Adaptive circular queue image steganography with RSA ... · circular queue image steganography with RSA cryptosystem ... Cryptography and steganography both are used ... Adaptive

Page 1

Perspectives in Science (2016) 8, 417—420

Available online at www.sciencedirect.com

ScienceDirect

jo ur nal homepage: www.elsev ier .com/pisc

Adaptive circular queue imagesteganography with RSA cryptosystem�

Mamta Jaina,∗, Saroj Kumar Lenkab, Sunil Kumar Vasisthaa

a Department of Computer Science and Engineering, Mody University of Science and Technology,Lakshmangarh, Rajasthan, Indiab Department of Information Technology, Mody University of Science and Technology, Lakshmangarh,Rajasthan, India

Received 27 January 2016; accepted 9 April 2016Available online 3 May 2016

KEYWORDSCircular queue;Steganography;Cryptography;RSA cryptosystem

Summary The major objective of the article is to supply the novel and efficient methodologyof digital image steganography that describes individuality regarding secret transmission usingthe adaptive circular queue least significant bits (LSBs) substitution. The data structure queue isemployed dynamically in resource distribution between multiple communication recipients andonce secret information transmitted asynchronously. Here, RSA cryptosystem is employed forsecret information confidentiality and authentication. The result of the cryptosystem organisedinto various blocks. In steganography method, organise the cover image into various circularqueues blocks. Dynamically adapted procedure is employed to assign secret cypher blocks to cir-cular queues for embedding. Authorised receiver will determine the right plain text using private

key in RSA decypherment. Performance analysis is evaluated by using MSE, PSNR and maximumembedding capacity. Results are higher as compared with several of existing algorithms of imagesteganography.© 2016 Published by Elsevier GmbH. This is an open access article under the CC BY-NC-ND license

rg/li

(http://creativecommons.o

� This article belongs to the special issue on Engineering and Mate-rial Sciences.

∗ Corresponding author. Tel.: +91 9460981894.E-mail addresses: [email protected] (M. Jain),

[email protected] (S.K. Lenka),[email protected] (S.K. Vasistha).

I

Iaataatat

http://dx.doi.org/10.1016/j.pisc.2016.04.0932213-0209/© 2016 Published by Elsevier GmbH. This is an open access artlicenses/by-nc-nd/4.0/).

censes/by-nc-nd/4.0/).

ntroduction

nformation and communication technology play a vitalppearance now days for implication of lots of IT deviceslong with security concerns. Sustaining secrecy of elec-ronic data over the World Wide Web in the era of internet,nd cloud computing is very crucial aspect. Cryptography

nd steganography both are used together to accomplishhe security challenges to the transmitted data over cloudnd mobile networking. Cryptography provides confiden-iality to the secret data at end to end communication

icle under the CC BY-NC-ND license (http://creativecommons.org/

Page 2

418

ass(cippkcqboq

efistroevaasatafdbtCiaiqspdi

P

F

Ttaus

P

•

•

•

•

•

Figure 1 Flow diagram of work methodology.

nd data centre for storage. Steganography is a method ofecret communication over transmission channel in whichecret information is embedded into various multimedia filesProvos and Honeyman, 2003). In steganography, an intruderannot leak or suspect the secret data that some secretnformation is passing through transmission channel. Theroposed method provides more than one level of securityaradigm. Formerly, the data is encrypted using RSA publicey cryptography algorithm and later, the encrypted data isoncealed into the LSBs of cover image using the circularueue substitutions, thus the strength of steganography cane increased with cryptography. Hiding the data into LSBsf cover image does not much affect its visual appearanceuality.

There are numerous procedures used to hide a vari-ty of multimedia secrets inside distinguish multimediales. Anderson and Petitcolas discussed some limitations integanography methods. They approached an informationheoretic method using Shannon’s theory for perfect secu-ity of data (Ross et al., 1998). In the another LSB procedure,ne bit of secret data is substituted at the 8th position ofvery byte of the coated file, if the entropy and correlationalues of stego image and the cover image show the equalityfter encyphering then it represents process is safe (Younesnd Jantan, 2008). Swain and Lenka (2015) proposed a newteganography technique based on LSB array. One of the fourrrays is obtained and formed on the basis of the length ofhe secret message. The different words of the secret datare mapped on the chosen array, where maximum matchound, there obscured the data and start indices are notedown. In the other method, author proposed a two-waylock matching procedure and the hop embedding schemeo hide a secret image data inside a cover image (Wang andhen, 2006). Nag et al. (2011) suggested a novel method

n a steganography system based on the affine encryptionlgorithm and embeds the secret data at the LSB positionn order to advise a solid security and imperceptible ocularuality to secret data. Our novel approach can be under-

tood by referring the following divisions. In division 2, theroblem formulation and work methodology is suggested, inivision 3, results and analysis are done, and finally the works concluded.

M. Jain et al.

roblem formulation and work methodology

low diagram of work methodology

he problem statement consists of circular queue LSB inser-ion and RSA algorithm to create a secure crypto-steganolgorithm, which is far more secure than many systems beingsed for the purpose of secretly sending the data. Fig. 1hows the work flow of this algorithm.

roposed algorithm

In our novel opted system, first of all we select a greyimage as a cover image and a secret message which willbe embedded in the cover image.

RSA encryption technique is used to encrypt the secretdata before embedding.

After RSA encryption, cypher text will be obtained. Nowcypher text will be divided in number of blocks and eachblock has 7 bytes instead of 8 bytes, since in full circularqueue one slot is empty hence one pixel will not used forembedding, it is a property of full circular queue. Nowcover image will be divided except some reserved locationi.e. to byte numbers 3045—4045 into 16 image blocks andeach image block has equal number of bytes or pixels.Now organise the 1st byte of all 16 image block in onecircular queue, 2nd byte of all 16 image block in secondcircular queue and so on. Starting index in circular queuefor embedding may be any pixel.

After that, dynamic adaptive procedure is used fordynamic selection of secret cypher data blocks and circu-lar queue for embedding the secret data and start indexfor embedding in circular queue. Subtract all the integervalues of first data byte of all the data blocks from 255.The data block whose first data byte is having minimumdifference value, taken as first data block for processing.After that, all data blocks are taken into considerationbased on their difference value in increasing order. Nowfor dynamic selection of circular queue for data embed-ding, we take the integer value of all the pixels of secondblock, which is the second pixel in all the circular queues.Now subtract all the integer values of all pixels of secondblock from 255. The circular queue whose second pixelis having maximum difference value, taken as first cir-cular queue for data embedding. After that, all circularqueues are taken into consideration based on their differ-ence value in decreasing order. For dynamic selection offirst embedding index in selected circular queue, we takethe starting four bits of first byte of selected secret datablock. Now by taking its integer value, we find the startindex for embedding in circular queue since in circularqueue we can start from any index for performing queueoperation.

Embedding the secret cypher text in cover image usingcircular queue, is done as follows.a. Embed the number of secret cypher message blocks,

secret cypher message block length, number of circu-

lar queues in cover image, number of circular queuesin cover image used for data embedding, length ofcircular queues, dynamic adaptive procedure gener-ated values for selection of circular queues of cover

Page 3

tosy

Cs

M

P

wi

aiStauitsiotidua3Mtoio

Adaptive circular queue image steganography with RSA cryp

image and secret message blocks, start index ofeach selected circular queue for embedding in somereserved location i.e. to byte numbers 3045—4045.

b. Convert the secret cypher text in various blocks ofseven bytes each and cover image in various circularqueues of sixteen pixels each.

c. Now by using dynamic adapted procedure, select oneblock of secret cypher message and assign them to onecircular queue of cover image, for embedding.

d. Embedding will be done in dynamically selected cir-cular queues sequentially. First pixel for embeddingin circular queue is identified by using dynamicallyselected start index. Here we can embed only 5th to8th bit LSB position in a pixel for better visual qualityof stego image.

e. Go to step (c). Continue this process until all thecypher data block is not empty and all secret cyphertext is not embedded in circular queues sequentiallyand send resultant stego image to the receiver.

• At receiver side, reverse mechanism is used to extract anddecrypt the secret data.

Results and analysis

Resultant simulated outcome using MATLAB for differentcover images and their stego images are being displayed inFig. 2. If histograms are also considered, then there is neg-ligible amount of difference between histogram of originalcover image and stego image. Histograms for various originalimages and their stego images are also shown in Fig. 2.

The PSNR, MSE and maximum embedding capacity valuesat divergent payloads for different images of various sizesare given in Table 1. PSNR is calculated in decibels (dB). A

high quality stego image should aspire for 40 dB and above (Liet al., 2011). PSNR outcome is defined by the mean squareerror (M.S.E) for two P × Q monochrome images, Where xas well as y are image coordinates, SGxy (stego image) and

boFv

Figure 2 (a), (e), (i), (m) are original cover images and (c), (g), (histograms of original cover images and (d), (h), (l), (p) are their ste

stem 419

Vxy (cover image), one of the images is approved a noisyurmise of the other is defined as:

.S.E = 1PQ

P∑x=1

Q∑y=1

(SGxy − CVxy) (1)

SNR = 10log10

{CV2

max

M.S.E

}(2)

here CVmax = the maximum 255 pixel value, for 8-bit covermages (Li et al., 2011).

Using Fig. 2, one can observe that there is no visualrtefacts with the stego images and histograms, it is look-ng exactly same as corresponding original cover images.teganography methods performance can be observed byhe three valuable specifications: secrecy, volume/capacity,nd visual imperceptibility (Li et al., 2011). Secrecy issed to protect data from unauthenticated attackers orntruders. The hiding capacity should be enough to obscurehe data in a cover image. Visual quality of stego imagehould be like that no one can claim about imperceptibil-ty (Cheddad et al., 2010). Fig. 3 shows the result analysisf proposed algorithm using various performance parame-ers. Using Table 1, results are analysed. By result analysis,t can be noticed that by increasing the cover image size andecreasing the secret data size PSNR value will be increasedp to 83.21 dB and MSE value will be decreased up to 0.0003s well as maximum embedding capacity is increased up to7%. So that performance will be high with respect to PSNR,SE and maximum embedding capacity value. Using Table 2,

he comparison of the proposed scheme is shown on the basisf minimum calculated PSNR, embedding capacity and visualmperceptibility with the different algorithms proposed byther researchers. Compared to other algorithms suggested

y different experts in this field, our approach is a strongerne and can be used for securing any kind of secret data.ig. 3 shows the result analysis of proposed algorithm usingarious performance measure parameters.

k), (o) are their stego images respectively, (b), (f), (j), (n) arego images histograms respectively.

Page 4

420 M. Jain et al.

Table 1 Observed capacity, MSE and PSNR value (different cover images of same/different size with various secret cypher dataof same/different size).Cover image(*.bmp)

Cover image size (inkilobytes)

Quantity of cypherembedded (in bytes)

Maximum embeddingvolume (kilo bytes)

Percentage ofembedding volume (%)w.r.t (image size)

MSE PSNR (indB)

Cameraman 262 256 89.25 34 0.0021 74.71Cameraman 262 1024 89.25 34 0.0054 70.21Baboon 262 256 86.61 33 0.0026 73.51Baboon 262 1024 86.61 33 0.0056 70.61Leena 262 256 84.23 32 0.0049 72.89Leena 262 1024 84.23 32 0.0038 70.91Barbara 262 256 85.67 33 0.0022 75.31Barbara 262 1024 85.67 33 0.0041 71.46Cameraman 1048 256 387.03 36 0.0003 83.21Cameraman 1048 1024 387.03 36 0.0010 78.21Baboon 1048 256 383.31 37 0.0004 82.03Baboon 1048 1024 383.31 37 0.0011 77.06Leena 1048 256 378.04 36 0.0004 82.63Leena 1048 1024 378.04 36 0.0010 78.52Barbara 1048 256 380.13 36 0.0003 82.89Barbara 1048 1024 380.13 36 0.0010 77.53

Figure 3 Result analysis of proposed algorith

Table 2 Comparison with other research.Name of the authorand schemes

MinimumcalculatedPSNR (dB)

Capacity Visualimperceptibility

Swain and Lenka(2015)

50.50 Good Better

Wang and Chen 44.20 Medium Good

C

IuLdtduosanht

as

R

C

L

N

P

A

S

W

161—164.

(2006)Nag et al. (2011) 30.48 Very low PoorProposed algorithm 70.21 Very good Best

onclusion

n this paper, a novel secret transmission scheme is proposedsing the notion of obscurity with respect to a circular queueSBs substitution. The secret message blocks are allocatedynamically by the sender to the image blocks with respecto circular queues, which increases security levels and givesynamic effect to proposed algorithm. Proposed algorithmses RSA public key cryptosystem to provide confidentialityf information at data centre end-to-end communication. Atteganography level, LSBs substitutions using circular queues

re used to protect data from leakage in transmission chan-el when resources are shared among multiple transmissionolders. By result and histogram analysis, it is concludedhat PSNR value is better as compared to some of the existing

Y

m using various performance parameters.

lgorithms and imperceptibility distortion cannot be mea-ured from the corresponding stego images.

eferences

heddad, A., et al., 2010. Digital image steganography survey andanalysis of current methods. Signal Process. 90, 727—752.

i, B., et al., 2011. A survey on image steganography and steganal-ysis. J. Inf. Hiding Multimed. Signal Process. 2 (2), 142—172.

ag, A., Singh, J.P., Khan, S., Ghosh, S., 2011. A Weighted LocationBased LSB Image Steganography Technique, vol. 191. SpringerACC 2011, CCIS 2, pp. 620—627.

rovos, N., Honeyman, P., 2003. Hide and seek: an introduction tosteganography. IEEE Secur. Priv. Mag. 1 (3), 32—44.

nderson, R.J., Petitcolas, F.A., 1998. On the limits of steganogra-phy. IEEE J. Sel. Areas Commun. 6 (4), 474—481, Special Issueon Copyright & Privacy protection.

wain, G., Lenka, S.K., 2015. A novel steganography technique bymapping words with LSB array. Int. J. Signal Imag. Syst. Eng.Indersci. 8 (1—2).

ang, R.Z., Chen, Y.S., 2006. High payload image steganographyusing two-way block matching. IEEE Signal Process. Lett. 13 (3),

ounes, M.A.B., Jantan, A., 2008. A new steganography approachfor image encryption exchange by using the LSB insertion. IJCSNSInt. J. Comput. Sci. Netw. Secur. 8 (6), 247—254.