A SECURE COLOR IMAGE STEGANOGRAPHY IN TRANSFORM DOMAIN

International Journal on Cryptography and Information Security (IJCIS), Vol.3, No.1, March 2013
DOI:10.5121/ijcis.2013.3103 17
A SECURE COLOR IMAGE STEGANOGRAPHY IN
TRANSFORM DOMAIN
Hemalatha S1
, U Dinesh Acharya2
, Renuka A3
, Priya R. Kamath4
1,2,3,4
Department of Computer Science and Engineering, Manipal Institute of Technology,
Manipal University, Manipal, Karnataka, India
1
hema.shama@manipal.edu;2
dinesh.acharya@manipal.edu;3
renuka.prabhu@manipa
l.edu;4
priyarkamath@gmail.com
ABSTRACT
Steganography is the art and science of covert communication. The secret information can be concealed in
content such as image, audio, or video. This paper provides a novel image steganography technique to hide
both image and key in color cover image using Discrete Wavelet Transform (DWT) and Integer Wavelet
Transform (IWT). There is no visual difference between the stego image and the cover image. The extracted
image is also similar to the secret image. This is proved by the high PSNR (Peak Signal to Noise Ratio),
value for both stego and extracted secret image. The results are compared with the results of similar
techniques and it is found that the proposed technique is simple and gives better PSNR values than others.
KEYWORDS
Steganography, DWT, IWT, PSNR, YCbCr
1. INTRODUCTION
Recently, the information hiding technique has developed rapidly in the field of information
security and has received significant attention from both industry and academia. It contains two
main branches: digital watermarking and steganography. The former is mainly used for copyright
protection of electronic products while, the latter is a way of covert communication. The main
purpose of steganography is to convey the information secretly by concealing the very existence
of information in some other medium such as image, audio or video. The content used to embed
information is called as cover object. The cover along with the hidden information is called as
stego-object [1]. In this paper image is the cover and secret information is also an image. Both
secret image and stego key are embedded in the cover image to get stego image.
The major objective of steganography is to prevent some unintended observer from stealing or
destroying the confidential information. There are some factors to be considered when designing
a steganography system: [1]
• Invisibility: Invisibility is the ability to be unnoticed by the human.
• Security: Even if an attacker realizes the existence of the information in the stego object it
should be impossible for the attacker to detect the information. The closer the stego
image to the cover image, the higher the security. It is measured in terms of PSNR. High
PSNR value indicates high security.
PSNR = dB, where L = maximum value, MSE = Mean Square Error.

18
MSE = , where X = original value, X’ = stego value and N = number of samples.
• Capacity: The amount of information that can be hidden relative to the size of the cover
object without deteriorating the quality of the cover object
• Robustness: It is the ability of the stego to withstand manipulations such as filtering,
cropping, rotation, compression etc.
The design of a steganographic system can be categorized into spatial domain methods and
transform domain methods [1].
In spatial domain methods, the processing is applied on the image pixel values directly. The
advantage of these methods is simplicity. The disadvantage is low ability to bear signal
processing operations. Least Significant Bit Insertion methods, Pallete based methods come under
this category.
In transform domain methods, the first step is to transform the cover image into different domain.
Then the transformed coefficients are processed to hide the secret information. These changed
coefficients are transformed back into spatial domain to get stego image. The advantage of
transform domain methods is the high ability to face signal processing operations. However,
methods of this type are computationally complex. Steganography methods using DCT (Discrete
Cosine Transforms), DWT, DFT (Discrete Fourier Transforms) come under this category.
1.1. Discrete Wavelet Transform
DWT is used for digital images. Many DWTs are available. Depending on the application
appropriate one should be used. The simplest one is haar transform. To hide text message integer
wavelet transform can be used. When DWT is applied to an image it is decomposed into 4 sub
bands: LL, HL, LH and HH. 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. But
sometimes distortion may be produced in the stego image and then other sub bands can be used
[1]. The decomposition of Lena image by 2 levels of 2D - DWT is shown in Figure 1.
Figure 1. 2 Level 2D – DWT
1.2. Integer Wavelet Transform
IWT is a more efficient approach to lossless compression. The coefficients in this transform are
represented by finite precision numbers which allows for lossless encoding. This wavelet
transform maps integers to integers. In case of DWT, if the input consists of integers (as in the
case of images), the resulting output no longer consists of integers. Thus the perfect

19
reconstruction of the original image becomes difficult. However, with the introduction of Wavelet
transforms that map integers to integers the output can be completely characterized with integers.
The LL sub-band in the case of IWT appears to be a close copy with smaller scale of the original
image while in the case of DWT the resulting LL sub-band is distorted slightly, as shown in
Figure 2.[2].
Figure 2 (a) Original image Lena. (b) One level DWT in sub band LL (c) One level IWT in sub-band LL.
If the original image (I) is X pixels high and Y pixels wide, the level of each of the pixel at (i,j) is
denoted by Ii,j.[3]
The IWT coefficients are given by
LLi,j = ( I2i, 2j + I 2i+1, 2j) /2 (1)
HLi,j = I2i+1, 2j - I2i, 2j (2)
LHi,j = I2i, 2j+1 - I2i, 2j (3)
HHi,j = I2i+1, 2j +1 - I2i, 2j (4)
The inverse transform is given by
I2i, 2j = LLi,j - HLi,j/2 (5)
I2i, 2j +1 = LLi,j + (HLi,j+1)/2 (6)
I2i+1, 2j = I2i, 2j +1 + LHi,j - HLi,j (7)
I2i+1, 2j+1 = I2i+1,2j + HHi,j - LHi,j (8)
where, 1 ≤ i ≤ X/2 , 1 ≤ j ≤ Y/2 and denotes floor value.
2. RELATED WORK
Color images are represented in different color spaces such as RGB (Red Green Blue), HSV
(Hue, Saturation, Value), YUV, YIQ, YCbCr (Luminance/Chrominance) etc. YCbCr is one of the
best representations for steganography because the eye is sensitive to small changes in luminance
but not in chrominance, so the chrominance part can be altered, without visually impairing the
overall image quality much. Y is luminance component and CbCr are the blue and red
chrominance components respectively. The values in one color space can be easily converted into
another color space using conversion formula [4].
S. M. Masud Karim, et al., [5] proposed a new approach based on LSB using secret key. The
secret key encrypts the hidden information and then it is stored into different position of LSB of
image. This provides very good security. XIE Qing et al.,[6] proposed a method in which the
information is hidden in all RGB planes based on HVS (Human Visual System). This degrades
the quality of the stego image. In the method proposed by Sunny Sachdeva et al., [7] the Vector
Quantization (VQ) table is used to hide the secret message which increases the capacity and also
stego size. Sankar Roy et al., [8] proposed an improved steganography approach for hiding text

20
messages within lossless RGB images which will suffer from withstanding the signal processing
operations. Minimum deviation of fidelity based data embedding technique has been proposed by
J. K. Mandal et al, [9] where two bits per byte have been replaced by choosing the position
randomly between LSB and up to fourth bit towards MSB. A DWT based frequency domain
steganographic technique, termed as WTSIC is also proposed by the same authors, [10] where
secret message/image bits stream are embedded in horizontal, vertical and diagonal components.
Anjali Sejul, et al, [4] proposed an algorithm in which binary images are considered to be secret
images which are embedded inside the cover image by taking the HSV (Hue, Saturation, Value)
values of the cover image into consideration. The secret image is inserted into the cover image by
cropping the cover image according to the skin tone detection and then applying the DWT. In this
method the capacity is too low.
El Safy et.al, [11], used an adaptive steganographic technique based on IWT, which improves the
hiding capacity and PSNR. Neda Raftari and Amir Masoud E. M. [12] used IWT and Munkres'
assignment algorithm which embeds secret image in frequency domain of cover image with high
matching quality. The improvement is obtained with higher computation. Saddaf Rubab et
al.,[13] proposed a complex method using DWT and Blowfish encryption technique to hide text
message in color image. In the paper by Kapre Bhagyashri et al, [14] a new singular value
decomposition (SVD) and DWT based water mark technique is proposed in full frequency band
in YUV color space. Nabin Ghoshal et al., uses a steganographic scheme for colour image
authentication (SSCIA) [15] where the watermark image is embedded using DFT.
The proposed work is the extension of our previous work [16] to color images in which the secret
image is transmitted without actually embedding in the cover image. Only the key is hidden in the
cover image. The steps for embedding are as follows:
• Obtain single level 2D DWT of the cover-image C and secret-image S.
• The resulting transformed matrix consists of four sub-bands CLL, CHL, CLH and CHH
and SLL, SHL, SLH and SHH obtained by transforming images C and S respectively.
• The sub-images CLL and SLL are subdivided into non-overlapping blocks BCk1 (1 ≤ k1
< nc) and BSi (1 ≤ i < ns) of size 2x2 where nc, ns are the total number of non-
overlapping blocks obtained from sub-images CLL and SLL respectively.
• Every block BSi, is compared with block BCk1. The pair of blocks which have the least
Root Mean Square Error is determined. A key is used to determine the address of the best
matched block BCk1 for the block BSi. Then inverse 2D DWT is applied to obtain C.
• The Key is then stored using one of the spatial domain techniques in the cover image C.
The simplest of the spatial domain techniques is LSB insertion algorithm.
• The resultant image is a stego-image.
•
The secret image can now be extracted from this image by following the steps mentioned below:
• From the stego-image G, obtain the secret key K1
• Transform the stego-image into single level 2D DWT.
• This transformation results in four sub-bands GLL, GHL, GLH and GHH.
• Divide the sub-band image GLL into 2x2 non-overlapping blocks. The secret key K1 is
used to obtain the blocks that have the nearest approximation to the original blocks in
secret image.
• The obtained blocks are then rearranged to obtain the sub-band image SLLnew.
Assuming SHLnew, SLHnew, SHHnew are zero matrices of dimension similar to
SLLnew , 2D IDWT (Inverse DWT) is obtained.
• The resultant image is the secret image that was originally intended to be embedded
within the cover-image.

21
3. PROPOSED METHOD
In the proposed method, the cover is 256x256 lena color image. The secret information is grey
scale image of size 128 x128. To transfer the secret image confidentially, the secret image itself is
not hidden, instead a key is generated and then the key is encrypted and Run Length Encoded.
The resultant key is hidden in the cover image using Integer Wavelet Transform (IWT). This
improves the security and also the capacity can be improved to some extent since the key is
compressed.
3.1. Key Generation
To generate the key following steps are performed.
• Represent the cover image C in YCbCr color space
• Obtain single level 2D DWT of secret-image S and Cr component of C.
• The resulting transformed matrix consists of four sub-bands SLL, SHL, SLH and SHH
and CLL, CHL, CLH and CHH obtained by transforming S and Cr component of C
respectively. After this the steps are same as our previous work, which are repeated again
for better understanding.
• The sub-images CLL and SLL are subdivided into non-overlapping blocks BCk1 (1 ≤ k1
< nc) and BSi (1 ≤ i < ns) of size 2x2 where nc, ns are the total number of non-
overlapping blocks obtained from sub-images CLL and SLL respectively.
• Every block BSi, is compared with block BCk1. The pair of blocks which have the least
Root Mean Square Error is determined. A key is used to determine the address of the best
matched block BCk1 for the block BSi. Then inverse 2D DWT is applied to obtain Cr
component.
• The key is then encrypted using simple exclusive or operation with a key and run length
encoded.
3.2. Key Embedding
The key obtained in the previous subsection is hidden in the cover image using IWT. The steps
are as follows:
• Find the integer wavelet transform of Cr component of the cover image.
• Replace the least significant bit planes of the higher frequency components of the
transformed image by the bits of the key.
• Obtain the inverse IWT of the resulting image to get the stego Cr component.
• Represent the resultant image in RGB color space to obtain stego image G.
The secret image can now be extracted from this image using the following steps:
3.3. Key Extraction
The steps are as follows:
• Represent the stego image G in YCbCr color space.
• Find the integer wavelet transform of Cr component of the stego image G
• Obtain the key from the least significant bit planes of the higher frequency components of
the transformed image. Convert back to RGB representation.
• Decompress the key and then decrypt it to get original key.

22
3.4. Secret Image Generation
To generate the secret image following steps are performed
• Transform the stego-image G into single level 2D DWT.
• This transformation results in four sub-bands GLL, GHL, GLH and GHH.
• Divide the sub-band image GLL into 2x2 non-overlapping blocks. The key is used to
obtain the blocks that have the nearest approximation to the original blocks in secret
image.
• The obtained blocks are then rearranged to obtain the sub-band image SLLnew.
Assuming SHLnew, SLHnew, SHHnew are zero matrices of dimension similar to
SLLnew, 2D IDWT is obtained.
• The resultant image is the secret image S.
4. EXPERIMENTAL RESULTS
The algorithm is tested in MATLAB with Cr component of the cover image. The results with
different cover images and secret images are shown. Original cover and secret images are shown
in Figure 3 and Figure 4 respectively. Football is hidden in lena, earth is hidden in peppers and
moon is hidden in baboon. The cover image size is 256x256 and secret image size is 128x128.
Figure 3. Color images that are used as cover images: (a) lena (b) baboon (c) peppers
Figure 4. Images which are used as secret images: (a) earth (b) football (c) moon
The stego and extracted secret images are shown in Figure 5 and Figure 6 respectively.
Figure 5. Stego images: (a) football, (b) earth, (c) moon as secret images

23
Figure 6. Extracted Secret images: (a) football (b) earth (c) moon
The PSNR in dB in all cases for stego and extracted secret images are tabulated in Tables 1 and 2
respectively.
Table 1. PSNR (in dB) of the stego image
Table 2. PSNR (in dB) of the extracted secret image
COVER IMAGE (256x256) SECRET IMAGE(128x128)
football earth moon
lena 37.5 34.1 30.7
peppers 30.4 28.6 26.7
baboon 37.2 27.6 36.5
Table 3 compares the PSNR values in the proposed method and that in the other four methods. In
all these the cover image considered is lena image and the secret images used are of comparable
sizes.
Table 3. Comparison of PSNR (in dB) of the stego image in different methods
TECHNIQUE PSNR
Mandal, J.K. et al. [9] 39.6
Mandal, J.K. et al. [10] 42.4
Kapre Bhagyashri, S. et al. [14] 36.6
Ghoshal, N. et al. [15] 33.2
PROPOSED 44.3
5. CONCLUSIONS
In this paper, we observe that the secret image can be regenerated without actually storing the
image itself. This approach results in high quality of the stego-image having high PSNR values
compared to other methods. Instead of taking the least significant bit plane to hide the key middle
bit planes can be considered to improve the security. One assumption made in this paper is that
COVER IMAGE (256x256) SECRET IMAGE(128x128)
football earth moon
lena 44.3 44.4 44.2
peppers 44.7 44.7 45.0
baboon 44.8 44.8 45.0

24
the encryption key is sent to the recepient by some means. Standard encryption techniques like
Blowfish or RC6 can be used to encrypt the key so that the security can be further improved. The
encryption key also has to be hidden in the cover image.
REFERENCES
[1] Katzenbeisser, S. and Petitcolas, F.A.P., (2000) Information Hiding Techniques for Steganography
and Digital Watermarking. Artech House, Inc., Boston, London.
[2] M. F. Tolba, M. A. Ghonemy, I. A. Taha, A. S. Khalifa, (2004) "Using Integer Wavelet Transforms in
Colored Image-Stegnography", International Journal on Intelligent Cooperative Information Systems,
Volume 4, pp. 75-85.
[3] Guorong Xuan et. al, (2002 ) “Distortionless Data Hiding Based on Integer Wavelet Transform”,
Electronics Lelters, Vol. 38, No. 25, pp. 1646-1648.
[4] Shejul, A. A., Kulkarni, U.L., (2011) “A Secure Skin Tone based Steganography (SSTS) using
Wavelet Transform”, International Journal of Computer Theory and Engineering, Vol.3, No.1, pp. 16-
22.
[5] Masud, Karim S.M., Rahman, M.S., Hossain, M.I., (2011) “A New Approach for LSB Based Image
Steganography using Secret Key.”, Proceedings of 14th International Conference on Computer and
Information Technology, IEEE Conference Publications, pp 286 – 291.
[6] Xie, Qing., Xie, Jianquan., Xiao, Yunhua., (2010) “A High Capacity Information Hiding Algorithm in
Color Image.”, Proceedings of 2nd International Conference on E-Business and Information System
Security, IEEE Conference Publications, pp 1-4.
[7] Sachdeva, S and Kumar, A., (2012) “Colour Image Steganography Based on Modified Quantization
Table.”, Proceedings of Second International Conference on Advanced Computing & Communication
Technologies , IEEE Conference Publications, pp 309 – 313.
[8] Roy, S., Parekh, R., (2011) “A Secure Keyless Image Steganography Approach for Lossless RGB
Images.”, Proceedings of International Conference on Communication, Computing & Security, ACM
Publications, 573-576.
[9] Mandal, J.K., Sengupta, M., (2011) “Steganographic Technique Based on Minimum Deviation of
Fidelity (STMDF).”, Proceedings of Second International Conference on Emerging Applications of
Information Technology, IEEE Conference Publications, pp 298 – 301.
[10] Mandal, J.K., Sengupta, M., (2010) “Authentication/Secret Message Transformation Through
Wavelet Transform based Subband Image Coding (WTSIC).”, Proceedings of International
Symposium on Electronic System Design, IEEE Conference Publications, pp 225 – 229.
[11] El Safy, R.O, Zayed. H. H, El Dessouki. A., (2009) “An Adaptive Steganographic Technique Based
on Integer Wavelet Transform”, IEEE conference publications, pp 111-117.
[12] Neda Raftari and Amir Masoud Eftekhari Moghadam, (2012), "Digital Image Steganography Based
on Integer Wavelet Transform and Assignment Algorithm", Sixth Asia Modeling Symposium, pp. 87-
92.
[13] Rubab, S., Younus, M., (2012) “Improved Image Steganography Technique for Colored Images using
Wavelet Transform.”, International Journal of Computer Applications, Volume 39– No.14, pp 29-32.
[14] Kapre Bhagyashri, S., Joshi, M.Y., (2011) “All Frequency Band DWT-SVD Robust Watermarking
Technique for Color Images in YUV Color Space.”, Proceedings of IEEE International Conference
on Computer Science and Automation Engineering, IEEE Conference Publications, pp 295 – 299.
[15] Ghoshal, N., Mandal, J.K.., (2011) “ A Steganographic Scheme for Colour Image Authentication
(SSCIA).”, Proceedings of International Conference on Recent Trends in Information Technology,
IEEE Conference Publications, pp 826 – 831.
[16] Kamath, Priya R., Hemalatha, S., Acharya, Dinesh U., Renuka, A., (2012) “High quality technique for
image steganography using discrete wavelet transform.”, Proceedings of National Conference on
Advances in Computer Science and Information Security, pp 89-92.

A SECURE COLOR IMAGE STEGANOGRAPHY IN TRANSFORM DOMAIN

More Related Content

Similar to A SECURE COLOR IMAGE STEGANOGRAPHY IN TRANSFORM DOMAIN (20)

Recently uploaded (20)

A SECURE COLOR IMAGE STEGANOGRAPHY IN TRANSFORM DOMAIN