International Journal of Emerging Trends in Engineering and Technology (IJETET)

Vol. I No. 1, 2011

ISSN 2248-9592
IJETET21

WAVELET BASED COLOR IMAGE

WATERMARKING SCHEME GIVING
HIGH ROBUSTNESS AND EXACT
CORELATION
Baisa L. Gunjal
Amrutvahini College of Engineering, Sangamner, Anagar, MS, India.
E_mail: hello_baisa@yahoo.com

ABSTRACT- A good quality watermarking scheme should have maximum PSNR, ideally Correlation Factor equals to 1
and should have maximum watermark information hiding capacity. In this paper, a robust digital color image
watermarking scheme based on Discrete Wavelet Transform is implemented. The RGB color space is converted into
YUV color space and then V component is selected for watermark embedding. The scrambled watermark is embedded in
middle frequency sub band. The security levels are increased by generating PN sequence depending on periodicity of
watermark image. The image scrambling is applied by Arnold Transform. The decomposition is done with Haar which
is simple, symmetric and orthogonal wavelet and the direct weighting factor is used in watermark embedding and
extraction process. The scheme results in exact recovery of watermark with standard database images of size 512x512,
giving Correlation Factor equals to 1. The PSNR with weighting factor 0.02 is up to 48.51 dBs. The presented scheme is
non blind and embeds binary watermark of 64x64 size.

Keywords - Discrete Wavelet Transform, Color Image, Scrambling, Color space.

1. INTRODUCTION
Digital image watermarking provides copyright protection to image by hiding appropriate information in
original image to declare rightful ownership [1]. Robustness, Perceptual transparency, capacity and Blind
watermarking are four essential factors to determine quality of watermarking scheme [4][5]. In spatial domain,
watermark is embedded by directly modifying pixel values of cover image. Least Significant Bit insertion is
example of spatial domain watermarking. These algorithms are simple in implementation. But problems with
such algorithms are: Low watermark information hiding capacity, Less PSNR, Less Correlation between
original and extracted watermark and less security, hence anybody can detect such algorithms. The Frequency
domain the watermarking inserted into transformed coefficients of image giving more information hiding
capacity and more robustness against watermarking attacks because information can be spread out to entire
image [1]. Digital Color Image Watermarking is essential in Digital age which is implemented here giving
robustness. The rest of the paper is organized as follows: Section 2 focuses on survey of existing digital gray
and color image watermarking algorithms. Section 3 focuses on RGB and YUV Color Spaces. In Section 4,
importance of Discrete Wavelet Transform is explored. In section 5, watermarking methodology is given.
Section 6 shows experimental results after implementation and Testing and the conclusion is drawn in section 7.

International Journal of Emerging Trends in Engineering and Technology (IJETET)
Vol. I No. 1, 2011
ISSN 2248-9592
IJETET22

2. SURVEY
LSB based watermarking in spatial domain is the straightforward method, But LSB based watermarking as
well as LSB based watermarking with pseudo random generator are not secured methods.[3]. In CWT,
Calculating wavelet coefficients at every possible scale is huge amount of work, and it generates a lot of data.
There is highly redundant information as per as the reconstruction of the signal is concerned. Due to the
attractive features of Discrete Wavelet Transform, researches have been focused on DWT [15]. Wang
Hongjun, Li Na have proposed a DWT based method [14] in which watermark was embedded in middle
frequency coefficient using as flexing factor with = |m| ,where m is mean value of all coefficients
watermarking embedded. But this method doesnt provide enough security. The method proposed in [14] using
DWT was extended in [15] to enhance security of algorithm by using Arnolds Transform pretreatment for
watermark. But this method can be extended to improve PSNR and security levels. A Integer Wavelet
Transform with Bit Plane complexity Segmentation is used with more data hiding capacity, for color images
[2]. But this method needs separate processing for R, G and B components of color image and technique is
time consuming. As given in [17] using DWT, host image is decomposed into 3 levels recursively. In level one
we get 4 sub bands. In level 2, each sub band of level 1 is divided to 4 sub bands to give total 16 sub bands.
Finally, each sub band of level 2 is again divided into 4 sub bands each to give total 64 sub bands. Then
Generic algorithm was applied to find the best sub band for watermark embedding to provide perceptual
transparency and robustness. But the process is too lengthy and time consuming. The common problem with
DCT watermarking is block based scaling of watermark image changes scaling factors block by block and
results in visual discontinuity.[1][6]. As given in [13], J. Cox et. al had presented Spread spectrum based
watermarking schemes, Chris Shoemaker has developed CDMA based Spread spectrum watermarking with
one scale DWT and got PSNR between 35-40 db for various attacks. As given in [10], Harsh Varma et. al
tested CDMA based watermarking scheme with spatial domain and frequency domain with DCT as well as
DWT. But these algorithms have low information hiding capacity.

3. RGB AND YUV COLOR SPACES
The RGB color space is converted into YUV Color space and then Watermark is embedded into V Component
of YUV color space. Initially color image is read and R, G, B components of original Cover Image are
separated. Then they are converted into YUV color Space using following equations.
(1)
International Journal of Emerging Trends in Engineering and Technology (IJETET)
Vol. I No. 1, 2011
ISSN 2248-9592
IJETET23

(2)
(3)
After embedding the watermark using DWT, YUV color space is converted back into RGB color space using
following equations.
R (4)
(5)
(6)
4. DISCRETE WAVELET TRANSFORM (DWT)
DWT has become researchers focus for watermarking as DWT is very similar to theoretical model of Human
Visual System (HVS). ISO has developed and generalized still image compression standard JPEG2000 which
substitutes DWT for DCT. DWT offers mutiresolution representation of a image and DWT gives perfect
reconstruction of decomposed image. Discrete wavelet can be represented as
(8)
For dyadic wavelets a
0
=2 and

b
0
=1, Hence we have,
j, k (9)
Image itself is considered as two dimensional signal. When image is passed through series of low pass and high
pass filters, DWT decomposes the image into sub bands of different resolutions [11][12]. Decompositions can
be done at different DWT levels. As shown in Fig: 1, At level 1, DWT decomposes image into four non
overlapping multiresolution sub bands: LLx (Approximate sub band), HLx (Horizontal sub band), LHx (Vertical
sub band) and HHx (Diagonal Sub band). Here, LLx is low frequency component whereas HLx, LHx and HHx
are high frequency (detail) components [7][8][9]. To obtain next coarser scale of wavelet coefficients after level
1, the sub band LL1 is further processed until final N scale reached. When N is reached, we have 3N+1 sub
bands with LLx (Approximate Components.) and HLx, LHx, HHx (Detail components) where x ranges from 1
to N. Three level image Embedding watermark in low frequency coefficients can increase robustness
significantly but maximum energy of most of the natural images is concentrated in approximate (LLx) sub band.




International Journal of Emerging Trends in Engineering and Technology (IJETET)
Vol. I No. 1, 2011
ISSN 2248-9592
IJETET24

Fig 1: Three Level Image Decomposition
Hence modification in this low frequency sub band will cause severe and unacceptable image degradation.
Hence watermark is not embedded in LLx sub band. The good areas for watermark embedding are high
frequency sub bands (HLx, LHx and HHx), because human naked eyes are not sensitive to these sub bands.
They yield effective watermarking without being perceived by human eyes. But HHx sub band includes edges
and textures of the image. Hence HHx is also excluded. Most of the watermarking algorithms have been failed
to achieve perceptual transparency and robustness simultaneously because these two requirements are
conflicting to each other. The rest options are HLx and LHx. But Human Visual System (HVS) is more sensitive
in horizontal than vertical. Hence Watermarking done in HLx region.

5. WATERMARKING METHODOLOGY
The following presented methodology is improvement of algorithm presented in 2010 by B.L.Gunjal et al,
given in [1] using Discrete Wavelet Transform with Arnold Transform.
The improvement is made for color image watermarking.

5.1 Watermark Embedding
A) Watermark Scrambling:
Watermark Scrambling is carried out through many steps to improve security levels. Different methods can be
used for image scrambling such as Fass Curve, Gray Code, Arnold Transform, Magic square etc. Here Arnold
Transform is used. The special property of Arnold Transform is that image comes to its original state after
certain number of iterations. These number of iterations is called Arnold Period or Periodicity of Arnold
Transform. The Arnold Transform of image is
(10)

LL3 HL3
HL2
HL1
LH3 HH3
LH2 HH2
LH1 HH1
International Journal of Emerging Trends in Engineering and Technology (IJETET)
Vol. I No. 1, 2011
ISSN 2248-9592
IJETET25

Where, (x, y) = {0,1,.....N} are pixel coordinates from original image. ( , ) are corresponding results after
Arnold Transform. The periodicity of Arnold Transform (P), is dependent on size of given image. From
equation: 3 we have,
=x+y (11)
=x+2*y (12)
If (mod ( , N) ==1 && mod ( , N) ==1) then P=N (13)

B) Embedding Algorithm:
First covert original color cover image into YUV color space and consider V_Component , say V_Cover
Image for embedding. Here, Watermark is 64X64 size gray scale image.
Step 1: Decompose the cover image using simple Haar Wavelet into four non overlapping multiresolution
coefficient sets: LL1, HL1, LH1 and HH1.
Step 2: Perform second level DWT on LL1 to give 4 coefficients: LL2, HL2, LH2 and HH2.
Step 3: Repeat decomposition for LL2 to give next level components: LL3, HL3, LH3 and HH3 as shown in
fig 1.
Step 4: Find Arnold periodicity P of watermark using equation 6.
Step 5: Determine KEY where . Then generate PN Sequence depending on KEY and find the
sum of random sequence say SUM.
Step 6: If SUM > T where, T is some predefined Threshold value, then find two scrambled images applying
Arnold Transform with KEY1 and KEY2, where, , ,
.
Now, Take absolute difference of two scrambled images to give Final Scrambled image.
Step 7: If SUM <T, then apply Arnold Transform directly to watermark image with KEY to get Final
Scrambled image.
Step 8: Add Final Scrambled image to HL3 coefficients of cover image as follows:
(14)
Where, K1 is weighting factor, New_HL3 (i, j) is newly calculated coefficients of level3, Watermark (i, j) is
Final Scrambled image.
Step 9: Take IDWT at Level3, Level2 and Level1 sequentially to get V_New Component i.e. V_Component
with watermark embedded into it.
International Journal of Emerging Trends in Engineering and Technology (IJETET)
Vol. I No. 1, 2011
ISSN 2248-9592
IJETET26

Step 10: Combine Y, U, and V_New Component to get YUV Image and convert into RGB color space. This
gives Watermarked Color Image,



























Key2=Key-Count
Yes
V_Cover Image
(V Component)
3 Levels
DWT
Watermark Arnold
Periodicity (P)
Inverse 3
Levels DWT
V_New
Component
HL3 Region
PN
Sequence
Final Scrambled
Image (F)
Key (K)
Embedding
Algorithm
Scrambled
Image (I2)
Scrambled
Image (I1)
Absolute Difference
between I1 and I2 gives F
Key1=Key + Count
If
SUM>T
Arnold Transform
No, Then find F
by Arnold Directly
R,G,B to Y,U,V
color space
Original Color
Cover Image
Convert to
R,G,B
Watermarked
Color Image
Fig 2: Watermark Embedding Process
Watermarked
Color Image
3 Levels
DWT (HL3)
Original Color
Cover Image
3 Levels
DWT (HL3)
Extracted
Watermark
Recovered
Watermark
Arnold
Transform
Extraction
Algorithm
RGB to YUV
Color Space
RGB to YUV
Color Space
V_Watermarked
Component
V_Cover
Component
Fig 3: Watermark Extraction Process
International Journal of Emerging Trends in Engineering and Technology (IJETET)
Vol. I No. 1, 2011
ISSN 2248-9592
IJETET27

5.2 Extraction Algorithm:
The proposed method is non blind. Hence the original image is required for extraction process. The simple
algorithmic steps are applied are given below. First consider Watermarked Color Image and Original Color
Cover Image. Convert them in to YUV Color space and consider their V_Components say V_Watermerked
Component and V_Cover Component respectively. Now follow steps given below for extraction process.
Step 1: Decompose V_Cover Component using Haar wavelet up to 3 levels to get HL3 Coefficients.
Step 2: Decompose V_Watermarked Component using Haar wavelet up to 3 levels to get HL3.
Step 3: Apply Extraction formula as follows:
(15)
If

Otherwise
Step 4: Perform Image Scrambling using Arnold Transform with KEY that we had used in embedding
process to recover the Watermark.

6. EXPERIMENTAL RESULTS AFTER IMPLEMENTATION AND TESTING
The project is implemented in Matlab and standard Lena Color image with 512x512 size as cover image and
64x64 size binary watermark images are used for testing. The performance Evaluation is done by two
performance evaluation metrics: Perceptual transparency and Robustness.
Perceptual transparency means perceived quality of image should not be destroyed by presence of watermark.
The quality of watermarked image is measured by PSNR

Bigger is PSNR, better is quality of watermarked image. PSNR for gray scale image with size M x N is:
(16)
Where e(x, y) is the difference between Original Image and Watermarked Image. With size M*N,
corresponding to color image, PSNR is expressed as,
(17)
Here, c indicates the color component which e belongs to. Robustness is measure of immunity of watermark
International Journal of Emerging Trends in Engineering and Technology (IJETET)
Vol. I No. 1, 2011
ISSN 2248-9592
IJETET28

against attempts to remove or destroy it by image modification and manipulation like compression, filtering,
rotation, scaling, collision attacks, resizing, cropping etc. It is measured in terms of correlation factor. The
correlation factor measures the similarity and difference between original watermark and extracted watermark.
It value is generally 0 to 1. Ideally it should be 1 but the value 0.75 is acceptable. Robustness is given by:
(18)
Here, N is number of pixels in watermark, wi is original watermark, wi is extracted watermark. Achieved
PSNR is 48.51 dB and =1, for weighting factor K1=0.02. The PSNR and for different weighting factors are
given in Table 2.















Table 1: Experimental results for standard database images with size 512x512




K1=0.02,
PSNR=48.5103

K1=0.03,
PSNR =48.474

K1=0.04,
PSNR=48.4482


K1=0.07,
PSNR=39.0016


K1=0.1,
PSNR =36.4954

Corelation ()
=1
Corelation ()
=1
Corelation ()
=1
Corelation ()
=1
Corelation ()
=1


Recovered
Watermark


Recovered
Watermark


Recovered
Watermark


Recovered
Watermark


Recovered
Watermark

Table 2: Experimental results for values of flexing factor K1 for color Lena image, size 512x512
7. CONCLUSION

















a) Original Cover
Image
b) Original
Watermark
c) Scrambled
Watermark
d) Recovered
Watermark
e) Watermarked
Image
International Journal of Emerging Trends in Engineering and Technology (IJETET)
Vol. I No. 1, 2011
ISSN 2248-9592
IJETET29

Experimental results have demonstrated that technique presented in this paper is very effective supporting more
security and exact correlation between original watermark and extracted watermark. As per ISOs norms, the still
Image Compression standard JPEG2000 has replaced Discrete Cosine Transform by Discrete Wavelet Transform.
This is the why more researchers are focusing on DWT. Here complete theory to implementation approach is
presented for DWT Based Robust Digital Color Image Watermarking which also achieves exact correlation.
The presented Digital color Image Watermarking methodology can be extended for videos for authentication
and copyright protection.
ACKNOWLEDGEMENT
Thanks to Board of Colleges and University Development (BCUD), University of Pune, Maharashtra, India, for
providing Research Grant for the research project Robust Digital Image Watermarking using Discrete Wavelet
Transform for period 2010-2012. I am also thankful to Amrutvahini College of Engineering, Sangamner,
Anagar, MS for providing me a strong platform for this research work.

REFERENCES
[1] B.L.Gunjal, R.R.Manthalkar, Discrete Wavelet Transform Based Strongly Robust Watermarking Scheme
for Information Hiding in Digital Images, Third International Conference- Emerging Trends in
Engineering and Technology,19-21 Nov 2010 , Goa, India, ISBN 978-0-7695-4246-1,
http://doi.ieeecomputersociety.org/10.1109/ICETET.2010.12.
[2] Cheng-qun Yin et al, Color Image Watermarking Algorithm Based on DWT-SVD, Proceeding of the
IEEE International Conference on Automation and Logistiocs August 18-21, 2007, China, PP: 2607-2611
[3] Wei-Min Yang et al, A Watermarking Algorithm Based on Wavelet and Cosine Transform for Color
Image, 2009 First International Workshop on Education Technology and Computer Science, pages: 899-
903
[4] S. Joo, Y. Suh, J. Shin, H. Kikuchi, and S. J. Cho., A new robust watermark embedding into wavelet DC
components, ETRI Journal, 24, 2002, pp. 401-404.
[5] Voloshynovskiy. S. S. Pereira and T. Pun. 2001. Attacks on Digital watermarks: classification, Estimation-
Based attacks and Benchmarks, Comm, Magazine. 39(8):118-126.
[6] Ahmad Nilchi1, Ayoub Taher, A New Robust Digital Image Watermarking Technique based on the
Discrete Cosine Transform and Neural Network, pp:1 7, April 2008.
International Journal of Emerging Trends in Engineering and Technology (IJETET)
Vol. I No. 1, 2011
ISSN 2248-9592
IJETET30

[7] M.S. Hsieh, and D.C. Tseng, Hiding digital watermarks using multi-resolution Wavelet transform,IEEE-
Transactions on industrial electronics, vol:48, No:5, pp:875-882, Oct, 2001.
[8] H. Guo, and N. Georganas, Multi-resolution Image Watermarking Scheme in the Spectrum Domain,
Proceeding of IEEE Canadian Conference on Electrical and Computer Engineering, pp. 873-878, May,
2002.
[9] Wang, S. and Y. Lin, Wavelet Tree Quantization for Copyright Protection Watermarking, IEEE
Trans. Image Processing, vol. 13, no. 2, pp: 154-164, 2004.
[10] Harsh Varma, Abhishek Singh, Raman Kumar ,Robustness of Digital Image Watermarking Techniques
against brightness and rotation attacks, (IJCSIS) International Journal of Computer Science and
Information Security, Vol No:1, 2009.
[11] V.R. Ratnaparkhe, R.R.Manthalkar and Y.V.Joshi, Texture Characterization of CT images Based on
Ridgelet Transform, ICGST International Journal on Graphics, Vision and Image Processing,
GVIP,Volume 8, 2008, pages:43-50.
[12] M.S. Joshi, R.R. Manthalkar, Y.V.Joshi, Image Compression Using Curvelet, Wavelet and Ridgelet
Transform: A Comparative study ,ICGST GVIP, ISSN 1687-398X, Volume 8, Issue III, October 2008,
Pages: 25-34
[13] Ingemar J. Cox, et. al, Secure Spread Spectrum Watermarking for Multimedia, IEEE Trans. on Image
Processing, 6, 12, 1673-1687, (1997).
[14] Wang Hongjun, Li Na, An algorithm of digital image watermark based on multiresolution wavelet
analysis, International Workshop on VLSI Design and Video Technology, Proceedings, 28-30 May 2005,
pp: 272- 275.
[15] Na Li; Xiaoshi Zheng; Yanling Zhao; Huimin Wu; Shifeng Li ,Robust Algorithm of Digital Image
Watermarking Based on Discrete Wavelet Transform, Electronic Commerce and Security, International
Symposium ,Volume, Issue , 3-5 Aug. 2008 , Page(s):942 945.
[16] Krishnan, N. Selvakumar, R.K. Rajapandian, S. Arul Mozhi, K.Nelson Kennedy Babu, A Wavelet
Transform Based Digital Image Watermarking and Authentication, Annual IEEE Conference, in India, 15-
17 Sept. 2006, pp: 1-6.
[17] Abu-Errub, A., Al-Haj, A.,Optimized DWT-based image watermarking, First International Conference
on Applications of Digital Information and Web Technologies, IEEE,2008, 4-6.