Scribd Upload a Document

Top of Form

Search Documents
Bottom of Form

Explore

ssinghjnp20

/ 6
Top of Form

Bottom of Form

Download this Document for Free

Volume 2 No. 1 ISSN 20798407

Journal of Emerging Trends in Computing and Information Sciences

2010-11 CIS Journal. All rights

reserved.http://www.cisjourn al.org

37

AN OVERVIEW OF TRANSFOR

M DOMAIN ROBUST DIGITALIM AGE WATERMA RKING

ALGORITH MS
Baisa L. Gunjal
Department of Computer Engineering,Amrutvahin i College of Engineering,Sangamner, Dist:Anagar, MS, India.hello_baisa@yaho o.com

R.R. Manthalkar
Department of Electronics and TelecommunicationSGG S Institute of Engineering and Technology, Nanded, MS, Indiarmanthalkar@yaho o.com

ABSTRACT

Internet and Multimedia technologies have become our daily needs. Hence it has become a common practice to createcopy, transmit and distribute digital data. Obviously, it leads to unauthorized

replication problem. Digital imagewatermarking provides copyright protection to image by hiding appropriate information in original image to declarerightful ownership. Aim of this paper is to

provide complete overview of Digital Image watermarking. The study focuseson quality factors essential for good quality watermarking, Performance evaluation metrics (PSNR and CorrelationFactors)

and possible attacks. Overview of several methods with spatial and Transform Domain watermarking is done withdetail mathematical formulae, their implementations, strengths and

weaknesses. The generalized algorithms are presentedfor DWT, CDMA based, DCTDWT combined approach. The Ridgelet Transform is also introduced. Comparative resultsof Digital Image

Watermarking using LSB, DCT and DWT are also presented. The paper recommends DWT basedtechniques for achieving Robustness in Digital Image Watermarking.
Keywords:

Replication, CDMA, DCT-DWT, Ridgelet Transform.

1. I
NTRODUCTION

Now-a-days, Internet and Multimediatechnologi es have become our daily needs. Hence it has become a common practice to create copy, transmit anddistribute digital data. Obviously, it leads to

unauthorizedreplicati on problem. Digital image watermarking providescopyright protection to image by hiding appropriateinformatio n in original image to declare rightfulownership [1]

There are four essential factors those arecommonly used to determine quality of watermarkingscheme. They are robustness, imperceptibility, capacity,and blindness.

Robustness: Watermark should be difficult to removeor destroy. Robust is a measure of immunity of watermark against attempts to image modification andmanipulation like compression, filtering,

rotation,scaling, collision attacks, resizing, cropping etc.

Imperceptibility: means quality of host image shouldnot be destroyed by presence of watermark.

Capacity: It includes techniques that make it possible toembed majority of information.

Blind Watermarking: Extraction of watermark fromwatermarked

image without original image is preferred because sometimes its impossible to availoriginal image.

2.

CLASSIFICATIO N

OFWATERMAR KING TECHNIQUES

Watermarking algorithms can be classified onseveral criteria as given below:

According to domain of watermark insertion:i) Spatial Domain Watermarking Techniquesii) Transform or Frequency Domain Techniques.

According to Watermark detection and extraction:i) Blind Watermarkingii) Non-blind Watermarking.The non-blind watermarking requires that originalimage to exist

for detection and extraction whereas blindtechniques do not require original image.

According to ability of watermark to resist attack:i) Fragile Watermarkingii)

Semi-fragile Watermarking

According to Visibility:i) Visible Watermarksii) No visible Watermarks

3.

PERFORMANCE EVALUATION OFDIGITAL IMAGE WATERMARKI NGALGORITHM S.

Performance Evaluation of

watermarkingalgorith m is done by two performance evaluation metrics:Perceptual transparency and Robustness. [3]. Perceptualtransparenc y means perceived quality of image should not be

destroyed by presence of watermark. The quality of watermarked image is measured by PSNR (Peak signal to Noise Ratio). Bigger is PSNR, better is quality of

Volume 2 No. 1 ISSN 20798407

Journal of Emerging Trends in Computing and Information Sciences

2010-11 CIS Journal. All rights reserved.http://www.cisjourn al.org

watermarked image. PSNR for image with size M x N isgiven by:

38

(1)Where, f (i, j) is pixel gray values of original image.f (i, j) is pixel gray values of

watermarked image. Max

I

isthe maximum pixel value of image which is equal to 255for gray scale image where pixels are represented with 8 bits. In general, Watermarked Images

with PSNR morethan 28 are acceptable. Robustness is measure of immunity of watermark against attempts to remove or destroy it by image modification and manipulation likecompression, filtering, rotation,

scaling, collision attacks,resizing, cropping etc. It is measured in terms of correlation factor. The correlation factor measures thesimilarity and difference between original watermark andextracted

watermark. It value is generally 0 to 1. Ideally itshould be 1 but the value 0.75 is acceptable. Robustness isgiven by:(2)Where, N is number of pixels in watermark, wi is originalwatermark,

wi is extracted watermark.

4.

DIFFERENT POSSIBLE ATTACKS

PSNR and Normalized Correlation

of Watermarked Image should be checked for differentattacks: Following are possible attacks: i] Noise Additionii] Rotation iii] Cropping iv] Scaling v] Resizing vi]Compression

5. OVERVIEW OF DIGITAL IMAGEWATER MARKING TECHNIQUES INSPATIAL AND TRANSFORM

DOMAIN5.1 Spatial Domain Watermarking

Early watermarking schemes were introduced inthe spatial domain, where watermark is added bymodifying pixel values of host image.

Least SignificantBit insertion is example of spatial domain watermarking.But such algorithms have low information hiding capacity,they can be easily discovered and quality of watermarkedimage

and extracted watermark is not satisfactory as pixelintensities are directly changed in these algorithms . Onesample Spatial Domain watermarking algorithm withLeast

Significant Bit insertion is given below. Anywatermarking algorithm has two parts: embeddingalgorithm and detection

(extraction) algorithm.

5.1.1.

Watermark Embedding

Figure: 1 Pixel of Cover image (Original Image),Watermark,

Watermarked Image and ExtractedWatermarkS tep 1

Read gray scale Cover Image and Watermark. Step 2 :Consider binary of pixel values of Cover

Imageand make its n Least Significant Bits 0e.g. For n=4, Binary of 143=>10001111 and Making 4LSB 0 =>10000000=>128 is decimal equivalent. Step 3: Consider binary of pixel values of Watermark andright

shift by k bits where k=8-n. For n=4, k will be 4.Binary of 36=>100100 and after right shift by 4:000010=>2 is decimal equivalent Step 4: Add result of step 1 and step 2 to givewatermarked

image. E.g. Add 128+2=>130. This gives pixel value of watermarked image=>10000010. Here, pixel values or cover image get changed. Hence Quality of watermarked image is degraded.

5.1.2. Watermark Extraction

Take pixels of watermarked Image and left shift by k bits where k=8-n. e.g. Left shift by 4=>00100000=>32.

This gives pixels of Extracted Watermark. Thesample value of Pixel of Cover image, Watermark,Watermar ked Image and Extracted Watermark are shownin figure.1.LSB based watermarking in spatial domain is

thestraightforward method, but once algorithm is discovered,watermark will be no more secured.[3].An improvementon LSB substitution is to use pseudo random generator todetermine pixels to be used for

embedding, based on givenseed or key. Security can be improved but algorithm is notstill such techniques are not completely secured [3].

5.2 Transform (Frequency)

DomainWatermar king

Volume 2 No. 1 ISSN 20798407

Journal of Emerging Trends in Computing and Information Sciences

2010-11 CIS Journal. All rights reserved.http://www.cisjourn al.org

The Frequency domain the watermark is insertedinto transformed coefficients of image

giving moreinformation hiding capacity and more robustness againstwatermarking attacks [1]. Watermarking in frequencydomain is more robust than watermarking in spatialdomain

because information can be spread out to entireimage. For frequency transform, we are having differentoptions like: Fourier Transform (FT), Short Time Fourier Transform (STFT), and Continuous Wavelet

Transform(CWT), Discrete Wavelet Transform (DWT), DiscreteCosine Transform (DCT) or Combination of DCT andDWT.

39 5.2.1 Digital Image Watermark

usingDiscrete Wavelet Transform.

DWT has become researchers focus for watermarking as DWT is very similar to theoretical modelof Human Visual System (HVS). ISO

has developed andgeneralized still image compression standard JPEG2000which substitutes DWT for DCT. DWT offersmutiresolution representation of a image and DWT gives perfect

reconstruction of decomposed image. Discretewavelet can be represented as(3)For dyadic wavelets a

0

=2 and
b0

=1, Hence we have, j, k (4)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

bedone at different DWT levels. DWT offers mutiresolutionreprese ntation of a signal. One Level DWTDecomposition is given in figure 2. Figure 2: One Level Image Decomposition

It has been widely accepted that maximum energy of mostof natural images is concentrated in approximate (LL)subband which is low frequency subband. Hencemodification to the coefficients of

these low frequencysubbands would cause severe and unacceptable imagedegradation. Hence, we do not embed watermark in LLsubband. The good areas for watermark embedding arehigh frequency sub bands

(vertical, horizontal anddiagonal components). The human naked eyes are notsensitive to these high frequency sub bands. So, effectivewatermark embedding is achieved without being perceived by

human eyes. The generalized DWT basedwatermarking is shown in figure-3. Actually, embeddingwill include specific algorithmic steps those are to beimplemented in wavelet domain.

Figure: 3 Generalized DWT based Watermarking Scheme

Many of the researchers have developedwatermarki ng algorithms using DWT. But most of them are failed to provide perceptual

transparency androbustness simultaneously, since these twowatermarking requirements are conflicting to eachother. Na Li et. all proposed a DWT based method[14] in which watermark was

embedded in middlefrequency coefficient using : (5)where, F(u,v) is wavelet coefficient, W is randomized binary watermarking, is flexing factor with

= |m|where m is mean value of all coefficients watermarkingembedd ed. But this method doesnt provide enoughsecurity.

The method proposed in [14] using DWT wasextended in [15] to enhance security of algorithm byusing Arnolds Transform pretreatment for watermark. The magic property of Arnold Transformis that, when it is

applied to image, after specificnumber of iterations, image comes to its originalstate. Arnold Transform is given by:(6)Where, (x,y) ={0,1,.....N} are pixel coordinates fromoriginal image. (x', y') are

corresponding results after Arnold Transform.

As given in [16] two phase watermark embedding process was carried out using DWT. Phase1: Visiblewatermark

logo embedding, Phase2: Feature extractedwatermark logo embedding. The algorithm was basedon Texture Based Watermarking.

In 2007, A Integer Wavelet Transform

with Bit Planecomplexity Segmentation is used with more datahiding capacity. [2]. But this method needs separate processing for R, G and B components of color image.

In 2008, as given in [17] using DWT, host image isdecomposed into 3 levels recursively. In level one we

An Overview of Transform Domain Robust Digital Image Watermarking Algorithms

Download this Document for FreePrintMobileCollectionsReport Document

Info and Rating

rplication CDMA dctdwt ridgelet transform. saeedullah81

Share & Embed More from this user

PreviousNext 1.

8 p.

6 p.

9 p. 2.

6 p.

10 p.

3.

5 p.

9 p.

8 p. 4.

9 p.

5 p.

10 p. 5.

7 p.

9 p.

4 p. 6.

4 p.

7 p. 7.

6 p.

3 p.

6 p. 8.

6 p.

5 p.

6 p. 9.

6 p.

Add a Comment
Top of Form

Bottom of Form

Upload a Document

Top of Form

Search Documents
Bottom of Form

Follow Us! scribd.com/scribd twitter.com/scribd facebook.com/scribd About Press Blog Partners Scribd 101 Web Stuff Scribd Store Support FAQ Developers / API Jobs Terms Copyright Privacy

Copyright 2011 Scribd Inc. Language: English