, it is recommended that the embedding is performed in the FRT of the audio signal to maintain a high detection correlation coefficient between the original watermark and the obtained watermark. A segment-based implementation of the proposed audio watermarking technique is also presented. This implementation succeeds in obtaining a high detection correlation coefficient in the presence of severe attacks. It is noticed from the results that in the presence of attacks, the SVD watermarking in the FRT domain with a phase angle of 5π/4 is better for watermark detection than watermarking using other angles in the FRT domain.
... Hanaa A. Abdallah et al 2011 [11] presented a blind wavelet-based image watermarking scheme. It was based on inserting the watermark bits into the coarsest scale wavelet coefficients of the third-level wavelet decomposition. ...
... The algorithm intended by Wang et al. (2007a, b) which resist the various desynchronizing attacks like rotation, scaling, translation and cropping etc. A blind image watermarking proposed by Abdallah and Hadhoud (2011) where the third level discrete wavelet transform (DWT) is applied on cover image and watermark is embedded into courser coefficients. The DWT and SVD based image watermarking is proposed in Run et al. (2012), Ghazy et al. (2015) and Zhang et al. (2011) to get better imperceptibility and robustness. ...
Article
Full-text available
Nowadays copyright protection is mandatory in the field of image processing to removes the illegitimate utilization and imitation of digital images. The digital image watermarking is one of the most reliable methods for protecting the illegal validation of data. In this paper, singular value decomposition based digital image watermarking scheme is proposed in complex wavelet transform (CWT) domain using intelligence algorithms like particle swarm optimization (PSO) and recently proposed Jaya algorithm. The watermark image is embedded into high frequency CWT subband of cover image. At the time of watermark embedding and extraction, optimization algorithms Jaya and PSO are applied to improve the robustness and imperceptibility by assessing the fitness function. The perceptual quality of watermarked image and robustness of extracted watermark image are verified under the filtering, rotation, scaling, Gaussian noise and JPEG compression attacks. From the comparative analysis it is proved that Jaya algorithm is better as compared to PSO algorithm under most types of attacks with higher magnitudes whereas identical under the lower magnitude of applied attacks. Moreover, using variety of cover images, it is found that, the elapse time and value of fitness function given by Jaya algorithm are also better as compared to PSO.
Chapter
This chapter devoted to the comprehensive survey on the current state-of-the-art in the field of medical image watermarking. In addition to this, we have introduced the wavelet based watermarking techniques because the main advantages this technique are space-frequency localization, multi-resolution representation, multi-scale analysis, reducing blocking artifact, adaptability, linear complexity and compatibility with the image coding standards JPEG 2000. Further, the potential challenges of existing techniques are also discussed followed by performance comparison of important wavelet domain medical image watermarking techniques are tabulated.
Article
Full-text available
Covert communication also better known as “Steganography” is the art of hiding information in ways that prevent the detection of hidden messages. This way, the very presence of communication becomes hidden. Data insertion is possible because the digital medium is ultimately consumed by human and the human hearing and visual systems are imperfect detectors. Discrete Wavelet Transform (DWT) is a widely used technique, exhibiting a strong similarity to the way the Human Visual System (HVS) processes. Given its suitability to model the HVS behavior, DWT can be used as a computationally efficient version of the frequency models for the HVS. Due to its excellent spatial and frequency localization properties, the DWT is suitable to identify the image areas where a disturbance can be easily hidden. In this paper we propose steganography technique based on the DWT. We hide the message in the high frequency band of the image, as eye is less sensitive to noise in high resolution bands.
Article
Full-text available
Several image compression standards (JPEG, MPEG, H.261) are based on the Discrete Cosine Transform (DCT). These standards do not specify the actual DCT quantization matrix. Ahumada & Peterson 1 and Peterson, Ahumada & Watson 2 provide mathematical formulae to compute a perceptually lossless quantization matrix. Here I show how to compute a matrix that is optimized for a particular image. The method treats each DCT coefficient as an approximation to the local response of a visual "channel." For a given quantization matrix, the DCT quantization errors are adjusted by contrast sensitivity, light adaptation, and contrast masking, and are pooled non-linearly over the blocks of the image. This yields an 8x8 "perceptual error matrix." A second non-linear pooling over the perceptual error matrix yields total perceptual error. With this model we may estimate the quantization matrix for a particular image that yields minimum bit rate for a given total perceptual error, or minimum perceptual error for a given bit rate. Custom matrices for a number of images show clear improvement over image-independent matrices. Custom matrices are compatible with the JPEG standard, which requires transmission of the quantization matrix.
Conference Paper
Full-text available
In this paper, we introduce a multiresolution watermarking method for copyright protection of digital images. The method is based on the discrete wavelet transform. A noise type Gaussian sequence is used as watermark. To embed the watermark robustly and imperceptibly, watermark components are added to the significant coefficients of each selected subband by considering the human visual system (HVS) characteristics. Some small modifications are performed to improve HVS model. The host image is needed in watermark extraction procedure and normalized correlation function (NCF) is used to measure similarities of extracted watermarks. It is shown that this method is robust against wide variety of attacks. Comparison with the existing methods shows the better performance of this suggested method.
Conference Paper
Full-text available
This paper considers the use of the image quality metric for still image compression systems comparison. Considering the fact that the conventional PSNR cannot sufficiently reflect the result of subjective assessment other quality measures have been considered to design the variable bit-rate coders. Distortion measures are developed for the purpose of predicting the subjective quality of pictures. Subjectively relevant distortion measures that mirror viewers' assessments of picture quality would make the task of designing and optimizing coding schemes considerably easier. One significant problem associated with these measures is that there is little information on how these metrics perform in comparison to each other. The purpose of this paper is to provide a rigorous evaluation of three metrics to assess the quality of compressed images. The compression system used in this evaluation is the classical JPEG coder. Both objective and subjective tests were performed on a 50 natural image data base with a panel of experimented and non-experimented observers. The results show that these metrics are highly correlated with the subjective quality grading but also depends on the complexity of the images under study. They can be improved by integrating all masking effects and typical artifacts of the compression method. Their combination as a multidimensional quality measure in which each dimension is related to a properly identified artifact, should be based on a subjective evaluation
Article
Covering classical cryptography, modern cryptography, and steganography, this volume details how data can be kept secure and private. Each topic is presented and explained by describing various methods, techniques, and algorithms. Moreover, there are numerous helpful examples to reinforce the reader's understanding and expertise with these techniques and methodologies. Features & Benefits: * Incorporates both data encryption and data hiding * Supplies a wealth of exercises and solutions to help readers readily understand the material * Presents information in an accessible, nonmathematical style * Concentrates on specific methodologies that readers can choose from and pursue, for their data-security needs and goals * Describes new topics, such as the advanced encryption standard (Rijndael), quantum cryptography, and elliptic-curve cryptography. The book, with its accessible style, is an essential companion for all security practitioners and professionals who need to understand and effectively use both information hiding and encryption to protect digital data and communications. It is also suitable for self-study in the areas of programming, software engineering, and security.
Article
SUMMARY In this paper, we propose two methods of digital watermark for image signals based on the wavelet transform. We classify wavelet coefficients as insignificant or significant by using zerotree which is defined in the embedded zerotree wavelet (EZW) algorithm (13). In the first method, information data are embed- ded as watermark in the location of insignificant coefficients. In the second method, information data can be embedded by thresh- olding and modifying significant coefficients at the coarser scales in perceptually important spectral components of image signals. Information data are detected by using the position of zerotree's root and the threshold value after the wavelet decomposition of an image in which data hide. It is shown from the numerical ex- periments that the proposed methods can extract the watermark from images that have degraded through several common signal and geometric processing procedures.
Article
Digital image watermarking techniques for copyright protection have become increasingly robust. The best algorithms perform well against the now standard benchmark tests included in the Stirmark package. However the stirmark tests are limited since in general they do not properly model the watermarking process and consequently are limited in their potential to removing the best watermarks. Here we propose a stochastic formulation of watermarking attacks using an estimation-based concept. The proposed attacks consist of two main stages: (a) watermark or cover data estimation; (b) modification of stego data aiming at disrupting the watermark detection and of resolving copyrights, taking into account the statistics of the embedded watermark and exploiting features of the human visual system. In the second part of the paper we propose a “second generation benchmark”. We follow the model of the Stirmark benchmark and propose the 6 following categories of tests: denoising attacks and wavelet compression, watermark copy attack, synchronization removal, denoising/compression followed by perceptual remodulation, denoising and random bending. Our results indicate that even though some algorithms perform well against the Stirmark benchmark, almost all algorithms perform poorly against our benchmark. This indicates that much work remains to be done before claims about “robust” watermarks can be made. We also propose a new method of evaluating image quality based on the Watson metric which overcomes the limitations of the PSNR.
Article
This paper presents a secure (tamper-resistant) algorithm for watermarking images, and a methodology for digital watermarking that may be generalized to audio, video, and multimedia data. We advocate that a watermark should be constructed as an independent and identically distributed (i.i.d.) Gaussian random vector that is imperceptibly inserted in a spread-spectrum-like fashion into the perceptually most significant spectral components of the data. We argue that insertion of a watermark under this regime makes the watermark robust to signal processing operations (such as lossy compression, filtering, digital-analog and analog-digital conversion, requantization, etc.), and common geometric transformations (such as cropping, scaling, translation, and rotation) provided that the original image is available and that it can be successfully registered against the transformed watermarked image. In these cases, the watermark detector unambiguously identifies the owner. Further, the use of Gaussian noise, ensures strong resilience to multiple-document, or collusional, attacks. Experimental results are provided to support these claims, along with an exposition of pending open problems.
Conference Paper
Digital imaging continues its triumphal march especially in geoscience and remote sensing. Digital information can be easily processed, manipulated and copied. Thus, the need for document ownership verification arises. In this study, we propose a watermarking method for spectral images. The watermark is embedded in a transform domain of the spectral image. The transform is based on the principal component analysis and on the wavelet transform. With the inverse transforms the watermarked spectral image is reconstructed. The experiments indicate that the watermark is robust against attacks in lossy compression but more sensitive against mean filtering.