You are on page 1of 5

STEGNOGRAPHY TECHNIQUE USING DCT QUANTIZATION AND

MODIFIED FIBONACCI SERIES DECODER

Dr. Shiv Kumar1 Shekhar Chander2,

1
Senior Professor, ACEIT Jaipur, 2 M.Tech Scholar, ACEIT Jaipur
Abstract Steganography receiver can be same or The Modified Fibonacci generating stego image.
is the art and science of different. number is generated It takes 7 bits (b7, b6, b5,
hiding information within
information. This thesis using the following b4, b3, b2, b1) LSB of
introduces steganography Second method is formula cover image by using
and focus on the Discrete steganography, where the modified Fibonacci
Cosine Transform (DCT)
secret message is hidden Ln = Ln-1 + Ln-2 L1 =2 representation as inputs
technique for hiding
messages in an image. The in another message, and L2 =1 and produces 3 bits of
proposed method modifies image, audio or video.. (1.1) outputs f3, f2, and f1.
the quantization table first. The word steganography Figure 1.1 shows the
Next, the secret message is To represent a number in
hidden in the cover-image is derived from the encoder/decoder
with its middle-frequency Greek words stegos the range [0,255] twelve generator.
of the quantized DCT meaning cover and Modified Fibonacci
coefficients modified, number are required.
using Modified Fibonacci
grafia meaning
Series Decoder (MFSD). writing [5] defining it
Further it enhances the as covered writing.
security by using public
key RSA encryption
algorithm on message 2. The Proposed Hiding
prior to hiding. Results of System The image pixel value
this work are compared will be represented as
with existing This stage includes two
steganography methods,
sum of non consecutive
the percent improvement is parts; cryptography and Modified Fibonacci
68.87634 higher using steganography. The number.For example
proposed method than LSB cryptography part Binary and Modified
in case study followed for includes encryption of
Polar Bear Image. Fibonacci number
the secret message by presentation of pixel Figure 1: Encoder/Decoder for
1. Introduction to using public key RSA value 26 is given below stego image generation
Stegnography encryption algorithm.
The steganography part Binary representation For each pixel the 7 least
Digital communication is includes hiding the 00011010 significant bits are
essential part in everyone encrypted secret message applied to the
life. Often it is desired in JPEG image. This Modified Fibonacci encoder/decoder, three
that communication work uses middle number representation Exclusive OR operations
should be secret Two coefficient to hide 000001010010 create three outputs f3,
techniques are available message using Modified f2 and f1
to accomplish this goal Fibonacci Series The motivation behind
the Cryptography [1] and Decoder (MFSD). using Modified f3=b1b5b6b7
the (1.2)
Fibonacci number is that
Steganography[2,3,4]. 3. Modified Fibonacci Modified Fibonacci bit is f2 =b1b3b4b7
Series Decoding less significant than that (1.3)
In Cryptography the (MFSD) Method of traditional binary bit.
sender encrypts the For example if the 5th bit f1 =b1b2b4b6
message with a key and This section describes (1.4)
of pixel value 26 is
the encrypted message is the Modified Fibonacci altered to 0 then it
TABLE 1 Procedure of
sent to receiver over an series decoding Method, becomes 10 in binary changing seven LSB of a pixel
insecure communication in this approach 3 bits representation while 19
channel. On the other are embedded by altering in Modified Fibonacci
side, the receiver only single bit in a pixel number representation
decrypts the received using Encoder/Decoder. which is more close to
encrypted message with First the image pixel is the original value.
a key and extracts the decomposed using
original message. The Modified Fibonacci A 7 bit encoder/decoder
key for sender and number representation. is proposed for
for three bits of message The proposed method DCT coefficients will be reconstructed image will
uses the JPEG image 0 and therefore, the not be too much
pre-processing method image quality is distorted.
upon the cover-image. degraded, especially
For this partition a cover- when the cover-image
image O into non undergoes a high
overlapping blocks of 8 compression ratio.
X 8 pixels, and then use
the DCT(Discrete cosine If there are many
transform) to transform quantized coefficients
each block into DCT equal to 0, then the
coefficients. The DCT message capacity will be
coefficients are then also decreased therefore,
scaled with a modified to increase message
quantization table.. This capacity and quality of
table is notably different image the quantization
from the quantization table of JPEG needs to
table of JPEG. This is be modified. To increase
because the secret quality of stego image
message will be secret message is hidden
embedded in the middle- by using modified
frequency part of the Fibonacci series with
quantized DCT encoder/decoder method,
coefficients. This new in this method a change
Where suppose f3, f2, f1 quantization table is in single bit in DCT
coefficients hides three
are the hidden message obtained by changing
which is embedded in the bit in single DCT
middle elements by
pixel. If f3, f2, f1 are the coefficients with
one in original minimal change in that
same as the hidden
quantization table. coefficient
message then there is no
need to alter that original
image pixel. If not, it is11 10 16 1 1 1 In Fig. 2, there are 26
needed to change that coefficients located in
12 14 1 1 1 1
original image pixel in a the middle part that are
way to cause the output13 1 1 1 1 69 set to be one. They are
of the encoder/decoder 1 p[0,4], p[0,5], p[0,6],
1 1 1 87 80
be equal to the hidden p[0,7], p[1,3], p[1,4],
message. This thesis 1 1 1 68 109 103 p[1,5], p[1,6], p[2,2],
allows alternation in only 1 p[2,3], p[2,4], p[2,5],
1 64 81 104 113
one bit in pixel that can p[3,1], p[3,2], p[3,3],
be any of the seven bits 1 78 87 103 121 120 p[3,4], p[4,0], p[4,1],
to make the outputs of92 95 98 112 100 103
p[4,2], p[4,3], p[5,0],
the encoder/decoder is p[5,1], p[5,2], p[6,0],
equal to the hidden Fig. 2. Modified quantization p[6,1], and p[7,0]. Here p
table is the modified
message. The procedure
that changes only one bit quantization table and
If using the standard p[a,b] is the value of the
is described in table 1.
quantization table to ath row and bth column
quantize and dequantize element of p. Based on
4. Message Hiding
DCT coefficients, then this quantization table,
Stage
majority of middle the secret messages can
frequency quantized be reserved and the
5. Test and Results Fig. 4. Polar Bear Image after Case 1 SNR using LSB steganography
Stegnography
and DCT for zero algorithms using
The testing of this character text file is Modified Fibonacci
Percentage Change for
example is 22.5561 and 39.0329 Series Decoder (MFSD)
SNR is calculated by the
performed by respectively which add complexity to
formula
taking the secret make it difficult to
text of varying Case 2 SNR using LSB extract information by
Percentage Change for
length and the and DCT for 1000 the intruder. Also a
SNR =
cover-image character text file is cryptography algorithm
(Polar Bear DCTSNR - LSBSNR 22.4301and 35.3966
*100 is used, which provides
Image) LSBSNR respectively more complexity. A
(256*173). Figure number of conclusions
3 shows the cover
(1.5) Figure 5 Showing Graph
derived from this study
of SNR versus No of
image before
TABLE 2 SNR BY character in text file for
encryption and EMBEDDING 1. Using cover images
LSB and Proposed DCT
hiding. Figure 4 DIFFERENTIAL LENGTH and a secret message
TEXT IN POLAR BEAR for Polar Bear Image
shows the stego- IMAGE with different lengths are
image after Average percentage that
extracting and Number of SNR By Using improvement for SNR
decryption. Table Character LSB Technique
in Text file is calculated by taking The SNR value for the
2 shows the SNR 0 22.5561 average of percentage color Polar bear image is
for each 1 22.5562 changes which is between (39- 35).
Technique (LSB 68.87634 for Polar Bear
2 22.556
or Proposed image The SNR value
3 22.5549
DCT) with decreases with the
4 22.5553
varying text increase in length of the
5 22.5554
length for Polar secret message.
Bear Image. 6 22.5553
7 22.5552
7. Future Scope
8 22.555
9 22.5549
1. The method of
10 22.5547 embedding is DCT
20 22.5526 transform insertion, in
30 22.5511 the future another
40 22.5488 embedding method
50 22.5471 should be employed like
60 22.5444 wavelet or Sparta
70 22.5421 spectrum methods.
80 22.5402
90 22.5391
2. The Encryption
Fig. 3.Polar Bear Image Before
method could be
100 22.5374
Stegnography replaced by other public
200 22.5216
keys ciphering
300 22.5096
Fig. 5. Graph Showing Result algorithm.
400 22.4943 SNR versus No of character in
text file for LSB and Proposed
500 22.4807 8. REFERENCES
DCT for Polar Bear Image
600 22.473
700 22.4628 6. Conclusions [1] W. Stallings,
800 22.45 Cryptography and
900 22.4411 This work is Network Security:
1000 22.4301 implemented using DCT Principles and Practice,
based image third ed., Pearson
Education, New Jersey,
2003.

[2] N. F. Johnson and S.


Katzenbeisser, .A
survey of steganographic
techniques,
Information Hiding,
pp.43-78. Artech House,
Norwood, MA,
Journal,2000.

[3] Lou, D. C. and Liu, J.


L.. Steganography
Method for Secure
Communications.
Elsevier Science on
Computers& Security,
21, 5:pp. 449-460, 2002.

[4] J. Fridrich and M.


Goljan,Practical
steganalysis of digital
images-state of the art,
Proc. SPIE Photonics
West, San Jose,
California, Vol. 4675,
pp. 1-13, Jan. 2002.

[5] Moerland, T.,


Steganography and
Steganalysis, Leiden
Institute of Advanced
Computing
Science,www.liacs.nl/ho
me/ tmoerl/privtech.pdf

You might also like