CHAPTER 1

INTRODUCTION

1.INTRODUCTION
In recent years, it is much easier to alter the content of digital images than before
using the popular image editing computer software. Some of these images are
irrational with respect to their content, making their fakery obvious to human.
However, many of the altered images can not be easily determined as real or
fake. Fake images are loosely defined as images that do not express the original
content . In other words, what human saw is not necessarily believable. Like it or
not, fake images are everywhere, such as those in movies, advertisements, etc.
There is no general model to detect them, since the production of fake images
leaves no visual clues of fakery. Fig. 1 gives some examples of fake images,
which show that fake images are consistent in their signal characteristic with real
images, and only their contents are exaggerated. To the fake images that are
produced for the sake of cheating, especially for political advantage, they can
barely be distinguished only based on their content.

Some similar concepts with image fakery have been introduced in the past few
years, such as digital forgery and image splicing . In digital forgeries are
referred to the manipulation and alteration of digital images, and method to
detect traces of resampling is proposed to expose digital forgeries. Furthermore,
some other statistical tools for detecting digital forgeries are also proposed and
analyzed in including techniques employing the inconsistences in digital camera
imaging techniques digital image sampling techniques, the direction of point light
, principal component analysis and higher-order wavelet statistics . In [an image
splicing model is proposed to combine different objects in images into a new
image. Aiming at detecting spliced images a statistical model using bicoherence
feature has been proposed in which was originally designed for detecting human
speech signal.
Practically, it is hard to say whether an image is real or fake only from viewing the
content of the image, because the purpose of creating fake images is to alter the
content of an image by adding, removing or replacing some objects in the image,
2

so that the altered image looks like a real image. For Fig. 1(a), it is illogical for a
flying hawk hanging a bomb, so it may be easily concluded that this image is
fake. The same conclusion can be applied to Fig. 1(b). Unfortunately, most of
the fake images can not be distinguished visually and appropriate methods
should be developed to detect them.

Fig 1.1:Fake Images

CHAPTER-2
AIM AND SCOPE

2 AIM AND SCOPE

It is much easier to alter the content of digital images than before,

using the popular image editing computer software. Some of these
images are irrational with respect to their content, making their fakery
obvious to human. However, many of the altered images can not be
easily determined as real or fake. Fake images are loosely defined as
images that do not express the original content . In other words, what
human saw is not necessarily believable. Like it or not, fake images are
everywhere, such as those in movies, advertisements, etc. There is no
general model to detect them, since the production of fake images leaves
no visual clues of fakery.some examples of fake images, which show that
fake images are consistent in their signal characteristic with real images,
and only their contents are exaggerated. To the fake images that are
produced for the sake of cheating, especially for political advantage, they
can barely be distinguished only based on their content. With the great
convenience of computer graphics and digital imaging, it becomes much
easier to alter the content of images than before without any visually
traces to catch these manipulations, i.e., many fake images are produced
whose content is feigned. Thus, the images can not be judged whether
they are real or not visually. Support vector machine is a technique for
universal data classification. In recent years, SVMs have been used for
many applications, such as pattern recognition, industrial engineering,
digital watermarking and so on. Generally, SVMs are deemed to be
easier and better to use than traditional neural network models
.

CHAPTER-3
DIGITAL IMAGE PROCESSING

3. DIGITAL IMAGE PROCESSING

3.1 BACKGROUND
Digital image processing is an area characterized by the need for
extensive experimental work to establish the viability of proposed solutions to a
given problem.

An important characteristic underlying the design of image

processing systems is the significant level of testing & experimentation that

normally is required before arriving at an acceptable solution. This characteristic
implies that the ability to formulate approaches &quickly prototype candidate
solutions generally plays a major role in reducing the cost & time required to
arrive at a viable system implementation.
3.2 WHAT IS DIP?
An image may be defined as a two-dimensional function f(x, y), where x
& y are spatial coordinates, & the amplitude of f at any pair of coordinates (x, y)
is called the intensity or gray level of the image at that point. When x, y & the
amplitude values of are all finite discrete quantities, we call the image a digital
image. The field of DIP refers to processing digital image by means of digital
computer. Digital image is composed of a finite number of elements, each of
which has a particular location & value. The elements are called pixels.
Vision is the most advanced of our sensor, so it is not surprising that
image play the single most important role in human perception. However, unlike
humans, who are limited to the visual band of the EM spectrum imaging
machines cover almost the entire EM spectrum, ranging from gamma to radio
waves. They can operate also on images generated by sources that humans are
not accustomed to associating with image.
There is no general agreement among authors regarding where image
processing stops & other related areas such as image analysis& computer
vision start. Sometimes a distinction is made by defining image processing as a
discipline in which both the input & output at a process are images. This is
7

limiting & somewhat artificial boundary. The area of image analysis (image
understanding) is in between image processing & computer vision.
There are no clear-cut boundaries in the continuum from image
processing at one end to complete vision at the other. However, one useful
paradigm is to consider three types of computerized processes in this
continuum: low-, mid-, & high-level processes. Low-level process involves
primitive operations such as image processing to reduce noise, contrast
enhancement & image sharpening. A low- level process is characterized by the
fact that both its inputs & outputs are images. Mid-level process on images
involves tasks such as segmentation, description of that object to reduce them
to a form suitable for computer processing & classification of individual objects.
A mid-level process is characterized by the fact that its inputs generally are
images but its outputs are attributes extracted from those images. Finally higherlevel

processing involves Making sense of an ensemble of recognized

objects, as in image analysis & at the far end of the continuum performing the
cognitive functions normally associated with human vision.
Digital image processing, as already defined is used successfully in a
broad range of areas of exceptional social & economic value.
3.3 WHAT IS AN IMAGE?
An image is represented as a two dimensional function f(x, y) where x
and y are spatial co-ordinates and the amplitude of f at any pair of coordinates
(x, y) is called the intensity of the image at that point.
3.3.1 GRAY SCALE IMAGE:
A gray scale image is a function I (xylem) of the two spatial coordinates
of the image plane.
I(x, y) is the intensity of the image at the point (x, y) on the image plane.
I(xylem) takes non-negative values assume the image is bounded by a
rectangle [0, a] [0, b]I: [0, a] [0, b] [0, info)
3.3.2 COLOR IMAGE
8

It can be represented by three functions, R (xylem) for red, G (xylem) for

green and B (xylem) for blue.
An image may be continuous with respect to the x and y coordinates
and

also in amplitude. Converting such an image to digital form requires that

the coordinates as well as the amplitude to be digitized. Digitizing the

coordinates values is called sampling. Digitizing the amplitude values is called
quantization.
3.4 COORDINATE CONVENTION
The result of sampling and quantization is a matrix of real numbers. We
use two principal ways to represent digital images. Assume that an image f(x, y)
is sampled so that the resulting image has M rows and N columns. We say that
the image is of size M X N. The values of the coordinates (xylem) are discrete
quantities. For notational clarity and convenience, we use integer values for
these discrete coordinates. In many image processing books, the image origin is
defined to be at (xylem)=(0,0).The next coordinate values along the first row of
the image are (xylem)=(0,1).It is important to keep in mind that the notation (0,1)
is used to signify the second sample along the first row. It does not mean that
these are the actual values of physical coordinates when the image was
sampled. Following figure shows the coordinate convention. Note that x ranges
from 0 to M-1 and y from 0 to N-1 in integer increments.
The coordinate convention used in the toolbox to denote arrays is
different from the preceding paragraph in two minor ways. First, instead of using
(xylem) the toolbox uses the notation (race) to indicate rows and columns. Note,
however, that the order of coordinates is the same as the order discussed in the
previous paragraph, in the sense that the first element of a coordinate topples,
(alb), refers to a row and the second to a column. The other difference is that the
origin of the coordinate system is at (r, c) = (1, 1); thus, r ranges from 1 to M and
c from 1 to N in integer increments. IPT documentation refers to the coordinates.
Less frequently the toolbox also employs another coordinate convention called
spatial coordinates which uses x to refer to columns and y to refers to rows. This
is the opposite of our use of variables x and y.

3.5 IMAGE AS MATRICES

The preceding discussion leads to the following representation for a
digitized image function:

f (xylem)=

f (0,0)

f(0,1)

..

f(0,N-1)

f (1,0)

f(1,1)

f(1,N-1)

f (M-1,0) f(M-1,1) f(M-1,N-1)

The right side of this equation is a digital image by definition. Each
element of this array is called an image element, picture element, pixel or pel.
The terms image and pixel are used throughout the rest of our discussions to
denote a digital image and its elements.
A digital image can be represented naturally as a MATLAB
matrix:
f (1,1) f(1,2) . f(1,N)

f=

f (2,1)

f(2,2) .. f (2,N)

.
.

f (M,1) f(M,2) .f(M,N)

Where f (1,1) = f(0,0) (note the use of a monoscope font to denote
MATLAB quantities). Clearly the two representations are identical, except for the
shift in origin. The notation f(p ,q) denotes the element located in row p and the
column q. For example f(6,2) is the element in the sixth row and second column
of the matrix f. Typically we use the letters M and N respectively to denote the
number of rows and columns in a matrix. A 1xN matrix is called a row vector
whereas an Mx1 matrix is called a column vector. A 1x1 matrix is a scalar.

10

Matrices in MATLAB are stored in variables with names such as A, a,

RGB, real array and so on. Variables must begin with a letter and contain only
letters, numerals and underscores. As noted in the previous paragraph, all
MATLAB quantities are written using mono-scope characters. We use
conventional Roman, italic notation such as f(x ,y), for mathematical expressions

3.6 READING IMAGES

Images are read into the MATLAB environment using function imread
whose syntax is
Imread (filename)
Format name

Description

Recognized extension

TIFF

Tagged

JPEG

Format
Joint Photograph Experts .jpg, .jpeg

GIF

Group
Graphics

BMP

Format
Windows Bitmap

PNG

Portable

XWD

Graphics
X Window Dump

Image

File tif, .tiff

Interchange .gif
.bmp
Network .png
.png

Table 1.1: Different image formats

Here filename is a spring containing the complete of the image

file(including any applicable extension).For example the command line
>> f = imread (8. jpg);
Reads the JPEG (above table) image chestxray into image array f. Note
the use of single quotes () to delimit the string filename. The semicolon at the
end of a command line is used by MATLAB for suppressing output If a
11

semicolon is not included. MATLAB displays the results of the operation(s)

specified in that line. The prompt symbol (>>) designates the beginning of a
command line, as it appears in the MATLAB command window.
3.7 DATA CLASSES
Although we work with integers coordinates the values of pixels
themselves are not restricted to be integers in MATLAB. Table above list various
data classes supported by MATLAB and IPT are representing pixels values. The
first eight entries in the table are refers to as numeric data classes. The ninth
entry is the char class and, as shown, the last entry is referred to as logical data
class.
All numeric computations in MATLAB are done in double quantities, so this
is also a frequent data class encounter in image processing applications. Class
unit 8 also is encountered frequently, especially when reading data from
storages devices, as 8 bit images are most common representations found in
practice. These two data classes, classes logical, and, to a lesser degree, class
unit 16 constitute the primary data classes on which we focus. Many ipt
functions however support all the data classes listed in table. Data class double
requires 8 bytes to represent a number uint8 and int 8 require one byte each,
uint16 and int16 requires 2bytes and unit 32.
Int 32 and single required 4 bytes each. The char data class holds
characters in Unicode representation. A character string is merely a 1*n array of
characters logical array contains only the values 0 to 1,with each element being
stored in memory using function logical or by using relational operators.

12

Name
Double

Description
Double _ precision, floating_ point

Uint16

numbers the Approximate.

unsigned 16_bit integers in the range

Uint 32

[0, 65535] (2byte per element)

unsigned 32_bit integers in the range

Int8

[0, 4294967295](4 bytes per element)

signed 8_bit integers in the range [-

Int 16

128,127] 1 byte per element)

signed 16_byte integers in the range

Int 32

[32768, 32767] (2 bytes per element)

Signed 32_byte integers in the range [2147483648, 21474833647] (4 byte
per element)
single _precision

Single

numbers
approximate

with

floating
values

range

(4

_point
In

bytes

the
per

elements)
Char
Logical

characters (2 bytes per elements)

values are 0 to 1 (1byte per element)
Table 1.2: Different Data types

3.8 IMAGE TYPES:

The toolbox supports four types of images:
1 .Intensity images;
2. Binary images;
3. Indexed images;
4. R G B images.
Most monochrome image processing operations are carried out using
binary or intensity images, so our initial focus is on these two image types.
Indexed and RGB colour images.

3.1 INTENSITY IMAGES:

13

An intensity image is a data matrix whose values have been scaled to

represent intentions. When the elements of an intensity image are of class unit8,
or class unit 16, they have integer values in the range [0,255] and [0, 65535],
respectively. If the image is of class double, the values are floating point
numbers. Values of scaled, double intensity images are in the range [0, 1] by
convention.

Fig 3.8.1: Intensity Image

3.8.2 BINARY IMAGES:

Binary images have a very specific meaning in MATLAB.A binary image
is a logical array 0s and1s.Thus, an array of 0s and 1s whose values are of data
class, say unit8, is not considered as a binary image in MATLAB .A numeric
array is converted to binary using function logical. Thus, if A is a numeric array
consisting of 0s and 1s, we create an array B using the statement.
B=logical (A)
If A contains elements other than 0s and 1s.Use of the logical function
converts all nonzero quantities to logical 1s and all entries with value 0 to logical
0s.Using relational and logical operators also creates logical arrays.To test if an
array is logical we use the I logical function:
islogical(c).

14

If c is a logical array, this function returns a 1.Otherwise returns a 0.

Logical array can be converted to numeric arrays using the data class
conversion functions.

Figure 2.8.2 Binary Image

3.8.3 INDEXED IMAGES:

An indexed image has two components:
A data matrix integer, x
A color map matrix, map
Matrix map is an m*3 arrays of class double containing floating point
values in the range [0, 1].The length m of the map are equal to the number of
colors it defines. Each row of map specifies the red, green and blue components
of a single color. An indexed images uses direct mapping of pixel intensity
values color map values. The color of each pixel is determined by using the
corresponding value the integer matrix x as a pointer in to map. If x is of class
double ,then all of its components with values less than or equal to 1 point to
the first row in map, all components with value 2 point to the second row and so
on. If x is of class units or unit 16, then all components value 0 point to the first
row in map, all components with value 1 point to the second and so on.

15

Figure 3.8.3: INDEXED IMAGE

3.8.4 RGB IMAGE:

An RGB color image is an M*N*3 array of color pixels where each color
pixel is triplet corresponding to the red, green and blue components of an RGB
image, at a specific spatial location. An RGB image may be viewed as stack of
three gray scale images that when fed in to the red, green and blue inputs of a
color monitor
Produce a color image on the screen. Convention the three images
forming an RGB color image are referred to as the red, green and blue
components images. The data class of the components images determines their
range of values. If an RGB image is of class double the range of values is [0, 1].
Similarly the range of values is [0,255] or [0, 65535].For RGB images of
class units or unit 16 respectively. The number of bits use to represents the pixel
values of the component images determines the bit depth of an RGB image. For
example, if each component image is an 8bit image, the corresponding RGB
image is said to be 24 bits deep.
Generally, the number of bits in all component images is the same. In
this case the number of possible color in an RGB image is (2^b) ^3, where b is a
16

number of bits in each component image. For the 8bit case the number is
16,777,216 colors.

FIGURE 3.8.4 RGB IMAGE

17

CHAPTER-4
WATERMARKING

18

4.1 ELEMENTS OF WATERMARKING SYSTEM

A watermarking system can be viewed as a communication system consisting
of three main elements: an embedded, a communication channel and a
detector. Watermark information is embedded into the signal itself, instead of
being placed in the header of a file or using encryption like in other security
techniques, in such a way that it is extractable by the detector. To be
more specific, the watermark information is embedded within the host
signal before the watermarked signal is transmitted over the communication
channel, so that the watermark can be detected at the receiving end, that is,
at the detector.
A general watermarking system is illustrated in Fig. 1. The dotted lines
represent the optional components, which may or may not be required
according to the application. First of all, a watermark Wo is generated by the
watermark generator possibly with a secret watermark generation key Kg.
The watermark Wo can be a logo, or be a pseudo- random signal.

Figure 4.1 A general watermarking system

Instead of directly embedding it into the host signal, the watermark Wo can
19

be pre- coded to optimize the embedding process, i.e. to increase

robustness against possible signal processing operations or imperceptibility
of the watermark. This is done by an information coder which may require the
original signal So.

The outcome of the information coding component is denoted by symbol

W that, together with the original signal So and possibly a secret key K, are
taken as input of the embedder. The secret key K is intended to differentiate
between authorized users and unauthorized users at the detector in the
absence of Kg. The embedder takes in W, So and K, so as to hide W within
So

in a most imperceptible way with the help of K, and produce the

watermarked signal Sw. Afterwards, Sw enters into the communication channel

where a series of unknown signal processing operations and attacks may
take place. The outcome of the communication channel is denoted by the
symbol Sw.
At the receiving end, the detector works in an inversely similar way as the
embedded, and it may require the secret key Kg, K, and the original signal
So. Then the detector reads Sw and decides if the received signal has the
legal watermark.
4.2 TYPES OF WATERMARKS

Watermarks and watermarking techniques can be divided into various

categories in various ways. Watermarking techniques can be divided into four
categories according to the type of document to be watermarked as follows:
Text Watermarking
Image Watermarking
Audio Watermarking
Video Watermarking

20

In other way, the digital watermarks can be divided into three different
types as

follows:

Visible watermark
Invisible-Robust watermark
Invisible-Fragile watermark
Visible watermark is a secondary translucent overlaid into the primary image.
The watermark appears visible to a casual viewer on a careful inspection.
The invisible- robust watermark is embedded in such a way that alternations
made to the pixel value are perceptually not noticed and it can be recovered
only with appropriate decoding mechanism. The invisible-fragile watermark is
embedded in such a way that any manipulation or modification of the image
would alter or destroy the watermark.
Also, the digital watermarks can be divided into two different types according
to the necessary data for extraction:

Informed (or private Watermarking): in which the original unwater

marked cover is required to perform the extraction process.

Blind (or public Watermarking): in which the original unwater marked

cover is not required to perform the extraction process.

4.3 WATERMARK REQUIREMENTS
In this section, we study a number of watermarking system requirements as
well as the tradeoffs among them.

Security: The security requirement of a watermarking system can differ

slightlydepending on the application. Watermarking security implies that the

watermark

should be difficult to remove or alter without damaging the host

signal. As all watermarking systems seek to protect watermark information,

without loss of generality, watermarking security can be regarded as the ability
to assure secrecy and integrity of the watermark information, and resist
21

malicious attacks

Imperceptibility: The imperceptibility refers to the perceptual transparency

of the watermark.

Ideally, no perceptible difference between the

watermarked and original signal should exist [4, 5]. A straightforward

way to reduce distortion during watermarking process is embedding the
watermark into the perceptually insignificant portion of the host signal [5].
However, this makes it easy for an attacker to alter the watermark
information without being noticed.

Capacity: Watermarking capacity normally refers to the amount of

information that can be embedded into a host signal. Generally speaking,
capacity requirement always struggle against two other important
requirements, that is, imperceptibility and robustness (Fig. 2). A higher
capacity is usually obtained at the expense of either robustness strength
or imperceptibility, or both.

Robustness

Figure 4.3 The tradeoffs among imperceptibility, Robustness, and capacity

22

Robustness: Watermark robustness accounts for the capability of the

watermark to survive

signal manipulations. Apart from malicious attacks,

common signal processing operations can pose a threat to the detection of

watermark, thus making it desirable to design a watermark that can survive
those operations. For example, a good strategy to robustly embed a watermark
into an image is to insert it into perceptually significant parts of the image.
Therefore, robustness is guaranteed when we consider the case of lossy
compression which usually discards perceptually insignificant data, thus data
hidden in perceptual significant portions is likely to survive lossy compression
operation. However, as this portion of the host signal is more sensitive to
alterations, watermarking may produce visible distortions in the host signal.
The exact level of robustness an algorithm must possess cannot be specified
without

considering the application scenario [6]. Not all watermarking

applications require a watermark to be robust enough to survive all attacks and

signal processing operations. Indeed, a watermark needs only to survive the
attacks and those signal processing operations that are likely to occur during
the period when the watermarked signal is in communication channel. In an
extreme case, robustness may be completely irrelevant in some case where
fragility is desirable.
4.4WATERMARKINGTECHNIQUES

Several different methods enable watermarking in the spatial domain. The

simplest (too simple for many applications) is just to flip the lowest-order bit of
chosen pixels. This works
any

modification. A

well only

more

robust

if

the image is not subject

watermark

can

be

embedded

to
by

superimposing a symbol over an area of the picture. The resulting mark may
be visible or not, depending upon the intensity value. Picture cropping, e.g., (a
common operation of image editors), can be used to eliminate the watermark.

Spatial watermarking can also be applied using color separation. In this way,
23

the watermark appears in only one of the color bands. This renders the
watermark visibly subtle such that it is

difficult to detect under regular

viewing. However, the mark appears immediately when the colors are
separated for printing. This renders the document useless for the printer
unless the watermark can be removed from the color band. This approach is
used commercially for journalists to inspect digital pictures from a photo-stock
house before buying unmarked versions.

Watermarking can be applied in the frequency domain (and other

transform domains) by

first

applying a transform like the Fast Fourier

Transform (FFT). In a similar manner to spatial domain watermarking, the

values of chosen frequencies can be altered from the original. Since high
frequencies will be lost by compression or scaling, the watermark signal is
applied to lower frequencies, or better yet, applied adaptively to frequencies
that contain important information of the original picture. Since watermarks
applied to the frequency domain will be dispersed over the entirety of the
spatial image upon inverse transformation, this method is not as susceptible to
defeat by cropping as the spatial technique. However, there is more a
tradeoff here between invisibility and decodability, since the watermark is
in effect applied indiscriminately across the spatial image. Table 1. shows a
small comparison between the two different techniques.

Computation Cost
Robustness
Perceptual Quality
Capacity

Spatial Domain
Low
Fragile
High Control
High (depend on the size of the

Frequency Domain
High
More Robust
Low Control
Low

Example of

image)
Mainly Authentication

Copy Rights

Applications
Table 4.4 Comparison between Watermarking Techniques

24

CHAPTER-5
FRAGILE WATERMARKING

5. FRAGILE WATERMARKING TECHNIQUE

25

5.1 FRAGILE WATERMARKING

A watermark is said to be fragile if the watermark hidden within the host
signal is destroyed as soon as the watermarked signal undergoes any
manipulation. When a fragile watermark is present in a signal, we can infer,
with a high probability, that the signal has not been altered.

Fragile

watermarking

authentication

has

an

interesting

variety

of

functionalities including tampering localization and discrimination between

malicious

and

non- malicious manipulations.

Tampering localization is

critical because knowledge of where the image has been altered can be
effectively used to indicate the valid region of the image, to infer the motive
and the possible adversaries. Moreover, the type of alteration may be
determined from the knowledge of tampering localization.
As to the fragile watermarks for authentication and proof of integrity, the
attacker is no longer interested in making the watermarks unreadable.
Actually, disturbing this type of watermark is easy because of its fragility.
The goal of the attackers is, conversely, producing a fake but legally
watermarked signal. This host media forgery can be reached by either
making undetectable modifications on the watermarked signal or inserting
a fake watermark into a desirable signal.
Now, it is necessary to formulate the unique features of fragile watermarking
systems in order to demonstrate what features are well sought after.

The features can also serve in theoretical analysis for making comparisons
26

among algorithms:

1)

High resolution tampering localization: This becomes an important

merit of fragile watermarking systems as it is one of the features that
makes watermarking outweighs cryptography in some applications. The
outcome of a detector can be as simple as authentic/tampered, but a
result indicating which portions in an image are tampered is more
desirable.

2)

Tampering detection with low false positive.

good

fragile

watermarking system should have a sound tamper indication stating

both statistical tampering probability and tampering localization with a
low false positive rate. If the tampering localization is required to be
accurate, the false positive rate must be kept low or localization
resolution is compromised. For example, the boundary of tampered region
may be obscure when false positive rate is not low enough, even if block
size is small.
3)

Geometric manipulation detectability. The watermark should be

correctly read by the detector in the intact portions after geometric
manipulations such as image cropping. Further, the ability of the detector
to indicate where the cropping took place is of crucial importance in some
applications.

4) Attack identification. With proper settings, the detector is also able to

estimate what kind of modification had occurred to an attacked image.
This includes the ability to di ffe re n ti a te g e o m e tri c a tta cks fro m
other a tta cks . It i mp li e s th a t cropping a part of the image will not result
in disturbing the whole watermark.

5)

Proper embedding sequence.

It implies that the selection of

dependency is limited to the previously watermarked portion of the image.

If localization is required, the dependency information for the to-bewatermarked pixel can only be chosen from the neighboring pixels from
27

the dependency selection. Because the to-be-watermarked pixel can only

depend on the content information that will not be changed later, otherwise
the watermark will not be recognized by the detector. Raster-scan and zigzag scan order are both widely used [3].

6)

Blind detection.

For practicality, watermark detectors should not

require an original copy, or there would be no necessity for watermarking

as the verification can be performed by simply comparing the received
image with the original one. The watermark extraction should naturally be
blind for practicality.
In the following section, we introduce a fragile watermarking scheme
which is immune to all forgery attacks and still retains high performance in
every aspect of the system.

5.2 FRAGILE WATERMARKING SCHEME

The combination of contextual and non-deterministic information has been

proved to be one of the most effective means to improve the security of a
watermarking system. The discussed scheme depends on both of the
contextual

information

and

non-deterministic

information

to

perform

watermarking. Contextual information, or dependency information, refers to the

information from other portions of the image. Non-deterministic information
usually relies on some randomly chosen parameters so that it can produce a
unique signature. In this way, the non-deterministic signatures of two identical
blocks at the same position of two images are different, even when they have
the

same neighborhood. As it is proved later, dependency information is

recognized as effective mean to counter basic forgery attacks while the addition
of non-deterministic information can thwart more advanced ones.

28

5.3. FEATURE CONSTRUCTION

In order to carry out fake image classification using SVM, first of all, we need to
construct a set of the features xi and it is described in detail in this section. Firstly,
a fragile digital water-marking scheme is introduced as an assistant, which is
used to find out the alteration of images. Then, the feature vector is constructed
by convoluting two vectors which are obtained from the detected alteration matrix.
In our fake image detection scheme, the altered area in an image should be
detected first. Here, we take use of a least significant bit (LSB) based fragile
watermarking scheme to watermark the original image, and then the watermarked
image is open to public. Digital watermarking has been developed over 10 years.
The main contributions are the robust water-marking which emphasizes the
existence of the watermark. On the other hand, the fragile watermarking received
less attention and it concerns with how to achieve the sensitivity to the image's
alteration. With the assistant of fragile watermarking, even a slight alteration to
the test images can be detected. In, a fragile watermarking scheme is introduced,
where a watermark W with the same size of the original image, i.e., mn, is
constructed to embed into the LSB plane of the original image I. Each watermark
element is inserted into the LSB of each image pixel. Then, a watermarked image
IW obtained. When the watermark is extracted form the test image I, there
will be difference between the extracted watermark We and the original
watermark W, if the watermarked image is altered.
Based on the difference, a matrix can be obtained as follow
A=XOR(W,We)
where A is a matrix composed of 0 and 1 with sizemn.It can be easily concluded
that when A(s,t)=1, the pixel I(s,t) In the test image is altered. On the other hand,
when A(s,t)=0, it is not altered. Here (s,t) is the pixel index with s=1,2,, m and
t=1,2,,n.In Fig.2, a sample detection process in shown
sample detection process in shown.Here, the
white pixels inFig. 2(b) denotes the altered pixels.

29

Figure 5.3 (a) Fake image of lena (b)Area detected with alteration

5.4 CONSTRUCTION OF FEATURE VECTORS

Once we get the matrix A, two column vectors U={u1,u2,,um}T and V={v1,v2,,
vn} T can be obtained as follows:

which show the number of the altered pixels along the row and the column of A
respectively. It is obvious that U and V reveal the altered area in the test image from
two directions. In the proposed scheme, we use two methods to construct the
feature vector. The first one is to joint the two vectors U and V as follows:

where + denotes the joint operator. The length l of the feature vector is m+n. The
other way to construct the feature vector is to convolute U and V, so that a feature
vector xi with length l=m+n1 is obtained as follows:
.

Wherexdenotes the convolution operator and the k-th element of the feature vector
xi can be obtained from the convolution equation as follows:

Fig shows a sample of U, V and U V curves. Once the feature vector is constructed,
the training set can be obtained as T={xi,yi}, where yi is the fakery indicator of the i-

th image, i.e.,either1or1.

Figure 5.4The curves of U, V and their convolution U V are shown in (a), (b)
and
(c) respectively.

CHAPTER-6
CONVOLUTION

6.1CONVOLUTION
Convolution is a mathematical way of combining two signals to form a third
signal.

It is the single most important technique in Digital Signal Processing.

Using the strategy of impulse decomposition, systems are described by a signal

called the impulse response.

Convolution is important because it relates the

three signals of interest: the input signal, the output signal, and the impulse
response.

This chapter presents convolution from two different viewpoints,

called the input side algorithm and the output side algorithm.

Convolution

provides the mathematical framework for DSP; there is nothing more important
in this book.

6.2THE DELTA FUNCTION & IMPULSE RESPONSE

An impulse is a signal composed of all zeros, except a single nonzero point. In

effect, impulse decomposition provides a way to analyze signals one sample at a
time. The input signal is decomposed into simple additive components, each of
these components is passed through a linear system, and the resulting output
components are synthesized (added). The signal resulting from this divide-andconquer procedure is identical to that obtained by directly passing the original
signal through the system.

While many different decompositions are possible,

two form the backbone of signal processing: impulse decomposition and Fourier
decomposition. When impulse decomposition is used, the procedure can be
described by a mathematical operation called convolution. Convolution also
applies to continuous signals, but the mathematics is more complicated.

Convolution is a formal mathematical operation, just as multiplication, addition,

and integration. Addition takes two numbers and produces a third number, while
convolution takes two signals and produces a third signal. Convolution is used in
the mathematics of many fields, such as probability and statistics.

In linear

systems, convolution is used to describe the relationship between three signals of

interest: the input signal, the impulse response, and the output signal.

6.3 CONVOLUTION PROPERTIES

6.3.1 IDENTITY
The result of a convolution with a delta function is the signal itself:
x(i) = S(i) * x(i).
The result for each individual sample can be computed by the sum of the sample
itself (weighted by 1) and all other samples weighted by 0, i.e., the sample value
itself
6.3.2 COMMUTATIVE
Changing the order of operands does change the result of the convolution operation.
That means that the distinction between impulse response and signal is of no
mathematical consequence in the context of convolution
h(i)*x(i)=x(i)*h(i).

6.3.3 ASSOCIATIVE
The associative property of convolution means that changing the order of
subsequent con-volution operations does not change the overall result. When
applying two or more filters to a signal, the output will be identical for every order of
filters.' This means that
g(i) * h(i)) * x(i) = g(i) * (h(i) * x(i)).

6.3.4 DISTRIBUTIVE
The order of different linear operations is irrelevant due to the distributive property

of the convolution, for example:

g(i) * (h(i) + x(i)) = g(i) * h{i) + g(i) * x(i).
This means that two signals, one computed by applying a filter to two different
signals and summing them together afterwards, the other computed by applying the
filter to the sum of the signals, are identical .

6.3.5 CIRCULARITY
The convolution with a periodic signal will result in a periodic output signal. The
periodic signal x(i) is the sum of the shifted (fundamental) periods with length N:
.

The multiplication of two spectra computed with the DFT will result in a circular
convolution; the result will be the convolution of the two periodically continued
sample blocks

CHAPTER-7
MATLAB FUNCTIONS

7.1 IMREAD
Read image from graphics files
7.1.1Syntax

A = imread(filename,fmt)
[X,map] = imread(filename,fmt)
[...] = imread(URL,...)
[...] = imread(...,idx) (CUR, ICO, and TIFF only)
[...] = imread(...,'frames',idx) (GIF only)
[...] = imread(...,ref) (HDF only)
[...] = imread(...,'BackgroundColor',BG) (PNG only)
[A,map,alpha] = imread(...) (ICO, CUR, and PNG only)

7.1.2 DESCRIPTION
The imread function supports four general syntaxes, described below.
The imread function also supports several other format-specific syntaxes.
See Special Case Syntax for information about these syntaxes.
A = imread(filename,fmt) reads a grayscale or true color image named
filename into A. If the f contains a grayscale intensity image, A is a twodimensional array. If the f i l e contains a true color (RGB) image, A is a
three-dimensional (m-by-n-by-3) array.
filename is a string that specifies the name of the graphics f i l e, and fmt is a
string that specifies the format of the f i l e. If the f i l e is not in the current
directory or in a directory in the MATLAB path, specify the fu l pathname of
the location on your system. If imread cannot find a f i l e named filename, it
looks for a f i l e named filename.fmt. See Formats for a list of a l the possible
values for fmt.
[X,map] = imread(filename,fmt) reads the indexed image in filename
into X and its associated colormap into map. The colormap values are
rescaled to the range [0,1]
.

7.2 IMSHOW
Display image

7.2.1Syntax
imshow(I)
imshow(I,[low,high])
imshow(RGB)
imshow(BW)
imshow(X,map)
imshow(filename)
himage=imshow(...)
imshow(..., param1, val1, param2, val2,...)
7.2.2 Description
imshow(I) displays the grayscale image I.
imshow(I,[low high]) displays the grayscale image I, specifying the display range for I
in [low high]. The value low (and any value less than low) displays as black; the
value high (and any value greater than high) displays as white. Values in between
are displayed as intermediate shades of gray, using the default number of gray
levels. If you use an empty matrix ([]) for [low high], imshow uses [min(I(:)) max(I(:))];
that is, the minimum value in I is displayed as black, and the maximum value is
displayed as white.
imshow(RGB) displays the truecolor image RGB.
imshow(BW) displays the binary image BW. imshow displays pixels with the value 0
(zero) as black and pixels with the value 1 as white.
imshow(X,map) displays the indexed image X with the colormap map. A color map
matrix may have any number of rows, but it must have exactly 3 columns. Each row
is interpreted as a color, with the first element specifying the intensity of red light, the
second green, and the third blue. Color intensity can be specified on the interval 0.0
to 1.0.
imshow(filename) displays the image stored in the graphics file filename. The file
must contain an image that can be read by imread or dicomread. imshow calls
imread or dicomread to read the image from the file, but does not store the image

data in the MATLAB workspace. If the file contains multiple images, imshow displays
the first image in the file. The file must be in the current directory or on the MATLAB
path.
himage = imshow(...) returns the handle to the image object created by imshow.
imshow(..., param1, val1, param2, val2,...) displays the image, specifying
parameters and corresponding values that control various aspects of the image
display. The following table lists all imshow parameters in alphabetical order.
Parameter names can be abbreviated, and case does not matter.

Parameter

Value

'Border'

Text string that controls whether imshow includes a border

around the image displayed in the figure window. Valid strings
are 'tight' and 'loose'.
Note: There can still be a border if the image is very small, or if
there are other objects besides the image and its axes in the
figure.
By default, the border is set to the value returned by
iptgetpref('ImshowBorder').

'Colormap'

2D, real, m-by-3 matrix specifying a colormap. imshow uses

this to set the figure's colormap property. Use this parameter to
view grayscale images in false color. If you specify an empty
colormap ([]), imshow ignores this parameter.

'DisplayRange'

Two-element vector [LOW HIGH] that controls the display range

of a grayscale image. See the imshow(I,[low high]) syntax
for more details about how to set this parameter.
Note

Including the parameter name is optional,

except when the image is specified by a filename. The

syntax imshow(I,[LOW

HIGH])

is equivalent to

imshow(I,'DisplayRange',[LOW HIGH]).

However,

Parameter

Value
the 'DisplayRange' parameter must be specified
when calling imshow with a filename, for example
imshow(filename,'DisplayRange'[LOW HIGH]).

'InitialMagnification'

A numeric scalar value, or the text string 'fit', that specifies

the initial magnification used to display the image. When set to
100, imshow displays the image at 100% magnification (one
screen pixel for each image pixel). When set to 'fit', imshow
scales the entire image to fit in the window.
On initial display, imshow always displays the entire image. If
the magnification value is large enough that the image would be
too big to display on the screen, imshow warns and displays the
image at the largest magnification that fits on the screen.
By default, the initial magnification parameter is set to the value
returned by iptgetpref('ImshowInitialMagnification').
If the image is displayed in a figure with its 'WindowStyle'
property set to 'docked', imshow warns and displays the image
at the largest magnification that fits in the figure.
Note: If you specify the axes position (using subplot or axes),
imshow

ignores any initial magnification you might have

specified and defaults to the 'fit' behavior.

When used with the 'Reduce' parameter, only 'fit' is allowed
as an initial magnification.
'Parent'

Handle of an axes that specifies the parent of the image object

that will be created by imshow.

'Reduce'

Logical value that specifies whether imshow subsamples the

image in filename. The 'Reduce' parameter is only valid for
TIFF images and you must specify a file name. Use this
parameter to display overviews of very large images.

Parameter

Value

'XData'

Two-element vector that establishes a nondefault spatial

coordinate system by specifying the image XData. The value can
have more than two elements, but only the first and last elements
are actually used.

'YData'

Two-element vector that establishes a nondefault spatial

coordinate system by specifying the image YData. The value can
have more than two elements, but only the first and last elements
are actually used.

7.2.3 CLASS SUPPORT

A true color image can be uint8, uint16, single, or double. An indexed image can be
logical, uint8, single, or double. A grayscale image can be logical, uint8, int16,
uint16, single, or double. A binary image must be of class logical.
For grayscale images of class single or double, the default display range is [0 1]. If
your image's data range is much larger or smaller than the default display range, you
might need to experiment with setting the display range to see features in the image
that would not be visible using the default display range. For all grayscale images
having integer types, the default display range is [intmin(class(I)) intmax(class(I))].

Examples
Display an image from a file
imshow('board.tif')
Display an indexed image.
[X,map] = imread('trees.tif');
imshow(X,map)

Display a grayscale image.

I = imread('cameraman.tif');
imshow(I)
Display the same grayscale image, adjusting the display range.
h = imshow(I,[0 80]);
7.3 BITAND
7.3.1 SYNTAX
C = bitand(A, B)
7.3.2 DESCRIPTION
C = bitand(A, B) returns the bitwise AND of arguments A and B, where A and B are
unsigned integers or arrays of unsigned integers.
Examples

Example 1
The five-bit binary representations of the integers 13 and 27 are 01101 and 11011,
respectively. Performing a bitwise AND on these numbers yields 01001, or 9:
C = bitand(uint8(13), uint8(27))
C =9

Example 2
Create a truth table for a logical AND operation:
A = uint8([0 1; 0 1]);
B = uint8([0 0; 1 1]);
TT = bitand(A, B)
TT =

7.4 BITOR
7.4.1 SYNTAX
C = bitor(A, B)

7.4.2 DESCRIPTION
C = bitor(A, B) returns the bitwise OR of arguments A and B, where A and B are
unsigned integers or arrays of unsigned integers.
Examples
Example 1
The five-bit binary representations of the integers 13 and 27 are 01101 and 11011,
respectively. Performing a bitwise OR on these numbers yields 11111, or 31.
C = bitor(uint8(13), uint8(27))
C =31
Example 2
Create a truth table for a logical OR operation:
A = uint8([0 1; 0 1]);
B = uint8([0 0; 1 1]);
TT = bitor(A, B)
TT =
0

7.4 Abs

7.4.1 SYNTAX
abs(X)
7.4.2 DESCRIPTION
abs(X) returns an array Y such that each element of Y is the absolute value of the
corresponding element of X.
If X is complex, abs(X) returns the complex modulus (magnitude), which is the same
as
sqrt(real(X).^2 + imag(X).^2)
Examples

abs(-5)
ans =5

abs(3+4i)
ans =5

7.5 IMNOISE
Add noise to image
7.5.1 SYNTAX
J=imnoise(I,type)
J=imnoise(I,type,parameters)
J=imnoise(I,'gaussian',m,v)
J=imnoise(I,'localvar',V)
J=imnoise(I,'localvar',image_intensity,var)
J=imnoise(I,'poisson')
J=imnoise(I,'salt&pepper',d)
J=imnoise(I,'speckle',v)

7.5.2 DESCRIPTION
J = imnoise(I,type) adds noise of a given type to the intensity image I. type is a
string that can have one of these values.
Value

Description

'gaussian'

Gaussian white noise with constant mean and variance

'localvar'

Zero-mean Gaussian white noise with an intensity-dependent variance

'poisson'

Poisson noise

'salt & pepper'

On and off pixels

'speckle'

Multiplicative noise

J = imnoise(I,type,parameters) Depending on type, you can specify additional

parameters to imnoise. All numerical parameters are normalized; they correspond to
operations with images with intensities ranging from 0 to 1.
J = imnoise(I,'gaussian',m,v) adds Gaussian white noise of mean m and variance
v to the image I. The default is zero mean noise with 0.01 variance.
J = imnoise(I,'localvar',V) adds zero-mean, Gaussian white noise of local variance
V to the image I. V is an array of the same size as I.
J = imnoise(I,'localvar',image_intensity,var) adds zero-mean, Gaussian noise to
an image I, where the local variance of the noise, var, is a function of the image
intensity values in I. The image_intensity and var arguments are vectors of the same
size, and plot(image_intensity,var) plots the functional relationship between noise
variance and image intensity. The image_intensity vector must contain normalized
intensity values ranging from 0 to 1.
J = imnoise(I,'poisson') generates Poisson noise from the data instead of adding
artificial noise to the data. If I is double precision, then input pixel values are
interpreted as means of Poisson distributions scaled up by 1e12. For example, if an
input pixel has the value 5.5e-12, then the corresponding output pixel will be
generated from a Poisson distribution with mean of 5.5 and then scaled back down
by 1e12. If I is single precision, the scale factor used is 1e6. If I is uint8 or uint16,
then input pixel values are used directly without scaling. For example, if a pixel in a

uint8 input has the value 10, then the corresponding output pixel will be generated
from a Poisson distribution with mean 10.
J = imnoise(I,'salt & pepper',d) adds salt and pepper noise to the image I, where d
is the noise density. This affects approximately d*numel(I) pixels. The default for d is
0.05.
J = imnoise(I,'speckle',v) adds multiplicative noise to the image I, using the equation
J = I+n*I, where n is uniformly distributed random noise with mean 0 and variance v.
The default for v is 0.04.
Note The mean and variance parameters for 'gaussian', 'localvar', and
'speckle' noise types are always specified as if the image were of class double
in the range [0, 1]. If the input image is of class uint8 or uint16, the imnoise
function converts the image to double, adds noise according to the specified type
and parameters, and then converts the noisy image back to the same class as the input.

7.2.3 CLASS SUPPORT

For most noise types, I can be of class uint8, uint16, int16, single, or double. For
Poisson noise, int16 is not allowed. The output image J is of the same class as I. If I
has more than two dimensions it is treated as a multidimensional intensity image
and not as an RGB image.

Fig:7.2.3 IMAGE OF SPECKLE NOISE

Fig 7.2.3: IMAGE OF POISSON NOISE

Fig:7.2.3 IMAGE OF SALT & PEPPER NOISE

CHAPTER-8
RESULT &ANALYSIS

RESULTS & ANALYSIS

Figure 8..1 Base image

Figure8.2 Marked Image

Figure 8.3 Watermark Image

Figure 8.4 : Marked Image after watermark

Figure 8.5 Noise Image

BIBLOGRAPHY

[1] DIGITAL Image Processing by RAFEL C.GONZALEZ& RICHARD

WOODS
[2] Digital Signal Processing by PROAKIS
[3] WIKIPEDIA.ORG
[4] R.D.Fiete,Photofakery.URLhttp://oemagazine.com/fromTheMagazine
/jan05/photofakery%.html .
[5] A.C.Popescu,H. Farid, Exposing digital
traces of

forgeries by detecting

resampling, IEEE Transactions on Signal Processing 53

(2) (2005) 758767.

[6]

.-T. Ng, S.-F. Chang, A model for image splicing, IEEE Int. Conf.
on Science and Engineering from the Chinese University

Image

Processing (ICIP), vol. 2, 2004, pp. 11691172.

[7] A. Popescu, H. Farid, Statistical tools for digital forensics, 6th
[8]

Inter-national

Workshop

Toronto,Cananda, University.

on

Information
His

current

Hiding,
research

Vol.

3200,

interests

include 2004, pp. 128 147.

[9] A. Popescu, H. Farid, Exposing digital forgeries in color filter array
inter- polated images, IEEE Transactions on Signal Processing
53(10)(2005) application in medicine 3948-3959
[10]
M. Johnson, H. Farid, Exposing digital forgeries by detecting inconsistencies in lighting, ACM Multimedia and Security Workshop, New York,
2006, pp. 19.
[11]
A.P. Farid, Exposing Digital Forgeries by Detecting Duplicated Image
Regions, Tech. Rep. TR2004-515, Department of Computer Science,
Dartmouth College, 2004.
[12]
H. Farid, S. Lyu, Higher-order wavelet statistics and their application to
digital forensics, IEEE Workshop on Statistical Analysis in Computer Vision
(in conjunction with CVPR), Madison, Wisconsin, 2003, pp. 1622.
[13]
T.-T. Ng, S.-F. Chang, Q. Sun, Blind detection of photomontage
using higher order statistics, IEEE Int. Symposium on Circuits and
Systems (ISCAS), vol. 5, 2004, pp. 688691.
[14]
H. Lu, R. Shen, F.-L. Chung, Fragile watermarking scheme for
image authentication, Electronics Letters 39 (12) (2003) 898900.
i. .
APPENDIX
ABOUT MATLAB

INTRODUCTION TO MAT LAB

MATLAB is a high-performance language for technical computing. It
integrates computation, visualization, and programming in an easy-to-use
environment where problems and solutions are expressed in familiar mathematical
notation. Typical uses include
MATLAB is a package that has been purpose-designed to make
computations easy, fast and reliable. It is installed on machines run by Bath
University Computing Services (BUCS), which can be accessed in the BUCS PC
Labs such as those in 1 East 3.9, 1 West 2.25 or 3 East 3.1, as well as from any
of the PCs in the Library. The machines which you will use for running MATLAB
are SUN computers that run on the UNIX operating system. However you will
need only a small number of UNIX commands when you are working with
MATLAB. There is a glossary of common UNIX commands in Appendix B.)
MATLAB started life in the 1970s as a user-friendly interface to certain
clever but complicated programs for solving large systems of equations. The idea
behind MATLAB was to provide a simple way of using these programs that hid
many of the complications. The idea was appealing to scientists who needed to
use high performance software but had neither the time nor the inclination (nor in
some cases the ability) to write it from scratch. Since its introduction, MATLAB has
expanded to cover a very wide range of applications and can now be used as a
very simple and transparent programming language where each line of code looks
very much like the mathematical statement it is designed to implement.

Basic MATLAB is good for the following

Computations, including linear algebra, data analysis, signal processing,
polynomials and interpolation, numerical integration (quadrature), and numerical
solution of differential equations.
Graphics, in 2-D and 3-D, including colour, lighting, and animation. It also
has collections of specialised functions, called toolboxes that extend its
functionality. In particular, it can do symbolic algebra, e.g. it can tell you that
(x+y)^2 is equal to x^2+2*x*y+y^2.

It is important not to confuse the type of programming that we shall do in

this course with fundamental programming in an established high-level language
like C, JAVA or FORTRAN. In this course we will take advantage of many of the
built-in features of MATLAB to do quite complicated tasks but, in contrast to
programming in a conventional high-level language, we shall have relatively little
control over exactly how the instructions which we write are carried out on the
machine. As a result, MATLAB programs for complicated tasks may be somewhat
slower to run than programs written in languages such as C. However the
MATLAB programs are very easy to write, a fact which we shall emphasis here.
We shall use MATLAB as a vehicle for learning elementary programming skills
and applications. These are skills which will be useful independently of the
language you choose to work in. Some students in First Year will already know
some of these skills, but we shall not assume any prior knowledge.

Typical uses of MATLAB

1. Math and computation
2. Algorithm development
3. Data acquisition
4. Modeling, simulation, and prototyping
5. Data analysis, exploration, and visualization
6. Scientific and engineering graphics
7. Application development, including graphical user interface building.

The main features of MATLAB

1) Advance algorithm for high performance numerical computation,
especially in the field matrix algebra.
2) A large collection of predefined mathematical functions and the ability to
define ones own functions.
3) Two-and three dimensional graphics for plotting and displaying data.
4) A complete online help system.
5) Powerful, matrix or vector oriented high level programming language for
individual applications.
6) Toolboxes available for solving advanced problems in several application
areas

Figure Features of MATLAB

Capabilities of MATLAB
MATLAB is an interactive system whose basic data element is an array that
does not require dimensioning. This allows you to solve many technical computing
problems, especially those with matrix and vector formulations, in a fraction of the
time it would take to write a program in a scalar non interactive language such as
C or FORTRAN.
The name MATLAB stands for matrix laboratory. MATLAB was originally
written to provide easy access to matrix software developed by the LINPACK and
EISPACK projects. Today, MATLAB engines incorporate the LAPACK and BLAS
libraries, embedding the state of the art in software for matrix computation.
MATLAB has evolved over a period of years with input from many users. In
university environments, it is the standard instructional tool for introductory and
advanced courses in mathematics, engineering, and science. In industry,
MATLAB is the tool of choice for high-productivity research, development, and
analysis.
MATLAB features a family of add-on application-specific solutions called
toolboxes. Very important to most users of MATLAB, toolboxes allow you to learn
and apply specialized technology. Toolboxes are comprehensive collections of
MATLAB functions (M-files) that extend the MATLAB environment to solve
particular classes of problems. Areas in which toolboxes are available include
signal processing, control systems, neural networks, fuzzy logic, wavelets,
simulation, and many others.

The MATLAB System

The MATLAB system consists of five main parts:

Development Environment
This is the set of tools and facilities that help you use MATLAB functions
and files. Many of these tools are graphical user interfaces. It includes the
MATLAB desktop and Command Window, a command history, an editor and
debugger, and browsers for viewing help, the workspace, files, and the search

path. This consists of

the MATLAB command window, created when you start a MATLAB session,
possibly supplemented with
a figure window which appears when you do any plotting,
an editor/debugger window, used for developing and checking MATLAB
programs, and

a help browser.

The command window is the one with the >> prompt. In version 7 it has three sub
windows,
the current directory window, listing all the files in your current directory,
the workspace window (click on the tab), listing all the variables in your
workspace, and
the command history window, containing all your commands going back
over multiple sessions.
You can do things like running M-files from your directory, plotting variables from
your workspace, and running commands from your command history or
transferring them to a new M-file, by clicking (or right-clicking or double-clicking)
on them.
To see the figure window in action, type
Z = peaks; surf(Z)
And then hit RETURN. (Note that it is crucial to get the punctuation in the
above command correct.) You should get a 3D plot of a surface in the Figure
window called Figure 1. (Note that peaks is a built-in function that produces the
data for this plot, and surf is a built-in function that does the plotting. When you
come to do your own plots you will be creating your own data of course.) You can
play with the buttons at the top of the Figure window (for example to rotate the
figure and see it from different perspectives).

To obtain the editor/debugger, type edit. You can then create files by typing
them into this window, and save them by clicking on File --> Save As. You must
call the file something like file1.m, i.e. save it with a .m extension to its name, so
that MATLAB recognizes it as an M-file. If the file is a script, you can run unit by
typing its name (without the .m). Create a script file containing the single line Z =
peaks; surf (Z), save it as file1.m, and run it.
To obtain help you can type help win, which produces a simple browser
containing the MATLAB manual. A much more sophisticated web-based help
facility is obtained by typing helpdesk, but at peak times this may be slow in
starting up. If you have MATLAB on your home PC you may well prefer the
helpdesk.
Finally, if you know the name of the command you are interested in, you
can simply ask for help on it at the MATLAB prompt. For example to find out about
the for command, type help for or doc for. Do each of these now: for is one of the
loop-constructing commands that will be used in the programs you write in this
course.
Use any help facility to find the manual page for the command plot, which resides
in the directorygraph2d. Read the manual page for plot.

The MATLAB Mathematical Function

This is a vast collection of computational algorithms ranging from
elementary functions like sum, sine, cosine, and complex arithmetic, to more
sophisticated functions like matrix inverse, matrix eigen values, Bessel functions,
and fast Fourier transforms.

The MATLAB Language

This is a high-level matrix/array language with control flow statements,
functions, data structures, input/output, and object-oriented programming features.
It allows both "programming in the small" to rapidly create quick and dirty throwaway programs, and "programming in the large" to create complete large and
complex application programs.

Graphics
MATLAB has extensive facilities for displaying vectors and matrices as
graphs, as well as annotating and printing these graphs. It includes high-level
functions for two-dimensional and three-dimensional data visualization, image
processing, animation, and presentation graphics. It also includes low-level
functions that allow you to fully customize the appearance of graphics as well as
to build complete graphical user interfaces on your MATLAB applications.

The MATLAB Application Program Interface (API)

This is a library that allows you to write C and Fortran programs that
interact with MATLAB. It includes facilities for calling routines from MATLAB
(dynamic linking), calling MATLAB as a computational engine, and for reading and
writing MAT-files.

M-Files
Script M-Files
The heart of MATLAB lies in its use of M-les. We will begin with a script Mle, which is simply a text file that contains a list of valid MATLAB commands. To
create an M-le, click on File at the upper left corner of your MATLAB window, then
select New, followed by M-le. A window will appear in the upper left corner of your
screen with MATLAB's default editor. (You are free to use an editor of your own
choice, but for the brief demonstration here, let's stick with MATLAB's. It's not the
most powerful thing you'll ever come across, but it's not a complete slouch either.)
In this window, type the following lines (MATLAB reads everything following a %
sign as a comment to be ignored):
%JUNK: A script file that computes sin(4),
%where 4 is measured in degrees.

t=4;%Define a variable for no particularly good reason

radiant=pi*t/180; %Converts 4 degrees to radians
s=sin(radiant) %No semicolon means output is printed to screen
Save this file by choosing File, Save As from the main menu. In this case,
save the file as junk. m, and then close or minimize your editor window. Back at
the command line, type simply help junk, and notice that the description you typed
in as the header of your script file appears on the screen. Now, type junk at the
prompt, and MATLAB will report that s=.0698. It has simply gone through your file
line by line and executed each command as it came to it. One more useful
command along these lines is type .Try the following:
>>type junk
The entire text of your file junk.m should appear on your screen. Since
you've just finished typing this stuff in, this isn't so exciting, but try typing, for
example, type mean. MATLAB will display its internal M-file mean.m. Per using
MATLAB's internal M-files like this is a great way to learn how to write them
yourself.
In fact, you can often tweak MATLAB's M-files into serving your own
purposes. Simply use type file name in the Command Window, then choose and
copy the M-file text using the Edit option. Finally, copy it into your own M-file and
edit it. (Though keep in mind that if you ever publish a routine you've worked out
this way, you need to acknowledge the source.)

Working in Script M-files

Even if you're not writing a program, you will often find that the best way
to work in MATLAB is through script M-files.
Certainly, this is a straightforward calculation that we could carryout step by
step in the Command Window, but instead, we will work entirely in an M-#file(see
template.m below).We begin our M-file with the commands clear, which clears the

workspace of all variables ,clc, which clears the Command Window of all
previously issued commands, and clf, which clears the figure window. In addition,
we delete the(assumed)diary file junk.m and restart it with the command diary
junk.out. Finally,
we set echo on so that the commands we type here will appear in the
command window. We now type in all commands required for plotting and
maximizing the function f(x).
%TEMPLATE:MATLABM-file containing a convenient
%workplace template.
clear; clc; clf;
delete junk.out
diary junk.out
echo on
%
fprime=diff(x*exp(-x^4/(1+x^2)))
pretty(fprime)
r = eval(solve(fprime))
%
Diary off
Echo off
The only new command in template.m is pretty(), which simply instructs
MATLAB to present the expression fprime in a more readable format. Once
template.m has been run, the diary filejunk.out will be created. Observe that the
entire session is recorded, both inputs and outputs. It's also worth noting that
while pretty()makes expressions look better in the MATLAB Command Window,
its output in diary files is regrettable.

Function M-files
The second type of M-file is called a function M-file and typically (though
not inevitably) these will involve some variable or variables sent to the M-file and
processed. Function M-files need neither accept input nor return output. Every
function M-file begins with the command function. Function M-files can have sub
functions (script M-files cannot have sub functions).

Debugging M-files
Since MATLAB views M-files as computer programs, it offers a handful of
tools for debugging. First, from the M-file edit window, an M-file can be saved and
run by clicking on the icon with the white sheet and downward-directed blue arrow
(alternatively, choose Debug, Run or simply type F5). By setting your cursor on a
line and clicking on the icon with the white sheet and the red dot, you can set a
marker at which MATLAB's execution will stop. A green arrow will appear, marking
the point where MATLAB's execution has paused. At this point, you can step
through the rest of your M-file one line at a time by choosing the Step icon
(alternatively Debug, Step or F6).
Unless you're a phenomenal programmer, you will occasionally write a
MATLAB program (M-file) that has no intention of stopping any time in the near
future. You can always abort your program by typing Control-c, though you must
be in the MATLAB Command Window for MATLAB to pay any attention to this.

File Management from MATLAB

There are certain commands in MATLAB that will manipulate file on its
primary directory. For example, if you happen to have the file junk.m in your
working MATLAB directory, you can delete it simply by typing delete junk.m at the
MATLAB command prompt. Much more generally, if you precede a command with
an exclamation point, MATLAB will read it as a unix shell command

of these

notes for more on Unix shell commands). So, for example, the three commands !
ls
junk.m more junk.m, and files serve to list the contents of the directory you happen

to be in, copy the file junk.m to the file more junk.m, and list the file again to make
sure it's there. Try it.

The Command Window

Occasionally, the Command Window will become too cluttered, and you will
essentially want to start over. You can clear it by choosing Edit, Clear Command
Window. Before doing this, you might want to save the variables in your
workspace. This can be accomplished with the menu option File, Save
Workspace As, which will allow you to save your workspace as a .mat file. Later,
you can open this file simply by choosing File, Open, and selecting it. A word of
warning, though: This does not save every command you have typed into your
workspace; it only saves your variable assignments

The Command History

The Command History window will open with each MATLAB session,
displaying a list of recent commands issued at the prompt. Often, you will want to
incorporate some of these old commands into a new session. A method slightly
less gauche than simply cutting and pasting is to right-click on a command in the
Command History window, and while holding the right mouse button down, to
choose Evaluate Selection. This is exactly equivalent to typing your selection into
the Command Window.

The MATLAB Workspace

As we've seen, MATLAB uses several types of data, and sometimes it can
be difficult to remember what type each variable in your session is. Fortunately,
this information is all listed for you in the MATLAB Workspace. Look in the upper
left corner of your MATLAB window and see if your Workspace is already open. If
not, choose View, Workspace from the main MATLAB menu and it should
appear. Each variable you define during your session will be listed in the
Workspace, along with its size and type

Miscellaneous Useful Commands

In this section I will give a list of some of the more obscure MATLAB
commands that Findparticularly useful. As always, you can get more information
on each of these commands by using MATLAB's help command.
strcmp(str1,str2) (string compare) Compares the strings str1 and str2 and
returns logical true (1)

ifthe two are identical and logical false (0)

otherwise.
char(input) Converts just about any type of variable input into a string
(character array).

num2str(num) Converts the numeric variable num into a string.

str2num(str) Converts the string str into a numeric variable. (See also
str2double().)

strcat(str1,str2,...) Horizontally concatenates the strings str1, str2, etc.

Graphical User Interface

Ever since 1984 when Apple's Macintosh computer popularized Douglas
Engelbart's mouse-driven graphical computer interface, users have wanted
something fancier than a simple command line. Unfortunately, actually coding this
kind of thing yourself is a full-time job. This is where MATLAB's add-on GUIDE
comes in. Much like Visual C, GUIDE is a package for helping you develop things
like pop-up windows and menu-controlled _les. To get started with GUIDE, simply
choose File, New, GUI from MATLAB's main menu.

SIMULINK
SIMULINK is a MATLAB add-on taylored for visually modeling dynamical
systems. To get started with SIMULINK, choose File, New, Model.

M-book
M-book is a MATLAB interface that passes through Microsoft Word,
apparently allowing for nice presentations. Unfortunately, my boycott of anything
Microsoft precludes the possibility of my knowing anything about it.

Further information on MATLAB

Look in the library catalogue for titles containing the keyword MATLAB. In
particular, the following are useful.

Getting started with MATLAB 6: a quick introduction for scientists and

engineers, by Rudra Pratap.

This is a user-friendly book that you might wish to buy.

Essential MATLAB for scientists and engineers, by Brian D. Hahn

The MATLAB 5 handbook, by Eva Part-Enander and Anders Sjoberg

The MATLAB handbook, by Darren Redfern and Colin Campbell

MATLAB Guide, by D.J. Higham and N.J. Higham

The MATLAB web-site http://www.mathworks.com is maintained by the company

The MathWorks which created and markets MATLAB.