International Journal of Science and Research (IJSR)

ISSN (Online): 2319-7064

Index Copernicus Value (2015): 78.96 | Impact Factor (2015): 6.391

Reversible Data Hiding on Encrypted Digital

Images
Kranti U. Patil1, J. V. Shinde2
1
Late G. N. Sapkal College of Engineering, Anjaneri, Nasik- 422212
2
Professor, Late G. N. Sapkal College of Engineering, Anjaneri, Nasik- 42221

Abstract: Recently, Reversible data hiding is a point of attention. Reversible data hiding in encrypted images is mostly used since it
maintains excellent property that the original cover image can be recovered losslessly after the embedded data is extracted. By using this
technique the confidentiality of image contents is also maintained. This paper proposes a novel reversible data hiding technique for
digital color image. This technique guarantees effective retrieval of hidden data in color image without degradation in the quality of
color image. The cryptographic methods are used to encrypt & decrypt the data to increase the security. The proposed method is
enhanced from text data to any data type like text, video, audio etc. In this system only digital image is considered as cover for data
hiding purpose.

Keywords: Reversible data hiding, Steganography, Cryptography, Encryption, Decryption

1. Introduction embedding the data. In the second group, frequency

coefficients of image are used for embedding purpose. In
In recent years, the technologies are developed so much that RDH, the lossless recovery of cover after the data extraction
most of the media’s use internet as the primary medium to is an important aspect. The application of cryptography with
transfer the data from one end to other end across the world. steganography can increase the capacity of information. In
Now-a-days, the data transmission is very simple and fast this paper, reversible data hiding on digital cover image is
using the internet. However, security threats are one of the proposed. In this paper data files are embedded after first 8
main problems with data transition over the internet. On bits of image. Here multiple data files are hidden in the cover
internet the personal data can be stolen or misused in many image. There are no limitations on embedded data size. With
ways. Therefore, data security is one of the most essential respect to cover image we can embed larger data files into
factors taken into consideration during the process of data that. The data used for embedding can be of any type like
transferring. image, audio, video etc. There is no restriction of data type.

Reversible data hiding (RDH) in images is a recently used 2. Literature Survey

technique, by which the original cover can be recovered
losslessly after the embedded data is extracted. Here the In W. Zhang, B. Chen, and N. Yu [2] method, a
image used for hiding the data is called cover and the image decompression algorithm for embedding the data is used.
generated after embedding data is called stego image. This They proved that using this construction they can achieve the
technique is mostly used in military, medical and law rate-distortion bound as long as the compression algorithm
forensics, where no distortion of the original cover is reaches the entropy. In this they have improved three RDH
allowed. schemes that are using binary features sequence as covers.
Using this system, embedding distortion can be reduced. It
In current trends, information security domain is a point of also improves reversible data hiding schemes for binary
interest. To overcome this, a technique of cryptography was JPEG images. This system did not work on gray scale covers
developed for keeping the communication secret. Many for designing recursive codes.
different methods for encryption and decryption of data have
been developed for keeping the data secret. Unfortunately, J. Fridrich, M. Goljan, and D. Rui’s [3] system, a general
keeping only the contents secret could not work. There is a framework for RDH is proposed. Extracted compressible
requirement of keeping the existence of message secret. The features of cover images are firstly introduced by them. In
technique developed for this purpose is called this system, they have reserved a space to hide data by
steganography. Steganography is an information hiding compressing the proper bit-planes having minimum
technique. Using steganography we can embed secret data redundancy. The lowest bit-planes which offer lossless
into digital cover media like images, video, audio, etc. file compression are used if the image is not noisy. In completely
formats. Steganography has several characteristics like noisy image some bit-planes are having strong correlations.
storage capacity, invisibility and resistance against attack. These bit-planes are used to vacate room space to store hash.
Due to these characteristics it is an efficient data hiding This system provides high capacity and security levels and
technology. can be used for authentication purpose of JPEG, audio file,
digitized holograms etc. But this system forces noisy images
Data hiding techniques generally falls into two groups i.e. to embed information in the higher bit-planes. Small images
spatial and frequency domain. In the first group, least having single bit-plane cannot offer enough space to hide
significant bits (LSBs) of image pixels are used for

Volume 5 Issue 12, December 2016

www.ijsr.net
Licensed Under Creative Commons Attribution CC BY
Paper ID: ART20163711 1271
 International Journal of Science and Research (IJSR)
ISSN (Online): 2319-7064
Index Copernicus Value (2015): 78.96 | Impact Factor (2015): 6.391
hash. This system has not enough capacity to embed large G. Xuan, J. Chen, J. Zhu, Y.Q. Shi, Z. Ni, and W. Su have
payload. proposed a integer wavelet transform based lossless data
hiding technique. This system hides the authentication
J. Tian has proposed a system which uses difference information. For preventing gray scale overflowing during
expansion method for embedding data. This system uses the data embedding, the histogram modification or integer
features which are compressed by expansion i.e. the modulo addition techniques are used. This method uses
differences between two neighboring pixels. Some second-generation wavelet transform IWT. The information is
differences are selected for expansion by one bit i.e. the hided into middle bit-plane and in the high frequency sub-
difference is multiplied by 2. Thus, LSB’s of the differences bands respectively. This makes the watermarked image
are all zero and this LSB’s can be used for embedding greatly as same as the original image. Also the PSNR value is
messages. The advantages of the system are: 1. Use of increased. The advantages of the system are: 1. High
compression and decompression causes no loss of data. 2. embedding capacity. 2. Security level is raised due to the use
This system is also applicable to audio and video data. 3. The of secret key during embedding of data. The disadvantages of
compressed location map and changeable bit streams of the system are: 1. only gray scale mapping is done. 2. Often
different numbers are encrypted which increases the security. multiple bit planes are needed to have enough space [8].
The disadvantages of the system are: 1. as there is division
by 2 there may be some round off errors. 2. Depends largely In this paper, V. Sachnev, H. J. Kim, J. Nam, S. Suresh, and
on the smoothness of natural image that’s why can’t be Y.-Q. Shi have proposed a system which gives reversible or
applied to images who’s capacity is zero or very low. 3. lossless watermarking for image without using a location
Degradation of visual quality due to bit replacements [4]. map. In this system data embedding depends on the
prediction errors. Prediction errors based on magnitude of its
Z. Ni, Y. Shi, N. Ansari, and S. Wei have proposed a system local variance can be recorded using sorting technique. Using
which uses histogram shift strategy for RDH. In this system, the sorted prediction errors and reduced size location map
the space is saved for embedding the data by shifting the bins whenever needed improves the data embedding capacity by
of histogram of gray values. The authors make use of zero decreasing the distortion. The histogram shift significantly
point and a peak point of given image histogram to embed reduces the size of location map. The double embedding
messages. In this system, the embedding capacity is the scheme in this system allows using each pixel for hiding data.
number of pixels with peak point. For embedding, the whole The advantages of the system are: 1. Capacity can be
image is searched for peak point. The advantages of the significantly increased. 2. Double embedding scheme is used.
system are: 1. Simple to implement. 2. Constant PSNR of 3. Less distortion [9].
48.0dB is obtained. 3. Distortions are quite invisible. 4.
Capacity is high. The disadvantages of the system are: 1. M. Johnson, P. Ishwar, V. M. Prabhakaran, D. Schonberg,
Capacity is limited by the frequency of peak-pixel values in and K. Ramchandran have tried to first encrypt the data and
histogram. 2. Time consuming as image is searched several then compressing it, so that the compressor don’t know the
times [5]. encryption key. The encrypted data is compressed using
distributed source coding principles. The key will be available
X. L. Li, B. Yang, and T. Y. Zeng have used a hybrid only to the decoder. They have shown that encrypted data can
algorithm which makes the combination of three techniques be compressed to same rate as that of original unencrypted
of PEE (i.e. Prediction-error Expansion), adaptive embedding data could have been. The perfect secrecy and original cover
and pixel selection. In the proposed system, depending on the recovery is obtain in this system [10].
threshold values the image pixels is divided into two parts.
Then the pixels are selected depending on their capacity- In the above methods, the data hiding is done but the size of
parameter and threshold. The smooth pixels are selected from the data hided is limited. Also, some of the systems have
two parts. Finally, data is embedded by modifying the considered gray-scale image as the cover image. So, we are
histograms that are derived from selected pixels. The trying to overcome these problems through our system.
advantages of the system are: 1. By decreasing the
modifications to pixel values, the system reduces the 3.Existing System
embedding impact. 2. More sharply distributed prediction-
error histogram can be obtained. 3. The visual quality of In the existing system, histogram shifting technique is used.
watermarked image is greatly improved [6]. The image is firstly divided into two planes A and B using
image partition technique. Then the LSBs of A are reversibly
L. Luo et al. have used an interpolation technique for embedded into B using standard RDH technique. The
developing their reversible image watermarking system. This estimating error is calculated so that data can be embedded
system can embed a large amount of converted data into into estimating error sequence using histogram shift
images with imperceptible modification. The interpolation technique. By using bidirectional histogram shift, some
errors which are residuals of this technique have greater messages can be embedded on each error sequence. The
decorrelation ability. The highly efficient reversible rearranged image is then encrypted using stream cipher to
watermarking scheme is developed by applying additive generate its encrypted version. The information for data hider
expansion to these interpolation-errors. The advantages of the is embedded into LSBs of first 10 pixels in encrypted image.
system are: 1. High image quality. 2. Greater embedding After image encryption, the data hider or any third person
capacity. 3. Less Computational cost [7]. cannot access the content of the original image without using
encryption key.

Volume 5 Issue 12, December 2016

www.ijsr.net
Licensed Under Creative Commons Attribution CC BY
Paper ID: ART20163711 1272
 International Journal of Science and Research (IJSR)
ISSN (Online): 2319-7064
Index Copernicus Value (2015): 78.96 | Impact Factor (2015): 6.391
Once the image is passed on to data hider, he can embed data digital image used for embedding the data files is called the
into image without getting access to original image. The data cover image and the image generated after embedding
hider can easily read the information in the LSBs of first 10 process is called stego-image. The following diagram shows
encrypted pixels. After getting the information about how the data hiding process:
many bit-planes and rows of pixels he can modify, the data
hider hides additional data into available bit-planes using
LSB substitution method. At the end of data embedding
process, data hider encrypts the additional data using data
hiding key to generate marked encrypted image. After that, no
one except data hiding key holder can extract the data.

Next, the marked encrypted image is passed on to database

manager. The database manager can extract data from either
encrypted or decrypted image.

Case 1: Data extraction from Encrypted image

In this case, the data manager can decrypt the LSB planes and
extract the data by directly reading the encrypted version
using data hiding key. Here both data embedding and data Figure 4.1: Reversible Data Hiding process
retrieval are done on encrypted image so the privacy of
original contents is maintained. 1) Platform: Microsoft Visual Studio 2010
Microsoft provides an integrated development environment
Case 2: Data extraction from Decrypted image called Microsoft visual studio. Microsoft visual studio is
In this case, the data extraction and image decryption is a used for developing the graphical user interface, web
sequential process. The database manager firstly decrypts the application, web sites etc. Visual studio also has code editor
image using encryption key then after he can retrieve the data which also provides syntax highlighting.
using data hiding key.
2) Proposed System
Disadvantages of Existing system: The proposed system architecture is as shown in the figure
a) It works only on text messages. below; it consists of image encryption, data embedding,
b)There is limitation on data size to be embedded on the image decryption, data extraction and image recovery block.
cover image. The cover image here is the original image used for hiding
c) The cover image considered here is gray-scale only. data. This system uses two keys one for encryption and one
d)Multiple data files cannot be embedded using this system. for data hiding. The encryption key is required for decryption
process, if the key doesn’t matched with encryption key
4. Implementation Details image decryption is not possible. Same for data hiding key,
data retrieval is possible only with data hiding key.
In this paper, we propose a reversible data hiding for digital
image. Here, we are going to propose the system for 3) Advantages
embedding multiple data files in the digital cover media. a) The cover used in this system is color image.
b) The data to be hidden on image can be of any format.
A reversible data hiding is a technique which gives lossless or c) Multiple data files can be added on cover image.
distortion-free data hiding. Using this technique, not only the d) No size limitations for data files to be embedded.
security of data is obtained but also the exact data and
original image after extraction is obtained. In this system the

Figure 4.2: Propose system architecture

Volume 5 Issue 12, December 2016

www.ijsr.net
Licensed Under Creative Commons Attribution CC BY
Paper ID: ART20163711 1273
 International Journal of Science and Research (IJSR)
ISSN (Online): 2319-7064
Index Copernicus Value (2015): 78.96 | Impact Factor (2015): 6.391
Modules Description 7. Now, concatenate input image byte array byInputImage
with zip file byte array byZipFiles.
1. Encrypted Image Generation: byInputImage.CopyTo(byOutput, 0);
byZipFiles.CopyTo(byOutput,byInputImage.Length)
Input: In this the image passed by first module, encryption 8. Save the bytes array byoutput on the image file.
key are the inputs.
Output: This module gives the data containing encrypted
Algorithm: Advanced Encryption standard Algorithm image as the output.
(Rijndael Algorithm)
AES is a symmetric block cipher algorithm. It uses same key 3. Data Extraction and Image Recovery:
for encryption and decryption process. The stages of AES
algorithm are as follows: Input: In this the data embedded image, data hiding key and
1. Substitute bytes encryption key is taken as input.
2. Shift rows
3. Mix Columns Algorithm: - To Extract data files and image
4. Add Round Key
The steps for data extraction and image recovery are as
Output: This module outputs the marked encrypted image. follows:
1. Separate the byte array byInputImage and byZipfiles from
2. Data Hiding In Encrypted Image: byoutput byte array with respect to byInputImage array
length.
Input: In this the marked encrypted image, the data hiding 2. Save the first part of byte array as bitmap image.
key, the data files to be embedded are taken as inputs. 3. Save the second part as zip files array.
4. Extract the zip files array to get the attached files.
Algorithm- The steps of algorithm are as follows:
1. Transform the image into a temp jpeg first. Algorithm- A.E.S Decryption Algorithm
2. Read all the bytes from input image and store in byte array
byInputImage. The stages of DES are same as that of AES algorithm with
byInputImage-> File.ReadAllBytes(InputImage) the exception that each step is inverse of its counterpart in the
encryption algorithm.
3. Calculate byte array length of byInputImage. 1. Inverse Shift rows
4. Copy all the selected files into a single folder. 2. Inverse Substitute bytes
5. Compress all the files in a zip file. 3. Inverse Add Round Key
ZipFile zpFiles = ZipUnzip.ZipFiles(arrFiles) 4. Inverse Mix Columns

6. Convert the zip file to byte array byZipFiles. Output: this module gives the original data files and image as
byte[] byZipFiles = smt.ToArray() output.

Table 5.1: Comparison of PSNR values in marked decrypted grey-scale and marked decrypted RGB images
Embedding 0.005 0.01 0.05 0.1
Rate(bpp) ES gray PS RGB ES gray PS RGB ES gray PS RGB ES gray PS RGB
Lena 65.151 67.80 62.1 65.54 55.28 57.31 51.809 54.54
Baboon 65.10 65.75 62.083 64.89 55.219 60.17 51.825 56.20
Peppers 65.120 65.87 62.128 63.80 55.214 60.79 51.831 56.16
Embedding 0.2 0.3 0.4 0.5
Rate(bpp) ES gray PS RGB ES gray PS RGB ES gray PS RGB ES gray PS RGB
Lena 49.31 51.72 47.42 50.16 46.373 49.51 46.36 48.76
Baboon 49.209 53.16 47.402 51.71 46.363 50.84 46.365 48.55
Peppers 49.205 52.404 47.404 50.36 46.36 48.28 46.37 46.87

5. Performance Evaluation and Results analysis PSNR values of previous method in marked
decrypted grey-scale images with PNSR values of our
5.1. Data Used proposed method in corresponding marked decrypted color
images. Table.5.1 shows comparisons result under various
In this paper digital color image of JPEG format and other embedding rates. Fig.5.1 plots the PSNR results of marked
formats like bmp, png are used as cover image. The data files decrypted image of Lena under given embedding rates. From
used for embedding are of txt, docx, pdf etc. format. The size the Fig.5.2, it can be observed that overall range of
of data is independent of the cover size. embedding rate, for all cases, our approach outperforms
state-of- the-art RDH algorithms in encrypted color images.
5.2. Results The original image and stego-image are as below:

We take standard images Lena, Baboon and Peppers to

demonstrate the feasibility of proposed method. Here we

Volume 5 Issue 12, December 2016

www.ijsr.net
Licensed Under Creative Commons Attribution CC BY
Paper ID: ART20163711 1274
 International Journal of Science and Research (IJSR)
ISSN (Online): 2319-7064
Index Copernicus Value (2015): 78.96 | Impact Factor (2015): 6.391
References
[1] Kede Ma, Weiming Zhang, Xianfeng Zhao, Nenghai
Yu, and Fenghua Li, “Reversible data hiding in
encrypted images by reserving room before encryption,”
IEEE Transactions on Information Forensics And
Security, vol. 8, no. 3, pp. 553-562, March 2013
[2] W. Zhang, B. Chen, and N. Yu, “Improving various
reversible data hiding schemes via optimal codes for
Figure 5.1: Lena image a. original image b. recovered image binary covers,” IEEE Trans.Image Process., vol. 21, no.
after data extraction 6, pp. 2991–3003, June. 2012.
[3] J. Fridrich, M. Goljan, and D. Rui, “'Invertible
Authentication”, In Proc. of SPIE Photonics West,
Security and Watermarking of Multimedia Contents III,
San Jose, California, USA, Vol. 3971, pp. 197-208,
January 2001.
[4] J. Tian, “Reversible data embedding using a difference
expansion,”IEEE Trans. Circuits Syst. Video Technol.,
vol. 13, no. 8, pp. 890–896,August. 2003.
[5] . Ni, Y. Shi, N. Ansari, and S. Wei, “Reversible data
hiding,” IEEE Trans. Circuits Syst. Video Technol., vol.
16, no. 3, pp. 354–362, March.2006.
[6] X. L. Li, B. Yang, and T. Y. Zeng, “Efficient reversible
watermarking based on adaptive prediction-error
expansion and pixel selection,” IEEE Trans. Image
Process., vol. 20, no. 12, pp. 3524–3533,
December.2011.
Figure 5.2: Graphical representation for PSNR comparison in
[7] L. Luo et al., “Reversible image watermarking using
test images
interpolation technique,” IEEE Trans. Inf. Forensics
Security, vol. 5, no. 1, pp. 187–193,March. 2010.
6. Conclusion and Future Scope [8] G. Xuan, J. Chen, J. Zhu, Y.Q. Shi, Z. Ni, and W. Su,
“Lossless Data hiding based on Integer Wavelet
In this paper, we proposed a reversible data hiding technique Transform”, In Proc. of IEEE International Workshop
to embed the secret data in color image with high security, on Multimedia Signal Processing. Marriott Beach Resort
imperceptibility and robustness. Due to proposed embedding St. Thomas, US Virgin Islands, 9-11 December 2002.
algorithm the data embedding capacity will be significantly [9] V. Sachnev, H. J. Kim, J. Nam, S. Suresh, and Y.-Q.
increased in this system. Multiple data files will be Shi, “Reversible watermarking algorithm using sorting
embedded in cover image. The security of the data is much and prediction,” IEEE Trans. Circuits Syst. Video
more increased due to the use of two keys: encryption key Technol., vol. 19, no. 7, pp. 989–999, July. 2009.
and data hiding key. No one can recognize the stego image [10] M. Johnson, P. Ishwar, V. M. Prabhakaran, D.
as the appearance looks same as cover image. All the data Schonberg, and K. Ramchandran, “On compressing
files are compressed and then embedded with cover, so more encrypted data,” IEEE Trans. Signal Process., vol. 52,
data can be embedded. no. 10, pp. 2992–3006, October. 2004.
[11] K. Hwang and D. Li, “Trusted cloud computing with
In future, we will extend this system considering audio or secure resources and data coloring,” IEEE Internet
video files as the cover. In this paper only digital image is Comput., vol. 14, no.5, pp.14–22, Sep./October.2010.
considered as cover. [12] X. Zhang, “Separable reversible data hiding in
encrypted image,” IEEE Trans. Inf. Forensics Security,
7. Acknowledgment vol. 7, no. 2, pp. 826–832, April. 2012.
[13] Y. Chakrapani and K. SoundaraRajan,“ Genetic
First and foremost, I would like to thank my guide, Prof. Ms. algorithm applied to fractal image compression,” in
J. V. Shinde, for his guidance and support. I will forever ARPN Journal of Engineering and Applied Sciences ,
remain grateful for the constant support and guidance vol. 4, no. 1, February.2009
extended by guide, in making this report. Through our many [14] MedisettyNagendra Kumar and S. Srividya,“Genetic
discussions, he helped me to form and solidify ideas. The Algorithm Based Color Image Steganogaphy Using
invaluable discussions I had with him, the penetrating Integer Wavelet Transform And Optimal Pixel
questions he has put to me and the constant motivation, has Adjustment Process,” in International Journal of
all led to the development of this project. Innovative Technology and Exploring Engineering
(IJITEE) ISSN: 2278-3075, Volume-3, Issue-5, October
I wish to express my sincere thanks to the Head of 2013.
department, Prof. N. R. Wankhede, also grateful thanks to [15] X. Zhang, “Reversible data hiding in encrypted
the M.E Co-coordinator Prof. Ms. J. V. Shinde and the images,” IEEE Signal Process. Lett., vol. 18, no. 4, pp.
departmental staff members for their support. 255–258, April. 2011.
Volume 5 Issue 12, December 2016
www.ijsr.net
Licensed Under Creative Commons Attribution CC BY
Paper ID: ART20163711 1275
 International Journal of Science and Research (IJSR)
ISSN (Online): 2319-7064
Index Copernicus Value (2015): 78.96 | Impact Factor (2015): 6.391
[16] W. Hong, T. Chen, and H.Wu, “An improved reversible
data hiding in encrypted images using side match,”
IEEE Signal Process. Lett., vol. 19, no. 4, pp. 199–202,
April. 2012.

Volume 5 Issue 12, December 2016

www.ijsr.net
Licensed Under Creative Commons Attribution CC BY
Paper ID: ART20163711 1276