RESISTING RST WATERMARKING ALGORITHM

FOR IMAGE CONTENT AUTHENTICATION

ABSTRACT: Our Paper aims to propose a
semi-fragile watermarking algorithm resisting to
RST (Rotation, Scaling, Translation). The
algorithm can be used to verify the authenticity
and integrity of image content. Firstly, the
algorithm generates watermarking information
by using the edge of the scaled image and
embeds watermarking information based on
human

visual

system.

Before

detecting

watermarking, the parameters of geometric

distortions are estimated and restored by using
the original moment information. Finally, users
compare the extracted watermarking information
with

the

reconstructed

watermarking

information of the watermarked image to

achieve authentication. The experiment results
show the watermarking algorithm has the
immunity to common operation (Compression,
Noise, Filtering, RST, etc). The watermarking
algorithm

can

also

achieve

accurate

Semi-fragile watermarking for Image Content

Authentication shows that based on HVS
(Human Visual System) adaptively embedded
watermark into host image in the wavelet
domain by using group quantization, and
compared the extracted watermarking
information with the original one to achieve
tampering authentication. Nevertheless, the
algorithm requires the original watermark,
which was logo image, for image authentication.
The paper, A new semi-fragile colour image
watermarking algorithm, which has made the
improvement to the earlier proposed algorithm,
has realized content-based adaptive digital
watermarking, and extracted watermark without
resorting to the original host image. However,
because it has not involved the asymmetrical
watermark, practicality and safety of the
algorithm was not strong. Generally speaking,
because robustness of image content
authentication take does not affect user to
distinguish image content as the principle, and
RST (Rotation, Scaling, Translation) does not
change people to distinguish image content, so
image content authentication watermark should
have the RST invariability, but the existing
algorithm did not take into account such a
situation.

authentication and tampering localization to

malicious processing (Cropping, Replacing).

Problem definition:
Relevance of the project: In a world today

INTRODUCTION : With the development of

digital watermarking technology authentication
watermarking technology which is used to
determine the image authenticity problem has
become a hotspot of current research. In recent
years, people have proposed unceasingly many
new algorithms for the image authentication
watermarking technology. The research paper

when images are exchanged between people on

a regular basis, it becomes imperative to have an
image authentication system in order to validate
the image data. Our project furthers this cause
by detecting changes made by an unauthorized
person in an image.

Scope: The image in question is watermarked

The main use of watermarking is to provide a

such that any alteration in the image would

level of certainty about the authenticity and/or

result in an altered watermark thereby helping in

ownership of a document.

detection of change in the image content.

Watermarking life-cycle phases:

REVIEW OF LITERATURE
A watermark is an image which appears on fine
papers or on some documents to prevent
counterfeiting.

There

are

two

types

of

watermarks: true watermarks and artificial

watermarks.

The information to be embedded is called a

A true watermark is applied during the paper

digital watermark, although in some contexts the

manufacturing process using a special tool

phrase digital watermark means the difference

called a dandy roll. The dandy roll is pressed

between the watermarked signal and the cover

against the paper pulp while it is drying, and

signal. The signal where the watermark is to be

marks on the dandy roll will transfer to the paper

embedded

pulp, creating an image. This image is called a

watermarking system is usually divided into

watermark because it is made while the paper

three distinct steps, embedding, attack and

pulp is still wet with water.

detection. In embedding, an algorithm accepts

is

called

the

host

signal.

the host and the data to be embedded and

An artificial watermark is applied during the

produces a watermarked signal.

printing process. Artificial watermarks are made

using specially formulated inks or varnishes

The watermarked signal is then transmitted or

which will only show up at certain angles or

stored, usually transmitted to another person. If

under certain conditions, such as black light.

this person makes a modification, this is called

These watermarks

cheaper than true

an attack. While the modification may not be

watermarks, and can be easily customized for

malicious, the term attack arises from copyright

individual uses. They are also easier to fake by

protection application, where pirates attempt to

skilled counterfeiters. Personal checks and

remove

official documents such as passports often use

modification.

artificial watermarks.

modifications, for example, lossy compression

are

the

digital
There

watermark
are

many

through
possible

of the data, cropping an image or video or

Watermarking is addition of an image or
pattern to paper by causing variations in paper
thickness by using various specialized machines.

intentionally adding noise.

Detection (often

called

an

Digital Watermarking gets its name from

algorithm which is applied to the attacked signal

watermarking. Digital watermarking is the

to attempt to extract the watermark from it. If

process of embedding information into a digital

the signal was unmodified during transmission,

signal, i.e. audio, pictures, video, etc. If the

then the watermark is still present and it can be

signal is copied, then the embedded information

extracted. In robust watermarking applications,

is also in the copy. The embedding takes place

the extraction algorithm should be able to

by manipulating the content of the digital data,

correctly produce the watermark, even if the

which means the information is not embedded in

modifications

fragile

the frame around the data. The hiding process

watermarking, the extraction algorithm should

has to be such that the modifications of the

fail if any change is made to the signal.

media are imperceptible.

were

extraction)

strong.

In

is

The first applications that came to mind were

The most important properties of digital

related to copyright protection of digital media.

watermarking

techniques

In the past duplicating art work was quite

robustness,

complicated and required a high level of

(reversibility) and complexity and possibility of

expertise for the counterfeit to look like the

verification. Based on these parameters the

original. However, in the digital world this is not

algorithms can be evaluated if a specific

true. Now it is possible for almost anyone to

algorithm has adequate properties and can be

duplicate or manipulate digital data and not lose

used for a certain application area.

security,

are

capacity,

transparency,
invertibility

data quality. Similar to the process when artists

creatively signed their paintings with a brush to

CLASSIFICATION:

claim copyrights, artists of today can watermark

their work by hiding their name within the
image. Hence, the embedded watermark permits
identification of the owner of the work. It is
clear that this concept is also applicable to other
media such as digital video and audio. Currently
the unauthorized distribution of digital audio
over the Internet in the MP3 format is a big
problem. In this scenario digital watermarking
may be useful to set up controlled audio

Watermarking vs Steganography

distribution and to provide efficient means for

copyright protection, usually in collaboration

Steganography is about concealing their very

with international registration bodies.

existence. It comes from Greek roots, literally

means

'covered

writing',

and

is

usually

image has been tampered with. Such a

interpreted to mean hiding information in other

technology might be important, for example, in

information.

admitting digital images as evidence in court.

Examples

include

sending

message to a spy by marking certain letters in a

Invisible watermarking, which is very resistant

newspaper using invisible ink, and adding sub-

to destruction under any image manipulation

perceptible echo at certain places in an audio

might be useful in verifying ownership of an

recording. As the purpose of steganography is

image suspected of misappropriation. Digital

having a covert communication between two

detection of the watermark would indicate the

parties whose existence is unknown to a possible

source of the image.

attacker, a successful attack consists in detecting

communication.

Visible watermarking: Visible watermarking is

Watermarking, as opposed to steganography, has

a visible translucent image, which is overlaid on

the (additional) requirement of robustness

the image. It could be your name, copyright,

against possible attacks. In most cases the

comment, website address, your logo, text or

information

graphical objects. Image filters, dates, photo

the

existence

of

hidden

this

using

steganographic

techniques is not related at all to the cover.

details and other EXIF information, which holds

These differences in goal lead to very different

the rights to the primary image can also be used

hiding techniques.

for image watermarking. Watermarking

processing allows the primary image to be
viewed, but still marks it clearly as the property

Classification of watermarking:
Imperceptible (Invisible) watermarking: In
watermarking, we traditionally seek high
fidelity, i.e. the watermarked work must look or
sounds like the original. Invisible watermarking
is the digital data that is added to audio, images
or video. But it cannot be perceived as such.
Because of its different applications, there are
two very different types of invisible
watermarking. Invisible watermarking, which is
destroyed when the image is manipulated
digitally in any way may be useful in proving
authenticity of an image. If the watermark is still
intact, then the image has not been "doctored." If
the watermark has been destroyed, then the

of the owning organization.

(Semi-) fragile vs Robust watermarks : The
aims of such watermarks are completely
different: A (semi-) fragile watermark is a mark
which is (highly) sensitive to a modification of
the stego-medium. A fragile watermarking
scheme should be able to detect any change in
the signal and identify where it has taken place
and possibly what the signal was before
modification. It serves at proving the
authenticity of a document. On the opposite, a
robust watermark should be stuck to the
document it has been embedded in, in such a
way that any signal transform of reasonable

strength cannot remove the watermark. Hence a

Applications of Invisible Watermarking:

pirate willing to remove the watermark will not

succeed unless they debase the document too

1. Rights Management: One of the traditional

applications of the watermark is copyright

much to be of commercial interest.

protection. The primary reason for using

watermarks is to identify the owner of the

Invisible vs Visible Watermarking

content by an invisible, hidden mark that is

imprinted into the image. In many cases, the

INVISIBLE

VISIBLE

WATERMARK

IMPERCEP

VISIBLY

encryption, where the encryption provides the

PERCEPTIBILI

TIBLE

MEANINGF

secure distribution method from the content

DISTORTI

UL

owners to the receivers and the watermark offers

ON

PATTERN

the content owners the opportunity to trace the

INTENTIO

USER

NAL

INTERVENT

ATTACKS

ION BASED

AND

WATERMAR

COMMON

SIGNAL

REMOVAL

TY

ROBUSTNESS

PROCESSI

PROTECTION

PASSIVE

ACTIVE

EXTRACTION

EXPLICIT

DIRECT

EXTRACTI

VIEWING

ON
MODULE

RESEARCH
STATUS

HOT

contents and detect the unauthorized use or

duplications. Without watermarking, there is no
way to extend the control of the content owner
once the content leaves the protected digital
domain and is released to the user.
The technical requirements for this application
are as follows:
The watermark does not incur visible (or
audible) artifacts to the ordinary users.
The watermark is independent of the data

NG

CURRENT

watermark is used in addition to the content

ONLY FEW
PAPERS

format.
The information carried by the watermark is
robust to content manipulations, compressions,
etc.
The watermark can be detected without the
unwatermarked original content.
The watermark can be identified by some kind
of keys that are used to identify large number
of individual contents uniquely.
2. Authentication and Tamper Proofing: The
objective in this case is not to protect the content
from being copied or stolen, but to provide a
method to authenticate the image and assure the

integrity

of

the

image.

The

technical

requirements are as follows:

Invisible to the ordinary users
Applicable to compressed image format (most
digital cameras use JPEG compatible format)
Sensitive to content manipulations, compression,
etc.
3. DVD

Application of Visible Watermarking:

Electronic Distribution: Unlike the other
digital watermarking technologies described
earlier, the visible reversible watermark is
visible. It is available as a commercial product.

Playback

and

Record

Control:

This unique form of watermarking technology

Watermarking technology can be viewed as a

by IBM allows the content owners to embed a

way to provide a secure data channel along with

visible shape or logo mark such as a companys

the contents without modifying the installed-

logo on top of the image. The mark is removed

base Consumer Electronics (CE) devices. The

(the watermark is reversed) only with the

embedded watermark is transparently passing

application of an appropriate decryption key and

through the conventional data path, and will

watermark remover software. With this visible

only be detected at the digital recorders. When

watermark on the image, the content becomes

the watermark detection is mandated in the

self-protective,

recorders, it can be used to trigger the copy

distribute the entire image as a sample to various

protection mechanism implemented in it. The

open media or to the Internet. When a user

watermarking data embedded into the video is

wants to use a clean copy of the image, all

difficult to remove without damaging the quality

he/she needs to do is to request a decryption key

of the content because it is carefully woven into

and pay some fee for it. This will reduce the

the visible part of the video data. A list of

security

thirteen essential technical requirements is

transmission per each buy/sell transaction.

shown below:
Transparency
Low cost digital detection
Digital detection domain
Generational copy control for one copy
Low false positive detection
Reliable detection
Watermark will survive normal video processing

in consumer use
Licensable under reasonable terms
Export/Import
Technical maturity
Data payload
Minimum impact on content preparation
Data rate

risk

and

and

content

the

owners

amount

of

can

data

VISIBLE WATERMARKING
TECHNIQUES
LSB MODIFICATION: One of the first
techniques for watermarking is the LeastSignificant-Bit modification. It is based on the
substitution of LSB plane of the cover image
with the given watermark. The idea behind this
watermarking technique is the following: if you
see you image as a matrix NxM (where N and M
are the dimension of the image) you can
represent the value of the pixel in the position

(i,j) as a binary number; this binary can be then

insert a watermark in the frequency domain of

divided in all of its bit, so that you will have a

an image we first apply Discrete Cosine

most significant bit (the one that contains quite a

Transform (DCT), perform the corresponding

lot of information, and a least significant bit that

transformation and in the end find the inverse

contains few information).

DCT. Such a watermark is very robust to most

If your image is for example in gray scale, you

of the common signal processing and geometric

can make changes to the value of the LSB

distortions.

without any perceptible distortion for the human

user therefore you can think of taking the LSB
of an image (the cover image) and change its
value in every pixel with the MSB of another
image, that we would like to embed in a
secret/non perceptible way in the cover image).
SPREAD SPECTRUM WATERMARKING:
The watermark should not be placed in
perceptually insignificant regions of the image
(or its spectrum), since many common signal
and

geometric

processes

affect

these

components. The problem then becomes how to

insert a watermark into the most perceptually
significant regions of the spectrum in a fidelity
preserving

fashion.

Clearly,

any

spectral

coefficient may be altered, provided such

modification is small. However, very small
changes are very susceptible to noise. To solve
this problem, the frequency domain of the image
is viewed as a communication channel and
correspondingly, the watermark is viewed as a
signal that is transmitted through it. Attacks and
unintentional signal distortions are thus treated
as noise that the immersed signal must be
immune to. The watermark is spread over many
frequency bins so that the energy in any one bin
is very small and certainly undetectable. To

USING AUTHENTICATION IN

TECHNIQUES OF ENCRYPTION /

WATERMARKING

AUTHENTICATION IN WATERMARKING
RANDOM NUMBER GENERATOR

Three-way handshaking technique:

A random number generator (often abbreviated

as RNG) is a computational or physical device
designed to generate a sequence of numbers or
symbols that lack any pattern, i.e. appear
random.
Physical methods: The earliest methods for
generating random numbers dice, coin
flipping, roulette wheels are still used today,
mainly in games and gambling as they tend to be
too slow for most applications in statistics and
cryptography. A physical

random

number

generator can be based on an essentially random

atomic or subatomic physical phenomenon
This is the most suitable technique that can be

whose unpredictability can be traced to the laws

used for image content authentication since the

of quantum mechanics. Sources of entropy

recipient can send an acknowledgement to the

include radioactive decay, thermal noise, shot

sender once the authentic image is received. If

noise, avalanche noise in Zener diodes, clock

acknowledgement is not received the sender will

drift, the timing of actual movements of a hard

realize that the image content has been altered

disk read/write head, and radio noise. However,

and thus both, the sender and receiver, will be

physical phenomena and tools used to measure

aware of a change in the image.

them

generally

feature

asymmetries

and

systematic biases that make their outcomes not

uniformly random. A randomness extractor, such
as a cryptographic hash function, can be used to
obtain uniformly distributed bits from a nonuniformly random source, though at a lower bit
rate.
Computational

methods:

Pseudo-random

number generators (PRNGs) are algorithms that

can automatically create long runs of numbers

generator based on that set of rules than to

with good random properties but eventually the

manually create passwords.

sequence repeats (or the memory usage grows

without bound). The string of values generated

HADAMARD TRANSFORM:

by such algorithms is generally determined by a

The Hadamard transform is an example of a

fixed number called a seed. One of the most

generalized class of Fourier transform. It

common PRNG is the linear congruential

performs

an

orthogonal,

symmetric,

generator.

involutional, linear operation on 2 real numbers

RANDOM PASSWORD GENERATOR:

(or complex numbers, although the Hadamard

A random password generator is software

program or hardware device that takes input
from a random or pseudo-random number
generator

and

automatically

generates

matrices

themselves

are

purely

real).The

Hadamard transform is based on the Hadamard

matrix which is a square array having entries
of+1 and -1 only.

password. Random passwords can be generated

The Hadamard transform can be regarded as

manually, using simple sources of randomness

being built out of size-2 discrete Fourier

such as dice or coins, or they can be generated

transforms (DFTs), and is in fact equivalent to a

using a computer.

multidimensional

DFT

of

size

. It decomposes an
While there are many examples of "random"
password generator programs available on the

arbitrary input vector into a superposition of

Walsh functions.

Internet, generating randomness can be tricky

and many programs do not generate random
characters in a way that ensures strong security.
A common recommendation is to use open
source security tools where possible, since they
allow independent checks on the quality of the
methods used. Note that simply generating a
password at random does not ensure the
password is a strong password, because it is
possible, although highly unlikely, to generate
an easily guessed or cracked password.
A password generator can be part of a password
manager. When a password policy enforces
complex rules, it can be easier to use a password

WALSH TRANSFORM:
It is a kind of non-sinusoidal orthogonal
transform. Walsh introduced a complete set of
orthogonal square wave functions, which can be
used to represent any arbitrary function. It is
defined for N=2^n.It requires that no. of samples
should be integer power of 2. The rows of
discrete Walsh transform matrix Wm of size M x
M is generated by sampling the Functions
having the sequence length less than or equal to
M-1, at equi-spaced M points, where M should
be an integer power of 2.

For M=8, the matrix Wm becomes

1
1
1
1
1
1
1
1

1
1
1
1
-1
-1
-1
-1

MD5 is a message digest algorithm developed

1
1
-1
-1
-1
-1
1
1

1
1
-1
-1
1
1
-1
-1

The Walsh transform matrix is obtained from the

Hadamard matrix by re-arranging the rows in
increasing order. Re-arranging the rows of the
Hadamard matrix to get an increasing order of
sign changes gives us the Walsh matrix.
WALSH-HADAMARD TRANSFORM:

by Ron Rivest at MIT. It is basically a secure

version of his previous algorithm, MD4 which is
a little faster than MD5. This has been the most
widely used secure hash algorithm particularly
in Internet-standard message authentication. The
algorithm takes as input a message of arbitrary
length and produces as output a 128-bit message
digest of the input. This is mainly intended for
digital signature applications where a large file
must be compressed in a secure manner before
being encrypted with a private (secret) key
under a public key cryptosystem. Assume we
have an arbitrarily large message as input and
that we wish to find its message digest. The
processing involves the following steps.

Image processing method uses a modified

Walsh-Hadamard transform to remove noise and
preserve image structure in a sampled image.
Image signals representative of the light value of
elements of the image are grouped into signal
arrays corresponding to blocks of image
elements. These signals are mapped into larger

(1) Padding: The message is padded to ensure that

its length in bits plus 64 is divisible by 512. That
is, its length is congruent to 448 modulo 512.
Padding is always performed even if the length
of the message is already congruent to 448
modulo 512. Padding consists of a single 1-bit

signal arrays such that one or more image

followed by the necessary number of 0-bits.

(2) Appending
length:
A
64-bit
binary

signals appear two or more times in each larger

representation of the original length of the

array. The larger arrays are transformed by

message is concatenated to the result of step.

Walsh-Hadamard combinations characteristic of

The expanded message at this level will exactly

the larger array into sets of coefficient signals.

be a multiple of 512-bits. Let the expanded

Noise is reduced by modifying--i.e., coring or

message be represented as a sequence of L 512-

clipping--and

coefficient

bit blocks Y0, Y1,..,Yq,..,YL-1 . IV and CV

signals so as to recover processed signals--less

represent initial value and chaining variable

noise--representative of each smaller signal

respectively.

inverting

selected

array.
MESSAGE DIGEST 5 (MD5) ALGORITHM

SHA-2. SHA-1 is very similar to SHA-0, but

corrects an error in the original SHA hash
specification that led to significant weaknesses.
The SHA-0 algorithm was not adopted by many
applications.

SHA-2

on

the

other

hand

(3) Initialize the MD buffer: The variables IV and

significantly differs from the SHA-1 hash

CV are represented by a fourword buffer

function. SHA-1 is the most widely used of the

(ABCD) used to compute the message digest.

existing SHA hash functions, and is employed in

Here each A, B, C, D is a 32-bit register and

several widely-used security applications and

they are initialized as IV to the following values

protocols.

in hexadecimal. Low-order bytes are put first.

Word A: 01 23 45 67; Word B: 89 AB CD EF;
Word C: FE DC BA 98; Word D: 76 54 32 10
(4) Process message in 16-word blocks: This is the
heart of the algorithm, which includes four
rounds of processing. It is represented by
HMD5 in and its logic is given in Figure 2. The
four rounds have similar structure but each uses
different auxiliary functions F, G, H and I.

SHA-1:
In cryptography, SHA-1 is a cryptographic hash
function designed by the National Security
Agency (NSA) and published by the NIST as a
U.S. Federal Information Processing Standard.
SHA stands for Secure Hash Algorithm. The
three SHA algorithms are structured differently
and are distinguished as SHA-0, SHA-1, and

3 IMPLEMENTATION

a. Software Requirement:
Matlab7.0
WindowsXP Professional
b. Hardware Requirement: Intel Core2 Duo
processor
c. Database Requirement: Image database.

 Implement the mentioned algorithm for image

content authentication.
Design a suitable GUI.
Perform adequate testing
REFERENCES
[1] Lintao Lv, Liang Hao, Hui Lv, Resisting
RST

algorithm

authentication,

for

2010

image

Second

content

International

Conference on Networks Security, Wireless

Communications and Trusted Computing.
[2] Dhananjay Theckdath, Digital Signal and
Image Processing.
[3]

Todor

Watermarking
System

Todorov,

Spread

Technique

Securing,

for

Spectrum
Information

International

Journal

Information Theories & Applications Vol.11.

[4] Digital Watermarking of Image, SGN1650/1656 Signal Processing Laboratory.
[5] Norishige Morimoto IBM Japan, Ltd., Tokyo
Research Laboratory, Digital Watermarking
Technology

with

Practical

Applications,

Informing Science Special Issue on Multimedia

Informing Technologies-Part 1 Vol. 2 No 4,
1999.
4. FURTHER WORK
We aim to do the following:

[6] Introduction to Visible Watermarking, IPR

Course: TA Lecture, 2002/12/18 NTU CSIE
R105.