CHAPTER 1

INTRODUCTION

1.1 Steganography
The word steganography is derived from the Greek words stegos meaning cover and grafia
meaning writing ,defining it as covered writing. In image steganography the information is hidden
exclusively in images. Steganography is the art and science of secret communication . It is the
practice of encoding/embedding secret information in a manner such that the existence of the
information is invisible. The original files can be referred to as cover text, cover image, or cover
audio. After inserting the secret image it is referred to as stego-medium. A stego-key is used for
hiding/encoding process to restrict detection or extraction of the embedded data.
1.2 Difference between Steganography and cryptography
Steganography differs from cryptography
Steganography
Hide the images inside the Cover medium, Many Carrier formats.
Breaking of steganography is known as Steganalysis.
Cryptography
Encrypt the image before sending to the destination, no need of carrier/cover
medium.Breaking of cryptography is known as Cryptanalysis.Watermarking and fingerprinting
related to steganography are basically used for intellectual property protection. A digital watermark
is a kind of marker covertly embedded in a noise-tolerant signal such as audio or image data. It is
typically used to identify ownership of the copyright of such signal. The embedded information in
a watermarked object is a signature refers the ownership of the data in order to ensure copyright
protection. In fingerprinting, different and specific marks are embedded in the copies of the work
that different customers are sup- posed to get. In this case, it becomes easy for the property owner to
find out such customers who give themselves the right to violate their licensing agreement when they
illegally transmit the property to other groups.
1.3 Evaluation of Steganography
The term steganography came into use in 1500s after the appearance of Trithemius
book on the subject Steganographia.The term stegna
1.3.1 Past
The word Steganography technically means covered or hidden writing. Its ancient origins can be
traced back to 440 BC. Although the term steganography was only coined at the end of the 15th
century, the use of steganography dates back several millennia. In ancient times, messages were
hidden on the back of wax writing tables, written on the stomachs of rabbits, or tattooed on the
scalp of slaves. Invisible ink has been in use for centuries for fun by children and students and for
serious undercover work by spies and terrorists.
1.3.2 Present
Majority of today steganographic systems uses multimedia objects like image, audio, video etc.
as cover media because people often transmit digital pictures over email and other Internet
communication. Modern steganography uses the opportunity of hiding information into digital
multimedia files and also at the network packet level.
 Hiding information into a medium requires following elements
 The cover medium(C) that will hold the secret message.
The secret message (M), may be plain text, digital image file or any type of data.
In modern approach, depending on the cover medium, steganography can be di- vided into five
types:
1. Text Steganography
2. Image Steganography
3. Audio Steganography
4. Video Steganography
5. Protocol Steganography
Text steganography
Hiding information in text file is the most common method of steganography. The
method was to hide a secret message into a text message. After coming of Internet and different
type of digital file formats it has decreased in importance. Text stenography using digital files is not
used very often because the text files have a very small amount of excess data.
Image steganography
Images are used as the popular cover medium for steganography. A image is embedded in a
digital image using an embed- ding algorithm, using the secret key. The resulting stego-image is
send to the receiver. On the other side, it is processed by the extraction algorithm using the same
key. During the transmission of stego- image unauthenticated persons can only notice the
transmission of an image but cant see the existence of the hidden image.
Audio steganography
Audio steganography is concerned with embedding information in an
innocuous cover speech in a secure and robust manner. Communication and transmission security
and robustness are essential for transmitting vital information to intended sources while denying
access to unauthorized persons. An audible, sound can be inaudible in the presence of another
louder audible sound .This property allows to select the channel in which to hide information.
Existing audio steganography software can embed messages in WAV and MP3 sound files. The list
of methods that are commonly used for audio steganography are listed and discussed below.
 LSB coding
 Parity coding
 Phase coding
 Spread spectrum
 Echo hiding
Video steganography
Video Steganography is a technique to hide any kind of files in any extension into
a carrrying Video file.
Protocol steganography
The term protocol steganography is to embed- ding information within network
protocols such as TCP/IP. We hide information in the header of a TCP/IP packet in some fields that
can be either optional or are never used.
1.4. Applications of Steganography
Secret Communications
The use steganography does not advertise secret communication and therefore avoids
scrutiny of the sender, message and recipient. A trade secret, blueprint, or other sensitive
information can be transmitted without alerting potential attackers.
Feature Tagging
Elements can be embedded inside an image, such as the names of individuals in
a photo or locations in a map. Copying the stego image also copies all of the embedded features
and only parties who possess the decoding stego key will be able to extract and view the features.
Copyright Protection
Copy protection mechanisms that prevent data, usually digital data, from being
copied. The insertion and analysis of water marks to protect copyrighted material is responsible
for the recent rise of interest in digital steganography and data embedding.
1.5 Objective
The project is carried out with the following objectives
 To hide the image or a secret image into an image which acts as a cover medium using
 LSB technique and visual secret sharing.
 The primary motivation of my current work is to increase SNR (signal to noise ratio)
and histogram.
 CHAPTER 2
LITERATURE SURVEY

In the year of 2013 Prema, G.; Natarajan, S.,[1] Investigated on Image steganography is an
emerging field of research for secure data hiding and transmission over networks. The proposed
system provides the best approach for Least Significant Bit (LSB) based steganography using
Genetic Algorithm (GA) along with Visual Cryptography (VC). Original message is converted
into cipher text by using secret key and then hidden into the LSB of original image. Genetic
Algorithm and Visual Cryptography has been used for enhancing the security. Genetic Algorithm
is used to modify the pixel location of stego image and the detection of this message is complex.
Visual Cryptography is used to encrypt the visual information. It is achieved by breaking the
image into two shares based on a threshold. The performance of the proposed system is
experimented by performing steganalysis and conducting benchmarking test for analysing the
parameters like Mean Squared Error (MSE) and Peak Signal to Noise Ratio (PSNR). The main
aim of this paper is to design the enhanced secure algorithm which uses both steganography using
Genetic Algorithm and Visual Cryptography to ensure improved security and reliability.

In the year of 2012 Nickfarjam, A.M.; Azimifar, Z.,[2] studied a novel approach for image
steganography by taking the advantages of Particle Swarm Optimization (PSO) and Least
Significant Bits (LSBs) replacement. This technique is based on hiding the Most Significant Bits
(MSBs) of secret image pixels in LSBs of a host image. The proposed method finds the best pixel
in order to embed. Authors define four feature functions and four corresponding coefficients to
rank the pixels. The features and the coefficients are defined based on the MSBs of host image.
Their method defines a special secret key for each host image based on PSO in which, each
particle represents a potential solution and authors can evaluate all of them. This novelty causes
better exploration of search space in order to find suitable pixel ranking and higher security.
The experimental outcomes show the superiority of this approach over the state-of-the-art
methods.
In the year of 2012 Sanchez, A.; Conci, A.; Zeljkovic, E.; Behlilovic, N.; Karahodzic, V.,[3]
presented the study on increasing range of data types which are exchanged over this network
(video, audio, text messages...), emphasize the security problem that this way of
communication has. The flood of multimedia contents in the structure of transmitted data has
made the appearance of images in this network quite normal. This revived the use of
steganography (hiding data within images) in order to hide information to avoid unauthorized
access. A much used technique for this purpose, the LSB (Least Significant Bits) technique,
still leads to visible changes in the original image, which was chosen to be the message
carrier. These differences make quite a path for a cryptanalyst to doubt the authenticity
(independence) of the picture itself. However, by using GA (Genetic Algorithm), the
differences between the original image and the image embedded with secret data can be
reduced. However, the difference between the original image and the image with embedded
information still remains, while the achieved improvements are paid with an increase of
computational complexity. Naturally a question arises: Can the image be embedded with
information in a way so that it does not undergo any changes? Most fast responses would be
that it is not possible. This paper shows that this is in fact possible.

In the year of 2012 Qiangfu Zhao; Akatsuka, M.; ChengHsiung Hsieh, [4] proposed the study
of an image morphing based method for information hiding. The basic idea is to hide a
secrete image into a morphed image which is obtained from the secrete image and another
reference image. To make this method practically useful, it is necessary to produce natural
morphed images. This is a necessary condition conceal the existence of the secret image. To
produce natural morphed images, we should choose a proper feature point set (FPS) for
morphing. This is a tedious work if we do it manually, because the number of possible FPSs
is very large. To solve the problem more efficiently, we adopted interactive genetic algorithm
(IGA) in this study and conducted experiments for generating facial images. Outcomes show
that, if we provide a relatively good initial FPS, IGA can finetune the FPS, and produce more
natural facial images with limited number of evaluations.

In the year of 2012 Khosravi, M.; SoleymanpourMoghaddam, S.; Mahyabadi, M.,[5]

investigated a novel steganographic method is proposed which is based on the spatial domain:
Least Significant Bit (LSB). The LSB matching method proposed by Mielikainen utilizes a
binary function to reduce the number of changed pixel values. While in this paper a bipolar
evaluating system is proposed to assess the performance of different orders for LSB
matching. Afterward a genetic algorithm strategy is employed to search for an optimal
solution among all the permutation orders. The experimental outcomes show that by
employing the proposed bipolar evaluating system, the distortion of the stego-image is
reduced while the probability of detection is decreased.
In the year of 2011 Brazil, A.L.; Sanchez, A.; Conci, A.; Behlilovic, N.,[6] The study of
Internet connects more people, increasing the need for transmitting secure information. One
way to protect the data sent over the web is to conceal the relevant information inside a
typical image, hiding the data from intruders. This paper proposes a hybrid heuristic,
combining a genetic algorithm and the path relinking metaheuristic to efficiently solve this
problem. Computational outcomes show that the proposed algorithm outperforms the LSB
(least significant bits) substitution technique, concerning the quality of solutions. In this way,
the inclusion of a path relinking procedure can significantly improve the performance of a
genetic algorithm for the problem considered.

In the year of 2011 Bhowal, K.; Sarkar, D.; Biswas, S.; Sarkar, P.P.,[7] presented the
Audio Steganography that is a method that ensures secured data transfer between parties
normally in internet community. In this paper, authors present a novel, principled approach to
resolve the remained problems of substitution technique of Audio Steganography.
In the first level, here authors first extract image data from an image file. In the second level,
authors use a powerful GA (Genetic Algorithm) based LSB (Least Significant Bit) Algorithm
to embed t h e image data into audio data. Here image data bits are embedded into random
and higher LSB layers, consequential in increased robustness against noise addition. On the
other hand, GA operators are used to reduce the distortion.
In the year of 2011 Mandal, J.K.; Khamrui, A.,[8] proposed a study of Image steganography
that is the art of hiding information onto the cover image. In this paper a Genetic Algorithm
based color image authentication/data hiding technique through steganographic approach, in
frequency domain using Discrete Fourier Transform (DFT) termed as GASFD, has been
proposed. 2×2 masks are taken from the source image in row major order where DFT is used
to transform original image (cover image) block from spatial domain to frequency domain.
Three bits of the hidden image are embedded per byte of the source image onto the rightmost
3 bits of each pixel excluding the first byte of each mask, as a effect large volume of
message/ image is embedded in frequency domain. 2×2 embedded image mask is transformed
from frequency domain to spatial domain using inverse DFT. Resulting image mask of size
32 bits are taken as initial population. New Generation and Crossover are applied on the
initial population to obtain stego image. In the process of embedding dimension of the hidden
image followed by the content of the message/hidden image are embedded. Reverse process
is followed during decoding. Genetic algorithm is used to enhance the security level.
Various statistical parameters computed are compared with the existing Discrete Cosine
Transformation based steganographic algorithms devised by Hashad A. I. et al. [9] which
shows that proposed GASFD obtained better outcomes in terms of large message embedding
and consistent PSNR.
 CHAPTER 3
EXISTING WORK

3.1 Proposed Secure Information Hiding

An information hiding system has been developed for confidentiality. However, in this
chapter, we study an image file as a carrier to hide image. Therefore, the carrier will be
known as cover-image, while the stego-object known as stego-image. The implementation
of system will only focus on Least Significant Bit (LSB) as one of the steganography
techniques as mentioned in below.

3.1.1 Least-Significant Bit (LSB) Technique

The least significant bit (in other words, the 8th bit) of some or all of the bytes inside an
image is changed to a bit of the secret image. Digital images are mainly of two types (i) 24 bit images
and (ii) 8 bit images. In 24 bit images we can embed three bits of information in each pixel, one in
each LSB position of the three eight bit values. Increasing or decreasing the value by changing the
LSB does not change the appearance of the image; much so the resultant stego image looks almost
same as the cover image. In 8 bit images, one bit of information can be hidden. A stego image is
obtained by applying LSB algorithm.
The hidden image is extracted from the stego image by applying the reverse process. If the LSB
of the pixel value of cover image C(i,j) is equal to the image bit m of secret massage to be
embedded, C(i,j) remain unchanged; if not, set the LSB of C(i, j) to m. The image embedding
procedure is given below

S(i,j) = C(i,j) - 1, if LSB(C(i,j)) = 1 and m = 0

S(i.j) = C(i,j), if LSB(C(i,j)) = m
S(i,j) = C(i,j) + 1, if LSB(C(i,j)) = 0 and m = 1
where LSB(C(i, j)) stands for the LSB of cover image C(i,j) and m is the next image bit to be
embedded.

S(i,j) is the stego image

As we already know each pixel is made up of three bytes consisting of either a 1 or a 0.
For example, suppose one can hide a image in three pixels of an image (24-bit colors). Suppose
the original 3 pixels are:
(11101010 11101000 11001011)
(01100110 11001010 11101000)
(11001001 00100101 11101001)
A steganographic program could hide the letter ”J” which has a position 74 into ASCII
character set and have a binary representation ”01001010”, by altering the channel bits of
pixels.
(11101010 11101001 11001010)
(01100110 11001011 11101000)
(11001001 00100100 11101001)
In this case, only four bits needed to be changed to insert the character successfully. The resulting
changes that are made to the least significant bits are too small to be recognized by the human
eye, so the image is effectively hidden. The advantage of LSB embedding is its simplicity and
many techniques use these methods.
LSB embedding also allows high perceptual transparency.
The following figure3.1 and 3.2 shows the mechanism of LSB technique

FIGURE 3.1: LSB insertion Mechanism

 FIGURE 3.2: LSB extraction Mechanism

3.1.2 Data Embedding

The embedding process is as follows.
Inputs Cover image, stego-key and the text file
Output stego image
Procedure
Step 1: Extract the pixels of the cover image.
Step 2: Extract the characters of the text file.
Step 3: Extract the characters from the Stego key.
Step 4: Choose first pixel and pick characters of the Stego key and place it in first component of
pixel.
Step 5: Place some terminating symbol to indicate end of the key. 0 has been used as a
terminating symbol in this algorithm.
Step 6: Insert characters of text file in each first component of next pixels by replacing it.
Step 7: Repeat step 6 till all the characters has been embedded.
Step 8: Again place some terminating symbol to indicate end of data.
Step 9: Obtained stego image.[17]

3.1.3. Data Extraction

The extraction process is as follows.
Inputs Stego image file, stego key
Output Secret text image.

Procedure:
Step 1: Extract the pixels of the stego image.
Step 2: Now, start from first pixel and extract stego key characters from first component of the
pixels.
Step 3: up to terminating symbol, otherwise follow
Step 4: If this extracted key matches with the key entered by the receiver, then follow
otherwise terminate the program.
Step 5: If the key is correct, then go to next pixels and extract secret image characters from
first component of next pixels. Follow Step 5 till up to terminating symbol, otherwise follow step
6.
Step 6: Extract secret image[18,20].
3.1.4. Image Encoding Algorithm
Inputs Image file, stego key and image file
Output Stego image.
1. The cover and secret images are read and converted into the unit8 type.
2. The numbers in secret image matrix are conveyed to 8-bit binary. Then the matrix is
reshaped to a new matrix a.
3. The matrix of the cover image is also reshaped to matrix b
4. Perform the LSB technique described above
5. The stego-image, which is very similar to the original cover image, is achieved by
reshaping matrix b.
6. While extracting the data, the LSB of the stego image is collected and they are
reconstructed into the decimal numbers. The decimal numbers are reshaped to the
secret image[10].
3.2 visual sharing scheme
Visual cryptography (VC), which was proposed by Naor and Shamir, allows the encryption
of secret information in the image form [1]. With the network is more and more popular, the
hackers utilize leak of the Internet to steal information that they want.
Therefore, secure data transmitting becomes very important. In the recent years, generally
using the traditional cryptology to avoid the data to be altered, but it needs complex
computation to decode.
 In order to reduce the computation and furthermore secure the data, Naor and Shamir
proposed a new cryptology called visual cryptography in 1994. Without huge calculation, it
can restore encrypted images by stacking two shares via human visual system to identify.
The first visual cryptography scheme is used for the black-and-white image in disorder to
embed the confidential image. These disordered images are called “shares” that one of them
may regard as the cipher text and the other treats as the key.
Hackers cannot decrypt the secret message from one share. Later on this theory, visual
cryptography can extend as (k, n) – threshold visual secret sharing scheme that divides n
transparencies into secret information. When be decoded, the owner must have k or more
shares to stack.
In 1998, Chen and Wu proposed a new visual cryptography scheme. It improves the
drawback of the conventional visual cryptography that two share images only can embed.
In a k-out-of-n scheme of VC, a secret binary image is cryptographically encoded into n
shares of random binary patterns. The n shares are Xeroxed onto n transparencies,
respectively, and distributed among n participants, one for each participant. No participant
knows the share given to another participant.
Any k or more participants can visually reveal the secret image by superimposing any k
transparencies together. The secret cannot be decoded by any k-1 or fewer participants
[4].There are many algorithms to encrypt the image in another image, but a few of them have
been in visual cryptography for colour image.
In this project, the different approach have been produced for the visual cryptography for
colour image, the proposed algorithm splits a secret image into two shares based on three
primitive colour components.
In this technique, A random key is used to choose the pixels randomly and embed the image.
This will make the image bits more difficult to find and hopefully reduce the realization of
patterns in the image [9]. Data can be hidden in the LSB of a particular colour plane (Red
plane) of the randomly selected pixel in the RGB colour space.
3.2.1 Embedding Algorithm
In this process of encoding method, a random key is used to randomized the cover image and
then hide the bits of a secret image into the least significant bit of the pixels within a cover
image.
The transmitting and receiving end share the stego key and random-key.
The random-key is usually used to seed a pseudo-random number generator to
select pixel locations in an image for embedding the secret image.
Inputs : Cover image, stego key and the image
Output : stego image
1. Read image from folder that is to be hidden.
2. Read the RGB colour image(cover image) into which the image is to be embedded.
3. Read the last bit of red pixel.
4. Initialize the random key and Randomly permute the pixels of cover image an reshape
into a matrix.
5. Initialize the stego-key and XOR with text file to be hidden and give image.
6. Insert the bits of the secret image to the LSB of the Red plane’s pixels.
7. Write the above pixel to Stego Image File[19].
 3.2.2 Extraction of Hidden Image

In this process of extraction, the process first takes the key image. These keys takes out the
points of the LSB where the secret image is randomly distributed. Decoding process searches
the hidden bits of a secret image into the least significant bit of the pixels within a cover image
using the random key.
In decoding algorithm the key must match. Then receiver can extract the embedded images exactly
using only the stego key.
The steps to extract the hidden images is given in the step, below the random key recovered
steps is also given in the below algorithm steps,
3.2.3 Image extraction algorithm
Inputs Stego image file, stego key,random key.
Output Secret image.
1) Open the Stego image file in read mode and from the Image file, read the RGB colour
of each pixel.
2) Extract the red component of the host image.
3) Read the last bit of each pixel.
4) Initialize the key that gives the position of the image bits in the red pixel that are
embedded randomnly.
5) For decoding, select the pixels and Extract the LSB value of red pixels.
6) Read each of pixels then content of the array converts into decimal value that is
actually ASCII value of hidden character.
7) ASCII values got from above is XOR with stego-key and gives image file, which we
hide inside the cover image
 CHAPTER 4
PROPOSED WORK

4.1 Performance Measure

As a performance measure for image distortion due to hiding of image, the well-known
peak-signal-to noise ratio (PSNR), which is categorized under difference distortion metrics, can be
applied to stego images. The PSNR block computes the peak signal-to-noise ratio, in decibels,
between two images. The higher the PSNR, the better the quality of the compressed, or
reconstructed image.
The mean-square error (MSE) and the peak signal-to-noise ratio (PSNR) are used to
compare image compression quality.
It is defined as:
PSNR = 10lg(Cmax)2/MSE.
MSE = mean square error,
which is given as:
MSE = 1/MN( (S-C)2).
Cmax = 255.
Where M and N are the dimensions of the image,
S is the resultant stego image, and C is the cover image.
PSNR values below 30 dB indicate low quality (i.e., distortion caused by embed- ding is high).
A high-quality stego image should strive for a PSNR of 40 dB.
 4.2 Implementation and Evaluation of above two techniques
We have implemented the above two techniques in MATLAB and the above
mentioned algorithms with respect to image steganography are not void of weak and strong points.
Consequently, it is important to decide the most suitable approach to be applied. As defined before,
there are several parameters to measure the performance of the steganographic system. Some
parameters are as follows:
Perceptibility does embedding information distort cover medium to a visually unacceptable level.
Capacity how much information can be hidden (relative to the change in
perceptibility) item.
Robustness to attacks can embedded data survive manipulation of the stego medium in an effort
to destroy, remove, or change the embedded data.

TABLE 4.1: Comparison of characters of above two techniques

SL.No Imperceptibility Robustness Capacity Tamper Resistance

Simple LSB High* Low High Low
LSB-VSS Encoding Higher* Low High High**

*: Indicates dependency on the used cover image

**: Indicates dependency on the used key and random seed

4.3 Introduction to the Discrete Wavelet Transform (DWT)

Image compression is used to reduce the image size and redundancy of the image
data.
The amount of data used to represent these image, therefore needs to be reduced. Image
compression deals with redundancy, the number of bits needed to represent on image by
removing redundant data.
Decreasing the redundancy is the main aim of the image compression algorithms. Image
compression technique, mostly used two dimensional (2D) image compression standards, such us
JPEG, JPRG-LS or JPRG2000 generally consider only intra brand Correlation.
Image compression is broadly classified into two categories namely Lossy and Lossless depending on
whether the original image can be recovered with fill mathematic precision from the compressed
image.

This is meant to be a brief, practical introduction to the discrete wavelet

transform (DWT), which aug-ments the well written tutorial paper by Amara Graps [1].
Therefore, this document is not meant to be comprehensive, but does include a discussion on
the following topics:

1. Qualitative discussion on the DWT decomposition of a signal.

2. Procedure for computing the forward and inverse DWT; and
3. The 2D DWT.
4.3.1 DWT decomposition

In Fourier analysis, the Discrete Fourier Transform (DFT) decompose a signal into sinusoidal
basis functions of different frequencies. No information is lost in this transformation; in other
words, we can completely recover the original signal from its DFT (FFT) representation.

In wavelet analysis, the Discrete Wavelet Transform (DWT) decomposes a signal into a set of
mutually orthogonal wavelet basis functions. These functions differ from sinusoidal basis
functions in that they are spatially localized – that is, nonzero over only part of the total signal
length. Furthermore, wavelet functions are dilated, translated and scaled versions of a a
common function φ, known as the mother wavelet. As is the case in Fourier analysis, the
DWT is invertible, so that the original signal can be completely recovered from its DWT
representation.

Unlike the DFT, the DWT, in fact, refers not just to a single transform, but rather a set of
transforms, each with a different set of wavelet basis functions. Two of the most common are
the Haar wavelets and the Daubechies set of wavelets. For example, Figures 1 and 2
illustrate the complete set of 64 Haar and Daubechies-4 wavelet functions (for signals of
length 64), respectively. Here, we will not delve into the details of how these were derived;
however, it is important to note the following important properties:

1. Wavelet functions are spatially localized;

2. Wavelet functions are dilated, translated and scaled versions of a common mother
wavelet.
3. Each set of wavelet functions forms an orthogonal set of basis functions.

IMAGE COMPRESSION AND TECHNIQUE

The term data image compression refers to the process of reducing the amount of data
required to represent a given amount of information.
A clear distinction must be made between data and information. Data redundancy is a central
issue in digital image compression.
There are two types if image compression technique Lossy technique and Lossless technique.
DCT is used in signal, image processing especially for Lossy compression because it has a
strong energy compaction to create predictions according to its local uniqueness.
The Lossy image compression did not give Proper vision of the image, but it gives good
compression ratio of the image.
DWT is used to separate the image into a pixel. DWT is used in signal and image processing
especially for lossless image compression.
DWT is also used for Lossy compression. The Lossless image compression is mostly used in
DWT Lossless image compression give the good quality of the image and also the
compression ratio of the image also good. The PSNR ratio of the image is also good in the
Lossless compression.
 Table for Comperision of image compression

TABLE 1 IMAGE COMPRESSION TECHNIQUE

S.NO IMAGE PAPER A PAPER B PAPER C PAPER D

NAME DWT DWT DWT DWT

1 Lena 39.05 29.50 17.65 34.65

2 Peppers 39.12 35.17 15.74 35.17
3 Boats 33.81 38.30 17.39 35.51
4 Goldhills 37.77 27.75 12.05 32.30
5 Barbara 31.61 24.15 19.59 32.04

4.3.2 DWT in one dimension

In this section, we describe the algorithm for computing the one-dimensional DWT
and its inverse.

4.3.3 Forward DWT

The (one-dimensional) DWT operates on a real-valued vector x of length 2n , n ∈ {2,

3, . . . }, and results in a transformed vector w of equal length. Figure 3(a) and (b) illustrate
the first two steps of the DWT for a vector of length 16. First, the vector x is filtered with
some discrete-time, low-pass filter (LPF) h of given length (in the Figures, we use length four
for illustration purposes) at intervals of two, and the resulting
 a) b)

Figure 3: (a) First step of the DWT for a signal of length 16: The original signal is
low-pass filtered in increments of two, and the resulting coefficients are grouped as the first
eight elements of the vector. (b) Second step of the DWT: The original signal is high-pass
filtered in increments of two, and the resulting coefficients are grouped as the last eight
elements of the vector.

The values are stored in the first eight elements of w. This step is illustrated in Figure
3(a). Second, the vector x is filtered with some discrete-time, high-pass filter (HPF) g of
given length (again, for illustration purposes, we use a filter of length four) at intervals of
two, and the resulting high-pass values are stored in the last eight elements of w. This step is
illustrated in Figure 3(b).
Note, qualitatively, how this procedure transforms the vector x. The low-pass part of
the vector w is essentially a down-sampled version (down-sampled by a factor of two) of the
original signal x, while the high-pass part of the vector w detects and localizes high
frequencies in x. If we were to stop here, the vector w would be a one-level wavelet
transform of x. We need not, however, stop here; the low-pass filtered part of w (first eight
elements for this example) can be further transformed using the identical procedure as
outlined above and shown in Figure 3. Figure 4, for example, illustrates a three-level, one-
dimensional DWT.
Note that in the final transform w3 , values L3 are the result of three consecutive low-pass
filters, values H3 are the result of two consecutive low-pass filtering operations followed by a
high-pass filter, values H2 are the result of a low-pass filter followed by a high-pass filter, and
values H1 are the result of one high-pass filter. Therefore, the highest frequencies will be
isolated and localized in values H1 of w3, intermediate frequencies will be isolated and
localized in values H2 of w3 , etc. Note that for lower frequencies, the resolution is decimated
by half for each level of the wavelet transform. Thus, the DWT operation implicitly
recognizes that lower frequencies cannot be localized at the same resolution as higher
frequencies (see Heisenberg Uncertainty Principle). In summary, the one-dimensional DWT
is a multi-resolutional frequency decomposition and localization of a one-dimensional,
discrete-time signal.

Figure 4: Three-level wavelet transform on signal x of length 16. Note that from w 1 to w2 ,
coefficients H1 remain unchanged, while from w2 to w3, coefficients H1 and H2 remain
unchanged.

4.3.4 Filter coefficients

Thus far, we have remained silent on a very important detail of the DWT –
namely, the construction of the low-pass filter h, and the high-pass filter g. Obviously, the
filter coefficients for h and g cannot assume arbitrary values, but rather have to be selected
carefully in order to lead to basis functions, such as those in Figures 1 and 2, with the
necessary properties of compactness (i.e. spatial localization) and orthogonality.

The derivation of filter coefficients is beyond the scope of this introduction; for an
understandable discussion and derivation see [2]. However, we note that filter coefficients for
h and g should be related as follows:

k
gk = (−1) hn−k−1 , k ∈ {0, . . . , n − 1}, (1)

where n denotes the length of the filter. For example, for filter lengths 2, 4 and 6:
 (a) (b)

Figure 5: Illustration of the inverse DWT for a one-level DWT w of length 16. First, the low-
pass and high-pass elements of w are interleaved. Then, (a) the inverse low-pass filter h−1 is
applied in increments of two, and (b) the inverse high-pass filter g−1 is applied in increments
of two.

4.3.5 Inverse DWT

To understand the procedure for computing the one-dimensional inverse DWT, consider
Figure 5, where we illustrate the inverse DWT for a one-level DWT of length 16 (assuming
filters of length four). Note that the two filters are now h− 1 and g−1 where,

h1 k− (
= hn−k−1 k ∈ {0, 2, . . . }

hk k ∈ {1, 3, . . . }

g−1 is determined from h−1 using equation (1).

To understand how to compute the one-dimensional inverse DWT for multi-level
DWTs, consider Figure 4. First, to compute w2 from w3, the procedure in Figure 5 is applied
only to values L3 and H3. Second, to compute w1 from w2 , the procedure in Figure 5 is
applied to values L2 and H2. Finally, to compute x from w1 , the procedure in Figure 5 is
applied to all of w1 – namely, L1 and H1.

4.4 DWT in two dimensions

In this section, we describe the algorithm for computing the two-dimensional

DWT through repeated appli-cation of the one-dimensional DWT. The two-dimensional
DWT is of particular interest for image processing and computer vision applications, and is a
relatively straightforward extension of the one-dimensional DWT discussed above.
 Figure 6: One-level, two-dimensional DWT. First, the one-dimensional DWT is
applied along the rows; second, the one-dimensional DWT is applied along the columns of
the first-stage result, generating four sub-band regions in the transformed space: LL, LH, HL
and HH.

Figure 6 illustrates the basic, one-level, two-dimensional DWT procedure. First, we

apply a one-level, one-dimensional DWT along the rows of the image. Second, we apply a
one-level, one-dimensional DWT along the columns of the transformed image from the first
step. As depicted in Figure 7 (left), the result of these two sets of operations is a transformed
image with four distinct bands : (1) LL, (2) LH, (3) HL and (4)
HH. Here, L stands for low-pass filtering, and H stands for high-pass filtering. The LL band
corresponds roughly to a down-sampled (by a factor of two) version of the original image.
The LH band tends to preserve localized horizontal features, while the HL band tends to
preserve localized vertical features in the original image. Finally, the HH band tends to isolate
localized high-frequency point features in the image.

As in the one-dimensional case, we do not necessarily want to stop there, since the
one-level, two-dimensional DWT extracts only the highest frequencies in the image.
Additional levels of decomposition can extract lower frequency features in the image; these
additional levels are applied only to the LL band of the transformed image at the previous
level. Figure 7 (right) for example illustrates the three-level, two-dimensional DWT on a
sample image.
 Figure 7: Two-dimensional wavelet transform: (left) one-level 2D DWT of sample
image, and (right) three-level 2D DWT of the same image. Note that the LH bands tend to
isolate horizontal features, while the HL band tend to isolate vertical features in the image.

4.4.1 Distortion Techniques

In this technique,store information by signal distortion and measure the devia-tion

from the original cover in the decoding process.Distortion techniques need knowledge of the
original cover image during the decoding process where the de-coder functions to check for
di erences between the original cover image and the distorted cover image in order to restore
the secret message.In this technique, a stego-image is created by applying a sequence of modi
cations to the cover im-age. This sequence of modi cations is use to match the secret message
required to transmit.The message is encoded at pseudo-randomly chosen pixels. If the stego-
image is di erent from the cover image at the given message pixel, the message bit is a 1.
otherwise, the message bit is a 0. The encoder can modify the 1 value pixels in such a
manner that the statistical properties of the image are not a ected.If an attacker interfere with
the stego-image by cropping, scaling or rotating, the receiver can easily detect it[4,12].

4.4.2 Characteristics feature of Data Hiding Techniques

Perceptibility does embedding message distort cover medium to a visually un-acceptable

level.
Capacity how much information can be hidden with relative to the change in
perceptibility.
Robustness to attacks can embedded data exist manipulation of the stego medium in an e
ort to destroy, or change the embedded data.
Tamper Resistance Beyond robustness to destruction, tamper-resistance refers to the di
culty for an attacker to alter a message once it has been embedded in a stego-image.[13]

4.4.3 Image Steganalysis

Steganalysis is the breaking of steganography and is the science of detecting hidden

information [14]. The main objective of steganalysis is to break steganography and the
detection of stego image. Almost all steganalysis algorithms depend on steganographic
algorithms introducing statistical di erences between cover and stego image.
Steganalysis are of three different types:

Visual attacks it discovered the hidden information, which helps to separate the image into
bit planes for further more analysis.
Statistical attacks Statistical attacks may be passive or active. Passive attacks involves
with identifying presence or absence of a secret message or embedding algorithm used.
Active attacks is used to investigate embedded message length or hidden message location or
secret key used in embedding.
Structural attacks The format of the data les changes as the data to be hidden is embedded,
identifying this characteristic structure changes can help us to nd the presence of image/text .

4.4.4 Steganalytic tools

There are several steganalytic tools available in market like PhotoTitle, 2Mosaic and
StirMark Benchmark etc. These three steganalytic tools can remove stegano-graphic content
from any image. This is achieved by destroying secret message by two techniques: break
apart and resample.
4.5 LEAST SIGNIFICANT BIT
Data hiding is a popularly used technique for secure communication. Data hiding is the
technique of embedding information into digital content without causing perceptual
 degradation. Watermarking, cryptography and steganography are three famous techniques
used in data hiding.
Watermarking is the process of hiding digital information in a carrier signal, where hidden
information does not need to contain a relation to the carrier signal. Digital watermarks can
be used to verify the authenticity or integrity of the carrier signal and also to show the identity
of its owners. It is prominently used for tracing copyright infringements and for banknote
authentication. Both steganography and digital watermarking employ steganographic
techniques to embed data covertly in noisy signals but their priorities differ. Steganography
provide priority to offer imperceptibility to human senses, whereas digital watermarking tries
to control the robustness as top priority.

Cover images
Pixels 1 Pixels 2 Pixels 3

1 0 0 1 0 1 1 0 1 0 1 1 1 0 0 1 0 1 1 0 0 1 1 0

Secret image
Pixels 1 Pixels 2 Pixels 3

1 0 1 0 1 0 0 1 1 1 0 1 0 1 1 0 0 1 0 0 1 1 0 1

Stego images
Pixels 1 Pixels 2 Pixels 3

1 0 0 1 0 1 0 1 1 0 1 1 1 0 1 0 0 1 1 0 0 0 1 1

The existing LSB techniques are considered easy and direct methods to hide an image in a
cover image. The information is embedded using the concept of sequence-mapping. Even
though LSB technique hides the information in such way that the humans do not recognize its
existence, still there exists the possibility of retrieving the secret information due to the
simplicity of the technique .

Therefore, we propose the randomization technique to improve the security of LSB scheme.
This method embeds the secret image bits into random pixels of the cover image. And thus
this allows overcoming the sequence mapping problem and making the system more secure.
The RC4 algorithm is used for implementing this randomization. The RC4 algorithm
generates the pixels of cover image in a random order and the secret image bits are embedded
into these pixels in the respective order. Since RC4 have very simple structure and can be
implemented efficiently RC4 is widely used. RC4 also offers simple key scheduling and
output generation process. Since the secret information is randomly dispersed it is more
difficult for an intruder to extract the hidden information. This algorithm makes use of a
stego-key which is required during embedding and extracting the secret information. This
stego-key can be made available at the receiver by embedding within the transmitting image
or can be supplied to the receiver separately. In the absence of stego-key it is very hard to
know the sequence in which cover image pixels are used for embedding the information.
Thus this RC4 based randomization improves the security of the system.
 CHAPTER 5
SIMULATION RESULT ANALYSIS

The implementation work is done using MATLAB tool and with the RGB color images. The
data hiding is performed in Red, Green, and Blue planes separately. It can also be applied on gray
scale images. Experiments are done with many images. Few samples are shown here.
The figure 5.1 shows the input image

Figure 5.1 Input Image

The figure 5.2 is the secret image input

Figure 5.2 Input Secret Image

The figure 5.3 and 5.4 shows the gray scale conversion image after preprocessing segment

Figure 5.3 Gray conversion of input image

Figure 5.4 Gray Conversion of secret image

 The figure 5.5 shows the steganography image after LSB steganography algorithm and four
level DWT.

Figure 5.5 steganography image

The figure 5.6 shows the steganography image after the visual secret sharing scheme applied

Figure 5.6 Image after VSS

 CHAPTER 6
CONCLUSION

Steganography is an effective way to hide sensitive information. In this project we have used
the LSB Technique and Visual Secret Sharing Technique on images to obtain secure stego image.
Our results indicate that the LSB insertion using VSS is better than simple LSB insertion in case
of lossless compression. The image resolution doesn’t change much and is negligible when we
embed the image into the image and the image is protected with the personal key image. So, it is
not possible to damage the data by unauthorized personnel. The algorithm is easy to be
implementing in both grayscale and color image. This project focuses on the approach like
increasing the security of the image and increasing PSNR and reducing the distortion rate.In future
work the objective is to improve SNR and histogram values.
 REFERENCES

[1] Prema, G.; Natarajan, S., "Steganography using Genetic Algorithm along with Visual
Cryptography for wireless network application," Information Communication and Embedded
Systems (ICICES), 2013 International Conference
on , vol., no., pp.727,730, 21-22 Feb. 2013.
[2] Nickfarjam, A.M.; Azimifar, Z., "Image steganography based on pixel ranking and
Particle Swarm Optimization," Artificial Intelligence and Signal Processing (AISP), 2012 16th
CSI International Symposium on , vol., no., pp.360,363, 2-3 May
2012.
[3] Sanchez, A.; Conci, A.; Zeljkovic, E.; Behlilovic, N.; Karahodzic, V., "A new approach to
relatively short message steganography," Telecommunications (BIHTEL), 2012 IX International
Symposium on , vol., no., pp.1,4, 25-27 Oct.2012.
[4] Qiangfu Zhao; Akatsuka, M.; Cheng-Hsiung Hsieh, "Generating facial images for
steganography based on IGA and image morphing," Systems, Man, and Cybernetics (SMC), 2012
IEEE International Conference on , vol., no., pp.364,369, 14-17 Oct. 2012.
[5] Khosravi, M.; Soleymanpour-Moghaddam, S.; Mahyabadi, M., "Improved pair-wise LSB
matching steganography with a new evaluating system," Telecommunications (IST), 2012 Sixth
International Symposium on , vol., no., pp.982,986, 6-8 Nov. 2012.
[6] Brazil, A.L.; Sanchez, A.; Conci, A.; Behlilovic, N., "Hybridizing genetic algorithms and
path relinking for steganography," ELMAR, 2011 Proceedings , vol., no., pp.285,288, 14-16 Sept.
2011.
[7] Bhowal, K.; Sarkar, D.; Biswas, S.; Sarkar, P.P., "Secured image transmission with GA
based Audio Steganography," India Conference (INDICON), 2011 Annual IEEE , vol., no.,
pp.1,4, 16-18 Dec. 2011.
[8] Mandal, J.K.; Khamrui, A., "A Genetic Algorithm based steganography in frequency
domain (GASFD)," Communication and Industrial Application (ICCIA), 2011 International
Conference on , vol., no., pp.1,4, 26-28 Dec. 2011.

[9] H. C. Huang, J. S. Pan, Y. H. Huang, F. H. Wang, and K. C. Huang, Progressive

watermarking techniques using genetic algorithms, Circuits, Systems, and Signal Processing, vol.
26, no. 5, pp. 671{687, 2007.
 [10] E. Kawaguchi and R. O. Eason, Principle and application of bpcs-steganography, Proc.
Of SPIE:Multimedia Systems and Applications, pp. 464{472, 1998.
[11] A. R. S. Marcal and P. R. Pereira, A steganographic method for digital images robust to
rs steganal-ysis, Lecture Notes in Computer Science, pp. 1192{1199, 2005.
[12] N .Provos, Steganography detection with stegdetect,
http://www.outguess.org/detection.php.
[13] A.Westfeld, F5-a steganographic algorithm, Proc. of the 4th International Workshop on
Information Hiding, Lecture Notes in Computer Science,2137.Springer-Verlag, pp. 289{302,
2001.
[14] A. Westfeld and A. P_tzmann, Attacks on steganographic systems, Proc. of Information
Hiding-Third International Workshop, 1999.
[15] A. Westfeld and A. P_tzmann, Attacks on steganographic systems, Lecture Notes in
Computer Science, pp. 61{76, 1999.
[16] D. C. Wu and W. H. Tsai, A steganographic method for images by pixel-value
differencing, Pattern Recognition Letters, pp. 1613{1626, 2003.
[17] X. Zhang and S. Z. Wang, Statistical analysis against spatial bpcs steganography,
Computer-Aided Design & Computer Graphics, pp. 395{406, 2003.
[18] X. Zhang and S. Z. Wang, Vulnerability of pixel-value differencing steganography to
histogram analysis and modification for enhanced security, Pattern Recognition Letters, pp.
331{339, 2004}.
[19] Reena M Patel,D J Shah , “Concealogram : Digital image in image using LSB insertion
method”, International journal of electronics and communication engineering &
technology(IJECET), 2013.
[21] Nadeem Akhtar, Pragati Johri, Shahbaaz Khan, “Enhancing the security and quality of
LSB based image steganography”, 2013 5th International Conference on Computational
Intelligence and Communication Networks.
[22] Mamta.Juneja , Parvinder S. Sandhu, “An improved LSB based Steganography with
enhanced Security and Embedding/Extraction”, 3rd International Conference on Intelligent
Computational Systems (ICICS'2013) January 26-27, 2013 Hong Kong (China).