A DWT, DCT AND SVD BASED WATERMARKING TECHNIQUE TO PROTECT THE IMAGE PIRACY

International Journal of Managing Public Sector Information and Communication Technologies (IJMPICT)
Vol. 4, No. 2, June 2013
DOI : 10.5121/ijmpict.2013.4203 21
A DWT, DCT AND SVD BASED WATERMARKING
TECHNIQUE TO PROTECT THE IMAGE PIRACY
Md. Maklachur Rahman1
1
Department of Computer Science and Engineering, Chittagong University of
Engineering and Technology, Bangladesh
mcr.rahman@gmail.com
ABSTRACT
With the rapid development of information technology and multimedia, the use of digital data is increasing
day by day. So it becomes very essential to protect multimedia information from piracy and also it is
challenging. A great deal of Copyright owners is worried about protecting any kind of illegal repetition of
their information. Hence, facing all these kinds of problems development of the techniques is very
important. Digital watermarking considered as a solution to prevent the multimedia data.
In this paper, an idea of watermarking is proposed and implemented. In proposed watermarking method,
the original image is rearranged using zigzag sequence and DWT is applied on rearranged image. Then
DCT and SVD are applied on all high bands LH, HL and HH. Watermark is then embedded by modifying
the singular values of these bands. Extraction of watermark is performed by the inversion of watermark
embedding process. For choosing of these three bands it gives facility of mid-band and pure high band that
ensures good imperceptibility and more robustness against different kinds of attacks.
KEYWORDS
DWT, DCT, SVD, Watermarking, Zigzag.
1.INTRODUCTION
In recent years, the increasing amount of applications using digital multimedia technologies has
emphasized the need to protect digital multimedia data from pirates. Authentication and
information hiding, copyright protection, content identification and proof ownership have also
become important issues. To accomplish these issues, watermarking technology is used.
Researchers are interested in the field of watermarking because of its significance. These kinds of
work in this field have lead to several watermarking techniques such as spatial domain and
transform domain. In transform domain it may discrete cosine transform (DCT), discrete wavelet
transform (DWT), singular value decomposition (SVD) and their cross relation. Watermarking is
a process embedding a piece of information into a multimedia content, such as image, audio and
video in such a way that it is imperceptible to a human, but easily detectable by computer. Before
the development of digital image watermarking, it was difficult to achieve copyright protection,
authentication, data hiding, content identification and proof ownership. But currently it is easy to
accomplish these kinds goal using watermarking techniques. So watermarking is very important
to us for these kinds of work. Every watermarking algorithm consists of an embedding and
extraction process that needs to hide desired information. The embedding and extraction process
are described below. Embedded watermark may have several properties such as imperceptibility
and robustness. If we cannot distinguish between host image and watermarked image called
imperceptibility. Basically imperceptibility depends on similarity between host image and
watermarked image. If it difficult to remove or destroy watermark from watermarked image then
it said to be robustness. Robustness measures how difficult to remove or destroy watermark from
watermarked image. If it is high then robustness is high. DCT based watermarking contains
the low frequency information so image contains all information that is similar to the

Vol. 4, No. 2, June 2013
22
original image. DWT based compression offers scalability so image can be divided into
four sub bands in every level of decomposed and by choosing of the sub-band to develop a hybrid
watermarking scheme for improving the robustness, imperceptibility and capacity and help to
develop a new hybrid method. In the first method (ref. 9) the original image is segmented into
blocks then find out spatial frequency of each block select reference image under certain
condition then apply DWT, DCT and SVD transformation. In the second method (ref. 3) the
original image is segmented into blocks then find out spatial frequency of each block select
reference image under certain condition then apply DWT and SVD transformation. In the
proposed method, since high band is considered so it fulfills the requirements imperceptibility
and robustness. Most of the domain transformation watermarking technique works with DCT,
DWT, SVD and their mixing algorithm such as DWT-DCT, DCT-SVD and so on. In this paper
we proposed a digital watermarking technique using DWT-DCT and SVD transformation. This
method is provided a good imperceptibility and high robustness against various kinds processing
attacks. The rest of the paper is organized as follows: Section 2, focuses on overview of
transforms for watermarking. Section 3, gives details the proposed methodology and
watermarking algorithms. In section 4, gives experimental results and compares. In section 5,
conclusion is drawn.
2. PRELIMINARIES
As stated earlier that transform domain based watermarking scheme is always a better choice than
spatial domain based watermarking scheme. This can be done by using different transformation
like DCT, SVD and DWT. In this section, we will briefly describe the DCT, DWT and SVD
transformations in below.
2.1. Discrete Wavelet Transform (DWT)
The basic idea of DWT in which a one dimensional signal is divided in two parts one is high
frequency part and another is low frequency part. Then the low frequency part is split into two
parts and the similar process will continue until the desired level. The high frequency part of the
signal is contained by the edge components of the signal. In each level of the DWT (Discrete
Wavelet Transform) decomposition an image separates into four parts these are approximation
image (LL) as well as horizontal (HL), vertical (LH) and diagonal (HH) for detail components. In
the DWT decomposition input signal must be multiple of 2n
. Where, n represents the number of
level. To analysis and synthesis of the original signal DWT provides the sufficient information
and requires less computation time. Watermarks are embedded in these regions that help to
increase the robustness of the watermark. A one level DWT decomposition process is shown in
Figure 1.
Image
Figure 1. One level DWT decomposition process
2.2. Discrete Cosine Transform (DCT)
The DCT is the most popular transform function used in signal processing. It transforms a signal
from spatial domain to frequency domain. Due to good performance, it has been used in JPEG
standard for image compression. It is a function represents a technique applied to image pixels in
LL1 HL1
LH1 HH1
L R

Vol. 4, No. 2, June 2013
23
spatial domain in order to transform them into a frequency domain in which redundancy can be
branded. DCT techniques are more robust compared to spatial domain techniques. Such
algorithms are robust against simple image processing operations like adjustment, brightness,
blurring, contrast and low pass filtering and so on. But it is difficult to implement and
computationally more expensive. The one-dimensional DCT is useful in processing one-
dimensional signals such as speech waveforms. For analysis of two-dimensional (2D) signals
such as images, we need a 2D version of the DCT. The 2D DCT and 2D IDCT transforms is
given by equation 1 and 2.
Formulae of 2-D DCT:
( , ) = ( ) ( ) ( , ) cos
(2 + 1)
2
∗
(2 + 1)
2
(1)
Formulae of 2-D inverse DCT:
( , ) = ( ) ( ) ( , ) cos
(2 + 1)
2
∗
(2 + 1)
2
(2)
Where,
( ), ( ) =
⎩
⎪
⎨
⎪
⎧ 1
, , = 0
2
, , = 1 − 1
2.3. Singular Value decomposition (SVD)
The singular value decomposition (SVD) matrix is very useful in computer vision as a
decomposition matrix and it is an efficient tool for image transformations. The SVD of a given
image F in the form of a matrix is defined as
= (3)
Where, S is the diagonal matrix that is
=
⎣
⎢
⎢
⎢
⎡
0 . 0 0
0 . 0 0
.
0
0
.
0
0
.
.
.
.
0
.
0
⎦
⎥
⎥
⎥
⎤
And U and V are the orthogonal matrices
= =

Vol. 4, No. 2, June 2013
24
=
, , … … . . , ≥ 0
The diagonal elements of matrix S are the singular values of matrix F and non-negative numbers.
3. PROPOSED METHODOLOGY AND WATERMARKING ALGORITHMS
In the proposed watermarking technique, A DWT, DCT and SVD based hybrid watermarking
technique is formulated. In this subsection, we have described the watermark embedding and
extraction process by using flowchart and algorithmically.
3.1. Watermark Embedding Procedure
Figure 2. Watermark embedding process

Vol. 4, No. 2, June 2013
25
3.2. Watermark Extraction Procedure
Figure 3. Watermark extraction process
3.3. Algorithm: Watermark Embedding
Step 1: Input Host image HI.
Step 2: Rearrange the host image HI by applying zigzag scanning process to get rearranged image
RI.
1 2 3 4
5 6 7 8
9 10 11 12
13 14 15 16
Original Matrix
Rearrange Matrix
Figure 4. Rearrange a matrix by zigzag process.
1 2 5 9
6 3 4 7
10 13 14 11
8 12 15 16
Zigzag process

Vol. 4, No. 2, June 2013
26
Step 3: Apply single level DWT on rearranged image RI to decompose it into four sub-bands LL,
HL, LH and HH.
Step 4: Select all high bands LH, HL and HH of RI. Apply DCT to all high bands LH, HL and
HH.
LL HL
LH HH
Figure 5. Selected bands for watermarking.
Step 5: Then apply SVD to all high bands LH, HL and HH to get SH1, SH2 and SH3.
Step 6: Input watermark image wi. Apply single level DWT to decompose it into four sub-bands
LL1, HL1, LH1 and HH1.
Step 7: Select all high bands LH1, HL1 and HH1 of wi. Apply DCT to all high bands LH1, HL1
and HH1.
Step 8: Then apply SVD to all high bands LH, HL and HH to get SW1, SW2 and SW3.
Step 9: Modify SH1, SH2 and SH3 by using equation = + ∗ ℎ , = 1 3.
Step 10: Construct modified SVD matrix LH11, HL11 and HH11.
Step 11: Apply inverse DCT to all high bands LH11, HL11 and HH11. Apply inverse DWT with
LL.
Step 12: Apply inverse zigzag process to arrange the original position of image and finally get
watermarked image WI.
3.4. Algorithm: Watermark Extraction
Step 1: Input Watermarked image WI.
Step 2: Rearrange the watermarked image WI by applying zigzag scanning process to get
rearranged image RI*.
Step 3: Apply single level DWT on rearranged image RI* to decompose it into four sub-bands
LL*, HL*, LH* and HH*.
Step 4: Select all high bands LH*, HL* and HH* of RI*. Apply DCT to all high bands LH*, HL*
and HH*.
Step 5: Then apply SVD to all high bands LH*, HL* and HH* to get SH1*, SH2* and SH3*.
Step 6: Modify SH1*, SH2* and SH3* by using equation = ( − )/ ℎ , =
1 3.
Step 7: Construct modified SVD matrix LH1*, HL1* and HH1*.
Step 8: Apply inverse DCT to all high bands LH1*, HL1* and HH1*.
Step 9: Apply inverse DWT to all bands to get watermark image.

Vol. 4, No. 2, June 2013
27
4. EXPERIMENTAL PERFORMANCE ANALYSIS
The proposed watermarking algorithm is simulated using MATLAB 9 with Processor Intel core 2
duo 2.2 GHZ and RAM 2 GB. The proposed watermarking algorithm is tested for the various host
and watermark images. Here some results are given. To evaluate the performance of the
proposed method, calculate PSNR (Peak Signal to Noise Ratio) and NCC (Normalized Cross
Correlation) values. PSNR is widely used to measure imperceptibility between the original image
and watermarked image. PSNR is defined by the eqn. (5). The similarity between the original
and extract watermark image use to represent how algorithm is robust against noise that is
calculated by NCC value. NCC is defined by eqn. (6).
= 10 log
255
(5)
Where,
=
1
×
[ ( , ) − ( , )]
=
∑ ∑ ( , ) ′( , )
∑ ∑ | ( , )|
(6)
Figure 6. Host images
Figure 7. Watermarked images with watermark copyright image

Vol. 4, No. 2, June 2013
28
Figure 8. Watermarked images with watermark cameramen image
PSNR = 40.5935
NCC = 0.9993
PSNR = 40.6685
NCC = 0.9991
PSNR = 40.0854
NCC = 0.9994
PSNR = 34.3189
NCC = 0.9997
Figure 9. Extracted copyright watermark Images without applying noise
PSNR = 36.5698
NCC = 0.9988
PSNR = 37.2255
NCC = 0.9979
PSNR = 36.4139
NCC = 0.9988
PSNR = 32.9252
NCC = 0.9994
Figure 10. Extracted CUET logo watermark Images without applying noise
Salt & Pepper noise Speckle Noise Poisson Gamma (0.6)

Vol. 4, No. 2, June 2013
29
Sharpening ( 100) Average Filtering
(13 x 3)
Median Filtering
(13 x 13)
Additive Gaussian
Noise (75%)
JPEG compression
(80:1)
Cropping
(25% area
remaining)
Resizing
(512 -> 128 -> 512)
Rotation 50o
Pixilation 3 Wrapping Histogram
Equalization
Motion Blur
Figure 11. Attacked watermarked image
Salt & Pepper noise Speckle Noise Poisson Gamma (0.6)
Sharpening ( 100) Average Filtering
(13 x 3)
Median Filtering
(13 x 13)
Additive Gaussian
Noise (75%)

Vol. 4, No. 2, June 2013
30
JPEG compression
(80:1)
Cropping
(25% area remaining)
Resizing
(512 -> 128 ->
512)
Rotation 50o
Pixilation 3 Wrapping Histogram
Equalization
Motion Blur
Figure 12. Extracted watermarks from attacked images
Table 1. Performance results in terms of Normalized Cross Correlation (NCC) values
Attacks
Normalized Cross Correlation (NCC) values
Existing methods Proposed Method
DWT-SVD
(Ref : 3)
DWT-DCT-SVD
(Ref : 9)
DWT-DCT-SVD
Average Filtering(13 x 3) 0.1198 -0.0928 0.9589
Median Filtering (13 x 13) -0.0852 -0.0852 0.9358
Additive Gaussian Noise (75%) 0.6749 0.6749 0.4255
JPEG compression (80:1) 0.9751 0.9751 0.9997
Cropping (25% area remaining) 0.8810 0.6120 0.9530
Resizing (512 -> 128 -> 512) 0.2570 0.2570 0.7391
Rotation 50o
0.8846 0.8846 0.9338
Pixilation 3 0.0871 -0.4185 0.9983
Wrapping 0.7299 -0.4559 0.8600
Histogram equalization 0.9182 0.9182 0.8416

Vol. 4, No. 2, June 2013
31
Motion blur -0.1854 -0.0363 0.9729
Sharpening 0.7240 0.7500 0.7727
Salt & Pepper noise --------- --------- 0.7889
Speckle Noise --------- --------- 0.9873
Poisson --------- --------- 0.9941
Gamma (0.6) --------- --------- 0.9041
From Table 1, it is experimental that the proposed DWT-DCT-SVD watermarking algorithm
gives more NCC values than the existing DWT-SVD and DWT-DCT-SVD method. That ensures
more robustness against different kinds of noise. And Figure 5 and 6 also shows the good PSNR
values that ensures more imperceptibility.
5. CONCLUSIONS
The proposed watermarking algorithm using DWT, DCT and SVD transformation that
contributes more robust in comparison with many watermarking algorithms. The watermarked
image quality is good in terms of imperceptibility. In this watermarking algorithm all high bands
LH, HL, HH are chosen which cover the mid bands LH, HL and pure high band HH that gives
more robust against different kinds of filtering noises and geometric noises. In future, the
proposed algorithm can be improved using full band DWT-DCT-SVD and further can be
extended to color images and video processing.
REFERENCES
[1] M. Calagna, H. Guo, L. V. Mancini and S. Jajodia, “A Robust Watermarking System Based on SVD
Compression”, Proceedings of ACM Symposium on Applied Computing (SAC 2006), Dijon, France,
pp. 1341-1347, 2006.
[2] Dr. M. A. Dorairangaswamy, “A Robust Blind Image Watermarking Scheme in Spatial Domain for
Copyright Protection”, International Journal of Engineering and Technology vol. 1, no.3, pp. 249 - 255,
August 2009.
[3] S. Murty. P, M. U. Bhaskar, P. N. Babu and P. R. Kumar, “A Semi-Blind Reference Watermarking
Scheme Using DWT-DCT-SVD for Copyright Protection”, International Journal of Computer Science
& Information Technology (IJCSIT) vol. 4, no 2, pp. 69-82 April 2012.
[4] F. Cayre, C. Fontaine and T. Furon, “Watermarking security: theory and practice”, Signal Processing
IEEE Transactions vol. 53, no. 10, pp. 3976–3987, Oct. 2005.
[5] S. K. Prajapati, A. Naik and A. Yadav, “Robust Digital Watermarking using DWT-DCT-SVD”,
International Journal of Engineering Research and Applications Vol. 2, Issue 3, May-Jun 2012,
pp.991-997.
[6] A. Sverdlov, S. Dexter and A. M. Eskicioglu, “Robust DCT-SVD Domain Image Watermarking for
Copyright Protection: Embedding Data in All Frequencies”, submitted to Multimedia Computing
and Networking 2005 Conference, San Jose, CA, January 16-20, 2005.
[7] C. C. Lai and C. C. Tsai, Digital Image Watermarking Using Discrete Wavelet Transform and Singular
Value Decomposition”, IEEE Trans. on Instrumentation and Measurement, vol. 59, no. 11 pp. 3060-
3063 2010.
[8] S. Mukherjee and A. K. Pal, “A DCT-SVD based Robust Watermarking Scheme for Gray scale Image”,
International Conference on Advances in Computing, Communications and Informatics (ICACCI-
2012).

Vol. 4, No. 2, June 2013
32
[9] S. Murty. P,M.U. Bhaskar and P.N.Babu, P. Rajesh Kumar, “A Semi-Blind Reference Watermarking
Scheme Using DWT-SVD for Copyright Protection”, The International Journal of Multimedia & Its
Applications (IJMA) Vol.3, No.3, pp.61-70.
[10] S. D. Lin and C. F.Chen,“A Robust DCT-Based Watermarking for Copyright Protection,” IEEE
Transactions on Consumer Electronics, vol. 46, no. 3, pp. 415-421, August 2000.
[11] S. Rezazadeh, and M. Rahmati, “A robust watermarking scheme based on wavelet transform and
block SVD,” 9th International Symposium on Signal Processing and Its Applications, pp. 1-4, 2007.
[12] V.Santhi, N. Rekha and S.Tharini “A Hybrid Block Based Watermarking Algorithm using DWT-
DCT-
SVD Techniques for Color Images”, proceedings of International Conference on Computing,
Communication and Networking, 2008. ICCCN- 2008.
[13] S. S. Kumar, B.C. Mohan and B.N.Chatterji, “An Oblivious Image Watermarking Scheme using
Singular
Value Decomposition,” IASTED International Conference on Signal and Image Processing
(ICSIP’07),
Honolulu, Hawaii, USA, August 20-22, 2007.
[14] B. Zhou and J. Chen, “A Geometric Distortion Resilient Image Watermarking Algorithm Based on
SVD”, Journal of Image and Graphics vol. 9, no. 4, pp. 506-512, 2004
[15] X. H. Ma and X. F. Shen, “A Novel Blind Grayscale Watermark Algorithm Based on SVD”, in
International Conference on Audio, Language and Image Processing, 2008, pp.1063-1068.
Author
Md. Maklachur Rahman will receive B.Sc. degree in Computer Science and
Engineering (CSE) from Chittagong University of Engineering and Technology
(CUET), Chittagong, Bangladesh in July, 2013. Currently he is a final year
student Dept. of CSE. His research interest includes Digital Image Processing,
Multimedia Security, Artificial Intelligence, Human Computer Interaction,
Digital Watermarking and Software Engineering.

A DWT, DCT AND SVD BASED WATERMARKING TECHNIQUE TO PROTECT THE IMAGE PIRACY

More Related Content

What's hot (17)

Similar to A DWT, DCT AND SVD BASED WATERMARKING TECHNIQUE TO PROTECT THE IMAGE PIRACY (20)

Recently uploaded (20)

A DWT, DCT AND SVD BASED WATERMARKING TECHNIQUE TO PROTECT THE IMAGE PIRACY