image seg final report

OBJECT DETECTION USING IMAGE SEGMENTATION AND BIOMETRIC INSPIRED STEGANOGRAPHY

DEPT OF ELECTRONICS AND COMMUNICATION, MSRIT

CHAPTER I

INTRODUCTION

1.1 GENERAL

The term digital image refers to processing of a two dimensional picture by a digital

computer. In a broader context, it implies digital processing of any two dimensional data. A

digital image is an array of real or complex numbers represented by a finite number of bits.

An image given in the form of a transparency, slide, photograph or an X-ray is first digitized

and stored as a matrix of binary digits in computer memory. This digitized image can then

be processed and/or displayed on a high-resolution television monitor. For display, the

image is stored in a rapid-access buffer memory, which refreshes the monitor at a rate of 25

frames per second to produce a visually continuous display.

1.1.1 THE IMAGE PROCESSING SYSTEM

Digitizer Mass Storage

Hard Copy

Device

Display

Image

Processor

Digital

Computer

Operator

Console

FIG 1.1 BLOCK DIAGRAM FOR IMAGE PROCESSING

SYSTEM



DIGITIZER

A digitizer converts an image into a numerical representation suitable for

input into a digital computer. Some common digitizers are

1. Microdensitometer

2. Flying spot scanner

3. Image dissector

4. Videocon camera

5. Photosensitive solid- state arrays.

IMAGE PROCESSOR

An image processor does the functions of image acquisition, storage,

preprocessing, segmentation, representation, recognition and interpretation and finally

displays or records the resulting image. The following block diagram gives the

fundamental sequence involved in an image processing system.

Problem

Domain

Knowledge

Base

Segmentation

Preprocessing

Image

Acquisition

Recognition &

interpretation

Representation &

Description

Result

FIG 1.2 BLOCK DIAGRAM OF FUNDAMENTAL SEQUENCE

INVOLVED IN AN IMAGE PROCESSING SYSTEM



As detailed in the diagram, the first step in the process is image acquisition by an

imaging sensor in conjunction with a digitizer to digitize the image. The next step is the

preprocessing step where the image is improved being fed as an input to the other

processes. Preprocessing typically deals with enhancing, removing noise, isolating

regions, etc. Segmentation partitions an image into its constituent parts or objects. The

output of segmentation is usually raw pixel data, which consists of either the boundary of

the region or the pixels in the region themselves. Representation is the process of

transforming the raw pixel data into a form useful for subsequent processing by the

computer. Description deals with extracting features that are basic in differentiating one

class of objects from another. Recognition assigns a label to an object based on the

information provided by its descriptors. Interpretation involves assigning meaning to an

ensemble of recognized objects. The knowledge about a problem domain is incorporated

into the knowledge base. The knowledge base guides the operation of each processing

module and also controls the interaction between the modules. Not all modules need be

necessarily present for a specific function. The composition of the image processing

system depends on its application. The frame rate of the image processor is normally

around 25 frames per second.

DIGITAL COMPUTER

Mathematical processing of the digitized image such as convolution, averaging,

addition, subtraction, etc. are done by the computer.

MASS STORAGE

The secondary storage devices normally used are floppy disks, CD ROMs etc.

HARD COPY DEVICE

The hard copy device is used to produce a permanent copy of the image and for

the storage of the software involved.



OPERATOR CONSOLE

The operator console consists of equipment and arrangements for verification of

intermediate results and for alterations in the software as and when require. The operator

is also capable of checking for any resulting errors and for the entry of requisite data.

1.1.2 IMAGE PROCESSING FUNDAMENTALS

Digital image processing refers processing of the image in digital form.

Modern cameras may directly take the image in digital form but generally images are

originated in optical form. They are captured by video cameras and digitalized. The

digitalization process includes sampling, quantization. Then these images are

processed by the five fundamental processes, at least any one of them, not necessarily

all of them.

IMAGE PROCESSING TECHNIQUES

This section gives various image processing techniques

FIG1.3 IMAGE PROCESSING TECHNIQUES

Image Enhancement

Image Restoration

Image Analysis

Image Compression

IP

Image Synthesis



IMAGE ENHANCEMENT

Image enhancement operations improve the qualities of an image like

improving the image‟s contrast and brightness characteristics, reducing its noise content,

or sharpen the details. This just enhances the image and reveals the same information in

more understandable image. It does not add any information to it.

IMAGE RESTORATION

Image restoration like enhancement improves the qualities of image but all

the operations are mainly based on known, measured, or degradations of the original

image. Image restorations are used to restore images with problems such as geometric

distortion, improper focus, repetitive noise, and camera motion. It is used to correct

images for known degradations.

IMAGE ANALYSIS

Image analysis operations produce numerical or graphical information based

on characteristics of the original image. They break into objects and then classify

them. They depend on the image statistics. Common operations are extraction and

description of scene and image features, automated measurements, and object

classification. Image analyze are mainly used in machine vision applications.

IMAGE COMPRESSION

Image compression and decompression reduce the data content necessary to

describe the image. Most of the images contain lot of redundant information,

compression removes all the redundancies. Because of the compression the size is

reduced, so efficiently stored or transported. The compressed image is decompressed

when displayed. Lossless compression preserves the exact data in the original image,

but Lossy compression does not represent the original image but provide excellent

compression.



IMAGE SYNTHESIS

Image synthesis operations create images from other images or non-image

data. Image synthesis operations generally create images that are either physically

impossible or impractical to acquire.

APPLICATIONS OF DIGITAL IMAGE PROCESSING

Digital image processing has a broad spectrum of applications, such as remote

sensing via satellites and other spacecrafts, image transmission and storage for business

applications, medical processing, radar, sonar and acoustic image processing, robotics

and automated inspection of industrial parts.

MEDICAL APPLICATIONS

In medical applications, one is concerned with processing of chest X-rays,

cineangiograms, projection images of transaxial tomography and other medical

images that occur in radiology, nuclear magnetic resonance (NMR) and ultrasonic

scanning. These images may be used for patient screening and monitoring or for

detection of tumors‟ or other disease in patients.

SATELLITE IMAGING

Images acquired by satellites are useful in tracking of earth resources;

geographical mapping; prediction of agricultural crops, urban growth and weather;

flood and fire control; and many other environmental applications. Space image

applications include recognition and analysis of objects contained in image obtained

from deep space-probe missions.

COMMUNICATION

Image transmission and storage applications occur in broadcast television,

teleconferencing, and transmission of facsimile images for office automation,

communication of computer networks, closed-circuit television based security

monitoring systems and in military communications.



RADAR IMAGING SYSTEMS

Radar and sonar images are used for detection and recognition of various

types of targets or in guidance and maneuvering of aircraft or missile systems.

DOCUMENT PROCESSING

It is used in scanning, and transmission for converting paper documents to a

digital image form, compressing the image, and storing it on magnetic tape. It is also

used in document reading for automatically detecting and recognizing printed

characteristics.

DEFENSE/INTELLIGENCE

It is used in reconnaissance photo-interpretation for automatic interpretation

of earth satellite imagery to look for sensitive targets or military threats and target

acquisition and guidance for recognizing and tracking targets in real-time smart-bomb

and missile-guidance systems.

1.2 OBJECTIVE

The main discussions and comparisons focus on spatial domain methods, frequency domain

methods and also adaptive methods. It will be shown that all of the Steganographic algorithms

discussed have been detected by Steganalysis and thus a robust algorithm with high embedding

capacity needs to be investigated. Simple edge embedding is robust to many attacks and it will

be shown that this adaptive method is also an excellent means of hiding data while maintaining a

good quality carrier. We intend to use human skin tone detection in a proposed edge embedding

Adaptive Steganographic method.

1.3 EXISTING SYSTEM

The existing systems hold the steganographed data in the known region of the image. The

usual steganography methods are prone to attacks and noise which may lead to perceptual

retrieval of the secret image or even the loss of data.



1.3.1 DISADVANTAGES OF EXISTING SYSTEM

Such images are highly prone to noise

There is a chance of losing the data itself.

1.3.2 LITERATURE SURVEY

1. Johnson, N. F. and Jajodia, S.: Exploring Steganography: Seeing the Unseen. IEEE

Computer, 31 (2): 26-34, Feb 1998.

In this paper the author says,

Residual carrier, balance, and quadrature error imperfections that are normally present in

phase/quadrature modulators limit the use of direct modulation techniques in some applications.

The classical precompensation techniques are revised and an adaptive solution proposed that

improves performance by some order of magnitude.

2. Jakubowski, J., Kwiatos, K., Chwaleba, A. and Osowski, S.: Higher Order Statistics and

Neural Network for Tremor Recognition. IEEE Transactions on Biomedical Engineering,

49 (2): February 2002.


This paper is concerned with the tremor characterization for the purpose of recognition.

Three different types of tremor are considered in this paper: the parkinsonian, essential, and

physiological. It has been proven that standard second-order statistical description of tremor is

not sufficient to distinguish between these three types. Higher order polyspectra based on third-

and fourth-order cumulants have been proposed as the additional characterization of the tremor

time series. The set of 30 quantities based on the polyspectra has been proposed and investigated

as the features for the recognition of tremor. The neural network of the multilayer perceptron

structure has been used as a classifier. The results of numerical experiments have proven high



efficiency of the proposed approach. The average error of recognition of three types of tremor

did not exceed 3%.

3. Areepongsa, S. Kaewkamnerd, N. Syed, Y. F. and Rao. K. R.: Exploring On

Steganography For Low Bit Rate Wavelet Based Coder In Image Retrieval System. IEEE

Proceedings of TENCON 2000. (3): 250-255. Kuala Lumpur, Malaysia. 2000.


An image retrieval system that provides an efficient retrieval, management and

transmission of selected image(s) from the database is proposed. The key point of research is

utilizing a steganographic technique to achieve the efficient use of resources by embedding

attributes into the image contents. To avoid the degradation of image quality, the attributes are

invisibly embedded in the edge representations of the compressed domain of the ZTE (zerotree

entropy)/modified SPIHT (set partitioning in hierarchical trees) wavelet based coder. The

evaluations of the proposed algorithm have shown several significant advantages. For example,

(1) fast transmission of the retrieved image to the receiver, (2) it allows searching based on the

retrieval images, (3) no reprocessing of the attributes for other applications, (4) no extra bits

required for the conventional thumbnail and (5) no extra bits for the attributes.

4. Kermani, Z. Z. and Jamzad, M.: A Robust Steganography Algorithm Based on Texture

Similarity using Gabor Filter. Proceedings of IEEE 5th International Symposium on Signal

Processing and Information Technology, 18-21 Dec. 2005, 578-582.


The main concern of steganography (image hiding) methods is to embed a secret image

into a host image in such a way that the host should remain as similar as possible to its original

version. In addition the host image should remain robust with respect to usual attacks. In this

paper we present a method that tries to cover all above mentioned concerns. The secret and host

images are divided into blocks of size 4 times 4. Each block in secret image is taken as a texture

pattern for which the most similar block is found among the blocks of the host image. The

embedding procedure is carried on by replacing these small blocks of the secret image with



blocks in host image in such a way that least distortion would be imposed on it. We have used

Gabor filter to measure the similarity between texture patterns. The location addresses of blocks

in host image which are replaced by blocks of secret image are saved. This data is converted to a

bit string and then modified by Hamming code. This bit string is embedded in DCT coefficients

of the modified host image using a key which is the seed of a random number generator. Our

experimental results showed a high level of capacity, robustness and minimum distortion on

standard images.

5. Marvel, L. M. and Retter, C. T.: A Methodology for Data Hiding Using Images.

Proceedings of IEEE Military Communications Conference (MILCOM98) Proceedings,

Boston, MA, USA, 18-21 Oct 1998,1044-1047.


We present a method of embedding information within digital images, called spread

spectrum image steganography (SSIS) along with its payload capacity. Steganography is the

science of communicating in a hidden manner. SSIS conceals a message of substantial length

within digital imagery while maintaining the original image size and dynamic range. The hidden

message can be recovered using the appropriate keys without any knowledge of the original

image. The capacity of the steganographic channel is described and the performance of the

technique is illustrated. Applications for such a data hiding scheme include in-band captioning,

hidden communication, image tamper proofing, authentication, invisible map overlays,

embedded control, and revision tracking.

1.4 PROPOSED METHOD

In the proposed system the steganography is done in the wavelet domain and hence the

low frequency domain is used for secure transmission.



1.4.1 ADVANTAGES:

The embedding is done in low frequency region and hence embedding is highly

secured

Scaling, Rotation, and RBA‟s are less prone to the low frequency region and hence

there will not be any loss of data.



CHAPTER2

PROJECT DESCRIPTION

2.1 INTRODUCTION

The science of hiding or embedding “data” in a transmission medium is known as

Steganography. It is actually a comprehension of two Greek words which mean “Covered

Writing”. Steganalysis is the science of attacking Steganography. It mimics the already

established science of Cryptanalysis. Note that a Steganography can create a Steganalysis merely

to test the strength of her algorithm. Its ultimate objectives, which are indefectibility, robustness

(i.e., against image processing and other attacks) and capacity of the hidden data (i.e., how much

data we can hide in the carrier file), are the main factors that distinguish it from other “sisters - in

science” techniques, namely watermarking and Cryptography. This paper provides an overview

of well-known Steganography methods. It identifies current research problems in this area and

discusses how our current research approach could solve some of these problems. We propose

using human skin tone detection in color images to form an adaptive context for an edge operator

which will provide an excellent secure location for data hiding. The simulation is done in

MATLAB and the Steganography process is shown.

.

2.2 BLOCK DIAGRAM

Fig 1.4 Block diagram of system

Skin tone

detection

Output image

Edge

detection Embedding

DWT

Input image



2.2.1 SKIN TONE DETECTION

Computer graphics and video signal transmission standards have givenbirth too many

color spaces with different properties. A wide variety of them have been applied to the problem

of skin color modeling. We will briefly review the most popular color spaces and their

properties.

2.2.2 RGB

RGB is a color space originated from CRT (or similar) display applications, when it was

convenient to describe color as a combination of three colored rays (red, green and blue). It is

one of the most widely used color spaces for processing and storing of digital image data.

However, high correlation between channels, significant perceptual non-uniformity mixing of

chrominance and luminance data make RGB not a very favorable choice for color analysis and

color based recognition algorithms.

2.2.3 NORMALIZED RGB

Normalized RGB is a representation that is easily obtained from the RGB values by a

simple normalization procedure:

(3.1)

Rr

R G B

Gg

R G B

Bb

R G B



As the sum of the three normalized components is known (r+g+b = 1), the third

component does not hold any significant information and can be omitted, reducing the space

dimensionality. The remaining components are often called ”pure colors”, for the dependence of

r and g on the brightness of the source RGB color is diminished by the normalization.

2.2.4 HSI, HSV, HSL - HUE SATURATION INTENSITY (VALUE,

LIGHTNESS)

Hue-saturation based color spaces were introduced when there was a need for the user to

specify color properties numerically. They describe color with intuitive values, based on the

artist‟s idea of tint, saturation and tone. Hue defines the dominant color (such as red, green,

purple and yellow) of an area; saturation measures the colorfulness of an area in proportion to its

brightness. The”intensity”,”lightness” or”value” is related to the color luminance. The

intuitiveness of the color space components and explicit discrimination between luminance and

chrominance properties made these color spaces popular in the works on skin color

segmentation. Several interesting properties of Hue were noted: it is invariant to highlights at

white light sources, and also, for matte surfaces, to ambient light and surface orientation relative

to the light source.

(3.2)

2.2.5 YCBCR

YCrCb is an encoded nonlinear RGB signal, commonly used by European television

studios and for image compression work. Color is represented by luma (which is luminance,

2

1

2arccos

min , ,1 3

1

3

R G R B

H

R G R B G B

R B GS

R G B

V R G B



computed from nonlinear RGB, constructed as a weighted sum of the RGB values, and two color

difference values Cr and Cb that are formed by subtracting luma from RGB red and blue

components

(3.3)

2.2.6 SKIN MODELING

The final goal of skin color detection is to build a decision rule that will discriminate

between skin and non-skin pixels. This is usually accomplished by introducing a metric, which

measures distance (in general sense) of the pixel color to skin tone. The type of this metric is

defined by the skin color modeling method.

2.3 PROBLEM DEFINITION

The main discussions and comparisons focus on spatial domain methods, frequency

domain methods and also adaptive methods. It will be shown that all of the Steganographic

algorithms discussed have been detected by Steganalysis and thus a robust algorithm with high

embedding capacity needs to be investigated. Simple edge embedding is robust to many attacks

and it will be shown that this adaptive method is also an excellent means of hiding data while

maintaining a good quality carrier. We intend to use human skin tone detection in a proposed

edge embedding adaptive Steganographic method.

0.299 0.587 0.114Y R G B

Cb B Y

Cr R Y



2.4 METHODOLOGIES

2.4.1 MODULE NAMES

1. Skin tone separation.

2. Discrete Wavelet transform(DWT).

3. Performance measurement.

2.4.2 MODULE DESCRIPTION

Skin tone separation.

Colour transformations are of paramount importance in computer vision. There

exist several colour spaces and here we list some of them3: RGB, CMY, XYZ, xyY,

UVW, LSLM, L*a*b*, L*u*v*, LHC, LHS, HSV, HSI, YUV, YIQ, YCbCr. Mainly two

kinds of spaces are exploited in the literature of biometrics which are the HSV and

YCbCr spaces. It is experimentally found and theoretically proven that the distribution of

human skin colour constantly resides in a certain range within those two spaces as

different people differ in their skin colour (e. g., African, European, Middle Eastern,

Asian, etc). A colour transformation map called HSV (Hue, Saturation and Value) can be

obtained from the RGB bases. Sobottka and Pitas defined a face localization based on

HSV.

Discrete Wavelet transform(DWT).

The experiments on the Discrete Cosine Transform (DCT) coefficients showed

promising results and redirected researchers‟ attention towards this type of image. In fact

acting at the level of DCT makes Steganography more robust and not as prone to many

statistical attacks. Spatial Steganography generates unusual patterns such as sorting of

colour palettes, relationships between indexed colours, exaggerated “noise”, etc, all of

which leave traces to be picked up by Steganalysis tools. This method is very fragile .



There is a serious conclusion drawn in the literature. “LSB encoding is extremely sensitive

to any kind of filtering or manipulation of the stego-image. Scaling, rotation, cropping,

addition of noise, or lossy compression to the stego-image is very likely to destroy the

message. Furthermore an attacker can easily remove the message by removing (zeroing)

the entire LSB plane with very little change in the perceptual quality of the modified

stego-image”

.

Performance measurement.

As a performance measurement for image distortion, the well-known Peak-

Signal-to-Noise Ratio (PSNR) which is classified under the difference distortion metrics

is applied on the stego images.

Advantages:

Preserves the quality of the original image.

A strong algorithm.

Improved quality.

Reduced implementation constraint.



CHAPTER 3

SOFTWARE SPECIFICATION

3.1 GENERAL

MATLAB (matrix laboratory) is a numerical computing environment and fourth-

generation programming language. Developed by Math Works, MATLAB

allows matrix manipulations, plotting of functions and data, implementation of algorithms, creation

of user interfaces, and interfacing with programs written in other languages,

including C, C++, Java, and Fortran.

Although MATLAB is intended primarily for numerical computing, an optional

toolbox uses the MuPAD symbolic engine, allowing access to symbolic computing capabilities. An

additional package, Simulink, adds graphical multi-domain simulation and Model-Based

Design for dynamic and embedded systems.

In 2004, MATLAB had around one million users across industry and

academia. MATLAB users come from various backgrounds of engineering, science,

and economics. MATLAB is widely used in academic and research institutions as well as

industrial enterprises.

MATLAB was first adopted by researchers and practitioners in control engineering,

Little's specialty, but quickly spread to many other domains. It is now also used in education, in

particular the teaching of linear algebra and numerical analysis, and is popular amongst scientists

involved in image processing. The MATLAB application is built around the MATLAB language.

The simplest way to execute MATLAB code is to type it in the Command Window, which is one

of the elements of the MATLAB Desktop. When code is entered in the Command Window,

MATLAB can be used as an interactive mathematical shell. Sequences of commands can be

saved in a text file, typically using the MATLAB Editor, as a script or encapsulated into

a function, extending the commands available.

http://en.wikipedia.org/wiki/Numerical_analysis

http://en.wikipedia.org/wiki/Fourth-generation_programming_language

http://en.wikipedia.org/wiki/Fourth-generation_programming_language

http://en.wikipedia.org/wiki/MathWorks

http://en.wikipedia.org/wiki/Matrix_(mathematics)

http://en.wikipedia.org/wiki/Function_(mathematics)

http://en.wikipedia.org/wiki/Algorithm

http://en.wikipedia.org/wiki/User_interface

http://en.wikipedia.org/wiki/C_(programming_language)

http://en.wikipedia.org/wiki/C%2B%2B

http://en.wikipedia.org/wiki/Java_(programming_language)

http://en.wikipedia.org/wiki/Fortran

http://en.wikipedia.org/wiki/Computer_algebra_system

http://en.wikipedia.org/wiki/Symbolic_computing

http://en.wikipedia.org/wiki/Simulink

http://en.wikipedia.org/wiki/Model_based_design

http://en.wikipedia.org/wiki/Model_based_design

http://en.wikipedia.org/wiki/Dynamical_system

http://en.wikipedia.org/wiki/Embedded_systems

http://en.wikipedia.org/wiki/Engineering

http://en.wikipedia.org/wiki/Science

http://en.wikipedia.org/wiki/Economics

http://en.wikipedia.org/wiki/Control_engineering

http://en.wikipedia.org/wiki/Linear_algebra

http://en.wikipedia.org/wiki/Numerical_analysis

http://en.wikipedia.org/wiki/Image_processing

http://en.wikipedia.org/wiki/Shell_(computing)

http://en.wikipedia.org/wiki/Shell_script

http://en.wikipedia.org/wiki/Functional_programming



MATLAB provides a number of features for documenting and sharing your work. You

can integrate your MATLAB code with other languages and applications, and distribute your

MATLAB algorithms and applications.

3.2 FEATURES OF MATLAB

High-level language for technical computing.

Development environment for managing code, files, and data.

Interactive tools for iterative exploration, design, and problem solving.

Mathematical functions for linear algebra, statistics, Fourier analysis,

filtering, optimization, and numerical integration.

2-D and 3-D graphics functions for visualizing data.

Tools for building custom graphical user interfaces.

Functions for integrating MATLAB based algorithms with external applications and

languages, such as C, C++, Fortran, Java™, COM, and Microsoft Excel.

MATLAB is used in vast area, including signal and image processing, communications,

control design, test and measurement, financial modeling and analysis, and computational. Add-on

toolboxes (collections of special-purpose MATLAB functions) extend the MATLAB

environment to solve particular classes of problems in these application areas.

MATLAB can be used on personal computers and powerful server systems, including

the Cheaha compute cluster. With the addition of the Parallel Computing Toolbox, the language

can be extended with parallel implementations for common computational functions, including

for-loop unrolling. Additionally this toolbox supports offloading computationally intensive

workloads to Cheaha the campus compute cluster. MATLAB is one of a few languages in which

each variable is a matrix (broadly construed) and "knows" how big it is. Moreover, the

fundamental operators (e.g. addition, multiplication) are programmed to deal with matrices when

required. And the MATLAB environment handles much of the bothersome housekeeping that

makes all this possible. Since so many of the procedures required for Macro-Investment Analysis

involves matrices, MATLAB proves to be an extremely efficient language for both

communication and implementation.



3.2.1 INTERFACING WITH OTHER LANGUAGES

MATLAB can call functions and subroutines written in the C programming

language or FORTRAN. A wrapper function is created allowing MATLAB data types to be passed

and returned. The dynamically loadable object files created by compiling such functions are

termed "MEX-files" (for MATLAB executable).

Libraries written in Java, ActiveX or .NET can be directly called from MATLAB and

many MATLAB libraries (for example XML or SQL support) are implemented as wrappers

around Java or ActiveX libraries. Calling MATLAB from Java is more complicated, but can be

done with MATLAB extension, which is sold separately by Math Works, or using an

undocumented mechanism called JMI (Java-to-Mat lab Interface), which should not be confused

with the unrelated Java that is also called JMI.

As alternatives to the MuPAD based Symbolic Math Toolbox available from Math Works,

MATLAB can be connected to Maple or Mathematica.

Libraries also exist to import and export MathML.

Development Environment

Startup Accelerator for faster MATLAB startup on Windows, especially on

Windows XP, and for network installations.

Spreadsheet Import Tool that provides more options for selecting and loading mixed

textual and numeric data.

Readability and navigation improvements to warning and error messages in the

MATLAB command window.

Automatic variable and function renaming in the MATLAB Editor.

Developing Algorithms and Applications

MATLAB provides a high-level language and development tools that let you quickly

develop and analyze your algorithms and applications.



http://en.wikipedia.org/wiki/Fortran

http://en.wikipedia.org/wiki/MEX_file



The MATLAB Language

The MATLAB language supports the vector and matrix operations that are fundamental to

engineering and scientific problems. It enables fast development and execution. With the

MATLAB language, you can program and develop algorithms faster than with traditional

languages because you do not need to perform low-level administrative tasks, such as declaring

variables, specifying data types, and allocating memory. In many cases, MATLAB eliminates the

need for „for‟ loops. As a result, one line of MATLAB code can often replace several lines of C

or C++ code.

At the same time, MATLAB provides all the features of a traditional programming language,

including arithmetic operators, flow control, data structures, data types, object-oriented

programming (OOP), and debugging features.

MATLAB lets you execute commands or groups of commands one at a time, without compiling

and linking, enabling you to quickly iterate to the optimal solution. For fast execution of heavy

matrix and vector computations, MATLAB uses processor-optimized libraries. For general-

purpose scalar computations, MATLAB generates machine-code instructions using its JIT (Just-

In-Time) compilation technology.

This technology, which is available on most platforms, provides execution speeds that rival those

of traditional programming languages.

Development Tools

MATLAB includes development tools that help you implement your algorithm

efficiently. These include the following:

MATLAB Editor

Provides standard editing and debugging features, such as setting breakpoints and single

stepping



Code Analyzer

Checks your code for problems and recommends modifications to maximize performance

and maintainability

MATLAB Profiler

Records the time spent executing each line of code

Directory Reports

Scan all the files in a directory and report on code efficiency, file differences, file

dependencies, and code coverage

Designing Graphical User Interfaces

By using the interactive tool GUIDE (Graphical User Interface Development

Environment) to layout, design, and edit user interfaces. GUIDE lets you include list boxes, pull-

down menus, push buttons, radio buttons, and sliders, as well as MATLAB plots and Microsoft

ActiveX® controls. Alternatively, you can create GUIs programmatically using MATLAB

functions.

3.2.2 ANALYZING AND ACCESSING DATA

MATLAB supports the entire data analysis process, from acquiring data from external

devices and databases, through preprocessing, visualization, and numerical analysis, to

producing presentation-quality output.

Data Analysis

MATLAB provides interactive tools and command-line functions for data analysis

operations, including:

Interpolating and decimating

Extracting sections of data, scaling, and averaging

Thresholding and smoothing

Correlation, Fourier analysis, and filtering



1-D peak, valley, and zero finding

Basic statistics and curve fitting

Matrix analysis

Data Access

MATLAB is an efficient platform for accessing data from files, other applications,

databases, and external devices. You can read data from popular file formats, such as Microsoft

Excel; ASCII text or binary files; image, sound, and video files; and scientific files, such as HDF

and HDF5. Low-level binary file I/O functions let you work with data files in any format.

Additional functions let you read data from Web pages and XML.

Visualizing Data

All the graphics features that are required to visualize engineering and scientific data are

available in MATLAB. These include 2-D and 3-D plotting functions, 3-D volume visualization

functions, tools for interactively creating plots, and the ability to export results to all popular

graphics formats. You can customize plots by adding multiple axes; changing line colors and

markers; adding annotation, Latex equations, and legends; and drawing shapes.

2-D Plotting

Visualizing vectors of data with 2-D plotting functions that create:

Line, area, bar, and pie charts.

Direction and velocity plots.

Histograms.

Polygons and surfaces.

Scatter/bubble plots.

Animations.



3-D Plotting and Volume Visualization

MATLAB provides functions for visualizing 2-D matrices, 3-D scalar, and 3-D

vector data. You can use these functions to visualize and understand large, often complex,

multidimensional data. Specifying plot characteristics, such as camera viewing angle,

perspective, lighting effect, light source locations, and transparency.

3-D plotting functions include:

Surface, contour, and mesh.

Image plots.

Cone, slice, stream, and isosurface.

3.2.3 PERFORMING NUMERIC COMPUTATION

MATLAB contains mathematical, statistical, and engineering functions to support all

common engineering and science operations. These functions, developed by experts in

mathematics, are the foundation of the MATLAB language. The core math functions use the

LAPACK and BLAS linear algebra subroutine libraries and the FFTW Discrete Fourier

Transform library. Because these processor-dependent libraries are optimized to the different

platforms that MATLAB supports, they execute faster than the equivalent C or C++ code.

MATLAB provides the following types of functions for performing mathematical

operations and analyzing data:

Matrix manipulation and linear algebra.

Polynomials and interpolation.

Fourier analysis and filtering.

Data analysis and statistics.

Optimization and numerical integration.

Ordinary differential equations (ODEs).

Partial differential equations (PDEs).



Sparse matrix operations.

MATLAB can perform arithmetic on a wide range of data types, including doubles,

singles, and integers.



CHAPTER 4

STEGANOGRAPHY METHODS

4.1 STEGANOGRAPHY EXPLOITING IMAGE FORMAT

Steganography can be accomplished by simply feeding into a Microsoft XP command window

the following half line of code:

C:\> Copy Cover.jpg /b + Message.txt /b Stego.jpg

This code appends the secret message found in the text file „Message.txt‟ into the JPEG image

file „Cover.jpg‟ and produces the stego-image „Stego.jpg‟. The idea behind this is to abuse the

recognition of EOF (End of file). In other words, the message is packed and inserted after the

EOF tag. When Stego.jpg is viewed using any photo editing application, the latter will just

display the picture and will ignore any data coming after the EOF tag. However, when opened in

Notepad for example, our message reveals itself after displaying some data. The embedded

message does not impair the image quality. Neither the image histograms nor the visual

perception can detect any difference between the two images due to the secret message being

hidden after the EOF tag. Whilst this method is simple, a range of Steganography software

distributed online applies it (e.g., Camouflage, JpegX, Hider, etc). Unfortunately, this simple

technique would not resist any kind of editing to the Stego image nor any attacks by Steganalysis

experts.

Another naïve implementation of Steganography is to append hidden data into the

image‟s Extended File Information (EXIF- a standard used by digital camera manufacturers to

store information in the image file, such as, the make and model of a camera, the time the picture

was taken and digitized, the resolution of the image, exposure time, and focal length). This is

metadata information about the image and its source located at the header of the file. Special

agent Paul Alvarez discussed the possibility of using such headers in digital evidence analysis to

combat child pornography. This method is not a reliable one as it suffers from the same

drawback as the EOF method. Note that it is not always the case to hide text directly without

encrypting it as we did here.



4.2 STEGANOGRAPHY IN THE SPATIAL DOMAIN

In spatial domain methods a Steganographer modifies the secret data and the cover

medium in the spatial domain, which is the encoding at the level of the LSBs. This method has

the largest impact compared to the other two methods even though it is known for its simplicity .

Embedding in the 4th LSB generates more visual distortion to the cover image as the hidden

information is seen as “non-natural”.

Potdar et al. used this technique in producing fingerprinted secret sharing Steganography

for robustness against image cropping attacks. Their paper addressed the issue of image cropping

effects rather than proposing an embedding technique. The logic behind their proposed work is

to divide the cover image into sub-images and compress and encrypt the secret data. The

resulting data is then sub-divided and embedded into those images portions. To recover the data

a Lagrange Interpolating Polynomial was applied along with an encryption algorithm. The

computational load was high, but their algorithm parameters, namely the number of sub-images

(n) and the threshold value (k) were not set to optimal values leaving the reader to guess the

values. Bear in mind also that if n is set, for instance, to 32 that means we are in need of 32

public keys, 32 persons and 32 sub-images, which turns out to be unpractical. Moreover, data

redundancy that they intended to eliminate does occur in their stego-image. Shirali-Shahreza

exploited Arabic and Persian alphabet punctuations to hide messages. While their method is not

related to the LSB approach, it falls under the spatial domain. Unlike English which has only two

letters with dots in their lower case format, namely “i” and “j”, Persian language is rich in that 18

out of 32 alphabet letters have points. The secret message is binarized and those 18 letters‟ points

are modified according to the values in the binary file. Colour palette based Steganography

exploits the smooth ramp transition in colours as indicated in the colour palette. The LSBs here

are modified based on their positions in the said palette index. Johnson and Jajodia were in

favour of using BMP (24-bit) instead of JPEG images. Their next-best choice was GIF files

(256-color). BMP as well as GIF based Steganography apply LSB techniques, while their

resistance to statistical counter attack and compression are reported to be weak . BMP files are

bigger in size than other formats which render them improper for network transmissions. JPEG

images however, were at the beginning avoided because of their compression algorithm which

does not support a direct LSB embedding into the spatial domain (Fridrich et al., claimed that

changes as small as flipping the LSB of one pixel in a JPEG image can be reliably detected).



The experiments on the Discrete Cosine Transform (DCT) coefficients showed promising

results and redirected researchers‟ attention towards this type of image. In fact acting at the level

of DCT makes Steganography more robust and not as prone to many statistical attacks. Spatial

Steganography generates unusual patterns such as sorting of colour palettes, relationships

between indexed colours, exaggerated “noise”, etc, all of which leave traces to be picked up by

Steganalysis tools. This method is very fragile. There is a serious conclusion drawn in the

literature. “LSB encoding is extremely sensitive to any kind of filtering or manipulation of the

stego-image. Scaling, rotation, cropping, addition of noise, or lossy compression to the stego-

image is very likely to destroy the message. Furthermore an attacker can easily remove the

message by removing (zeroing) the entire LSB plane with very little change in the perceptual

quality of the modified stego-image”. Almost any filtering process will alter the values of many

of the LSBs. By inspecting the inner structure of the LSB, Fridrich et al., claimed to be able to

extract hidden messages as short as 0.03bpp (bit per pixel). Xiangwei et al., stated that the LSB

methods can result in the “pair effect” in the image histograms. This “pair effect” phenomenon is

empirically observed in Steganography based on the modulus operator. This operator acts as a

means to generate random (i.e., not sequential) locations to embed data. It can be a complicated

process or a simple one like testing in a raster scan if a pixel value is even then embed, otherwise

do nothing. Avcibas et al., applied binary similarity measures and multivariate regression to

detect what they call “telltale” marks generated by the 7th and 8th bit planes of a stego image.

4.3 STEGANOGRAPHY IN THE FREQUENCY DOMAIN

New algorithms keep emerging prompted by the performance of their ancestors (Spatial

domain methods), by the rapid development of information technology and by the need for an

enhanced security system. The discovery of the LSB embedding mechanism is actually a big

achievement. Although it is perfect in not deceiving the HVS, its weak resistance to attacks left

researchers wondering where to apply it next until they successfully applied it within the

frequency domain. DCT is used extensively in Video and image (i.e., JPEG) lossy compression.

Each block DCT coefficients obtained is quantized using a specific Quantization Table (QT).

This matrix shown in Figure 1 is suggested in the Annex of the JPEG standard. The logic behind

choosing such a table with such values is based on extensive experiments that tried to balance the



tradeoff between image compression and quality factors. The HVS dictates the ratios between

values in the QT.

FIGURE 1.5 JPEG suggested Luminance

Quantization Table used in DCT lossy compression. The value 16 (in bold-face) represents the

DC coefficient and the other values represent AC coefficients.

The aim of quantization is to loosen up the tightened precision produced by DCT while

retaining the valuable information descriptors. Most of the redundant data and noise are lost at

this stage hence the name lossy compression.

The quantization step is specified by:

where x and y are the image coordinates, f �(� x ,� y ) denotes the result function, f (W x ,W y )

is an 8x8 non overlapping intensity image block and is a floor rounding operator. T (Wx,Wy)

represents a quantization step which, in relationship to JPEG quality, is given by:



where, � � QT � x ,� y is the quantization table depicted in (Figure 1) and Q is a quality factor.

JPEG compression then applies entropy coding such as the Huffman algorithm to compress the

resulted T(W x ,W y ) . The above scenario is a discrete theory independent of Steganography.

Xiaoxia and Jianjun presented a Steganographic method that modifies the QT and inserts the

hidden bits in the middle frequency coefficients. Their modified QT is shown in Figure 2. The

new version of QT gives them 36 coefficients in each 8x8 block to embed their secret data into,

which yields a reasonable payload. Their work was motivated by a prior published work by

Chang et al., Steganography based on DCT JPEG compression goes through different steps as

shown in Figure 3.

FIGURE 1.6 The modified Quantization Table .

FIGURE 1.7 Data Flow Diagram showing a general process of embedding in the frequency domain.



Most of the techniques here use a JPEG image as a vehicle to embed their data. JPEG

compression uses DCT to transform successive sub-image blocks (8x8 pixels) into 64 DCT

coefficients. Data is inserted into these coefficients‟ insignificant bits. However, altering any

single coefficient would affect the entire 64 block pixels. Since the change is operating on the

frequency domain instead of the spatial domain there will be no visible changes in the cover

image.

According to Raja et al., Fast Fourier Transform (FFT) introduces round off errors, thus it

is not suitable for hidden communication. Johnson and Jajodia included it among the used

transformations in Steganography. Choosing which values in the 8x8 DCT coefficients block to

alter is very important as changing one value will affect the whole 8x8 block in the image. The

JSteg algorithm was among the first algorithms to use JPEG images. Although the algorithm

stood strongly against visual attacks, it was found that examining the statistical distribution of

the DCT coefficients yields a proof for existence of hidden data. JSteg is easily detected using

the X2-test, which is a non-parametric (a rough estimate of confidence) statistical algorithm used

in order to detect whether the intensity levels scatter in a uniform distribution throughout the

image surface or not. If one intensity level has been detected as such, then the pixels associated

with this intensity level are considered as corrupted pixels or in our case have a higher

probability of having embedded data. Moreover, since the DCT coefficients need to be treated

with sensitive care and intelligence, the JSteg algorithm leaves a serious statistical signature.

Wayner stated that the coefficients in JPEG compression normally fall along a bell curve and the

hidden information embedded by JSteg distorts this.

Manikopoulos et al., discussed an algorithm that utilizes the Probability Density Function

(PDF) used to generate discriminator features fed into a neural network system to detect hidden

data in this domain. OutGuess, developed by Provos and Honeyman, was a better alternative as it

uses a pseudo-random-number generator to select DCT coefficients. The X2-test does not detect

data that is randomly distributed. Strangely enough the developer of OutGuess himself suggests a

counter attack against his algorithm. Provos and Honeyman, suggest applying an extended

version of X2-test to select Pseudo-randomly embedded messages in JPEG images. Andreas

Westfeld based his “F5” algorithm on subtraction and matrix encoding. Neither X2-test nor its

extended versions could break this solid algorithm. Unfortunately, F5 did not survive attacks for



too long. Fridrich et al., proposed Steganalysis that does detect F5 contents, disrupting F5‟s

survival.

For the Discrete Wavelet Transform (DWT), the reader is directed to Chen‟s work.

Abdul-Aziz, and Pang , use vector quantization called Linde-Buzo-Gray (LBG) coupled with

Block codes known as BCH code and 1-Stage discrete Haar Wavelet transforms. They reaffirm

that modifying data using a wavelet transformation preserves good quality with little perceptual

artifacts.

The DWT based embedding technique is still in its infancy, Paulson reports that a group

of scientists at Iowa State University are focusing on the development of an innovative

application which they called “Artificial Neural Network Technology for Steganography

(ANNTS)” aimed at detecting all present Steganography techniques including DCT, DWT and

DFT. The Inverse Discrete Fourier Transform (IDFT) encompasses round-off error which

renders DFT improper for Steganography applications.

4.4 PERFORMANCE MEASURE

As a performance measurement for image distortion, the well known Peak-Signal-to-

Noise Ratio (PSNR) which is classified under the difference distortion metrics can be applied on

the stego images. It is defined as:

where MSE denotes the Mean Square Error which is given as:

and holds the maximum value in the image, for example:



x and y are the image coordinates, M and N are the dimensions of the image, Sxy is the

generated stego image and Cxyis the cover image.

Many authors in the literature consider Cmax =255 as a default value for 8-bit images. It

can be the case, for instance, that the examined image has only up to 253 or fewer

representations of gray colours. Knowing that Cmax is raised to the power of 2 results in a severe

change to the PSNR value. Thus we define Cmax as the actual maximum value rather than the

largest possible value. PSNR is often expressed on logarithmic scale in decibels (dB). PSNR

values falling below 30dB indicate a fairly low quality (i.e., distortion caused by embedding can

be obvious); however, a high quality stego should strive for 40dB or higher.



CHAPTER 5

EDGE DETECTION

5.1 INTRODUCTION

Edges characterize boundaries and are therefore a problem of fundamental importance in

image processing. Edges in images are areas with strong intensity contrasts a jump in intensity

from one pixel to the next. Edge detecting an image significantly reduces the amount of data and

filters out useless information, while preserving the important structural properties in an image.

There are many ways to perform edge detection. However, the majority of different methods

may be grouped into two categories, gradient and Laplacian. The gradient method detects the

edges by looking for the maximum and minimum in the first derivative of the image. The

Laplacian method searches for zero crossings in the second derivative of the image to find edges.

An edge has the one-dimensional shape of a ramp and calculating the derivative of the image can

highlight its location.

5.2 VARIOUS METHODS

Sobel

Canny

Perewitt

Laplacian

SOBEL

Based on this one-dimensional analysis, the theory can be carried over to two-dimensions

as long as there is an accurate approximation to calculate the derivative of a two-dimensional

image. The Sobel operator performs a 2-D spatial gradient measurement on an image. Typically



it is used to find the approximate absolute gradient magnitude at each point in an input grayscale

image. The Sobel edge detector uses a pair of 3x3 convolution masks, one estimating the

gradient in the x-direction (columns) and the other estimating the gradient in the y-direction

(rows). A convolution mask is usually much smaller than the actual image.

CANNY

Edges characterize boundaries and are therefore a problem of fundamental importance in

image processing. Edges in images are areas with strong intensity contrasts – a jump in intensity

from one pixel to the next. Edge detecting an image significantly reduces the amount of data and

filters out useless information, while preserving the important structural properties in an image.

This was also stated in my Sobel and Laplace edge detection tutorial, but I just wanted

reemphasize the point of why you would want to detect edges.

INTRODUCTION

The Canny edge detection algorithm is known to many as the optimal edge detector.

Canny's intentions were to enhance the many edge detectors already out at the time he started his

work. He was very successful in achieving his goal and his ideas and methods can be found in

his paper, "A Computational Approach to Edge Detection". In his paper, he followed a list of

criteria to improve current methods of edge detection. The first and most obvious is low error

rate. It is important that edges occurring in images should not be missed and that there be NO

responses to non-edges. The second criterion is that the edge points be well localized. In other

words, the distance between the edge pixels as found by the detector and the actual edge is to be

at a minimum. A third criterion is to have only one response to a single edge. This was

implemented because the first 2 were not substantial enough to completely eliminate the

possibility of multiple responses to an edge.

Based on these criteria, the canny edge detector first smoothes the image to eliminate and

noise. It then finds the image gradient to highlight regions with high spatial derivatives. The



algorithm then tracks along these regions and suppresses any pixel that is not at the maximum

(non maximum suppression). The gradient array is now further reduced by hysteresis. Hysteresis

is used to track along the remaining pixels that have not been suppressed. Hysteresis uses two

thresholds and if the magnitude is below the first threshold, it is set to zero (made a non edge). If

the magnitude is above the high threshold, it is made an edge. And if the magnitude is between

the 2 thresholds, then it is set to zero unless there is a path from this pixel to a pixel with a

gradient above T2.

STEP 1

In order to implement the canny edge detector algorithm, a series of steps must be

followed. The first step is to filter out any noise in the original image before trying to locate and

detect any edges. And because the Gaussian filter can be computed using a simple mask, it is

used exclusively in the Canny algorithm. Once a suitable mask has been calculated, the Gaussian

smoothing can be performed using standard convolution methods. A convolution mask is usually

much smaller than the actual image. As a result, the mask is slid over the image, manipulating a

square of pixels at a time. The larger the width of the Gaussian mask, the lower is the detector's

sensitivity to noise. The localization error in the detected edges also increases slightly as the

Gaussian width is increased. The Gaussian mask used in my implementation is shown below.



STEP 2

After smoothing the image and eliminating the noise, the next step is to find the edge

strength by taking the gradient of the image. The Sobel operator performs a 2-D spatial gradient

measurement on an image. Then, the approximate absolute gradient magnitude (edge strength) at

each point can be found. The Sobel operator uses a pair of 3x3 convolution masks, one

estimating the gradient in the x-direction (columns) and the other estimating the gradient in the

y-direction (rows). They are shown below:

The magnitude, or EDGE STRENGTH, of the gradient is then approximated using the

formula:

|G| = |Gx| + |Gy|

STEP 3

Finding the edge direction is trivial once the gradient in the x and y directions are known.

However, you will generate an error whenever sumX is equal to zero. So in the code there has to

be a restriction set whenever this takes place. Whenever the gradient in the x direction is equal to

zero, the edge direction has to be equal to 90 degrees or 0 degrees, depending on what the value

of the gradient in the y-direction is equal to. If GY has a value of zero, the edge direction will

equal 0 degrees. Otherwise the edge direction will equal 90 degrees. The formula for finding the

edge direction is just:

Theta = invtan (Gy / Gx)



STEP 4

Once the edge direction is known, the next step is to relate the edge direction to a

direction that can be traced in an image. So if the pixels of a 5x5 image are aligned as follows:

x x x x x

x x x x x

x x a x x

x x x x x

x x x x x

Then, it can be seen by looking at pixel "a", there are only four possible directions when

describing the surrounding pixels - 0 degrees (in the horizontal direction), 45 degrees (along the

positive diagonal), 90 degrees (in the vertical direction), or 135 degrees (along the negative

diagonal). So now the edge orientation has to be resolved into one of these four directions

depending on which direction it is closest to (e.g. if the orientation angle is found to be 3

degrees, make it zero degrees). Think of this as taking a semicircle and dividing it into 5 regions.

Therefore, any edge direction falling within the yellow range (0 to 22.5 & 157.5 to 180 degrees)

is set to 0 degrees. Any edge direction falling in the green range (22.5 to 67.5 degrees) is set to

45 degrees. Any edge direction falling in the blue range (67.5 to 112.5 degrees) is set to 90

degrees. And finally, any edge direction falling within the red range (112.5 to 157.5 degrees) is

set to 135 degrees.



STEP 5

After the edge directions are known, non maximum suppression now has to be applied.

Non maximum suppression is used to trace along the edge in the edge direction and suppress any

pixel value (sets it equal to 0) that is not considered to be an edge. This will give a thin line in the

output image.

STEP 6

Finally, hysteresis is used as a means of eliminating streaking. Streaking is the breaking up of an

edge contour caused by the operator output fluctuating above and below the threshold. If a single

threshold, T1 is applied to an image, and an edge has an average strength equal to T1, then due to

noise, there will be instances where the edge dips below the threshold. Equally it will also extend

above the threshold making an edge look like a dashed line. To avoid this, hysteresis uses 2

thresholds, a high and a low. Any pixel in the image that has a value greater than T1 is presumed

to be an edge pixel, and is marked as such immediately. Then, any pixels that are connected to

this edge pixel and that have a value greater than T2 are also selected as edge pixels. If you think

of following an edge, you need a gradient of T2 to start but you don't stop till you hit a gradient

below T1.



CHAPTER 6

EMBEDDING IN THE SKIN TONE COLOURSPACE

6.1 GENERAL

For adaptive image content retrieval in sequences of images (e.g., GIF, Video) we can

use color space transformations to detect and track any presence of human skin tone. The latter

emerged from the field of Biometrics, where the threefold RGB matrix of a given image is

converted into different colour spaces to yield distinguishable regions of skin or near skin tone.

Colour transformations are of paramount importance in computer vision. There exist several

colour spaces and here we list some of them3: RGB, CMY, XYZ, xyY, UVW, LSLM, L*a*b*,

L*u*v*, LHC, LHS, HSV, HSI, YUV, YIQ, YCbCr. Mainly two kinds of spaces are exploited in

the literature of biometrics which is the HSV and YCbCr spaces. It is experimentally found and

theoretically proven that the distribution of human skin colour constantly resides in a certain

range within those two spaces as different people differ in their skin colour (e. g., African,

European, Middle Eastern, Asian, etc). A colour transformation map called HSV (Hue,

Saturation and Value) can be obtained from the RGB bases. Sobottka and Pitas defined a face

localization based on HSV. They found that human flesh can be an approximation from a sector

out of a hexagon with the constraints: S min =0.23, S Max =0.68, H min =0 and H max =50.

6.2 COLOUR MAPPING

The other utilized colour mapping, YCbCr (Yellow, Chromatic blue and Chromatic red),

is another transformation that belongs to the family of television transmission color spaces. Hsu

et al., introduced a skin detection algorithm which starts with lighting compensation, they detect

faces based on the cluster in the (Cb/Y)-(Cr/Y) subspace. Lee et al., showed that the skin-tone has

a center point at (Cb, Cr) = (-24, 30) and demonstrated more precise model.

Based on the literature, highlighted earlier in sections 2.1, 2.2, 2.3 and 2.5, we can

conclude and point to the following facts:



Algorithms F5 and Outguess are the most reliable methods although they violate the

second order statistics as mentioned previously. Both utilize DCT embedding.

Embedding in the DWT domain shows promising results and outperforms the DCT

domain especially in surviving compression. A Steganographer should be cautious when

embedding in the transformation domains in general. However, DWT tends to be more

tolerant to embedding than DCT. Unlike JPEG the newly introduced image coding

system JPEG20004 allows for wavelets to be employed for compression in lieu of the

DCT. This makes DWT based Steganography the future central method.

Without loss of generality, edge embedding maintains an excellent distortion free output

whether it is applied in the spatial, DCT or DWT domain. However, the limited payload

is its downfall.

Most Steganographic methods do not use the actual elements of the image when hiding a

message. These elements (e.g., faces in a crowd) can be adjusted in perfectly

undetectable ways.

6.3 STEGANOFLAGE

As of now the investigation and evaluation of the ideas of the earlier techniques are going

on. We aim to embed within the edge directions in the 2D wavelet decomposition. In this way

we are guaranteed a high quality stego image. To tackle the problem of edge limited payload we

choose video files. Spreading the hidden data along the frames of the video will compensate for

the drawback of the edge embedding technique.

6.4 PROPOSED FRAMEWORK

We anticipate that Computer Vision can play a role here. Successful face localization

algorithms for colour images exploit the fact that human skin tone can be localized within a

certain range in the transform colour domain (i.e., RGB to YCbCr, HSV or Log-opponent).

Steganography can benefit from this in such a way that permits us to track and embed into the

edge of sequential appearances of human skin in the frames (e.g., faces in crowd, an athlete



exercising, etc). We can also adjust the human skin tone values, within the permissible value

ranges, to embed secret data without introducing artifacts on the carrier image.

6.5 RETRIEVAL APPLICATION

Video files indexing and content based retrieval applications have attracted a lot of

attention during the last few years and they still are areas of active research. The core of our

proposal is to find salient spatial features in image frames. We perform skin tone detection to

embed secret data in videos for the following reasons:

1) When the embedding is spread on the entire image (or frame), scaling, rotation or cropping

will result in the destruction of the embedded data because any reference point that can

reconstruct the image will be lost. However, skin tone detection in the transformed colour space

ensures immunity to geometric transforms.

2) Our suggested scheme modifies only the regions of the skin tone in the colour transformed

channel, this is done for imperceptibility reasons.

3) The skin-tone has a centre point at Cb, Cr components, it can be modelled and its range is

known statistically, therefore, we can embed safely while preserving these facts. Moreover, no

statistical breach occurs whether it is of first order or second order type.

4) If the image (or frame) is tampered with by a cropping process, it is more likely that our

selected region will be in the safe zone, because the human faces generally demonstrate the core

elements in any given image and thus protected areas (e.g., portraits).

5) Our Steganographic proposal is consistent with the object based coding approach followed in

MPEG4 and MPEG7 standards (the concept of Video Objects (VOs) and their temporal

instances, Video Object Planes (VOPs) is central to MPEG video).

6) Intra-frame and Inter-frame properties in videos provide a unique environment to deploy a

secure mechanism for image based Steganography. We could embed in any frame (e.g., 100) an



encrypted password and a link to the next frame holding the next portion of the hidden data in

the video. Note this link does not necessarily need to be in a linear fashion (e.g., frames

100�12�3...�n).

7) Videos are one of the main multimedia files available to public on the net thanks to the giant

free web-hosting companies (e.g., YouTube, Google Videos, etc). Every day a mass of these files

is uploaded online and human factors are usually present.

6.6 UNALTERED QUALITY OF THE IMAGE

Figure 4 shows how the proposed method preserves the quality of the original image.

Table 2 shows the in comparison of our approach to F5 and S-Tools which are known as strong

algorithms. The table was generated using the images shown in Figure 5. F5 and S-Tools are

available online6. S-Tools performance was discussed in our work.



FIGURE 1.8 Our proposal in action. Set A, B&C: (left) Original test images and (right) Stego

images hiding UU template. Bottom: data to hide (University of Ulster‟s logo - 47x48).

TABLE 1: Comparisons of Stego images‟ quality

Method PSNR (dB)

Set A

Steganoflage 76.917

S-Tools 68.7949

F5 53.4609

Set B

Steganoflage 71.449

S-Tools 68.144

F5 53.221

Set C

Steganoflage 70.1268

S-Tools 68.9370

F5 48.7112



CHAPTER 7

IMPLEMENTATION

7.1 GENERAL

Matlab is a program that was originally designed to simplify the implementation of

numerical linear algebra routines. It has since grown into something much bigger, and it is used

to implement numerical algorithms for a wide range of applications. The basic language used is

very similar to standard linear algebra notation, but there are a few extensions that will likely

cause you some problems at first.

7.2 CODE IMPLEMENTATION

%%%%%%%%%%%%% Biometric Inspired Digital Image Steganography %%%%%%%%%%%%%%

%%%%%%%%%%%%%%%%%%%%%%%%%% spatial domain %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

warning off;

% clear all;

% close all;

clc;

%%%%%%%%%%%%%%%%%%%%%%%%%%%% start %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

inputimage=imread('01.jpg');

hideimage=imread('Template.bmp');

sizeofhideimage=size(hideimage)

sizeinputimage = size(inputimage);

%%%%%%%%%%%%%%%%%%%%%%%% skin tone detection %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

[skinpixels noskinpixels] = skintone_detection(inputimage);

%%%%%%%%%%%%%%%%%%%%%%%% edge detection %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

[skinedges noedgepixels] = edge_detection(skinpixels);

%%%%%%%%%%%%%%%%%%%%%%%%%% template preparation %%%%%%%%%%%%%%%%%%%%%%%%%%%

template = image_template(hideimage);



%%%%%%%%%%%%%%%%%%%%%%%% verification %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

notemplatepixels = size(template,1) * size(template,2);

if noedgepixels < notemplatepixels

msgbox('No of pixels in template is more than no of pixels in skin edges','HEAVY DATA','error')

break

end

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% stegnography %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

embededimage02 = stegno_graphy(inputimage,skinedges,notemplatepixels,template);

%%%%%%%%%%%%%%%%%%%%%%%%%% performance measure %%%%%%%%%%%%%%%%%%%%%%%%%%%%

MSE1 = 0;

inputimage00 = im2double(inputimage);

embededimage03 = im2double(embededimage02);

for i = 1:size(inputimage00,1)

for j = 1:size(inputimage00,2)

MSE(i,j) = (embededimage03(i,j) - inputimage00(i,j))^2;

MSE1 = MSE1 + MSE(i,j);

end

end

MSE2 = MSE1 / (size(inputimage00,1) * size(inputimage00,2));

inputimage_max = max(max(inputimage00));

inputimage_max = (inputimage_max(1))^2;

PSNR = 10*(log10((inputimage_max / MSE2)));

disp('Peak Signal to Noise Ratio in Spatial Domain is...');

disp(ceil(PSNR));

% pause(5);

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% end %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%



EDGE DETECTION

function [inputimage03 count] = edge_detection(inputimage02)

%%%%%%%%%%%%%%%%%%%%%%%%%%% start %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

inputimage03 = edge(inputimage02,'canny');

subplot(222),imshow(inputimage03),title('\color{blue}SKIN EDGES')

count = 0;

for i = 1:(size(inputimage03,1) * size(inputimage03,2))

if inputimage03(i) == 1

count = count + 1;

end

end

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% end %%%%%%%%%%%%



OUTPUT



SKINTONE DETECTION

function [inputimage02 count] = skintone_detection(inputimage00)

%%%%%%%%%%%%%%%%%%%%%%%%%% start %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

figure('name','1. SKIN TONE DETECTION','numbertitle','off')

subplot(221),imshow(inputimage00),title('\color{blue}INPUT IMAGE')

inputimage01 = rgb2ycbcr(inputimage00);

inputimage_y = inputimage01(:,:,1);

inputimage_cb = inputimage01(:,:,2);

inputimage_cr = inputimage01(:,:,3);

subplot(222),imshow(inputimage_y),title('\color{blue}LUMINANCE "Y"')

subplot(223),imshow(inputimage_cb),title('\color{blue}CROMINANCE "Cb"')

subplot(224),imshow(inputimage_cr),title('\color{blue}CROMINANCE "Cr"')

count = 0;

for i =1:size(inputimage_cr,1)

for j = 1:size(inputimage_cr,2)

if inputimage_cr(i,j) >= 140

inputimage02(i,j) = 255;

count = count + 1;

else

inputimage02(i,j) = 0;

end

end

end

figure('name','2. SKIN EDGE DETECTION','numbertitle','off')

subplot(221),imshow(inputimage02),title('\color{blue}SKIN PIXELS')

%%%%%%%%%%%%%%%%%%%%%%%%%%%% end %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%



OUTPUT



EMBEDDING PROCESS

function embededimage02 = stegno_graphy(inputimage,edge1,notemplatepixels,template)

%%%%%%%%%%%%%%%%%%%%%%%%%%%%% start %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

[inputimage05,property] = rgb2ind(inputimage,255);

embededimage = [];

k = 1;

for i = 1:size(edge1,1)

for j = 1:size(edge1,2)

if edge1(i,j) == 255 && k >= notemplatepixels

embededimage(i,j) = inputimage05(i,j) + template(k);

k = k + 1;

else

embededimage(i,j) = inputimage05(i,j);

end

end

end

embededimage01 = uint8(embededimage);

embededimage02=ind2rgb(embededimage01,property);

figure('name','3. STEGNOGRAPHY','numbertitle','off')

imshow(embededimage02),title('\color{blue}STEGO IMAGE')

%%%%%%%%%%%%%%%%%%%%%%%%%%% end %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%



OUTPUT

THE PSNR IN SPATIAL DOMAIN - 38

THE PSNR IN FRQUENCY DOMAIN - 27



CHAPTER 8

APPLICATIONS

Applications:

Digital Imaging.

Adaptive steganography.

Digital steganography.



CHAPTER 9

CONCLUSION AND REFERENCES

9.1 CONCLUSION

Digital Steganography is a fascinating scientific area which falls under the umbrella of

security systems. We have presented in this work some background discussions on algorithms of

Steganography deployed in digital imaging. The emerging techniques such as DCT, DWT and

Adaptive Steganography are not an easy target for attacks, especially when the hidden message

is small. That is because they alter bits in the transform domain, thus image distortion is kept to a

minimum. Generally these methods tend to have a lower payload compared to spatial domain

algorithms. In short there has always been a trade off between robustness and payload. Our

proposed framework, Steganoflage, is based on edge embedding in the DWT domain using skin

tone detection in RGB sequential image files. We chose to use the latter to compensate for the

limited capacity that edge embedding techniques demonstrate. We use the actual elements of the

image when hiding a message. This leads to many exciting and challenging future research

problems.

9.2 REFERENCES

1. Johnson, N. F. and Jajodia, S.: Exploring Steganography: Seeing the Unseen. IEEE Computer,

31 (2): 26-34, Feb 1998.

2. Judge, J.C.: Steganography: Past, Present, Future. SANS Institute publication, December 1,

2001.Retrieved from: http://www.sans.org/reading_room/whitepapers

/stenganography/552.php

3. Provos, N. and Honeyman, P.: Hide and Seek: An Introduction to Steganography. IEEE

Security and Privacy, 01 (3): 32-44, May-June 2003.



4. Moulin, P. and Koetter, R.: Data-hiding codes. Proceedings of the IEEE, 93 (12): 2083- 2126,

Dec. 2005.

5. Sadkhan, S. B.: Cryptography: Current Status and Future Trends. IEEE International

Conference on Information & Communication Technologies: From Theory to Applications.

Damascus. Syria: April 19 - 23, 2004.

6. Simmons, G. J.: The Prisoners‟ Problem and the Subliminal Channel. Proceedings of

CRYPTO83- Advances in Cryptology, August 22-24. 1984. pp. 51.67.

7. Kurak, C. and McHugh, J.: A cautionary note on image downgrading. Proceedings of the

Eighth Annual Computer Security Applications Conference. 30 Nov-4 Dec 1992 pp. 153-159.

8. Thomas, T. L.: Al Qaeda and the Internet: The Danger of “Cyberplanning”. Parameters, US

Army War College Quarterly - Spring 2003. Retrieved from:

http://www.carlisle.army.mil/usawc/Parameters /03spring/thomas.pdf on 22-Nov-2006.

9 Petitcolas, F.A.P.: “Introduction to Information Hiding”. In: Katzenbeisser, S and Petitcolas,

F.A.P (ed.) (2000) Information hiding Techniques for Steganography and Digital Watermarking.

Norwood: Artech House, INC.

10. Bender, W., Butera, W., Gruhl, D., Hwang, R., Paiz, F.J. and Pogreb, S.: Applications for

Data Hiding. IBM Systems Journal, 39 (3&4): 547-568. 2000

11. Hernandez-Castro, J. C., Blasco-Lopez, I. and Estevez-Tapiador, J. M.: Steganography in

Games: A general methodology and its application to the game of Go. Computers & Security,

25(2006): 64- 71.

12. Jakubowski, J., Kwiatos, K., Chwaleba, A. and Osowski, S.: Higher Order Statistics and

Neural Network for Tremor Recognition. IEEE Transactions on Biomedical Engineering, 49 (2):

February 2002.



13. Areepongsa, S. Kaewkamnerd, N. Syed, Y. F. and Rao. K. R.: Exploring On Steganography

For Low Bit Rate Wavelet Based Coder In Image Retrieval System. IEEE Proceedings of

TENCON 2000. (3):250-255. Kuala Lumpur, Malaysia. 2000.

14. Kruus, P., Scace, C., Heyman, M. and Mundy, M.: A survey of Steganographic Techniques

for Image Files. Advanced Security Research Journal. V(I): 41- 51, Winter 2003.

15. Alvarez, P.: Using Extended File Information (EXIF) File Headers in Digital Evidence

Analysis. International Journal of Digital Evidence, 2 (3). Winter 2004.

16. Lin, E. T. and Delp, E. J.: A Review of Data Hiding in Digital Images. Retrieved on

1.Dec.2006 from Computer Forensics, Cyber crime and Steganography Resources, Digital

Watermarking Links and Whitepapers, Apr 1999.

17. Kermani, Z. Z. and Jamzad, M.: A Robust Steganography Algorithm Based on Texture

Similarity using Gabor Filter. Proceedings of IEEE 5th International Symposium on Signal

Processing and Information Technology, 18-21 Dec. 2005, 578- 582.

18. Potdar, V. M., Han, S. and Chang, E.: Fingerprinted Secret Sharing Steganography for

Robustness against Image Cropping Attacks. Proceedings of IEEE's 3rd

International Conference

on Industrial Informatics (INDIN), Perth, Australia, 10-12 August 2005.

19. Shirali-Shahreza, M. H. and Shirali-Shahreza, M.: A New Approach to Persian/Arabic Text

Steganography. Proceedings of 5th IEEE/ACIS International Conference on Computer and

Information Science (ICIS-COMSAR 2006), 10-12 July 2006, 310- 315.

20. Marvel, L. M. and Retter, C. T.: A Methodology for Data Hiding Using Images. Proceedings

of IEEE Military Communications Conference (MILCOM98) Proceedings, Boston, MA, USA,

18-21 Oct 1998, 1044-1047.

image seg final report

Documents

image segmentation

digitized image

image dissector

resulting image

term digital image

functions of image acquisition

digital processing

output of segmentation