The proposed video watermarking scheme consists of two stages

1) Watermark Embedding process

2) Watermark Extraction process

a. Watermark Embedding Process

Before embedding watermark pixels into the input video sequences, the following process should
carried out to enhance the security of the hiding information as well as to improve the efficiency
of our proposed approach. The process includes,

1) Shot segmentation of video sequences

The original input video sequence is first segmented into non-overlapping units, called shots that
depict different actions. Each shot is characterized by no significant changes in its content which
is determined by the background and the objects present in the scene. Here, proposed method
uses Discrete Cosine Transform and correlation measure to identify the number of frames
involved in each shot.

2) Bit plane slicing of grayscale image

Bit-Plane Slicing is a technique in which the image is sliced at different planes. Instead of
highlighting gray level
images, highlighting the contribution made to total image appearance by specific bits might be
desired.

3) Pixel permutation

After the bit plane slicing process, the sliced images are allowed to permute each pixel value to
enhance the
security of the hiding information.
 4) Decomposition of image using DWT

Wavelets are functions defined over a finite interval and having an average value of zero. The
basic idea of the wavelet transform is to represent any arbitrary function as a superposition of a
set of such wavelets or basis functions. The figure 1 shows the watermarking embedding process
as shown below.

a. Watermark Embedding Algorithm

Step1: Segment the original input video sequence into number shots using shot segmentation
technique. Then, identify the number of frames involved in each segmented shots for embedding
purpose.

Step2: Grayscale image. Slice the grayscale watermark image into sliced images using bit plane
slicing.

Step3: Permute the sliced images using pixel permutation technique to obtain the permuted
grayscale image.
 Fig.1. Watermark Embedding process

Step4: Decompose the components of each partitioned frame into four sub-bands such as HH,
HL,LH and LL with the aid of the DWT to attain the transformed frames.
Step5: Choose the low frequency sub-bands (HL,LH) from the transformed frames to embed the
permuted
Step6: Find the similarity matrix of the permuted image to embed into the chosen sub-bands. The
upper part of the similarity matrix is embedded into the HL sub-band and the lower part of the
similarity matrix into the LH sub-band.
Step7: The HL and LH sub-bands used to embed the permuted watermark image are divided into
four parts as per the similarity matrix. The lower part embedding part of the similarity matrix of
the HL and LH bands is chosen for embedding the two similar parts of the watermark image.
Step8: In the HL sub-band, the upper part of the similarity matrix is embedded with the
following steps:

Calculate the mean value and the maximum value mean (Up), max (Ep) of the chosen
embedding part Ep .

Step9: Embed the watermark pixels 0 or 1 in a zigzag manner in the chosen embedding part,
since the watermark is the grayscale image.
Step10: Similarly, the lower part Lp of the similarity matrix is embedded into the LH sub-band.
Also, each permuted image is embedded into all the frames of every shot.
Step11: Divide all the embedded frames with the embedding strength to enhance the quality of
the image.
Step12: Map the modified sub-bands into its original position and apply the inverse discrete
wavelet transform to attain the watermarked video sequence.
b. Watermark Extraction process The figure 2 shows the watermarking extraction process as
shown below.

Fig.2. Watermark Extraction process

1). Watermark Extraction Algorithm

Step1: Extract the components of all the partitioned frames for extracting Segment the
watermarked video sequence into a number of non-overlapping shot using the shot segmentation
technique. Then, identify the number of frames involved in each segmented shots for the
extraction process.
Step2: Extract the components of all the partitioned frames for extracting the embedded
watermark pixels.
Step3: Decompose the components of the frames with the aid of the discrete wavelet transform
into four sub-bands HH, HL, LH and LL.
Step4: Select the low frequency sub-bands (HL, LH) from the transformed frames to extract the
watermark grayscale image.
Step5: Extract the watermark pixels from the embedding part in a zigzag manner from the HL
and the LH sub-bands with the aid of the following steps. If theembedded pixel value is greater
than the mean pixel value, then the extracted pixel value is one. If it is lesser, then the extracted
pixel is zero.

Step6: Form the matrix with the size of the watermark image and the extracted pixels are placed
init to attain the watermark image.
Step7: Obtain the watermark image W [i', j'] I by applying the reverse process of permutation and
bit plane slicing.

Video watermarking

A digital watermark is a kind of marker covertly embedded in a noise-tolerant signal such as

audio or image data. It is typically used to identify ownership of the copyright of such signal.
"Watermarking" is the process of hiding digital information in a carrier signal; the hidden
information should,[1] but does not need to contain a relation to the carrier signal. Digital
watermarks may be used to verify the authenticity or integrity of the carrier signal or to show the
identity of its owners. It is prominently used for tracing copyright infringements and
for banknote authentication. Like traditional watermarks, digital watermarks are only perceptible
under certain conditions, i.e. after using some algorithm, and imperceptible anytime else.[2] If a
digital watermark distorts the carrier signal in a way that it becomes perceivable, it is of no
use.[2] Traditional Watermarks may be applied to visible media (like images or video), whereas
in digital watermarking, the signal may be audio, pictures, video, texts or 3D models. A signal
may carry several different watermarks at the same time. Unlike metadata that is added to the
carrier signal, a digital watermark does not change the size of the carrier signal.

The needed properties of a digital watermark depend on the use case in which it is applied. For
marking media files with copyright information, a digital watermark has to be rather robust
against modifications that can be applied to the carrier signal. Instead, if integrity has to be
ensured, a fragile watermark would be applied.

Both steganography and digital watermarking employ steganographic techniques to embed data
covertly in noisy signals. But whereas steganography aims for imperceptibility to human senses,
digital watermarking tries to control the robustness as top priority.

Since a digital copy of data is the same as the original, digital watermarking is a passive
protection tool. It just marks data, but does not degrade it nor controls access to the data.

One application of digital watermarking is source tracking. A watermark is embedded into a

digital signal at each point of distribution. If a copy of the work is found later, then the
watermark may be retrieved from the copy and the source of the distribution is known. This
technique reportedly has been used to detect the source of illegally copied movies.

Applications

Digital watermarking may be used for a wide range of applications, such as:

Copyright protection
Source tracking (different recipients get differently watermarked content)
Broadcast monitoring (television news often contains watermarked video from international
agencies)

Digital Watermarking life cycle

The information to be embedded in a signal is called a digital watermark, although in some
contexts the phrase digital watermark means the difference between the watermarked signal and
the cover signal. The signal where the watermark is to be embedded is called thehost signal. A
watermarking system is usually divided into three distinct steps, embedding, attack, and
detection. In embedding, an algorithm accepts the host and the data to be embedded, and
produces a watermarked signal.

Then the watermarked digital signal is transmitted or stored, usually transmitted to another
person. If this person makes a modification, this is called an attack. While the modification may
not be malicious, the term attack arises from copyright protection application, where third parties
may attempt to remove the digital watermark through modification. There are many possible
modifications, for example, lossy compression of the data (in which resolution is diminished),
cropping an image or video, or intentionally adding noise.

Detection (often called extraction) is an algorithm which is applied to the attacked signal to
attempt to extract the watermark from it. If the signal was unmodified during transmission, then
the watermark still is present and it may be extracted. In robust digital watermarking
applications, the extraction algorithm should be able to produce the watermark correctly, even if
the modifications were strong. In fragile digital watermarking, the extraction algorithm should
fail if any change is made to the signal.

Robustness

A digital watermark is called fragile if it fails to be detectable after the slightest modification.
Fragile watermarks are commonly used for tamper detection (integrity proof). Modifications to
an original work that clearly are noticeable, commonly are not referred to as watermarks, but as
generalized barcodes.

A digital watermark is called semi-fragile if it resists benign transformations, but fails detection
after malignant transformations. Semi-fragile watermarks commonly are used to detect
malignant transformations.

A digital watermark is called robust if it resists a designated class of transformations. Robust

watermarks may be used in copy protection applications to carry copy and no access control
information.

Perceptibility

A digital watermark is called imperceptible if the original cover signal and the marked signal are
perceptually indistinguishable.

A digital watermark is called perceptible if its presence in the marked signal is noticeable (e.g.
Network Logo, Content Bug, Codes, Opaque images.)

This should not be confused with perceptual, that is, watermarking which uses the limitations of
human perception to be imperceptible.

Capacity

The length of the embedded message determines two different main classes of digital
watermarking schemes:

The message is conceptually zero-bit long and the system is designed in order to detect the
presence or the absence of the watermark in the marked object. This kind of watermarking
scheme is usually referred to as zero-bit or presence watermarking schemes. Sometimes, this
type of watermarking scheme is called 1-bit watermark, because a 1 denotes the presence
(and a 0 the absence) of a watermark.
 The message is an n-bit-long stream , with

or and is modulated in the watermark. These kinds of schemes usually are

referred to as multiple-bit watermarking or non-zero-bit watermarking schemes.

Embedding method

A digital watermarking method is referred to as spread-spectrum if the marked signal is obtained

by an additive modification. Spread-spectrum watermarks are known to be modestly robust, but
also to have a low information capacity due to host interference.

A digital watermarking method is said to be of quantization type if the marked signal is obtained
by quantization. Quantization watermarks suffer from low robustness, but have a high
information capacity due to rejection of host interference.

A digital watermarking method is referred to as amplitude modulation if the marked signal is

embedded by additive modification which is similar to spread spectrum method, but is
particularly embedded in the spatial domain.

Evalution and Benchmarking

The evaluation of digital watermarking schemes may provide detailed information for a
watermark designer or for end-users, therefore, different evaluation strategies exist. Often used
by a watermark designer is the evaluation of single properties to show, for example, an
improvement. Mostly, end-users are not interested in detailed information. They want to know if
a given digital watermarking algorithm may be used for their application scenario, and if so,
which parameter sets seems to be the best.

Camera

Epson and Kodak have produced cameras with security features such as the Epson PhotoPC
3000Z and the Kodak DC-290. Both cameras added irremovable features to the pictures which
distorted the original image, making them unacceptable for some applications such as forensic
evidence in court. According to Blythe and Fridrich, "[n]either camera can provide an
undisputable proof of the image origin or its author"
A secure digital camera (SDC) was proposed by Mohanty, et al. in 2003 and published in
January 2004. This was not the first time this was proposed. Blythe and Fridrich also have
worked on SDC in 2004 for a digital camera that would use lossless watermarking to embed
a biometric identifier together with a cryptographic hash.

Discrete Wavelet Transform

A discrete wavelet transform (DWT) is any wavelet transform for which the wavelets are
discretely sampled. As with other wavelet transforms, a key advantage it has over Fourier
transforms is temporal resolution: it captures both frequency and location information (location
in time).

EXAMPLES
Haar wavelets

The first DWT was invented by the Hungarian mathematician Alfrd Haar. For an input
represented by a list of numbers, the Haar wavelet transform may be considered to simply
pair up input values, storing the difference and passing the sum. This process is repeated
recursively, pairing up the sums to provide the next scale: finally resulting in differences
and one final sum.

Daubechies wavelets

The most commonly used set of discrete wavelet transforms was formulated by the Belgian
mathematician Ingrid Daubechies in 1988. This formulation is based on the use ofrecurrence
relations to generate progressively finer discrete samplings of an implicit mother wavelet
function; each resolution is twice that of the previous scale. In her seminal paper, Daubechies
derives a family of wavelets, the first of which is the Haar wavelet. Interest in this field has
exploded since then, and many variations of Daubechies' original wavelets were developed.

The Dual-Tree Complex Wavelet Transform (WT)

The Dual-Tree Complex Wavelet Transform (WT) is relatively recent enhancement to the
discrete wavelet transform (DWT), with important additional properties: It is nearly shift
invariant and directionally selective in two and higher dimensions. It achieves this with a
redundancy factor of only ubstantially lower than the undecimated DWT. The
multidimensional (M-D) dual-tree WT is non separable but is based on a computationally
efficient, separable filter bank (FB).[2]

Others

Other forms of discrete wavelet transform include the non- or undecimated wavelet
transform (where down sampling is omitted), the Newland transform (where an ortho
normal basis of wavelets is formed from appropriately constructed top-hat filters in frequency
space). Wavelet packet transforms are also related to the discrete wavelet transform. Complex
wavelet transform is another form.

Applications

The discrete wavelet transform has a huge number of applications in science, engineering,
mathematics and computer science. Most notably, it is used for signal coding, to represent a
discrete signal in a more redundant form, often as a preconditioning for data compression.
Practical applications can also be found in signal processing of accelerations for gait analysis, in
digital communications and many others.

It is shown that discrete wavelet transform (discrete in scale and shift, and continuous in time) is
successfully implemented as analog filter bank in biomedical signal processing for design of
low-power pacemakers and also in ultra-wideband (UWB) wireless communications.

One level of the transform

The DWT of a signal is calculated by passing it through a series of filters. First the samples
are passed through a low pass filter with impulse response resulting in aconvolution of the
two:
The signal is also decomposed simultaneously using a high-pass filter . The outputs giving the
detail coefficients (from the high-pass filter) and approximation coefficients (from the low-pass).
It is important that the two filters are related to each other and they are known as a quadrature
mirror filter.

However, since half the frequencies of the signal have now been removed, half the samples can
be discarded according to Nyquists rule. The filter outputs are then subsampled by 2 (Mallat's
and the common notation is the opposite, g- high pass and h- low pass):

This decomposition has halved the time resolution since only half of each filter output
characterises the signal. However, each output has half the frequency band of the input so the
frequency resolution has been doubled.

Block diagram of filter analysis

With the subsampling operator

the above summation can be written more concisely.

However computing a complete convolution with subsequent downsampling would waste

computation time.
The Lifting scheme is an optimization where these two computations are interleaved.

Cascading and Filter banks

This decomposition is repeated to further increase the frequency resolution and the
approximation coefficients decomposed with high and low pass filters and then down-sampled.
This is represented as a binary tree with nodes representing a sub-space with a different time-
frequency localisation. The tree is known as a filter bank.

A 3 level filter bank

At each level in the above diagram the signal is decomposed into low and high frequencies. Due
to the decomposition process the input signal must be a multiple of where is the number
of levels.

For example a signal with 32 samples, frequency range 0 to and 3 levels of decomposition, 4
output scales are produced:

Level Frequencies Samples

to 4

to 4
2 to 8

1 to 16

Frequency domain representation of the DWT

Digital watermarking is the act of hiding a message related to a digital signal (i.e. an image,
song, video) within the signal itself. It is a concept closely related to steganography, in that they
both hide a message inside a digital signal. However, what separates them is their goal.
Watermarking tries to hide a message related to the actual content of the digital signal, while in
steganography the digital signal has no relation to the message, and it is merely used as a cover
to hide its existence.

Watermarking has been around for several centuries, in the form of watermarks found initially in
plain paper and subsequently in paper bills. However, the field of digital watermarking was only
developed during the last 15 years and it is now being used for many different applications. In
the following sections I will present some of the most important applications of digital
watermarking, explain some key properties that are desirable in a watermarking system, and give
an overview of the most common models of watermarking as presented in the book by Ingemar
Cox, Matthew Miller, Jeffrey Bloom, Jessica Friedrich and Ton Kalker . These basic models will
be further illustrated by the use of example watermarking systems that were developed in
Matlab. All images used in this essay, except those used to present the results of the example
watermarking systems are taken from this book.

Watermarking applications

The increasing amount of research on watermarking over the past decade has been largely driven
by its important applications in digital copyrights management and protection.

One of the first applications for watermarking was broadcast monitoring. It is often crucially
important that we are able to track when a specific video is being broadcast by a TV station. This
is important to advertising agencies that want to ensure that their commercials are getting the air
time they paid for. Watermarking can be used for this purpose. Information used to identify
individual videos could be embedded in the videos themselves using watermarking, making
broadcast monitoring easier.
Another very important application is owner identification. Being able to identify the owner of a
specific digital work of art, such as a video or image can be quite difficult. Nevertheless, it is a
very important task, especially in cases related to copyright infringement. So, instead of
including copyright notices with every image or song, we could use watermarking to embed the
copyright in the image or the song itself.

Transaction tracking is another interesting application of watermarking. In this case the

watermark embedded in a digital work can be used to record one or more transactions taking
place in the history of a copy of this work. For example, watermarking could be used to record
the recipient of every legal copy of a movie by embedding a different watermark in each copy. If
the movie is then leaked to the Internet, the movie producers could identify which recipient of
the movie was the source of the leak.

In the area of image watermarking, some commercial watermarking products already emerged
on the market. Digimarc Corporation's Picture Marc is available as a tool in Adobe Photoshop
image processing program. The detector can find watermarks in the images if they were
embedded into the image by the Digimarc product.

In this thesis, our aim is to investigate the robustness of some selected watermarking algorithms.
In the following chapters, first, we provide a basic look into the problem of watermarking by
means of its applications and requirements.Next, we set out to give the major ideas about
watermarking and try to establish the framework of the watermarking system by giving real
world examples and some experimental results to strengthen the underlying ideas. We handle the
basic building blocks of classical watermarking systems; namely embedding, the channel(attacks
and distortions) and detection, one by one and try to explain their properties, limitations, etc.

Finally, copy control is a very promising application for watermarking. In this application,
watermarking can be used to prevent the illegal copying of songs, images of movies, by
embedding a watermark in them that would instruct a watermarking-compatible DVD or CD
writer to not write the song or movie because it is an illegal copy.

Watermarking properties
Every watermarking system has some very important desirable properties. Some of these
properties are often conflicting and we are often forced to accept some trade-offs between these
properties depending on the application of the watermarking system.

The first and perhaps most important property is effectiveness. This is the probability that the
message in a watermarked image will be correctly detected. We ideally need this probability to
be 1.

Another important property is the image fidelity. Watermarking is a process that alters an
original image to add a message to it, therefore it inevitably affects the image's quality. We want
to keep this degradation of the image's quality to a minimum, so no obvious difference in the
image's fidelity can be noticed.

The third property is the payload size. Every watermarked work is used to carry a message. The
size of this message is often important as many systems require a relatively big payload to be
embedded in a cover work. There are of course applications that only need a single bit to be
embedded.

The false positive rate is also very important to watermarking systems. This is the number of
digital works that are identified to have a watermark embedded when in fact they have no
watermark embedded. This should be kept very low for watermarking systems.

Lastly, robustness is crucial for most watermarking systems. There are many cases in which a
watermarked work is altered during its lifetime, either by transmission over a lossy channel or
several malicious attacks that try to remove the watermark or make it undetectable. A robust
watermark should be able to withstand additive Gaussian noise, compression, printing and
scanning, rotation, scaling, cropping and many other operations.

Watermarking models
There are several ways in which we can model a watermarking process. These can be broadly
classified in one of two groups. The first group contains models which are based on a
communication-based view of watermarking and the second group contains models based on a
geometric view of watermarking. In the rest of this essay, I only refer to image watermarking
because I only concentrated on images during the development of example watermarking
systems.

While digital watermarking for copyright protection is a relatively new idea, the idea of data
hiding dates back to the ancient Greeks and has progressively evolved over the ages. An
excellent survey of the evolution of data hiding technologies can be found . The inspiration of
current watermarking technology can be traced to paper watermarks which were used some 700
years ago for the purpose of dating and authenticating paper. It was 1988, when the term digital
watermark is first used by Komatsu and Tominaga . Watermarks in the context of digital images
first appeared in 1990. The first Information Hiding Workshop, which included digital
watermarking as one of its primary topics , was held in 1996. Beginning in 1999, SPIE began
devoting a conference focused on Security and Watermarking of Multimedia Contents . The
Copy Protection Technical Working Group (CPTWG) evaluated some watermarking systems for
their possible usage as a method of protecting DVD video. The Secure Digital Music Initiative
(SDMI), adopting the technology of the Verance Corporation, made watermarking a basic
component of their portable device specification. The same technology is also used by some
Internet music distributors like Liquid Audio. Two projects, VIVA and Talisman that are funded
by the European Union, tested watermarking schemes for their suitability for broadcast
monitoring.

Watermarking without side-information

As described earlier, some communication-based watermarking models do not take advantage of

the channel side-information. In this kind of models, the image is simply considered as another
form of channel noise that distorts the message during its transmission. This can be seen in
Figure 3. The watermark embedder encodes a message using a watermark encoder and a key.
This is then added to the original image and transmitted over the communication channel which
adds some noise. The watermark detector at the other end receives the noisy watermarked image
and tries to decode the original image using a key.

Fig.1.1 Standard model for watermarking with no side-information

An example watermarking system that illustrates watermarking without side-information was

implemented in Matlab.

Blind embedding and linear correlation detection

This system is an example of blind embedding, which does not exploit the original image
statistics to embed a message in an image. The detection is done using linear correlation. This
system is a 1-bit watermarking system, in other words it only embeds one bit (a 1 or 0) inside the
cover image. The algorithm for the embedder and the detector is as follows:

Embedder:

1.

Choose a random reference pattern. This is simply an array with the same dimensions as the
original image, whose elements are drawn from a random Gaussian distribution in the interval [-
1, 1]. The watermarking key is the seed that is used to initiate the pseudo-random number
generator that creates the random reference pattern.

Calculate a message pattern depending on whether we are embedding a 1 or a 0. For a 1, leave

the random reference pattern as it is. For a 0, take its negative to get the message pattern.

Scale the message pattern by a constant ? which is used to control the embedding strength. For
higher values of ? we have more robust embedding, at the expense of loosing image quality. The
value used at the initial experiment was ? = 1.

Add the scaled message pattern to the original image to get the watermarked image.
Detector:

1. Calculate the linear correlation between the watermarked image that was received and the
initial reference pattern that can be recreated using the initial seed which acted as the
watermarking key.

Decide what the watermark message was, according to the result of the correlation. If the linear
correlation value was above a threshold, we say that the message was a 1. If the linear correlation
was below the negative of the

threshold we say that the message was a 0. If the linear correlation was between the negative and
the positive threshold we say that no message was embedded.

An example of the embedding process can be seen in Figure 4. The top left image is the original
image, the bottom left image is the reference pattern and the watermarked image resulting from
embedding a 1, with ?=1, is seen on the right. As we can see, there is no perceptual difference
between the original and the watermarked image.

Finally, we present an evaluation of 9 of the existing watermarking algorithms from the

literature. The algorithms that we examine differ from each other in their some basic properties
such as embedding domain, embedding rules, host feature selection, etc. We give plots, figures
about their robustness against a number of attacks implemented generally by Stirmark v3.1 [95,
96] and MATLAB and try to end up with generalizations and comments on the relationship
between their algorithmic properties and their performance in terms of robustness.

Multiscale gradient estimators such as theMulti-scale Derivative of Gaussian (DOG) are

frequently used to extract the image features at different scales. These are used in applications
such as image matching, image segmentation and quality assessment.In other applications,
however it is desired to modify the multiscale gradients of the image. In this case, the resulting
modified gradient vectors are then inverse-transformed to the image domain, resulting in changes
in the original image. These applications include: image denoising (total variation and nonlinear

anisotropic diffusion filtering methods) contrast enhancement and watermarking. The wavelet-
based image derivative estimation technique, proposed by Mallat, is widely used to obtain the
image gradient at different scales. In this method, the horizontal and vertical wavelet subbands
are used to estimate the horizontal and vertical derivatives of different scaled versions of an
image. The wavelet transform has the advantage of being both computationally efficient and
invertible. However, this type of derivative estimation has the following problem: the wavelet-
based derivative vector is a function of only the horizontal and vertical wavelet coefficients (and
not the diagonal ones). This is because, a wavelet with p vanishing moments can only estimate
the pre-smoothed version of the pth-order derivative, and the pre-smoothing filter (embedded in
the derivative operator) removes the diagonal wavelet components. Thus, estimating the
derivative vector by the traditional wavelet-based approach may decrease the estimation
accuracy.

To solve this problem, the shearlet transform has been used to estimate the derivative vectors at
different scales of the image.This transform captures the anisotropic information of the image
more efficiently than the wavelet transform. However, its highly redundant nature makes it
unsuitable for some applications, such as image watermarking. In image watermarking
application if the values of the derivative vectors, obtained from the shearlet transform of the
image, are modified (due to watermark embedding), the modified vectors may be different from
those obtained after inverse-transforming them to the image domain and then transforming back
to the derivative domain.

In this chapter, to take advantage of the computational efficiency and invertibility of the wavelet
transform and to employ all the detail wavelet subbands in estimating the image derivatives, we
propose a derivative transform that obtains the image gradient for different scaled versions of the
image, using wavelet transform. This transform obtains the derivative vectors as a linear function
of the multi-scale wavelet detail coefficients. At each pixel, the proposed forward derivative
transform maps the diagonal, as well as the horizontal and vertical, wavelet coefficients to the
horizontal and vertical derivative components. The proposed inverse derivative transform is
designed based on the criterion that any change in the derivative vectors should result in the
minimum possible change in the original wavelet coefficients. In image watermarking
applications, this renders a watermark that was embedded in the image derivative domain to be
as invisible as possible in the image domain.

The rest of this chapter is organized as follows: gives a brief overview of the traditional wavelet-
based image derivative estimation approach.The proposed framework for multi-scale image
derivative estimation is presented. The forward and inverse derivative transforms are also
described in this section. As a special case, we focus on obtaining the multiscale first and higher-
order finite difference (FD) derivatives of an image. Potential applications of the proposed
transform in texture feature extraction, edgedetection, image quality assessment and image
watermarking are discussed.

Due to the quick spread of Internet facilities all over the world in these modern years, the
necessity of digital technology has been evidently amplified. Digital technology includes digital
images, digital media, text etc. Since most of the organizations, publishers, artists post their
models, images, secrets via internet; security has to be provided to greater extent. Security
nowadays has become a question among the users who publish their resources. Security can be
of different forms such as detecting the mall practices, blocking the unauthorized users,
providing measures to prove the ownership etc. Security system that will handle all the attacks
should be developed. Digital Image Processing has paved the way for digital images to be
processed and handled. Here images to be published are taken as the input. Using certain
developed techniques, transforms or applications those images are protected from malicious
attackers.

Steganography is the knack of concealing information inside digital carriers. The information
need to be protected is embed or hide within a digital carrier so that the information reaches its
destination without any security issues. Watermarking is a small part in steganography.
Steganography is used for data encryption where as Watermarking is used for data hiding within
a host carrier.
PROPOSED REVERSIBLE WATERMARKING SCHEME

The proposed scheme is inspired by the pairwise RW of [2]. Data is embedded into the 2D
prediction error histogram computed for the rhombus predictor (see [1]) after the usual image
splitting and sorting. The main novelties are a context-based pixel classication stage and the
adaptive pairing. The major details of the scheme follow.

Watermarking may be either visible or invisible. If only authentication is needed then it is wise
to go for visible watermarking but if both authentication and security are needed then better go
for invisible watermarking methods. Robustness is a parameter used in watermarking technique.
A watermarking technique is robust if and only if the watermarking withstands the general image
processing attacks such as rotation, scaling, noise, etc.

A novel chaos primarily based watermarking theme for image authentication and tamper
detection is projected . This theme provides each integrity and legitimacy for digital
watermarking. Extracting the proper watermark is just potential if somebody has correct keys.
Since chaotic maps are sensitive to initial values, they're used as key during this theme. someone
with wrong keys won't be ready to forge the watermark. so as to thwart counterfeiting attacks it's
essential to interrupt constituent wise freedom, this theme employs chaotic maps to interrupt the
corresponding position relation between pixels within the watermarked image and therefore the
watermark. Provides hi-fi and is capable of localizing changed regions in watermarked image.
Watermark embedding research has been initiated in 1990. With the redundancy of the medium
as design and influence, digital watermarking technology is to use the digital embedding method
to hide the watermarking information into the digital products of image, visible and video. Seen
from the field of signal process, the watermarking signal being embedded into carrier is as a
feeble signal to add into a strong background. As long as the intensity of watermarking is lower
than the contrast restriction of human visible system (HVS) or the apperceive restriction of
human audio system (HAS), the watermarking signal won't be felt by HVS or HAS. With these
unique characters and important application, digital watermarking technology has been got more
and more attention. Discrete wavelet transform in frequency domain is preferred among all the
other techniques because of its unique characteristics such as perceptivity, security and
robustness . Robust image watermarking against noise and pepper attack is clearly defined.
Biometric images using salient region-based authentication watermarking is proposed for self-
recovery and tamper detection and also it recover the damaged data of original biometric images
with hidden information based on tampering detection. An adaptive digital image watermarking
algorithm based on non-linear wavelet transform and Morphological Haar Wavelet Transform
used. Human Visual System (HVS) is adaptively embedded into the original image in different
resolutions. The extraction of watermark can be done in the decrypted domain. Stream cipher is
used to propose encryption algorithm. The phase DWT and inverse DWT (IDWT) in the
encrypted domain, perform an analysis of data development and quantization errors under the
frame. The problem of data development can be solved by reducing data development. In the
case that multilevel DWT/IDWT can be performed with less data growth in homomorphic
encrypted domain. supported appropriate performance for those content-preserving geometric
deformations and image process operations, as well as JPEG compression, low pass filtering,
cropping and RBAs. Scaling the dimensions of pictures, freelance of the element position within
the image plane, proof against cropping, sturdy to interpolation errors throughout geometric
transformations, and customary image process operations.. A second-order statistics (SOS)-based
image quality metric, that considers the feel masking result and therefore the variation sensitivity
in Karhunen-Love rework domain. It improves the robustness in return . The gradient direction
watermarking can be used for uniform quantization for the direction of gradient vectors. The
watermark bits are entrenched by quantizing the angles of large gradient vectors at multiple
wavelet scales .

Discrete Wavelet Transform

Wavelet analysis consists of decomposing a signal or an image into a hierarchical set of

approximations and information. The levels in the hierarchy often correspond to those in a
dyadic scale. From the signal analyst's point of view, wavelet analysis is a decomposition of the
signal on a family of analysing signals, which is usually an orthogonal function method. From an
algorithmic point of view, wavelet analysis offers a harmonious compromise between
decomposition and smoothing techniques. Unlike conventional techniques, wavelet
decomposition produces a family of hierarchically organized decompositions. The selection of a
suitable level for the hierarchy will depend on the signal and experience. Often the level is
chosen based on a desired low-pass cut-off frequency. At each level j, we build the j-level
approximation Aj, or approximation at level j, and a deviation signal called the j-level detail Dj,
or detail at level j. We can consider the original signal as the approximation at level 0, denoted
by A0. The words approximation and detail are justified by the fact that A1 is an approximation
of A0 taking into account the low frequencies of A0, whereas the detail D1 corresponds to the
high frequency correction.

One way of understanding this decomposition consists of using an optical comparison.

Successive images A1, A2, A3 of a given object are built. We use the same type of photographic
devices, but with increasingly poor resolution.

The Matlab algorithm for discrete wavelet transform (DWT) is, in fact, a classical scheme in the
signal processing community, known as a two channel sub-band coder using conjugate
quadrature filters. The decomposition algorithm starts with signal s, next calculates the
coordinates of A1 and D1, and then those of A2 and D2, and so on. The reconstruction algorithm
called the inverse discrete wavelet transform (IDWT) starts from the coordinates of AJ and DJ,
next calculates the coordinates of AJ-1, and then using the coordinates of AJ-1 and DJ-1
calculates those of AJ-2.

WATERMARKING PROCESS

A. Watermark Embedding Algorithm

? Colour image of any size/format is taken as the host image and the watermark.

? The host image is separated into three basic colour components.

? Each component is employed with single level 2-dimentional discrete wavelet transform.

? Low-Low sub-band of single level decomposition is further decomposed into two level 2-
dimentional discrete wavelet transform .

? Repeat the above step till the decomposition reaches its fifth level.

? The watermark is employed with Arnold's cat map transform which shuffles the original
watermark with the same pixels.

? Encrypt the shuffled watermark with the secret key which is the weight factor in
watermarking.

? Separate the encrypted shuffled watermark into three basic colour components.

? Embed each of decomposed host image component with the separated colour component
of encrypted shuffled watermark.

? Inverse discrete wavelet transform is used at last to obtain the watermarked image.

B. Watermark Extraction Algorithm

? Watermarked image from embedding process is the host image to be extracted.

? The host image is separated into three basic colour components.

? Each component is employed with single level 2-dimentional discrete wavelet transform.

? Low-Low sub-band of single level decomposition is further decomposed into two level 2-
dimentional discrete wavelet transform.

? Repeat the above step till the decomposition reaches its fifth level.

? Extraction process is applied to the decomposed host image components. Shuffled watermark
components are extracted from this process.
? The shuffled watermark components are employed with inverse Arnold's cat map transform to
get the original watermark without any image quality degradation.