Computer Engineering and Intelligent Systems www.iiste.

org
ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online)
Vol 3, No.4, 2012

80
Secure Compressed Image Transmission using Self Organizing
Feature Maps

G. Mohiuddin Bhat

(Corresponding Author)
University Science Instrumentation Centre, University of Kashmir, Hazratbal, Srinagar;
gmbhat_ku@yahoo.co.in

Asifa Baba
School of Technology, Islamic University of Science & Technology, Awantipora; asifababa@gmail.com

Abstract: Due to the widespread use of multimedia applications, the data communication channels feel
short of bandwidth with regard to channel capacity requirements. As such, the need for improved image
compression techniques, together with image security, is increasing day by day. In this paper, the concept of
compressed image security has been explored. The input image data is applied to the image partitioning and
vectorization block where the whole image is divided into 4x4 non-overlapping blocks. Each block serves
as a vector of 16 elements for the Self Organizing Feature Map (SOFM) network by which the indexes of
the codewords are determined. These indexes are coded in a binary stream of 0s and 1s using a variable
length Entropy Coding Scheme. These long strings of 0s and 1s are scrambled by a typical scrambler in
order to secure the image data from the unauthorized receiver. The simulation results demonstrate the
improved coding efficiency of the proposed method, when compared with JPEG, in addition to providing
the message security. The proposed scheme achieves a compression ratio upto 38:1.

Keywords: SOFM, Entropy Coding, Codewords, Image Security, Scrambler, encryption, JPEG, Arithmetic
Coding..

Computer Engineering and Intelligent Systems www.iiste.org
ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online)
Vol 3, No.4, 2012

81
1. Introduction: Message security is the science and study of methods for protecting message signals
from unauthorized disclosure and modification. The message signals include digital data from computers
and other communication systems, digital image data, analog messages including audio signals like speech
signals, video signals and other instrumentation signals. Secure image transmission is of paramount
importance in defense, surveillance and other strategic applications. The compressed image data can be
encrypted before transmission over an insecure channel. The image information can thus be denied to an
unauthorized receiver thereby safeguarding eavesdropping. However, an intended receiver equipped with
proper Key and the decryption technique should be able to recover the original image from the received
encrypted data.
Cryptography is the science of deliberately disguising the signals using typical ciphers so that they assume
the form of noise signal for an unauthorized receiver. The noise-like (encrypted) signals can be decrypted
or deciphered back to recover the original message signals [1]. Scramblers are a class of substitution
ciphers and have been found to be suitable for various security requirements such as those used by cable
and satellite TV operators and mobile phone service providers [2,3].
In this paper secure compressed image compression scheme is presented where the input image data is first
compressed using Self Organizing Feature Maps (SOFM) based technique. The compressed image data is
then scrambled using a typical scrambler so that it appears noise like for an unauthorized receiver [4]. The
Simulation results for the compressed image transmission and reception using the proposed techniques
have been presented in the paper for standard grayscale images.

2 Proposed technique of secure image transmission using SOFM based Image Compression:
Block Diagram of the proposed compressed Image Encryption technique is shown in Fig. 1. The input
image data is applied to the image partitioning and vectorization block where the whole image is divided
into 4x4 non-overlapping blocks. Each block serves as a vector of 16 elements for the Self Organizing
Feature Map (SOFM) network by which the indexes of the codewords are determined. These indexes are
coded in a binary stream of 0s and 1s using a variable length Entropy Coding Scheme. These long strings
Computer Engineering and Intelligent Systems www.iiste.org
ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online)
Vol 3, No.4, 2012

82
of 0s and 1s are scrambled by a typical scrambler shown in Fig. 2 in order to secure the image data from
the unauthorized receivers.
3. Self Organizing Feature Maps:
Self Organizing Feature Map (SOFM) has formed a basis for a great deal of research into applying network
models to the problem of codebook design in Vector Quantization [5]. The SOFM introduced by Kohonen
is an unsupervised learning method which has both clustering and visualization properties and creates a
correspondence between the input space of stimuli and the output space constituted of the codebook
elements (the code words or neurons) [6]. The learning algorithm ensures that the most highly activated
node as well as its neighbors move towards a sample presented to the network. The networks are self
organizing in that nodes tend to attain weight vectors that capture characteristics of the input vector space,
with the neighborhood relation translating into proximity in Euclidean space, even if the initial values of
weight vectors are arbitrary. In the SOFM algorithm, the vector X is used to update not only the winning
class but also its neighboring classes according to the following rule:
For each vector X in the training
1. X is classified according to:
) 1 ..( .......... min
j i i
W X W X if C X =

2. The weights W
j
are then updated according to:
( )
( ) ( ) ( ) ( )
( ) ( )
) 2 ...(
,
,
1
2

+
= +
t C N C if t W
t C N C if t W X lr t W
t W
i j j
i j j j
j

Where W

is the feature vector, lr is the learning parameter in the range of 0-1 and N (C
i
, t) is the set of
classes, which are in the neighborhood of the winning class C
i
at time t. The subscript j represents the
index of all vectors in the neighborhood of the i
th
feature vector.

4. Binary Arithmetic Coding:
In Arithmetic coding (AC) Scheme, a one to one correspondence between source symbols and codewords
does not exist; instead, an entire sequence of source symbols (or message) is assigned a single arithmetic
Computer Engineering and Intelligent Systems www.iiste.org
ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online)
Vol 3, No.4, 2012

83
codeword. As the number of symbols in the message increases, the interval used to represent it becomes
smaller and the number of information bits required to represent the interval becomes larger [7].

5. Scrambler for Data:
Fig. 2 shows the circuit diagram of the proposed scrambler. Here tap-gain values are continuously changed
with the help of a PN-sequence. Inputs Q
A
, Q
B
, Q
C
and Q
D
to the AND gates are the outputs of a 15-bit PN-
sequence generator. The value of a Q
i
=1 means that the corresponding shift register stage is effective in the
generation of the Key K. Similarly Q
i
=0 means that a
i
=1. For unscrambling the received message, the
receiver has to know the value of N, the PN-sequence and its starting point. Thus, the unscrambling
becomes very difficult even for smaller values of N, thus minimizing the problem of error propagation

6. Simulation Results:
The proposed Algorithm based on SOFM and Arithmetic Coding has been implemented using MATLAB-
7.02 and the proposed algorithm has been simulated on various grayscale images of size 256x256 with 8
bits per pixel over a PC with Intel Pentium IV, 2.9 GHz and 256MB RAM under Windows-XP operating
system. The Lena and Couple images are used for training the initial set and codebook design. The
performance of the proposed technique is tested for images Einstein and Woman, which are outside the
training sequence. The performance is measured for various codebook sizes of 2
n
where n is an integer
varying from 5 to 8, and then compression efficiency is measured in terms of compression ratio (CR) which
is defined as:
CR =
(no.o] pxcIs n 4x4 bIock)x
(No.o] bts pc pxcI n thc ognuI mugc)
bts cqucd cpcscntng cuch codcbook ndcx
(S)
CR =
16x8
n
(4)
The quality of the decoded image is measured in terms of Peak-Signal-to-Noise-Ratio (PSNR) which is
defined as:
PSNR = 1u log
10
_
2SS
2
1
HN
(
1
(i, ]) -
2
(i, ]))
2
N-1
=0
M-1
]=0
_ (S)
Computer Engineering and Intelligent Systems www.iiste.org
ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online)
Vol 3, No.4, 2012

84

Where NxM is the image size, f
1
(i, j) is the random field of the original field intensities and f
2
(i, j) denotes
the random field of the reconstructed image intensities. Fig. 3 outlines the simulation results of the image
transmission over the proposed secure image transmission systems. Variation of CR with PSNR and
Codebook size has been plotted. Standard grayscale image including Lena, Couple, Einstein were first
compressed using SOFM based image compression technique. The compressed image data has been
subsequently scrambled using the modified scrambler as shown in Fig. 2. The Scrambled image
corresponding to the given standard image woman is shown in Fig. 4 along with the original and the
reconstructed images.
The waveforms of the Key bits, Arithmetic Coded bits and the scrambled image data has been presented in
Fig. 5. Further, the histogram of the original and encrypted image is shown in Fig. 6, and the Correlation
analysis of the plain image and encrypted image has been presented in Fig. 7.

7. Comparison with JPEG:
The performance of the proposed technique for image compression has been compared for the standard
image Woman with the performance obtained by using JPEG standard using 8x8 default quantization for
compression. The results of the comparison have been presented in Fig. 8. It can be observed from the
figure that the proposed method outperforms the DCT based JPEG by up to 1 dB.

8. Conclusion:
The paper presents an interesting technique for fast mage data transmission with data security against eaves
dropping. An SOFM technique has been used for data compression while as a typical scrambler has been
used for data encryption for message security. The proposed technique is shown to be very efficient for
secure image data compression with reduced bit error propagation. The transmission and reception of
encrypted image using the proposed technique has been tested experimentally on standard grayscale images.
The proposed technique provides a compression ratio of about 38:1 which outperforms the conventional
JPEG standard for image compression. The simulation of the proposed system has been carried out using
Computer Engineering and Intelligent Systems www.iiste.org
ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online)
Vol 3, No.4, 2012

85
MATLAB 7.02 and the performance results have been presented in the paper. The performance of the
proposed secure image transmission system has been found to be satisfactory.

References:

1. Shaaban, E., Data Compression Basics,, 2002, Retrieved from-
http://meseec.ce.rit.edu/eecc694-spring2000/694-5-9-2000.pdf
2. M. Buer and J. Wallace, Integrated Security for digital video broadcast, IEEE Trans. Cons. Elect.,
Aug. 1996, vol. 42, No. 3, pp. 500-503.
3. Dorothy Elizabeth and Robling Denning, Cryptography and Data Security, Addison Wesley, 1982.
4. H. Feistel, Cryptography and Computer Privacy, Scientific America, May 1973, vol. 228, pp. 15-23,
5. Wu, Chung-Ping & Kuo, C.-C.J., Design of integrated multimedia compression and encryption
systems, IEEE Transactions on Multimedia, 2005, Vol. 7, No. 5, pp 828 839.
6. P. Wayner, A redundancy reducing Cipher, Cryptologia, April 1988, vol. XII, pp. 107.
7. J. B. Kam & G. I. Davida, Structured design of substitutional permutation encryption networks,
IEEE Trans. Computers, Oct. 1979, vol. C-28, pp. 747-753.



Fig. 1 Proposed Image Encryption using SOFM based Image Compression


Image
partition and
SOFM based
Vector
quantization
Arithmetic
coding
Code book
Input
Image

Scrambler Coded
Bit stream
Computer Engineering and Intelligent Systems www.iiste.org
ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online)
Vol 3, No.4, 2012

86



Fig. 2 Proposed Image Data Scrambler


Fig. 3 (a) Variation of Compression ratio with PSNR (b) Variation of Compression ratio with
Codebook Size

QA
QD
QC
QB
QA
QD
QC
QB
Decrypted Compressed
Image data
Compressed image data






S
H
I
F
T









S
H
I
F
T



Cryptogram
Computer Engineering and Intelligent Systems www.iiste.org
ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online)
Vol 3, No.4, 2012

87


(a) (b) (c)
Fig. 4 SOFM based Secure Image transmission using standard Image Woman
(a) Original Image (b) Scrambled Image (c) Reconstructed Image


(a)



(b)


(c)

Fig. 5 Various Waveforms generated: (a) Key bits (b) Arithmetic Code (c) Scrambled Image data

Computer Engineering and Intelligent Systems www.iiste.org
ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online)
Vol 3, No.4, 2012

88


Fig. 6 Histograms of Original and Encrypted Image Woman

Fig. 7 Correlation analysis of original image and encrypted image


Fig. 8 Performance comparison of the proposed method with JPEG for image woman


International Journals Call for Paper
The IISTE, a U.S. publisher, is currently hosting the academic journals listed below. The peer review process of the following journals
usually takes LESS THAN 14 business days and IISTE usually publishes a qualified article within 30 days. Authors should
send their full paper to the following email address. More information can be found in the IISTE website : www.iiste.org


Business, Economics, Finance and Management PAPER SUBMISSION EMAIL
European Journal of Business and Management EJBM@iiste.org
Research Journal of Finance and Accounting RJFA@iiste.org
Journal of Economics and Sustainable Development JESD@iiste.org
Information and Knowledge Management IKM@iiste.org
Developing Country Studies DCS@iiste.org
Industrial Engineering Letters IEL@iiste.org

Physical Sciences, Mathematics and Chemistry PAPER SUBMISSION EMAIL
Journal of Natural Sciences Research JNSR@iiste.org
Chemistry and Materials Research CMR@iiste.org
Mathematical Theory and Modeling MTM@iiste.org
Advances in Physics Theories and Applications APTA@iiste.org
Chemical and Process Engineering Research CPER@iiste.org

Engineering, Technology and Systems PAPER SUBMISSION EMAIL
Computer Engineering and Intelligent Systems CEIS@iiste.org
Innovative Systems Design and Engineering ISDE@iiste.org
Journal of Energy Technologies and Policy JETP@iiste.org
Information and Knowledge Management IKM@iiste.org
Control Theory and Informatics CTI@iiste.org
Journal of Information Engineering and Applications JIEA@iiste.org
Industrial Engineering Letters IEL@iiste.org
Network and Complex Systems NCS@iiste.org

Environment, Civil, Materials Sciences PAPER SUBMISSION EMAIL
Journal of Environment and Earth Science JEES@iiste.org
Civil and Environmental Research CER@iiste.org
Journal of Natural Sciences Research JNSR@iiste.org
Civil and Environmental Research CER@iiste.org

Life Science, Food and Medical Sciences PAPER SUBMISSION EMAIL
Journal of Natural Sciences Research JNSR@iiste.org
Journal of Biology, Agriculture and Healthcare JBAH@iiste.org
Food Science and Quality Management FSQM@iiste.org
Chemistry and Materials Research CMR@iiste.org

Education, and other Social Sciences PAPER SUBMISSION EMAIL
Journal of Education and Practice JEP@iiste.org
Journal of Law, Policy and Globalization JLPG@iiste.org
New Media and Mass Communication NMMC@iiste.org
Journal of Energy Technologies and Policy JETP@iiste.org
Historical Research Letter HRL@iiste.org
Public Policy and Administration Research PPAR@iiste.org
International Affairs and Global Strategy IAGS@iiste.org
Research on Humanities and Social Sciences RHSS@iiste.org
Developing Country Studies DCS@iiste.org
Arts and Design Studies ADS@iiste.org
[Type a quote from the document or the
summary of an interesting point. You can
position the text box anywhere in the
document. Use the Drawing Tools tab to change
the formatting of the pull quote text box.]
Global knowledge sharing:
EBSCO, Index Copernicus, Ulrich's
Periodicals Directory, JournalTOCS, PKP
Open Archives Harvester, Bielefeld
Academic Search Engine, Elektronische
Zeitschriftenbibliothek EZB, Open J-Gate,
OCLC WorldCat, Universe Digtial Library ,
NewJour, Google Scholar.
IISTE is member of CrossRef. All journals
have high IC Impact Factor Values (ICV).