Professional Documents
Culture Documents
watermarking technique have been developed in past to protect the data from ill
egal copies, any forgery with data or prevent copyright. This technique is propo
sed for English language text either in use textual or image watermark. There ar
e many ways to forgery with data as like insertion, deletion or recorded a copy
of data. The existing text watermarking algorithms are not robust against random
tampering attacks (insertion, deletion or reordering of words). Watermarks comp
osed of both image and text make the text secure and has better robustness [1].T
ext document watermarking is shown in fig 1.
Fig1. Text is hidden in image In this figure we add text in image that is not po
ssible to find out data. Attacker confuse at this point, Is this only image or i
mage with text?
International Journal of Computer Application and Technology
(IJCAT)
Volume 1 Issue 1 (April 2014)
ISSN: 2349-1841
17 www.technopublications.com
There are two types of digital watermarking: visible (perceptible) and invisible
(imperceptible). Invisible watermarking, watermarks are embedded in such a way
that they are visible when the content is viewed. Invisible watermarks cannot be
seen but recovering of watermark is possible with an appropriate decoding algor
ithm.
II. RELATED WORK
Kumar et al (2012) proposed a unique technique which combines and uses DCT DWT &
SVD algorithm for the watermarking technique. They have used the base image for
the content storage and the text for the merging part. The approach is quite ap
preciable and its algorithm is explained is quite clearly explained in the paper
. The only drawback in this paper is that it has not been used for the medical i
mage which has been provided as the future scope of this paper [1]. Surya Pratap
Singh (2012) introduces a combinational algorithm for the color and gray scale
image watermarking. This paper focuses on time based watermarking. The time base
d watermarking indicates the significance of how effectively the content has bee
n encrypted and merged into the base content. They describe that if the time of
retrieval is more, the performance is good. The work is appreciable and can be u
seful for the future research works [2]. Sharma et al (2012) presents a new prop
osal for hiding a logo-based watermark in color still image. This dodge is based
on averaging of central frequency coefficients of block Discrete Cosine Transfo
rm (DCT) coefficients of an image. It is unique from earlier dodge based on midd
le frequency coefficient by mean of high redundancy, to nurture malicious attack
s. Here we propose algorithm of aliquot watermarking technique based on DCT (Dis
crete Cosine Transformation) using mid band for robustness. Through adjusting th
e block DCT coefficient of the image the watermarks are numbered [3].
Fezollari et al (2012) concludes that some of these algorithms are robust agains
t some attacks but not against some others.SVD based algorithms are new and dema
nd continuous improvements. In their approach they provided some SVD based algor
ithms and furthermore suggested an idea of a DCT SVD based algorithm. The modifi
ed method is applicable for some attacks compare with Sverdlov Method. Moreover
a digital watermarking scheme was implemented in the Xilinx Virtex II Pro platfo
rm in which the alternation of the watermarking algorithm from encoder to decode
r is sufficient for the scheme to be dynamic [4]. Priyanka Sharma (2012) discuss
es about the quality management while the operation performance of watermarking.
She studied that there are a lot of algorithms which can be useful for the wate
rmarking but it is important to maintain the quality of the base content. She ha
s used the term entropy to find the quality of the base image. Basically it is t
he disorder which occur when so ever we perform the operation [5]. Qing-Cheng Li
(2012) presents a novel text watermarking technique but for the Chinese text on
ly. In his approach he has discussed the bit stream pattern of the text on the b
asis of which the merging can be done .His approach also describes the pictograp
hic approach of the text and the visual ability of the user. The approach is alt
hough has an effective thinking but the problem is as the Chinese language is so
sophisticated, it fits there but not with every language. His algorithm has bee
n designed especially for Chinese characters and hence this algorithm cannot be
used for global language [6]. Aziz Shahid et al (2010) presented a zero text wat
ermarking scheme in the international conference of 2010. According to them, Exi
sting text watermarking algorithms are not robust against random insertion and d
eletion attacks on text. With the increasing volume of attack, the survival of w
atermark in text becomes challenging and hence they developed a novel text water
marking algorithm that can be used for copyright protection of textual contents.
They compared their results with other existing algorithm of the same contrast
and their results are found to be effective enough to get proceeded for modifica
tion [7]. Makarand L. Mali (2012) presented a watermarking scheme on the basis o
f Natural Language Processing. It was a fantastic idea to introduce Neural Netwo
rks into the contrast of encryption. The Neural Network generates weight for eac
h and every input provided to it rather than taking everything as an input strea
m. The pattern changing of neural network is quite similar to SVM as it also con
verts the entire input according its simplification and then proceeds. Hence his
method is quite effective and can be considered for future development process
[8].
Nidhi Divecha (2012) presented a watermarking scheme based on the wavelet quanti
zation method which is again an appreciable effort in this filed. DWT stands for
Discrete Wavelet Transformation and it converts the entire data scenario into w
aves. Proceeding
International Journal of Computer Application and Technology
(IJCAT)
Volume 1 Issue 1 (April 2014)
ISSN: 2349-1841
18 www.technopublications.com
the texts as wave is a unique method in this type of implementation. The effort
done by Nidihi had only one drawback , she did not mention the type of wavelet t
ransformation she is using as there are a lot of wavelet transformation like Dab
uchi, Symlet and others and hence her method can be tried with the above mention
ed wavelet family members [9]. Fahim Irfan Alam (2012) introduced the concept of
signature in his scheme of watermarking. The signature system is bit similar to
the private and public key concept in which the public key is visible to all bu
t it requires a private key to get unlocked. If this method is tried under invis
ible watermarking concept, it is fine but if it is used as a visible watermarkin
g concept, the motive of hiding of data remains untouched as the user would be a
ble to identify easily that some data is hidden behind the encrypted text [10].
III. TEXT WATERMARKING BASED ON DCT
DCTs are used to convert data into the summation of a series of cosine waves osc
illating at different frequencies. They are widely used in image and audio compr
ession. They are very similar to Fourier Transforms, but DCT involves the use of
just Cosine functions and real coefficients, whereas Fourier Transformations ma
ke use of both Sins and Cosines and require the use of complex numbers. DCTs are
simpler to calculate. Both Fourier and DCT convert data from a spatial-domain i
nto a frequency-domain and their respective inverse functions convert things bac
k the other way. The JPEG (Joint Photographic Experts Group) format uses DCT to
compress images. Frequency components of DCT block should be high frequency, bec
ause of the heavy quantization of coefficients during JPEG compression. Hence it
is better to embed the watermark in mid or high frequency DCT components. If th
e embedding factor M is chosen small, embedding the watermark in lowest frequenc
y components will be more desirable, because these components are the ones that
are least likely to be quantized in JPEG compression. The flow chart of the algo
rithm is shown in figure 2.
Fig. 2: The DCT algorithm flow chart
IV. TEXT WATERMARKING BASED ON DWT
Wavelet domain is a promising domain for watermark embedding. Wavelet refers to
small waves. Discrete Wavelet Transform is based on small waves of limited durat
ion and varying frequency [3]. This is a frequency domain technique in which fir
stly cover image is transformed into frequency domain and then its frequency coe
fficients are modified in accordance with the transformed coefficients of the wa
termark and watermarked image is obtained which is very much robust. DWT decompo
ses image hierarchically, providing both spatial and frequency description of th
e image. Then the image is decomposing in three directions i.e. vertical, horizo
ntal and diagonal in results to separate different components name as LL, HL, LH
, and HH. Here letter L and H refers to applying either low pass frequency opera
tion or high pass frequency operations to the rows and the second letter refers
to the
International Journal of Computer Application and Technology
(IJCAT)
Volume 1 Issue 1 (April 2014)
ISSN: 2349-1841
19 www.technopublications.com
filter applied to the columns of the cover image [4]. As shown in fig 3.
Fig.3 Two-level DWT Algorithm for carrying out DWT: Step1. By using DWT, we deco
mpose the original image in some sub-bands. Step2. In this step our work is to c
hoose a Sub-band for embedding watermark that should be most suitable. Step3. Th
e Wavelet coefficients of the selected sub-band are modified according to the wa
termark image. Step4. After embedding watermark, watermarked image is obtained.
V. OTHER TECHNIQUES
There are many other techniques excluding DCT and DWT for watermarking which are
as follows:
a) Singular Value Decomposition (SVD)
Singular Value Decomposition transform is a linear algebra transform which is us
ed for factorization of a real or complex matrix with numerous applications in v
arious fields of image processing [5]. As a digital image can be represented in
a matrix form with its entries giving the intensity value of each pixel in the i
mage, SVD of an image M with dimensions m x m is given by: M=USVT Where, U and V
are orthogonal matrices and S known as singular matrix is a diagonal matrix car
rying non-negative singular values of matrix M. There are two main properties of
SVD to employ in digital watermarking schemes [6]:
1. Quality of image does not affect by Small variations in singular values
2. Singular values of an image have high stability so; they do not change after
various attacks.
b) LEAST SIGNIFICANT BIT (LSB) This is invisible digital watermarking technique,
in which each 8-bit pixel s least significant bit is overwritten with a bit from
the watermark. In a digital image, information can be inserted directly into eve
ry bit of image information or the more busy areas of an image can be calculated
so as to hide such messages in less perceptible parts of an image. For example,
assume the image uses a 24-bit RGB color model (Figure 4). The first image pixe
l has a teal green color (0x008080), while the first watermark byte is 0xDC. Fir
st, the image pixel is divided into its constituent bytes of red, green, and blu
e as shown in figure 4. 0x008080 => r: 0x00, g: 0x80, b: 0x80 Then the watermark
byte is divided into pairs of bits. Next embed each bit pair into bits 1 and 0
of each color byte. So the red byte (0x00) becomes 0x03, the green byte (0x80) b
ecomes 0x81, and the blue byte (0x80) becomes 0x83. Then use these modified byte
s to recolor the image pixel. The new pixel color, however, seems to be the same
teal color (Figure 4).
Fig. 4 Example of LSB
Fig. 4 Encoding of watermark bits
International Journal of Computer Application and Technology
(IJCAT)
Volume 1 Issue 1 (April 2014)
ISSN: 2349-1841
20 www.technopublications.com
VI. COMPARITIVE RESULT
Comparative result of DCT and DWT need GUI first of all. So we develop DCT GUI a
nd DWT GUI as shown in fig 5.1 and fig 5.2:
Fig.5.1 GUI for DWT Based Watermarking
Fig.5.2 GUI for DCT Based Watermarking There is evolution of results on the basi
s of three images as shown in GUI. Table I show results for imperceptibility.