Comparative Study of

LSB and DWT based

Steganography
combined with Arnold
Encryption for Image
Security

Piyush Gupta
M.Tech. [Embedded System Technology]
Amity University
 Abstract
The transfer of images over the internet has led to a security concern where confidential
images need to be protected from unauthorised access. In the growing technological era,
almost everyone shares his/her personal information including images over the internet
with other users or to a database repository which attract hackers to take undue advantage
from this information. Security provided to images like blue print of company projects,
secret images of concern to the army or of company’s interest using image processing
techniques have proven to be beneficial. In this work, two steganographic methods i.e.
LSB and DWT are combined with Arnold Transformation. The results of Arnold
transformed LSB and DWT images are analysed and compared on the basis of image
quality parameters. Our focus is to scramble the image information to a level where it
becomes difficult for intruder to get actual transferred image.
 Contents
Abstract..........................................................................................................................................1
List of Figures................................................................................................................................4
Chapter 1.......................................................................................................................................5
Introduction...................................................................................................................................5
1.1 Image Processing...................................................................................................................5
1.2 Applications of Image Processing..........................................................................................6
1.3 Issues in Image Processing....................................................................................................7
1.3.1 Image Compression........................................................................................................7
1.3.2 Image Enhancement.......................................................................................................7
1.3.3 Image Security................................................................................................................7
1.3.4 Image Recognition.........................................................................................................8
1.3.5 Image Visualization........................................................................................................8
1.4 Image Security.......................................................................................................................8
1.5 Techniques used for Image Security......................................................................................9
1.6 Motivation of the work.......................................................................................................10
1.7 Project Outline....................................................................................................................10
Literature Survey........................................................................................................................12
2.1 Advanced Encryption Standard and Least Significant Bit Algorithm....................................12
Figure 2.1 AES encryption algorithm.............................................Error! Bookmark not defined.
2.2 Encryption using Discrete Cosine Transform (DCT).............................................................13
2.3 Reversible Data Hiding Technique.......................................................................................13
2.4 Haar Discrete Wavelet Transform.......................................................................................14
2.5 Data Encryption Standard (DES) Algorithm.........................................................................16
2.6 Secret Fragment Visible Mosaic Image................................................................................16
2.7 Modified Advanced Encryption Standard............................................................................18
2.8 Arnold Transformation........................................................................................................19
2.9 Conclusion of Literature Survey..........................................................................................20
Chapter 3.....................................................................................................................................21
Project Objective.........................................................................................................................21
3.1 Problem Statement.............................................................................................................21
3.2 Project Objective.................................................................................................................21
Chapter 4.....................................................................................................................................22
Secure Information Hiding.........................................................................................................22
4.1 Arnold’s cat map.................................................................................................................22
4.1.1 Properties of Arnold Transformation (Γ)......................................................................23
4.2 Least-Significant Bit (LSB) Technique...................................................................................24
4.2.1 LSB Embedding Process..............................................................................................26
4.2.2 LSB Extraction Process................................................................................................26
4.3 Discrete Wavelet Transform (DWT) Technique...................................................................27
4.3.1 DWT Embedding Procedure.........................................................................................28
4.3.2 DWT Extraction Procedure..........................................................................................29
Chapter 5.....................................................................................................................................31
Implementation, Results and Discussion...................................................................................31
5.1 Block Diagram of Proposed Work........................................................................................31
5.2.4 Comparison of timings of Arnold-LSB vs. Arnold-DWT.............................................41
......................................................................................................................................................41
5.3 Discussion............................................................................................................................41
Chapter 6.....................................................................................................................................43
Conclusion...................................................................................................................................43
References....................................................................................................................................44
 List of Figures
Figure 1 Two-Level Security Mechanism.....................................................................................10
Figure 2 AES encryption algorithm..............................................................................................12
Figure 3 Detailed 2D Haar Wavelet Transform.............................................................................15
Figure 4 Example of 2D Haar Wavelet Transform........................................................................15
Figure 5 Flow diagram of mosaic image approach........................................................................17
Figure 6 Hybrid Approach for Image Security..............................................................................18
Figure 7 Conceptual Diagram of Arnold-LSB Technique.............................................................19
Figure 8 Arnold Transformation on a Grey Scale Cameraman Image...........................................23
Figure 9 LSB Steganographed Image............................................................................................25
Figure 10 LSB Insertion Mechanism.............................................................................................26
Figure 11 LSB Extraction Mechanism..........................................................................................27
Figure 12 DWT Steganographed Image........................................................................................28
Figure 13 DWT Embedding Mechanism.......................................................................................29
Figure 14 DWT extraction mechanism.........................................................................................30
Figure 15 Flow diagram of Proposed Mechanism at sender side...................................................31
Figure 16 Flow diagram of Proposed Mechanism at Receiver side...............................................32
Figure 17 Input image...................................................................................................................33
Figure 18 Block divided and Arnold Transformed Image.............................................................33
Figure 19 Block Swapped Image..................................................................................................34
Figure 20 Cover, Secret, Stego and Extracted Image of Proposed Work......................................35
Figure 21 Outputs of Arnold-LSB and Arnold-DWT Techniques.................................................36
Figure 22 Plot of MSE values of Arnold-LSB vs Arnold-DWT....................................................37
Figure 23 Plot of PSNR values of Arnold-LSB vs Arnold-DWT..................................................39
Figure 24 Plot of Entropy values of cover and stego images for LSB and DWT Techniques.......40
Figure 25 Plot of Average timing of Arnold-LSB vs. Arnold-DWT.............................................41
 Chapter 1

Introduction
1.1 Image Processing
Image processing is a method to convert an image into a digital form and perform
operations on it, in order to get an enhanced image or to extract some useful information
from it. It is a type of signal dispensation in which input is image, like video frame or
photograph and output may be image or characteristics associated with that image. Image
Processing can also be defined as a technique to enhance raw images received from
cameras/sensors placed on satellites, space probes and aircrafts or pictures taken in
normal day-to-day life for various applications. Various techniques have been developed
in Image Processing during the last four to five decades. Most of the techniques are
developed for enhancing images obtained from unmanned spacecraft, space probes and
military reconnaissance flights. Image Processing systems are becoming popular due to
easy availability of powerful personnel computers, large-sized memory devices, graphics
software etc. [1] It is a rapidly growing area of computer science and its growth has been
fuelled by technological advances in digital imaging, computer processors and mass
storage devices. Fields which traditionally used analog imaging are now switching to
digital systems, for their flexibility and affordability. Important examples are medicine,
video production, photography, remote sensing, and security monitoring. These and other
sources produce huge volumes of digital image data every day, more than could ever be
examined manually. Digital image processing is concerned primarily with extracting
useful information from images. Ideally, this is done by computers, with little or no
human intervention [2]. The purpose of image processing has been divided into five
groups. They are:
1.      Visualization - Observe the objects that are not visible.
2.      Image sharpening and restoration - Creating a better image.
3.      Image retrieval - Seek for the image of interest.
4.      Measurement of pattern – Measures various objects in an image.
5.      Image Recognition – Distinguish the objects in an image.
1.2 Applications of Image Processing
As optics, imaging sensors, and computational technology advanced, image processing
has become more commonly used in many different areas. Some areas of application of
digital image processing include image enhancement for better human perception, image
compression and transmission, as well as image representation for automatic machine
perception. Most notably, digital image processing has been widely deployed for defence
and security applications such as small target detection and tracking, missile guidance,
vehicle navigation, wide area surveillance, and automatic/aided target recognition. One
goal for an image processing approach in defence and security applications is to reduce
the workload of human analysts in order to cope with the ever increasing volume of
image data that is being collected. A second, more challenging goal for image processing
research is to develop algorithms and approaches that will significantly aid the
development of fully autonomous systems capable of decisions and actions based on all
sensor inputs. Broad areas of digital image processing applications include:

• Medical applications (Gamma ray imaging, PET scan, X Ray Imaging, Medical CT,
UV imaging)
• Restorations and enhancements
• Digital cinema
• Image transmission and coding
• Color processing
• Remote sensing
• Robot vision
• Hybrid techniques
• Facsimile
• Pattern recognition
• Registration techniques
• Multidimensional image processing
• Image processing architectures and workstations
• Video processing
• Programmable DSPs for video coding
• High-resolution display
• High-quality color representation
• Super-high-definition image processing
• Impact of standardization on image processing [1, 4,27].

1.3 Issues in Image Processing

1.3.1 Image Compression: A modern trend in image storage and transmission is to use
digital techniques. Digitizing a television signal requires high bandwidth. So for
applications such as teleconferencing, one wants to use a channel of 64 kilobits per
second. For other applications such as videophone and mobile videophone, even lower
channel bandwidths (e.g., 9kilobits per second) are desirable. For example, how can one
compress the bit rate from 100 megabits per second to 1 kilobit per second without severe
loss of image quality?
1.3.2 Image Enhancement: In enhancement, one aims to process images to improve their
quality. An image may be of poor quality because its contrast is low, or it is noisy, or it is
blurred, etc. Many algorithms have been devised to remove these degradations. The
difficult problem is how to remove degradations without hurting the signal. For example,
noise-reduction algorithms typically involve local, averaging or smoothing which,
unfortunately, will blur the edges in the image. Adaptive methods have been investigated-
e.g., smoothing less near the edges. However, they are generally effective only if the
degradation is slight. A challenging problem is then how to enhance severely degraded
images [3].
1.3.3 Image Security: Now-a-days images are transmitted by using the web for various
applications like legal document storage system, military databases, educational data
system, confidential documents etc. As these images may contain secret and confidential
information hence security of these images is extremely important it means that they
should be protected from an attackers or hackers while transmitting. In other words, these
images should be robust against leakages. There are number of techniques which can be
used for secure data transmission, out of this image encryption and data hiding is
generally used techniques [10].
1.3.4 Image Recognition: Typically, a recognition system needs to classify an unknown
input pattern into one of a set of prespecified classes. The task is fairly easy if the number
of classes is small and if all members in the same class are almost exactly the same.
However, the problem can become very difficult if the number of classes is very large or
if members in the same class can look very different. Thus, a most challenging problem is
how to recognize generic objects. For example, how can one construct a system to
recognize "chairs"?
1.3.5 Image Visualization: Commonly, visualization is considered as a part of computer
graphics. The main task is to generate images or image sequences based on three-
dimensional object and scene models. A challenging problem is how to model dynamic
scenes containing non rigid objects (such as clothing, hair, trees, waves, clouds, etc.). The
models have to be realistic, and yet the computation cost has to be reasonable [3].

1.4 Image Security

Today, the Internet has become a necessity and the information in the form of messages,
images and videos are sent through internet. All the information can be sent with a variety
of electronic media by e-mail, personal blogs, chat applications, and others which are
prone to several attacks. With the rapid development of Internet usage today, one of the
things that have received less attention is the unawareness of security and secrecy of data.
The survey result of Norton Report 2013 shows about 62% there is nothing called as
"online privacy" in the world today due to the information open on the internet. We can
send and receive any information from all over the world by various communication
networks. But not many people realize that the process of sending and receiving messages
is vulnerable to eavesdropping, tapping, hijacking, and many other things. Especially
much of that information is confidential and should not be known or intercepted by
others. Along with the growth of technology, indications of cyber-crime also increased.
There are several forms of attacks on data and information, such as hackers, crackers,
Trojans, etc. Therefore, today many systems reinforced with various efforts to maintain
data security and overcome these attacks [9].
With rapid advancement in Multimedia technologies during the recent years,
communication and information exchange have become much easier and faster but at the
same time the issues related to data security and confidentiality have become a major
concern of the time. To cater to this need of information security, a number of hidden and
secret communication techniques have been developed [2]. The security of data over the
internet is a crucial thing especially if this data is personal or confidential. The
transmitted data can be intercepted during its journey from device to another. For that
reason, we are willing to develop a simple method to secure data.

1.5 Techniques used for Image Security

Cryptography: Cryptography is the study of algorithms of sending messages in
encrypted form (not understood) so that only the authorized recipients can decrypt
message and read it. Cryptography system is classified into symmetric-key system and
asymmetric-key system. Symmetric-key system uses a single key that both the sender and
the receiver have, and asymmetric-key system uses two different keys (a public-key and a
private-key), where the public key is known to everyone and the private key is known to
only the authorized recipient of messages [12].

Steganography: The term steganography refers to the art of hidden communications,

encoding/embedding hidden information in cover media in such a way that it is a difficult
job for a unauthorized person to see that there is something hidden in the cover media.
The output is an image called stego-image that is similar to the cover media. This stego
image is then sent to the receiver where the receiver retrieves the hidden message by
applying the de-steganography. A stego-key is used for embedding/encoding process to
restrict decoding or extraction of the embedded data in cover media. The modern age
steganography is usually implemented computationally, where multimedia files are used
as cover media. A good Steganographic method has three features, good hiding capacity,
good imperceptibility and the last is robustness. The digital image Steganographic
methods generally depends upon various image scrambling techniques in order to further
improve the level of security of the hidden information. Image scrambling techniques
scramble the pixels of an image in such a manner that the image becomes chaotic and
indistinguishable. Hence, the message remains highly secured against unauthorized
access.
For enhancing the security of images, above mentioned techniques i.e.
cryptography and steganography are combined together to give better security rather than
using these alone. In the hybrid approach of cryptography and steganography of an image
security scheme where two levels of security are being used to secure the embedded
information, adds more complexity for steganalysis [12]. The combined concept of
cryptography and steganography in one system to secure the transmitted data over an
insecure channel is shown in figure 1. This is a two-step process

Figure 1 Two-Level Security Mechanism

1.6 Motivation of the work

Recent advances in the field of information technology and communication has allowed
content sharing including images among remote users. The number of attackers and the
tools available for them are increasing each day which makes the information exchanged
over internet vulnerable to various types of attacks such as brute force attack, chosen
plaintext attack, cipher text only attack. This necessitates safeguarding information in-
transit from unauthorised users. In this project, we are interested in securing image
content from intruders. Many contributions by researchers have been made to secure
images from illegitimate users. The motivation of our project is to study and work
towards designing a secure and efficient mechanism for images in-transit.

1.7 Project Outline

Chapter 1: This chapter gives an introduction to the Image Processing, discussing the
applications and issues of Image Processing. Further the chapter contains a brief
introduction to the Image Security and Techniques used to obtain image security,
motivation to the work and the work presented in this report.
Chapter 2: The second chapter consists of a detailed description of the work taken up,
that is the literature survey of related work done on it so far followed by the conclusion of
literature survey.

Chapter 3: The third chapter elaborates on the problem statement and Project objectives.

Chapter 4: The fourth chapter elaborates on the Arnold Cat Map, LSB and DWT
steganography discussing its key characteristics and the various parameters.

Chapter 5: The final chapter gives the implementation details of the work done in this
report and the observations made thereafter. The chapter further contains comparative
plots of some parameters for both attempted techniques. The simulation results of the
designs are also given in this chapter.

Chapter 6: This chapter concludes the report.

 Chapter 2

Literature Survey

2.1 Advanced Encryption Standard and Least Significant Bit Algorithm

The research designed the application of steganography using Least Significant Bit (LSB)
in which the previous message is encrypted using the Advanced Encryption Standard
algorithm (AES) and it can restore the previously hidden data. The messages in this form
application and hidden text on media digital image so as not to arouse suspicion. The
result of research shown the steganography is expected to hide the secret message, so the
message is not easy to know other people who are not eligible.

Figure 2 AES encryption algorithm

Least Significant Bit (LSB) is one method that is widely used in steganography because it
is simple computation and hidden message relative safe. Hiding message strategy with
this method is replacing bits of data messages with the lowest bit in the media / cover.
The quality of the digital image is changing after inserted message. Changes in quality
depend on the size of the message inserted. While the size of digital image also changed
after the insertion process of the message [9, 14].
AES provides high security but takes much time in execution and LSB technique is easy
and fast but it have drawback if the message have a great size, it will affect the image and
it can be noticeable [11].

2.2 Encryption using Discrete Cosine Transform (DCT)

In this technique, the image to be encrypted is decomposed into 8X8 blocks; these blocks
are transformed from the spatial domain to frequency domain by the DCT. Then, only
selected DCT coefficients i.e. the DCT coefficients correlated to the higher frequencies of
the image block are encrypted. For encryption the DCT coefficients are XORed with
pseudorandom bit, Pseudorandom bit is generated by Non-Linear Shift back Register.
The bits generated by Non- Linear Shift back Register cannot be predicted so
cryptanalysis becomes difficult. To enhance the security further, the unencrypted DCT
coefficients are shuffled, since some information may also be stored in DCT coefficient
correlating to lower frequency. While encrypting selected DCT coefficients alone will
provide complete perceptual encryption, it would be possible for an attacker to gain
information about the image from the other coefficients, especially in images that have a
lot of edges. This technique is used for faster image encryption. Here partial image
encryption is used to decrease the time required for encryption and decryption. Instead of
encrypting the whole image only the selected portion of the image are encrypted. To add
additional security to the image in the cryptosystem, after encrypting the selected higher
coefficients, all the coefficients can be scrambled. The idea behind scrambling all the
coefficients is to withstand any type of attack [13]. This technique makes the encryption
faster and reduces the time complexity.

2.3 Reversible Data Hiding Technique

During Recent days, huge importance is given to Reversible Data Hiding (RDH)
techniques in images, since it assures the excellent feature of retrieving the original cover
image information in a lossless manner after embedded secret information is extracted,
while protecting the image contents confidentiality. RDH technique is a lossless data
hiding scheme for color cover images with very important and confidential information
content. This method, exploits least computational complexity by using histogram
shifting technique to hide secret bits in confused color cover image, where confusion in
cover image is created by using Quad tree decomposition. An enhanced histogram
shifting method is used for color images by embedding the secret information into three
different planes of RGB image in such a way that, it not only improves the quality of
image and also achieves high embedding capacity. This algorithm hides the pay load by
using histogram shifting technique into non overlapped block of pixels. A histogram
shifting based reversible data hiding technique for color images has been proposed, which
could result with high visual quality and as well as high embedding capability. This
method was found reliable to hide the data and retrieve the data by recovering the cover
image as well. The most useful feature of this algorithm is that, it could be applied for
even DICOM images [7]. This technique improves the quality of image and also achieves
high embedding capacity but it exploits the least computational complexity.

2.4 Haar Discrete Wavelet Transform

Next technique is an advanced technique for encrypting data using Advanced Encryption
System (AES) and hiding the data using Haar Discrete Wavelet Transform (HDWT).
One-fourth of the image carrying the details of the image in a region and other three
regions carrying a less details of the image then the cipher text is concealed at most two
Least Significant Bits (LSB) positions in the less detailed regions of the carrier image, if
the message doesn’t fit in the first LSB only it will use the second LSB. This algorithm
covers almost all type of symbols and alphabets.
Wavelet transform has the capability to offer some information on frequency-time
domain concurrently. In this transform, time domain is passed through high-pass and low-
pass filters to extract high and low frequencies respectively. This process is repeated for a
number of times and each time a section of the signal is drawn out. DWT analysis splits
signal into two classes (i.e. Approximation and Detail) by signal decomposition for
different frequency bands and scales. DWT employs two function sets: scaling and
wavelet which associate with low and high pass filters orderly.
 Figure 3 Detailed 2D Haar Wavelet Transform

Decomposition follows the manner of dividing time separability. Meanly, only half of the
samples in a signal are sufficient to represent the whole signal, doubling the frequency
separability. Haar wavelet operates on data by calculating the sums and differences of
adjacent elements. This wavelet operates first on adjacent horizontal elements and then on
adjacent vertical elements. One important feature of the Haar wavelet transform is that the
transform is equal to its inverse. Each transform computes the data energy in relocated to
the top left hand corner. After each transform is performed the size of the square which
contains the most important information is reduced by a factor of 4 as seen in Figure 4.

Figure 4 Example of 2D Haar Wavelet Transform

HDWT aims to decrease the complexity in image steganology while providing less image
distortion and lesser detectability [11].
2.5 Data Encryption Standard (DES) Algorithm

Here, securing of the image by encryption is done by DES algorithm using the key image.
The encrypted image is hidden in another image by using LSB techniques, so that the
secret’s very existence is concealed. The decryption can be done by the same key image
using DES algorithm. The algorithm is designed to encipher and decipher blocks of data
consisting of 64 bits under control of a 64-bit key. Deciphering must be accomplished by
using the same key as for enciphering, but with the schedule of addressing the key bits
altered so that the deciphering process is the reverse of the enciphering process. Each byte
(pixel) of all the three matrices (R, G, B matrices of payload) are encrypted using DES
algorithm and an image comprised of encrypted pixels is formed. The key used to encrypt
each pixel is of 10-bit length and is obtained from the pixels of key image. The pixel
values of red, green and blue intensities of each pixel of key image are combined to get a
24-bit value. The first ten bits are selected as the key to encrypt the red intensity pixel of
payload image. The middle ten bits will be the key to encrypt the green intensity pixel of
payload and finally the last ten bits is the key to encrypt blue intensity pixel of payload
image. So the size of key image must be same as that of payload. If not, then the key
image will get resized. Each pixel (24-bit) of the key image is split into three keys (10- bit
each). LSB technique is used to embed information in a cover image [15].

2.6 Secret Fragment Visible Mosaic Image

Another method is data hiding in which secret image is hidden behind a carrier image but
due to data hiding cover image has experienced eternal problems. A new kind of mosaic
image is being created known as secret fragment visible mosaic image for hiding the
existence of secret image with the help of target image. Mosaic image generation contains
following steps:
1. Divide original secret image into number of small tiles.
2. For every tile, find another image with similar contents.
3. Replace all tiles with their similar images for image mosaic generation.
This method has two main stages, namely: Creation of mosaic image; Recovery of secret
image. In the first phase of mosaic image, creation of mosaic image is done in highly
equipped way in which created mosaic image contains color corrected fragments of target
image. This approach consists of four stages: Fit the tile images of secret image in the
corresponding blocks of user selected target image; Transform the color characteristics of
each tile image of secret image to become that of corresponding target blocks in target
image, so mosaic image looks similar like target image.; A new thing involved in this
work is rotating of each tiles of secret image with minimum RMSE with respect to the
blocks of target image for better fitting of tiles into created mosaic image.; The most
important step is embedding of relevant information into the created mosaic image for
further recovery of the original secret image in lossless manner. The Recovery of secret
image also consists of four steps: To recover original secret image from created mosaic
image, extract the embedded information from the created mosaic image.; Reverse
transform the color characteristics of each tile image of secret image with respect to the
corresponding target block.; Reverse rotate each tile image into a direction with respect to
its corresponding target block.; Recovering the secret image using extracted information.

Figure 5 Flow diagram of mosaic image approach

This scheme is advantageous over conventional schemes to protect the privacy during
transmission. In order to control overflows and underflows, the color transformation
scheme is used [10]. It was found that the method gives high quality mosaic image and
accurate extraction of secret image but was very complex.
2.7 Modified Advanced Encryption Standard

There is a modification in AES as Modified AES (MAES) which presents a new shift row
transformation that enhances the security of AES. MAES is byte oriented. The 128 bit
plaintext is considered as sixteen 8 bit bytes. To be able to perform some mathematical
transformations on bytes MAES defined the concept of state. A state is 4*4 matrix in
which each entry is byte. The new shift row transformation used in MAES is performed
on the rows of the state. It cyclically shifts the bytes of each row depending upon the first
element of state. If the value in first row and first column of state matrix is even then first
row remains unchanged; each bytes in second, third and fourth are cyclically shifted left
over different numbers.

Figure 6 Hybrid Approach for Image Security

The calculated and analysed histograms of several encrypted images as well as original
images which have totally different content. The histogram of encrypted image is totally
different from that of original image. It was clear from the histograms of plain and cipher
image that there is no resemblance between original and encrypted image. Hence does not
provide any clue to employ statistical attack on proposed image encryption procedure [8].
MAES is very secure but takes large execution time and is very complex.
2.8 Arnold Transformation

Another approach provides two levels of security through a two-step process, Image to be
transmitted is scrambled in a random order generated by 2D Arnold Cat Map after that
encrypted message is hidden behind a cover image using basic LSB method, rather than
hiding the message bits directly in cover image. Image scrambling techniques scramble
the pixels of an image in such a manner that the image becomes chaotic and
indistinguishable. Hence, the message remains highly secured against unauthorized
access. Scrambled message is passed to the second layer, where user enters a stego- key
and a cover image, and then the scrambled message is hidden in the cover image through
stego-key. The two key are required to retrieving back the data at receiver’s side that have
been hidden inside the cover image.

Figure 7 Conceptual Diagram of Arnold-LSB Technique

Experimental study points out that this system is better than basic LSB method in terms
of higher visual quality [12].
2.9 Conclusion of Literature Survey

In the above mentioned techniques, it is found that methods used in them are not suitable
for image encryption for two main reasons. The first reason is that the image size is
almost always much greater than that of text, which needs much time to directly encrypt
the image data using the software implementations of traditional cryptosystems. The
second reason is laying on its sensitivity. The decrypted text must be exactly equal to the
original text. However, this requirement is not necessary for image data. Because of
human perception property and the high redundancy of image data, a decrypted image
containing small distortion is usually acceptable. DCT is faster but selected DCT
coefficients are providing complete perceptual encryption that will allow attacker to gain
information about the image from the other coefficients present in the edges of an image.
Use of visible mosaic image provided high quality mosaic image and gives accurate
extracted image but is very complex. In different stenographic techniques mentioned
above, LSB is the most simple and fast technique but it have a drawback that if the
message have a great size, it will affect the image and it can be noticeable. RDH
technique improves the quality of image and also achieves high embedding capacity but
Quad tree decomposition and histogram shifting method makes it slow. HDWT technique
is less complex, fast and gives less image distortion and lesser detectability.
The Arnold Transformation has suggested a new and efficient way to deal with
the intractable problem of fast and highly secure image encryption. This chaotic system is
rich in significance and in implication because of sensitivity to change initial conditions,
control parameters, random-like behaviour, repeated processing and very high diffusion
and confusion properties that are desirable for cryptography. Among all the above
mentioned techniques it is found that Arnold Transformation and HDWT techniques were
found to be fast and provide good security. In addition to this, HDWT provided better
quality than some other steganographic techniques. Combining both techniques would
give better security and can help in improving the quality of an image [23].
 Chapter 3

Project Objective
3.1 Problem Statement

Techniques discussed in the previous chapter have certain limitations in terms of quality,
complexity etc. hence can’t be used for all image processing applications. Among
different techniques discussed above, it is observed that while making an image secure
there was a significant difference in the quality of the input and output image. Due to
many cryptographic and steganographic processes, the quality of the image reduces which
results in loss of important information from the image. Safeguarding images in-transit
and maintaining quality of image is the main focus of our work.

3.2 Project Objective

 To improve the security (Confidentiality) of images in-transit.

 To hide Arnold Transformed image in cover image using two different
Steganographic Techniques i.e. Least Significant Bit Technique and Discrete Haar
Wavelet Transform.
 To compare both approaches i.e. Arnold combined with LSB technique and
Arnold combined with HDWT technique using qualitative and quantitative
parameters. These parameters include PSNR, MSE, Entropy values and timing for
encryption and decryption process.
 To compare the basic parameters of Steganographic process of an Arnold
transformed image.
 Chapter 4

Secure Information Hiding

4.1 Arnold’s cat map

Arnold’s cat map (ACM) or Arnold transform (AT), proposed by Vladimir Arnold in
1960, is a chaotic map which when applied to a digital image randomizes the original
organization of its pixel locations and the image becomes chaotic. However, it has a
period k and if iterated k number of times, the original image reappears. The generalized
form of Arnold's cat map can be given by

( xy '' )=(21 11 )( xy )( mod N )

Where x, y ϵ {0, 1, 2......, N-1} and N is the size of digital image. Applying ACM to a
digital image randomizes the original organization of its pixel locations and the image
becomes chaotic (figure 8). Now scrambled image can be transmitted over the
communication channels without revealing any information to the unauthorized receivers
or sniffers.

Arnold's cat map is the transformation given by the formula

Г: (x, y) (2x + y, x + y) mod 1

Equivalently, in matrix notation, this is

Г x = 2 1 x mod 1= 1 1 1 0 x mod 1
([ ]) [ ][ ]
y 1 1 y 0 1 1 1 y [ ][ ][ ]
That is, with a unit equal to the width of the square image, the image is sheared one unit
up, then two units to the right, and all that lies outside that unit square is shifted back by
the unit until it is within the square.
 Figure 8 Arnold Transformation on a Grey Scale Cameraman Image

Above figure 8 shows the Arnold Transformation applied to an image (Cameraman.tif)

for ten numbers of iterations. This Arnold transformed image is then sent through the
public communication channel to confuse attackers. In the Decryption process, this
Arnold transformed image is then applied to Inverse Arnold Transform which gives
original image back as shown in figure 8.

4.1.1 Properties of Arnold Transformation (Γ)

 Γ is invertible because the matrix has determinant 1 and therefore its inverse has

integer entries.
 Γ is area preserving.
 Γ has a unique hyperbolic fixed point (the vertices of the square). The linear
transformation which defines the map is hyperbolic: its eigenvalues are irrational
numbers, one greater and the other smaller than 1 (in absolute value), so they are
associated respectively to an expanding and a contracting eigen space which are also
the stable and unstable manifolds. The Eigen space is orthogonal because the matrix
is symmetric. Since the eigenvectors have rationally independent components both
the Eigen spaces densely cover the torus. Arnold's cat map is a particularly well-
known example of a hyperbolic toral automorphism, which is an automorphism of
a torus given by a square unimodular matrix having no eigenvalues of absolute value
1.
 The set of the points with a periodic orbit is dense on the torus. Actually a point is
pre-periodic if and only if its coordinates are rational.
 Γ is topologically transitive (i.e. there is a point, whose orbit is dense, this happens
for example for any points on the expanding Eigen space).
 Γ is ergodic and mixing.
 Γ is an Anosov diffeomorphism and in particular it is structurally stable [24,25].

4.2 Least-Significant Bit (LSB) Technique

LSB technique is a technique in which the least significant 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.e. 24 bit images and 8 bit images. In 24 bit images, three bits of information in
each pixel, one in each LSB position of the three eight bit values is changed to a bit of
secret image. 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. The hidden image is
extracted from the stego-image by applying the reverse process [16, 21].

If the LSB of the pixel value of cover image C(i,j) is equal to the message 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
message 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 message
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 bit in three pixels of an image. Suppose the original
3 pixels are:

(11101010 11101000 11001011)

(01100110 11001010 11101000)

(11001001 00100101 11101001)

A steganographic program could hide the bit having 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 recognised
by the human eye, so the message is effectively hidden. The advantage of LSB
embedding is its simplicity and many techniques use these methods [22]. LSB embedding
also allows high perceptual transparency. The following figure 4.2 shows the output of
LSB technique. The secret image is hidden in cover image by replacing LSB bit of cover
image. The output of the LSB technique is the stego image which is same as cover image
as shown in figure. Then hidden bits are extracted from stego image to give back the
secret image [19].
 Figure 9 LSB Steganographed Image

4.2.1 LSB Embedding Process

The cover and secret images are first read and converted into the unit8 type. The numbers
in secret image matrix are conveyed to 8-bit binary. Then the matrix is reshaped to a new
matrix a. The matrix of the cover image is also reshaped to matrix b. Embedding process
first extracts the pixels of the cover image and the secret image. It also extracts the
characters from the Stego key. In next step, it chooses first pixel and picks characters of
the Stego key and places it in the first component of pixel. Embedding process uses some
terminating symbol to indicate end of the key. Then, it insert bits of secret image in first
component of pixels of cover image by replacing it and repeats the above steps till all the
bits of secret image has been embedded. Again it places some terminating symbol to
indicate end of bits which has to be hidden. Obtained image is the stego image. The stego
image, is very similar to the original cover image, achieved by reshaping matrix b [20].
Figure 10 below shows the LSB insertion mechanism which hides the secret image in a
cover image using stego key. The output of the LSB embedding is stego image.
 Figure 10 LSB Insertion Mechanism

4.2.2 LSB Extraction Process

In extraction process, first the pixels of the stego image are extracted. Now, the process
starts from the first pixel and extracts the stego key characters from first component of the
pixels of stego image. This step follows up to terminating symbol. If the extracted key
matches with the key entered by the receiver then go to next pixel otherwise terminate the
program. Next pixel is followed only if the key is correct, and extract secret bit from first
component of next pixels. This step is followed up till the terminating symbol, otherwise
continue extracting the bits of secret image [21, 22]. While extracting the bits, 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 [20]. Figure 11 below shows the
mechanism where stego image and stego key are used as input to the extraction process;
the output is the extracted secret image.
 Figure 11 LSB Extraction Mechanism

4.3 Discrete Wavelet Transform (DWT) Technique

Discrete Wavelet Transform is used for digital images. Many DWTs are available.
Depending on the application most appropriate one should be used. If we apply DWT on
an image, it divides the image in frequency components. When DWT is applied to an
image it is decomposed into four sub bands: LL, HL, LH and HH. The LL part contains
the most significant features. So if the information is hidden in LL part the stego image
can withstand compression or other manipulations [5]. Sometimes distortion may be
produced in the stego image and then other sub bands can be used. The low frequency
components are approximate coefficients holding almost the original image and high
frequency components are detailed coefficients holding additional information about the
image. These detailed coefficients can be used to embed secret image. In embedding
process, first we convert cover image in wavelet domain. After the conversion we
manipulate high frequency component to keep secret image data. These secret image data
further retrieved in extraction procedure to serve the purpose of steganography. Figure 11
shows the outputs of DWT technique where cover image is divided into four sub bands,
the bits of secret image are embedded in the high frequency sub band of cover image.
Output of the process is a stego image which is same as cover image as shown in the
figure 12. The bits from stego image are extracted to give original secret image back
using [6].
 Figure 12 DWT Steganographed Image

4.3.1 DWT Embedding Procedure

In this process, insertion of secret message onto cover object is carried out. An m × n
carrier image and a secret image act as input. The cover image (Ic) is first read and its size
is calculated. In next step Secret image (Im) is read to work as a message vector. Then
cover image (Ic) is decomposed by using discrete Haar wavelet transform. Components
with detailed coefficients like horizontal and vertical coefficients of wavelet
decomposition are modified with secret bits which have to be hidden. Then, Inverse DWT
is applied which gives output as m×n stego image. Figure 13 below shows the flow
diagram of the DWT embedding process where Cover image is divided into four sub
bands using DWT and high frequency coefficients are selected. These coefficients are
then modified based on the secret image bits and IDWT technique is applied to give stego
image as output.
 Figure 13 DWT Embedding Mechanism

4.3.2 DWT Extraction Procedure

In extraction process, extraction of secret message is carried out. An m × n carrier image

and an m×n stego-image act as input. At first step, the cover images (Ic) and stego image
(Is) are read. Then Ic and Is are decomposed by using Haar wavelet transform. Then
message vector of all ones is generated. In next step, the correlation between the original
and modified coefficients is calculated. If the correlation value is greater than mean
correlation value, the message vector bit is turned to 0. This message vector is then
prepared to display secret image. Figure 14 below shows the flow diagram of DWT
extraction process where DWT is applied to both; cover and stego image, then the
correlation is found between these images and its mean is calculated. Bits where
correlation value is greater than mean of correlation are replaced by 0. The output would
be the original secret image [6,26].
 Figure 14 DWT extraction mechanism

Chapter 5

Implementation, Results and Discussion

5.1 Block Diagram of Proposed Work

Figure 15 Flow diagram of Proposed Mechanism at sender side

Above figure 15 shows the flow diagram of the system at the sender side to make an
image secure with the help of Arnold’s Cat Map and Steganographic technique. In the
Encryption process, to make an image secure Input image is first divided into different
number of blocks; these blocks are then swapped just to confuse the intruders. Then
Arnold Transformation is applied to these different blocks. Output of this Arnold
Transformation is the scrambled image whose pixels are randomised according to the
number of iterations. This scrambled image is then hidden in some cover media as shown
in figure to hide the existence of the secret image using Steganographic techniques. Here,
we are using LSB and HDWT techniques to embed Arnold transformed scrambled image,
then the output of these Steganographic processes would be stego image which will look
like cover image. This Stego image is the output of the encryption process.
 Figure 16 Flow diagram of Proposed Mechanism at Receiver side

Figure 16 shows the decryption process where stego image works as input image. From
the stego image, Secret Scrambled image is extracted using de-Steganographic technique
i.e. Inverse LSB or IDWT Technique. Then the blocks of this extracted image are
unswapped to their original positions. This image is then Inverse transformed using
Inverse Arnold Transformation. The output would be the original secret image.

5.2 Results
To compare the results of Arnold Transformation combined with LSB technique and
Arnold Transformation combined with DWT technique, various outputs at each step are
shown below. Figure 17 shows an input image of size 256 X256 in which the LSB and
DWT techniques along with Arnold Transformation are implemented.
 Figure 17 Input image

Figure 18 below represents the block- divided input image after Arnold Transformation
has been applied to it.

Figure 18 Block divided and Arnold Transformed Image

Swapping of divided blocks of image in Figure 18 is performed to enhance the security of
image further. The swapped image is shown in Figure 19 below.

Figure 19 Block Swapped Image

Figure 20 shows the output of the Arnold transformation combined with Steganographic
process which includes cover image i.e. peppers.png of size 256X256, the secret image
which is to be hidden in the cover image is the Arnold transformed image is shown in
second part of the figure, third part shows the output of the steganographic process i.e. the
stego image. This stego image is then sent through public communication channel to hide
the existence of secret image. Fourth part of the figure shown below is the extracted
secret image which is the same Arnold transformed image which was hidden in the cover
image.
 Figure 20 Cover, Secret, Stego and Extracted Image of Proposed Work

Figure 21 shown below is the output of both the techniques. Figure at the left side is the
output of the Arnold transformation combined with LSB technique and the figure at the
right side is the output of the Arnold transformation combined with DWT technique. We
can see that the output of the Arnold combined with DWT gives much better results than
the Arnold combined with LSB. These techniques will be compared on the basis of
various parameters such as PSNR, MSE, Entropy values and the average timings of both
the techniques will also be compared for encryption and decryption processes.
 Figure 21 Outputs of Arnold-LSB and Arnold-DWT Techniques

5.2.1 Comparison of MSE values of Arnold-LSB vs. Arnold-DWT

MSE is the averaged pixel-by-pixel squared difference between the input-image and the
output-image. Mathematically, MSE is expressed as:
M N
1 2
MSE= ∑ ∑ [ C ( i , j )−S ( i, j ) ]
M × N i=1 j=1
Where, M and N are the rows and columns of the input image respectively, and C(i, j) and
S(i, j) means the pixel value at position (i, j) in the input image and the corresponding
output image, respectively [12].
Table 1 gives the measured values of MSE for different types of grayscale input images
of size 256×256. It is observed that for all the input images there is large difference
between the MSE values of both the techniques. MSE of the Arnold transform combined
with LSB technique is found to be much larger than the MSE of the Arnold transform
combined with DWT technique. Figure 22 shown below is the plot of the MSE values of
Arnold-LSB vs. Arnold-DWT where considerable change can be seen which shows that
Arnold combined with DWT is far better than Arnold combined with LSB.
 TABLE 1: MSE values of Arnold-LSB and Arnold-DWT
Image 256X256 Arnold-LSB Arnold-DWT
Fruit.bmp 0.0061 0.0017
Baboon.bmp 0.0065 0.0016
Elephant.gif 0.0064 0.0017
Bird.gif 0.0060 0.0016
Bank.jpg 0.0060 0.0018
Einstein.jpg 0.0062 0.0016
Lena.png 0.0064 0.0017
Obama.png 0.0056 0.0017
Rangoli.tif 0.0041 0.0014
Cameraman.tif 0.0058 0.0017

Figure 22 Plot of MSE values of Arnold-LSB vs Arnold-DWT

5.2.2 Comparison of PSNR values of Arnold-LSB vs. Arnold-DWT

The PSNR is expressed in dB’s and can be calculated using MSE as
P2
PSNR=10 × log ( )
MSE
Where, P is the peak signal value of the input image, and
P = max(C(i, j), S(i, j)) [12]

Table 2 gives the measured values of PSNR for different types of grayscale input images
of size 256×256. It is observed that for all the input images there is considerable amount
of changes between the PSNR values of both the techniques. Since MSE and PSNR are
inversely proportional to each other, PSNR of the Arnold transform combined with LSB
technique is found to be smaller than the PSNR of the Arnold transform combined with
DWT technique. Figure 23 shown below is the plot of the PSNR values of Arnold-LSB
vs. Arnold-DWT. PSNR is greater than 75db, this indicates good performance of the
proposed system (Arnold-DWT) which was only 70db in the previous technique (Arnold-
LSB). This shows that Arnold combined with DWT gives better results than Arnold
combined with LSB.
Table 2: PSNR values of Arnold-LSB and Arnold-DWT
Image 256X256 Arnold-LSB Arnold-DWT
Fruit.bmp 70.2819 75.7726
Baboon.bmp 70.0225 76.1066
Elephant.gif 70.0632 75.9172
Bird.gif 70.3477 76.1406
Bank.jpg 70.3834 75.6886
Einstein.jpg 70.2348 75.9725
Lena.png 70.0615 75.8588
Obama.png 70.6675 75.8236
Rangoli.tif 71.9857 76.5451
Cameraman.tif 70.5057 76.8450
 Figure 23 Plot of PSNR values of Arnold-LSB vs Arnold-DWT

5.2.3 Comparison of Entropy Values of Cover and Stego Images

The randomness of pixels is measured by entropy, which is computed by:
255
E( v )=−∑ p ( mi ) log 2 p ( mi )
i=0

The entropy of pixels v={ m0 , m1 , … .m255 } with equal probability p ( m i ) equals 8, which is
the perfect value. The entropy of stego image must be close to the perfect value as much
as possible [28].

The entropy for stego images is also tested and it was close to 8 in case of DWT
steganography. Table 3 lists the entropies of both cover and stego image using various
input images for both LSB and DWT techniques, where

E1 is the entropy of cover image, and

E2 is the entropy of stego image.

The entropy of cover image is found as 7.6114 since we have taken same cover image for
both techniques using different input images which are to be hidden i.e. Arnold
transformed images. Figure 24 shows the plot of entropy values of cover and stego
images for LSB and DWT techniques, which shows that the entropy values of DWT
technique is far better than that of LSB values.

Table 3: Entropy values of cover and stego images for LSB and DWT Techniques
Image 256X256 LSB DWT
Fruit.bmp E1=7.6114 E1=7.6114
E2=7.5239 E2=7.6993
Baboon.bmp E1=7.6114 E1=7.6114
E2=6.5646 E2=7.6475
Elephant.gif E1=7.6114 E1=7.6114
E2=7.2486 E2=7.6785
Bird.gif E1=7.6114 E1=7.6114
E2=6.6908 E2=7.6501
Bank.jpg E1=7.6114 E1=7.6114
E2=7.4028 E2=7.6842
Einstein.jpg E1=7.6114 E1=7.6114
E2=6.8538 E2=7.6583
Lena.png E1=7.6114 E1=7.6114
E2=7.2179 E2=7.6842
Obama.png E1=7.6114 E1=7.6114
E2=7.3058 E2=7.7053
Rangoli.tif E1=7.6114 E1=7.6114
E2=6.5867 E2=7.7063
Cameraman.tif E1=7.6114 E1=7.6114
E2=7.0404 E2=7.6821

Figure 24 Plot of Entropy values of cover and stego images for LSB and DWT Techniques
5.2.4 Comparison of timings of Arnold-LSB vs. Arnold-DWT
Table 4 shows the average timings for encryption and decryption processes of both the
techniques. It was found that Arnold-LSB technique takes little less time than that of
Arnold-DWT technique which is shown in Figure 25 below.

Table 4: Encryption and Decryption Timing of Arnold-LSB vs. Arnold-DWT

Process Arnold-LSB (seconds) Arnold-DWT (seconds)
Encryption 3.92815138 4.36446484
Decryption 2.73531262 3.05837376

Encryption Decryption

Figure 25 Plot of Average timing of Arnold-LSB vs. Arnold-DWT

5.3 Discussion
From the above results it is concluded that the Arnold-DWT transformed image provides
better quality output image as compared to the Arnold-LSB transformed image. There is a
significant improvement between the measured values of PSNR and MSE values of
Arnold-DWT over Arnold-LSB. The changes in the entropies of stego image using LSB
and DWT also show that DWT provides much better results in comparison to LSB
technique. However, the time taken by Arnold-DWT transformation is found to be
slightly higher than Arnold-LSB transformation.
Chapter 6
 Conclusion

In this report a comparison on two steganographic processes i.e. LSB and DWT
techniques is being done which are combined with Arnold transformation. All image
quality parameters described in this report have been calculated for different types of
images. Output results have been evaluated for comparison of both the techniques. These
result shows that Arnold transformed DWT technique provides better quality over Arnold
transformed LSB technique as there is a significant increase of 7.96% and 9.03% in the
PSNR and Entropy values respectively and 72.08% of decrement in the MSE value of
Arnold Transformed DWT technique with respect to Arnold Transformed LSB technique.
Input image was also divided into different blocks which slightly enhances the security of
the Arnold transformed image.
The performance of a steganographic technique is rated by three parameters; (i)
hiding capacity, (ii) imperceptibility and (iii) robustness. The hiding capacity means the
maximum amount of information that can be hidden in an image. It can also be
represented as the number of bits per pixel. It is analysed that the size of the cover image
must be greater than or equal to the size of the image which is to be embedded. The
perceptual imperceptibility of the embedded image is indicated by comparing the cover
image and the stego image so that their visual differences, if any, can be determined. The
Visual score test of imperceptibility is performed to compare the stego image obtained in
DWT and LSB technique and 98% result is found in the favour of DWT technique.
Robustness refers to how well the steganographic system resists the extraction of data.
Robustness measures the ability of embedded data to withstand against various changes in
the image. Robust image steganography algorithm is based on maximum entropy.
Measured values of Entropy obtained in this work indicate that Arnold-DWT is more
robust than Arnold-LSB technique. All these parameters show that Arnold-DWT is
provide better results than Arnold-LSB technique.

robustness, which refers to how well the steganographic

system resists the extraction of hidden data
robustness, which refers to how well the steganographic
system resists the extraction of hidden data
robustness, which refers to how well the steganographic
system resists the extraction of hidden dat

References
[1] Rao, Raghuveer M., and Manoj K. Arora. "Overview of image processing."
In Advanced image processing techniques for remotely sensed hyperspectral data, pp. 51-
85. Springer Berlin Heidelberg, 2004.
[2] Soni, Megha, Asst Prof Anand Khare, and Asst Prof Saurabh Jain. "A SURVAY OF
DIGITAL IMAGE PROCESSING AND ITS PROBLEM." International Journal of
Scientific and Research Publications (2014): 168.
[3] Huang, T. S., and K. Aizawa. "Image processing: some challenging
problems." Proceedings of the National Academy of Sciences 90, no. 21 (1993): 9766-
9769.
[4] Billingsley, Fred C. "Applications of digital image processing." Applied Optics9, no. 2
(1970): 289-299.
[5] Mandal, Pratap Chandra. "A Study of Steganography Technique using Discrete
Wavelet Transform." Journal of Global Research in Computer Science 5, no. 5 (2014).
[6] Banik, Barnali Gupta, and Samir Kumar Bandyopadhyay. "Secret sharing using 3
level DWT method of image steganography based on Lorenz chaotic encryption and
visual cryptography." In Computational Intelligence and Communication Networks
(CICN), 2015 International Conference on, pp. 1147-1152. IEEE, 2015.
[7] Devi, Anitha, and K. B. ShivaKumar. "Protection of confidential color image
information based on Reversible Data Hiding technique (PCCIRT)." In 2015
International Conference on Computing and Network Communications (CoCoNet), pp.
742-747. IEEE, 2015.
[8] Saini, Jaspal Kaur, and Harsh K. Verma. "A hybrid approach for image security by
combining encryption and steganography." In Image Information Processing (ICIIP),
2013 IEEE Second International Conference on, pp. 607-611. IEEE, 2013.
[9] Ahmad, Syukri Sayyid. "Steganography for inserting message on digital image using
least significant bit and AES cryptographic algorithm." In Cyber and IT Service
Management, International Conference on, pp. 16. IEEE, 2016.
[10] Phutane, Mr Indrajeet, and Sanjay Nalbalwar. "A New Method For Secret Image
Transmission via Secret Fragment Visible Mosaic Image."
[11] Houssein, Essam H., Mona AS Ali, and Aboul Ella Hassanien. "An image
steganography algorithm using Haar Discrete Wavelet Transform with Advanced
Encryption System." In Computer Science and Information Systems (FedCSIS), 2016
Federated Conference on, pp. 641-644. IEEE, 2016.
[12] Bhardwaj, Rupali, and Divya Khanna. "Enhanced the security of image
steganography through image encryption." In 2015 Annual IEEE India Conference
(INDICON), pp. 1-4. IEEE, 2015.
[13] Divya, V. V., S. K. Sudha, and V. R. Resmy. "Simple and secure image encryption."
International Journal of Computer Science 9, no. 6 (2012): 286.
[14] Radhadevi, P., and P. Kalpana. "Secure image encryption using AES." International
Journal of Research in Engineering and Technology Volume: 1 Issue 2 (2012).
[15] Nivedhitha, R., and T. Meyyappan. "Image security using steganography and
cryptographic techniques." International Journal of Engineering Trends and Technology
1, no. 3 (2012): 366-371.
[16] Anderson, Ross J., and Fabien AP Petitcolas. "On the limits of
steganography." IEEE Journal on selected areas in communications 16, no. 4 (1998):
474-481.
[17] Wategaonkar, S. R., Vrushali Gaikwad, Suvarna Kalghuge, Ashvini Gandhare, and
Prajakta Bawakar. "AN IMPROVED LSB BASED STEGANOGRAPHY TECHNIQUE
FOR RGB COLOUR IMAGES."
[18] Cheddad, Abbas, Joan Condell, Kevin Curran, and Paul Mc Kevitt. "Digital image
steganography: Survey and analysis of current methods." Signal processing 90, no. 3
(2010): 727-752.
[19] Krenn, Robert. "Steganography and steganalysis." (2004).
[20] Juneja, Mamta. "Data Hiding Algorithm for Bitmap Images Using
Steganography." communications 1: 4.
[21] Devi, Kshetrimayum Jenita. "A Secure Image Steganography Using LSB Technique
and Random Pseudo Random Encoding Technique." PhD diss., 2013.
[22] Kavitha, Kavita Kadam, Ashwini Koshti, and Priya Dunghav. "Steganography using
least significant bit algorithm." International Journal of Engineering Research and
Applications 2, no. 3 (2012): 338-341.
[23] Thapliyal, Pragati, and Madhu Sharma. "A image encryption scheme using chaotic
maps." International Journal 4, no. 1 (2015).
[24] Svanström, Fredrik. "Properties of a generalized Arnold’s discrete cat map." (2014).

[25] Li, Min, Ting Liang, and Yu-jie He. "Arnold transform based image scrambling
method." In 3rd International Conference on Multimedia Technology. 2013.
[26] Hussainbabu, G. "Secure Data Transmission Using Cryptography and
Steganography."
[27] Saxena, Ekta, and Mrs Neha Goswami. "Automatic Object Detection in Image
Processing: A Survey."
[28] Abdul-Majeed, Ansam Osama. "Chaotic Scheme for Image Encryption Based on
Arnold Cat's Map." International Journal of Computer Science and Information
Security 12, no. 3 (2014): 26.