www.tjprc.org editor@tjprc.

org
International Journal of Computer Science Engineering
and Information Technology Research (IJCSEITR)
ISSN(P): 2249-6831; ISSN(E): 2249-7943
Vol. 4, Issue 2, Apr 2014, 71-80
TJPRC Pvt. Ltd.

IMAGE STEGANOGRAPHY USING MANDELBROT FRACTAL
HARDIKKUMAR V. DESAI
1
& APURVA A. DESAI
2
1
Research Scholar, Singhania University, Pacheri Bari, Jhunjhunu, Rajasthan, India
2
Professor & Head, Department of Computer Science, Veer Narmad South Gujarat University, Surat, Gujarat, India

ABSTRACT
The development of Information and Communication Technology (ICT) represents an important part of the
society; it draws significant attention towards security and integrity of data. Definition of passing secret data in a traditional
way is now changed. ICT, leads to researcher to develop applications which are used to communicate secretly. There are
various technologies which are used for data hiding. Steganography, Cryptography and Water marking are the popular
techniques available to hide data securely. The paper focuses on various steganographic techniques as well algorithm is
developed to achieve steganography to hide image.
KEYWORDS: Steganography, Hide & Seek, Jsteg, Outguess0.1, Outguess0.2, F3, F4, F5, Histogram
INTRODUCTION
Steganography
The term steganography refers to the art of covert communications. Steganography can take data confidentiality to
a whole new level, since it embed message within another object by tweaking its properties, making the existence of the
messages practically undetectable. One of the oldest examples of Steganography dates back to around 440 BC in Greek
History. Herodotus, a Greek historian from the 5
th
century BC, revealed .Some examples of its use in his work entitled
The Histories of Herodotus. One elaborate example suggests that Histaeus, ruler of Miletus, tattooed a secret message on
the shaven head of one of his most trusted slaves. After the hair had grown back, the slave was sent to Aristagorus where
his hair was shaved and the message that commanded a revolt against the Persians was revealed
[1]
.Astonishingly the
method was still used by some German spies at the be-ginning of the 20th century
[2]
. Herodotus also tells how Demeratus,
a Greek at the Persian court, warned Sparta of an imminent invasion by Xerxes: he removed the wax from a writing tablet,
wrote his message on the wood underneath and then covered the message with wax. The tablet looked exactly like a blank
one (it almost fooled the recipient as well as the customs men).
There are various steganographic techniques, HIDE and SEEK is applicable to all data objects that contain
redundancy. People often transmit digital pictures over email and other Internet communication, and JPEG is one of the
most common formats for images. Moreover, steganographic systems for the JPEG format seem more interesting because
the systems operate in a transform space and are not affected by visual attacks
[3]
. Visual attacks mean that you can see
steganographic messages on the low bit planes of an image because they overwrite visual structures; this usually happens in
BMP images. Neil F. Johnson and Sushil Jajodia
[4]
showed that steganographic systems for palette-based images leave
easily detected distortions. While, other technique is JSteg, was developed by Derek Upham, it embeds the message data
within the LSBs of the DCT coefficients, rather than its pixel values
[5]
. Whereas, in OutGuess 0.1 is designed by Neils
Provos
[6]
,

improved the JSteg algorithm by scattering the embedding locations over the entire image according to PRNG
on image. This is very similar to the way that the randomized embedding approach improved the Hide & Seek algorithm.
72 Hardikkumar V. Desai & Apurva A. Desai

Impact Factor (JCC): 6.8785 Index Copernicus Value (ICV): 3.0

Out Guess 0.1 algorithm is considered much more secure than the Hide & Seek and JSteg algorithms because it not only
embedded the information in a more discrete area of the image (DCT coefficients), but it also scattered the locations of
embedding by using a PRNG to shuffle the ordering of the coefficients. However, after the release of the algorithm,
steganalysts were able to find a serious error in the technique that left statistical artifacts in the stegogrammes. As a result,
Neils Provos
[6]
created a revised version of the OutGuess 0.1 algorithm, called OutGuess 0.2. It was ensure that the
statistical properties of the cover image were maintained after embedding, such that stegogrammes looks statistically
similar to a clean image. Further, F3 Technique is an alternative to the OutGuess 0.2 algorithm, Andreas Westfeld designed
an algorithm called F3
[7]
. It was considered even more secure. The reason for this is that it did not instantiate the same
embedding process as the JSteg and OutGuess algorithms. Instead of avoiding embedding in DCT coefficients equal to 1,
the F3 algorithm permitted embedding in these regions, at the same time as it would still avoid embedding in zeros and the
DCT coefficients. The algorithm still embedded the message data sequentially .Another change with this algorithm was
that it did not embed directly in the least significant bits of the DCT coefficients, but instead took the absolute value of the
coefficients first, before comparing them to the message bits. If both the absolute value of the coefficient, and the message
bit were the same, then no changes are made. If they are different, then the absolute value of the DCT coefficient is
reduced by 1. The main drawback with F3 was the reality that it effectively embedded more zeros than ones as a result of
the shrinkage mechanism. This meant that when the statistical properties of the stegogrammes are examined through its
some object of embedding became visible. This is much the same as what happened in the JSteg implementation except a
slightly different pattern is derived. In addition to this, steganalysts also found that more odd coefficients existed in F3
stegogrammes than even coefficients. This now meant that there were two weaknesses that could be examined when
viewing the histogram of a suspect image. Though, another technique is F4 was developed to remove these properties such
that the histogram would appear similar to that of a clean image. Moreover the F5 algorithm
[8]
is predominantly the same
as the F4 algorithm, at least in terms of its strategy for encoding the message data. However, the F5 algorithm was
designed in an attempt to improve on the F4 algorithm by minimizing the disturbance caused when embedding the message
data. This was achieved by introducing matrix encoding, and the algorithm was the first known stego-system to make use
of this technique. Hamming coding is then used to embed potentially more than one bit per value by making no more than
1 change to the coefficients. This is denoted to the number of bits that are to be embedded
[9]
.The F5 algorithm is therefore
much more secure than the F3 and F4 algorithms.

Figure 1(a): Image to be Hide (b) Cover/Carrier Image (c) Cover /Carrier Image (with Stegogramme)
Fractals
Image Steganography Using Mandelbrot Fractal 73

www.tjprc.org editor@tjprc.org
Fractals were first described in the 1970s by 1BM mathematician Benoit Mandelbrot. He found geometry to be
incomplete. It could not describe the enormous and irregular shape of a mountain. It had no formal representation of the
appearance of a cloud. The new geometry that he developed could do all this. It was a description of the beautiful yet
irregular and fragmented patterns around us. The term Fractal was coined by Mandelbrot from the Latin Fractus,
an adjective for the irregular and the fragmented. Essentially, they replicate themselves by fragmentation.
LITERATURE REVIEW
The earliest work was done by Davern and Scott in
[10]
, who divide the domain blocks of the image into two sets.
They then use standard fractal image compression techniques to select a domain block that matches the range block,
however they choose a block from one of the two domain sets depending on whether the data bit they are embedding is a
one or a zero.
A great variety of steganographic methods on fractal compression principles
[11,12,13]
are described in the literature,
but greatest robustness is ensured by means of the methods
[12,13]
since they directly manipulate the code of compressed
image. Building in secrecy increase the given approaches will provide high level of protection.
Fractal generation exploits the concepts of iterated function systems (IFS). This consists of a collection of
contractive transformations in an iterative process until convergence, which is known as an attractor
[14,15]
. Fractal
compressed images employ partitioned iterated function systems (PIFS) to find self similarity within images along with
approximating the corresponding uncompressed images
[16,17]
. Fractal compression utilizes the theory of signal compression
in order to preserve memory space. In order to generate a fractal image, the transformed coefficients are stored in a fractal
code table for decoding the image instead of storing or transmitting image pixel values directly. Companies such as
Microsoft and IBM have investigated and use fractals as a means to store the large file size images in practice.
LIMITATIONS OF PREVIOUS ALGORITHMS
Hide & Seek and JSteg algorithm follows the sequential least significant bit for embedding. While, OutGuess 0.1
& OutGuess 0.2 are better as compare to Hide & Seek and JSteg because it scatters the locations for embedding by using a
PRNG to shuffle the ordering. Moreover, F3 is advanced compared to all others technique as it embedded in DCT
coefficients equal to 1 and avoids embedding in 0. The algorithm still embedding the message data sequentially.
Further, F4 is enhanced but it results in shrinkage mechanism. But, in F5 disturbance is minimized by using the matrix
encoding scheme. Whereas Davern Scott, Barnsley use fractal image compression technique in which self similarity of
structure in image is found through fractal compression for data hiding.
PROPOSED ALGORITHM
Embedding Algorithm
In this technique we generate Mandelbrot fractal location on input image .Fractal is generated by performing a
series of transforms on a line segment from (0, 0) to (1, 0).
= * + (1)
The affine transform in eq. (1) shows how the parametrical fractal transforms are performed. The a-d coefficients
of the matrix multiplication do the scaling and rotation operations, followed by the addition of the translation, e and f,
74 Hardikkumar V. Desai & Apurva A. Desai

Impact Factor (JCC): 6.8785 Index Copernicus Value (ICV): 3.0

coefficients. All six coefficients are functions of the input geometric parameters, p, which govern the geometry of the
generated fractal image. In order to plot a line segment in the final image, these operations must be performed twice,
once to generate the starting coordinates of a line segment and once to generate the ending coordinates. These operations
on a line segment are repeated in a recursive manner, until the specified number of iterations is reached. Mandelbrot fractal
is generated to define targeted locations on image and is divided into four parts by changing the pixel value.
The Mandelbrot set is the set of values of c in the complex plane for which the orbit of 0 under iteration of the complex
quadratic polynomial remains bounded

[18]
.

Z
n+1
=
+ c
(2)
That is, a complex number c is part of the Mandelbrot set if, when starting with z
0
= 0 and applying the iteration
repeatedly, the absolute value of z
n
remains bounded however large n gets.
Pixel values are change by using bit manipulation. To create a stegogramme the second image pixel values are
change accordingly and pixels are distributed into four pairs. Embedding process begins with finding the appropriate pixel
location and embeds four pairs anticlockwise till all the pixels are padded.
Step 1: Get the image A.
Step 2: Generate Mandelbrot Fractalto defined targeted region in Image
Step 3: Make the partition of image in four parts
Step 4: Get the image B to embed.
Step 5: Make a pair of two bits.
Step 6: Join the pairs of image B into imageA to generate stegogramme C.
Extracting Algorithm
In the extraction process we take the embedded image and recognized targeted region by generating Mandelbrot
fractal on image. Pixels are embedded in four groups in anticlockwise manner, first process each of group individually to
extract the value. After getting all the values from different groups and pad them together to extract the embedded image.
Step 1: Get the embedded image C.
Step 2: Generate Mandelbrot fractal to define the locations of stegogramme.
Step 3: extract the value from different groups.
Step 4: pad the region in two groups.
Step 5: Display the embedded imageB.


EXPERIMENTAL RESULTS
Image Steganography Using Mandelbrot Fractal 75

www.tjprc.org editor@tjprc.org
The results are based on comparison of images, histogram of images in gray scale and histogram of RGB images.
In the following (figure 2(a), (c)), (figure 5(a), (c)), (figure 8(a), (c)) and (figure 11 (a), (c)) represent the comparison of
original image and embedded image .we analyze that the image containing the stegograme is not easy to recognize by
naked eye detection that data is embedded, so it is difficult for steganalysis to recognize that image have stegogramme.
Whereas (figure 2(b), (figure 5(b)), (figure 8(b) and (figure 11 (b) is the image to be hidden.
A histogram is a graphical representation of the distribution of data. It is an estimate of the probability
distribution of a continuous variable and was first introduced by Karl Pearson
[19]
. A histogram is a representation of
tabulated frequencies, shown as adjacent rectangles, over discrete intervals (bins), with an area equal to the frequency of
the observations in the interval. The height of a rectangle is also equal to the frequency density of the interval, i.e., the
frequency divided by the width of the interval. The total area of the histogram is equal to the number of data. A histogram
may also be normalized displaying relative frequencies. It then shows the proportion of cases that fall into each of several
categories, with the total area equaling the categories are usually specified as consecutive, non-overlapping intervals of a
variable. The categories (intervals) must be adjacent, and often are chosen to be of the same size
[20]
. Stangor [21]
suggested that the rectangles of a histogram are drawn so that they touch each other to indicate that the original variable is
continuous.
Histogram comparison of gray scale is shown in comparison (figure 3(a),(b),(c)), (figure 6(a),(b),(c)),
(figure 9(a),(b),(c)) and (figure 12(a),(b),(c)). we found thatthe histogram distribution of data is same the difference is that
embedded image is double in size as compare to original once. Same result for the histogram generated with RGB image

Figure 2 (a): Tree Image (Original) (b) Image to Hide (c) Tree Image (Embedded)

Figure 3 (a): Tree Image Gray Scale (Original) (b) Tree Gray Scale (Embedded) (c) Tree Gray Scale (Extracted)
76 Hardikkumar V. Desai & Apurva A. Desai

Impact Factor (JCC): 6.8785 Index Copernicus Value (ICV): 3.0


Figure 4(a): Tree Image RGB (Original) (b) Tree Image RGB (Embedded) (c) Tree Image RGB (Extracted)

Figure 5(a): Aero Image (Original) (b) Image to Hide (c) Aero Image (Embedded)

Figure 6 (a): Aero Image Gray Scale (Original) (b) Aero Gray Scale (Embedded) (c) Aero Gray Scale (Extracted)

Figure 7 (a): Aero Image RGB (Original) (b) Aero Image RGB (Embedded) (c) Aero Image RGB (Extracted)
Image Steganography Using Mandelbrot Fractal 77

www.tjprc.org editor@tjprc.org

Figure 8 (a): Aqua Image (Original) (b) Image to Hide (c) Aqua Image (Embedded)

Figure 9 (a): Aqua Image Gray Scale (Original) (b) Aqua Image Gray Scale (Embedded)
(c) Aqua Image Gray Scale (Extracted)

Figure 10 (a): Aqua Image RGB (Original) (b) Aqua Image RGB (Embedded) (c) Aqua Image RGB (Extracted)

Figure 11 (a): Traffic Image (Original) (b) Image to Hide (c) Traffic Image (Embedded)]
78 Hardikkumar V. Desai & Apurva A. Desai

Impact Factor (JCC): 6.8785 Index Copernicus Value (ICV): 3.0


Figure 12 (a): Traffic Image Grayscale (Original) (b) Aqua Image Grayscale (Embedded)
(c) Aqua Image Grayscale (Extracted)

Figure 13 (a): Traffic Image RGB (Original) (b) Traffic Image RGB (Embedded)
(c) Aqua Image Traffic (Extracted)

CONCLUSIONS
Steganography is very popular technique for data hiding in todays world, After analyze different techniques,
we proposed that the new technique ofsteganography by keeping in the mind of various attack like visual , structural and
statistical attacks to produce best stego-grame.Steganography proved to be very best technique to embed the data is much
important for researcher as well as for the stego-analyst, steganography aiming to be ahead of the field such that covert
communications may exist. There is then a call for steganalysis to catch up such that covert communications are minimized
as much as possible.
REFERENCES
1. Herodotus, the Histories, chap. 5 book entitled Terpsichore, chap. 7 book entitled Polymnia. London, England:
J. M. Dent & Sons, Ltd, 1992, ISBN 0-460-87170-6, translated by George Rawlinson, introduction by Hugh
Bow-den.
2. B. Newman, Secrets of German Espionage. London: Robert Hale Ltd, 1940.
3. A. Westfeld and A. Pfitzmann, Attacks on Stegano- graphic Systems, Proc. Information Hiding 3rd Intl
Work- shop , Springer Verlag, 1999, pp. 6176.
4. N.F. Johnson and S. Jajodia, Steganalysis of Images Created Using Current Steganographic Software, Proc. 2
nd

Intl Workshop in Information Hiding, Springer-Verlag, 1998, pp. 273289.
5. T. Zhang and X. Ping, A Fast and Effective Steganalytic Technique Against JSteg-like Algorithms, Proc. 8th
Image Steganography Using Mandelbrot Fractal 79

www.tjprc.org editor@tjprc.org
ACMSymp. Applied Computing, ACM Press, 2003.
6. N. Provos. "Defending Against Statistical Steganalysis", Proceedings of the 10
th
USENIX Security Symposium,
vol. 10, pp. 323-335, 2001.
7. A. Westfeld. "F5 - A Steganographic Algorithm: High Capacity Despite Better Steganalysis", Lecture Notes in
Computer Science, vol. 2137, pp. 289-302, 2001.
8. A. Westfeld. "F5 - A Steganographic Algorithm: High Capacity Despite Better Steganalysis", Lecture Notes in
Computer Science, vol. 2137, pp. 289-302, and 2001.
9. M. Leivaditis. "Statistical Steganalysis", Masters thesis, Department of Computing, University of Surrey,
2007.
10. Davern , P. and Scott, M., Fractal based image steganography, Proc. of the First Intl. Workshop on Information
Hiding, Lecture Notes in Computer Science 1174, 279-294 (1996).
11. Bas, P., Chassery, J. M., Davoine, F. Using the Fractal Code to Watermark Images // Proc ICIP98, 1998. 1.
P. 469-473.
12. Li, C., Wang, S. Digital Watermarking Using Fractal Image Coding // IEICE Trans. Fund., 2000. 6.
P.1286-1288.
13. Liao, P., Chen, C., Chen, C., Pan, J. Interlacing Domain Partition for Fractal Watermarking // IIH-MSP06,
2006. P. 441-444.
14. Barnsley, M. F., Fractals Everywhere, Academic Press, Boston (1988).
15. Barnsley, M. F., Fractal image compression, Notices of the AMS 43(6), 657-662 (1996).
16. Jacquin, A. E., Image coding based on a fractal theory of iterated contractive image transformations,
IEEE Trans. on Image Processing 1(1), 18-30 (1992).
17. Jacquin, A. E., Fractal image coding: a review, Proc. of the IEEE 81(10), 1451-1465 (1993).
18. http://math.bu.edu/DYSYS/explorer/def.html
19. Pearson, K. (1895). "Contributions to the Mathematical Theory of Evolution. II. Skew Variation in Homogeneous
Material". Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering
Sciences 186: 343414. Bibcode:1895RSPTA.186..343P. doi:10.1098/rsta.1895.0010.
20. Howitt, D. and Cramer, D. (2008) Statistics in Psychology. Prentice Hall
21. Charles Stangor (2011) "Research Methods For The Behavioral Sciences". Wadsworth, Cengage Learning. ISBN
9780840031976.