A Novel Visual Cryptographic Steganography Technique by Mohit Goel
1.
International Journal of Computer, Electronics & Electrical Engineering (ISSN: 2249 - 9997)Volume 2– Issue 2 A Novel Visual Cryptographic Steganography Technique Mohit Kumar Goel *, Dr. Neelu Jain # * Dept. of Electronics and Electrical Comm., PEC University of Technology, Chandigarh mohitgoel4u@gmail.com # Dept. of Electronics and Electrical Comm., PEC University of Technology, Chandigarh neelujain@pec.ac.inAbstract— With development in technologies, the amount of data being exchanged on internet is increasing exponentially. The securityof information can be achieved by cryptography and steganography. Cryptography hides the contents of message by converting it to anunreadable cipher. Steganography hides the existence of message by embedding data in some other digital media like image or audiofiles. The paper proposes a security system which is combination of both the techniques. In proposed system data is firstly encryptedusing RSA encryption algorithm and then embedded in an image using DCT based steganographic method. The experimental resultshows that proposed system has better PSNR value in comparison to other techniques like LSB, Modulus arithmetic steganography. Italso maintainstaisfactory security as secret message can’t be extracted without knowing the decoding algorithm.Keywords- steganography; RSA encryption; data hiding; discrete cosine transform. conceal the secret messages within some image, music or I. INTRODUCTION audio file so that it is not visible to others. Image Due to increasing the technologies, security systems are steganography schemes can be divided into two categories:very popular in many areas. The information comes in Spatial Domain and Frequency Domain.numerous forms and requires secure communication. Suchsecure communication ranges from bank transactions, A. Spatial domain steganographycorporate communications and credit card purchases. Spatial domain techniques embed messages in theEncryption and steganography are the preferred techniques intensity of the pixels directly [6][7][8]. Least Significantfor protecting the transmitted data. Encryption hides the Bit (LSB) is the first most widely used spatial domaincontents of the message, but cannot hide the message data steganography technique. It embeds the bits of a message initself. However, encryption clearly marks a message as the LSB of the image pixels [9][10]. But the problem withcontaining “interesting” information, and the encrypted this technique is that if the image is compressed then themessage becomes subject to attack. The commonly used embedded data may be lost. Thus, there is a fear for loss ofencryption schemes include DES (Data Encryption data that may have sensitive information [11]. LSB has beenStandard) [1], AES (Advanced Encryption Standard) [2] and improved by using a Pseudo Random Number GeneratorRSA [3]. DES, an encryption standard that was used by (PRNG) and a secret key in order to have private access tomany national governments, successfully withstood attacks the embedded information [12]. The embedding processfor many years. However, E. Biham and A. Shamir mention starts with deriving a seed for a PRNG from the usera cryptanalytic attack that can break DES in only a few password and generating a random walk through the coverminutes [4]. Another example of a broken encryption image that makes the steganalysis hard. Another recentalgorithm is WEP. WEP was designed to provide improvement based on random distribution of the messageconfidentiality to users on wireless networks. A. was introduced by M. Bani Younes and A. Jantan [13]. InStubblefield illustrates how WEP can be broken within this method they utilize an encryption key to hidehours [5]. DES and WEP are examples of two encryption information about horizontal and vertical blocks where thealgorithms that were thought to be secure at the time of their secret message bits are randomly concealed. SSB-4design, but were broken in the future when attackers had steganography approach introduced by Rodrigues, Rios andmore powerful computational resources. So, in many cases Puech is about changing the 4th bit of a pixel in the originalit is desirable to send information without being notice by image according to the bit message. Then modify the otheranyone that information has been sent. Steganography bits (1st, 2nd, 3rd and/or 5th) to minimize the difference www.ijceee.org Page 39
2.
International Journal of Computer, Electronics & Electrical Engineering (ISSN: 2249 - 9997)Volume 2– Issue 2between the changed pixel value and the original one [14]. 3) Hash Functions: Uses a mathematicalThe 4th digit is a significant bit and if the image is transformation to irreversibly “encrypt”compressed the embedded information is not destroyed [15]. information.Tu C. and Tran T D. argued that the difference must be A. RSA encryption algorithmequal or less than four (i.e., ±4) [16]. The 4th bit was chosenbecause it satisfies that changing of ±4 units in the channel RSA is a Public key cryptography named after itscolor value is imperceptible to human eyes, and it is the inventors: Ronald Rivest, Adi Shamir and Leonard Adleman.most significant bit which provides the minimum change in RSA can be used for encryption as well as for authenticationthe pixel values. Modulus arithmetic steganography [3]. An example of Alice and Bob, who want to useproposed by Sayuthi Jaafar and Azizah A Manaf has asymmetric RSA algorithm for secure communication iscalculated last four bits of each pixel by mod-16 operation. shown in fig. 1. For encryption purpose, Alice would encryptThen these bits are replaced with data bits [8]. In this the the message using Bob’s Public key and send the cipher textamount of the data that can be embedded is more but stego to Bob. Upon receiving the cipher text, Bob, who is owner ofimage has less PSNR value than LSB and SSB-4 corresponding private key, can then decrypt the messagetechniques. with his private key. For authentication purposes, Alice would encrypt (or sign) the message using her own privateB. Frequency domain steganography key. Other people such as Bob can verify the authenticity of the message by using Alice’s Public key, which is the only In frequency domain, images are first transformed and key that matches the signing private key.then the message is embedded in the image [17][18][19].When the data is embedded in frequency domain, the hiddendata resides in more robust areas, spread across the entireimage, and provides better resistance against statisticalattacks. There are many techniques used to transform imagefrom spatial domain to frequency domain. The mostcommon frequency domain method usually used in imageprocessing is the 2D discrete cosine transform [20][21]. Inthis technique the image is divided into 8×8 blocks andDCT transformation on each block is performed. The databits are embedded in the low frequency coefficients of DCT.SSB-4 & DCT steganography proposed by Nedal M. S.Kafri and Hani Y Suleiman uses DCT approach with SSB-4technique [21]. Steganography with cryptography can be combined sothat, even if an attacker does realize that a message is sent,he would still have to decode it [26]. Piyush Marwaha andParesh Marwaha use DES encryption and LSBsteganography for data security [25]. In this paper wepropose a method which uses RSA encryption and LSB-DCT steganography for data security. II. BACKGROUND OF CRYPTOGRAPHY Figure 1. RSA Encryption In cryptography, the message is scrambled to make itmeaningless and unintelligible unless the decryption key is The steps for RSA algorithm are:available. It makes no attempt to disguise or hide the 1) Select two prime numbers p, q.encoded message. Basically, cryptography offers the ability 2) Calculate n= p × q and (n)= (p-1)(q-1)of transmitting information between persons in a way that 3) Select integer ‘e’ such thatprevents a third party from reading it. Cryptography canalso provide authentication for verifying the identity of gcd ( (n),e)=1; 1<e < (n)someone or something. There are several ways of 4) Calculate d such that d × e=1mod( (n))classifying cryptographic algorithms. The three types of 5) Now Public key (PU) is {e, n} and Privatealgorithms are: Key (PR) is {d, n}. 6) At sender side, message (M) to be sent is 1) Secret Key Cryptography: Uses a single key for converted into cipher text (C) as follows: both encryption and decryption. C= Me mod n (1) 2) Public Key Cryptography: Uses one key for 7) At receiver side, cipher text is converted to original encryption and another for decryption. message as follows: M= Cd mod n (2) www.ijceee.org Page 40
3.
International Journal of Computer, Electronics & Electrical Engineering (ISSN: 2249 - 9997)Volume 2– Issue 2 III. LSB-DCT STEGANOGRAPHY DCT is performed on each block. Then scan the DCT block in zigzag way and extract the embedded data. LSB-DCT steganography image (I) is divided into 8x8blocks and two dimensional (2-D) is performed on each IV. PROPOSED METHODblock. The 2-d DCT is calculated as follow: The challenge in this work was to find a way to 1 7 7 π(2x +1)u π(2y +1) camouflage a secret message in an image withoutF(u, v) = C(u)C(v) f (x, y)cos cos (3) perceptible degrading the image quality and to provide 4 x=0 y=0 16 16 better resistance against the steganalysis process. The data isfor x=0,..., 7 and y=0,..,7 first converted into cipher text using RSA encryption and then hided into lower frequency components of image using 1 / 2 for k = 0 LSB-DCT steganography.where C ( k ) = 1 otherwise A. Embedding algorithm In DCT block lower frequency cofficents are at upper Steps of embedding algorithm are given as follow:left positions and high frequency coefficients are lower rightpositions. Now image is compressed by quantization. Input: An M×N size cover image and data to be concealed.Quantization is achieved by dividing each element in theDCT coefficient block by the corresponding value in the Output: Stego image.standard quantization matrix shown in fig. 2 and the result is 1) Encrypt the plain text using encryption key.rounded to the nearest integer. As eye is not able to discern 2) Divide the cover image into 8×8 blocks.the change in high frequency components so these can be 3) Perform 2-D DCT on each block.compressed to larger extent. Lower right side components of 4) Perform quantization on each block.quantization matrix are of high value so that after 5) Perform zigzag scan to convert 8×8 block into onequantization high frequency components become zero. dimensional array. 16 11 10 16 24 40 51 61 6) Replace the LSB of DCT coefficients with data 12 12 14 19 26 58 60 55 bits. 14 13 16 24 40 57 69 56 7) Convert 1-D zigzag array back to 8×8 block. 8) Perform Inverse DCT on each block. 14 17 22 29 51 87 80 62 Q= 9) Combine all the blocks to form stego image. 18 22 37 56 68 109 103 77 24 35 55 64 81 104 113 92 49 64 78 87 103 121 120 101 72 92 95 98 112 100 103 99 Figure 2. Quantization Matrix Although the DCT coefficients have been decorrelatedby DCT transform to some extent, DCT coefficients in thesame block are still not independent, which is called asintra-block correlation [16]. While neglecting the impact ofblock edge, the general trend in magnitude of the blockcoefficients in each block is non-increasing along zigzagscan order. After block DCT coefficients are arranged byzigzag scan pattern, dependencies among neighboringcoefficients in both horizontal and vertical directions can beconveniently investigated [23]. Now data is embedded inone dimensional zigzag array a) If data bit is ‘0’, then make the DCT coefficient even or, b) If the data bit is ‘1’, then make the DCT coefficient odd After embedding data zigzag array is again convertedinto 8×8 block. These blocks are dequantized and inverse Figure 3. Proposed MethodDCT is performed. The entire 8×8 blocks are combined toform the stego image which is then sent to receiver. B. Extraction algorithm At the receiver side the stego-image is received in spatial Steps for extraction algorithm are given as follows:domain. Now stego image is divided into 8×8 blocks and Input: An M×N size Stego image. www.ijceee.org Page 41
4.
International Journal of Computer, Electronics & Electrical Engineering (ISSN: 2249 - 9997)Volume 2– Issue 2Output: Secret message. (c) Original Flower.jpg (d) Stego Flower.jpg 1) Divide the stego image into 8×8 blocks. 2) Perform 2-D DCT on each block. 3) Perform quantization on each block. 4) Perform zigzag scan to convert 8×8 block into one dimensional array. 5) Check the DCT coefficient. a) If DCT coefficient is even then data bit is 0 or, b) If DCT coefficient is odd then data bit is 1. (e) Original Building.jpg (f) Stego Building.jpg 6) Concatenate the bits to obtain cipher message. 7) Decrypt the cipher text using decryption keys and display original message on screen. V. EXPERIMENTAL RESULTS Since the visual detection of stego images is dependingon the nature of the image [24] so, varieties of imagecategories are utilized in the experiments. The experimentalimage data set consists of 100 JPEG images, which were (g) Original Tree.jpg (h) Stego Tree.jpgtaken by digital camera. We focused on short messages withlength of 3000 bits because they are the most challenging to Figure 4. Original Images and Stego Images using DCT steganographydetect [24]. Comparative analysis of LSB, Modulus The comparative analysis of PSNR value of differentarithmetic (mod-16), and proposed method has been done steanography technique, is given in table 1, shows thaton the basis of Peak signal to noise ratio (PSNR). To proposed steganography method has better image quality ofcalculate PSNR, first MSE is calculated as follows: stego image than other techniques. 1 m −1 n −1 2 Table 1. Comparative analysis of PSNR values of different steganography MSE = I (i , j ) − K ( i , j ) ( 4) techniques mn i =0 j =0 PSNR Value Where MSE is the Mean Squared Error of Originalimage (I) and stego image (K). Thereafter PSNR value is Image Modulus RSA & LSB- LSBcalculated as follow: (mod-16) DCT 55.87 MAX i2 MAX i Human.jpg 52.10 49.23 PSNR = 10. log10 = 20. log10 (5) MSE MSE 56.36 Flower.jpg 53.54 50.53 Where, MAXi is the maximum pixel value of the image. 54.59In other words MAXi = 2b − 1, where b is the bit depth of Building.jpg 52.43 48.77the original image. PSNR computes the peak signal to noise 55.57ratio, in decibels, between two images. This ratio is used as Tree.jpg 53.46 50.46quality measurement between two images. V. CONCLUSION In this paper we used mixed approach cryptography and steganography is used for data security. By using RSA encryption, ASCII codes corresponding to characters of plain text are converted into 16 bits encrypted codes. Hence it becomes difficult to get original text without knowing decryption keys. Then cipher data is hided into cover image. Average PSNR value of 55 is obtained for 100 images using (a) Original Human.jpg (b) Stego Human.jpg proposed method. The obtained experimental results indicate that, the proposed method is a good and acceptable scheme for data security. Furthermore, by embedding information in the least significant bits of the DCT domain, the hidden message resides in more robust areas, spread across the entire stego image, and provides better resistance against statistical attacks than other techniques. The future work may focus on the improvement and further development in this technique. www.ijceee.org Page 42
5.
International Journal of Computer, Electronics & Electrical Engineering (ISSN: 2249 - 9997)Volume 2– Issue 2 VI. REFERENCES [15] J. Fridrich, and M. Goljan, “Practical steganalysis: state-of the-art”, In Proceeding of SPIE Photonics West, Electronic Imaging 2002, vol.[1] DES Encryption Standard (DES), National Bureau of Standard (U.S.). 4675, pp. 1-13, 2002. Federal Information Processing Standards Publication 46, National [16] Tu C. and Tran T D. “Context based entropy coding of block Technical Information Service, Springfield, VA, 1997. transform coefficients for image compression”, IEEE Transaction on[2] Daemen J., and Rijmen, V. “Rijndael: The Advanced Encryption Image Processing, vol. 11, No.11, November, 2002. Standard”, Dr. Dobb’s Journal, March 2001. [17] Wenqiong Yu, “Blind Detection for JPEG Steganography”,[3] R. Rivest, A. Shamir, and L. Adleman, “A method for obtaining International Conference on Networking and Information digital signatures and public-key cryptosystems”. Communication of Technology, pp. 128-132, July 2010. the ACM, pp. 120-126, 1978. [18] Chung, K.L., Shen, C.H. and Chang, “A novel SVDand VQ-based[4] E. Biham, A. Shamir “Differential cryptanalysis of DES-like image hiding scheme”.Pattern Recognition Letters, 22, pp. 1051 – cryptosystem”, Journal of cryptography, vol. 4, pp. 63-72, Jan. 1991. 1058, 2001.[5] A. Stubblefield, J. Loannidis and A. D. Rubin, “A Key Recovery [19] Iwata M., Miyake K. and Shiozaki, “Digital Steganography Utilizing Attack on the 802.11b Wired Equivalent Privacy Protocol”, ACM Features of JPEG Images”, IEICE Transfusion Fundamentals, E87-A, transaction on Information and System Security, vol. 7, pp. 319-332, 4, pp. 929 – 936, 2004. May 2004. [20] M. Kharrazi, H. Sencar and N. Memon, “Performance study of[6] Chan, C.K. and Cheng. L.M. “Hiding data in image by simple LSB common image steganography and steganalysis techniques,” substitution. Pattern Recognition”, 37, pp. 469 – 474, 2004. Communications of the SPIE and IS&T, 15, No.4, pp. 1017-9909,[7] Chang,C.C and Tseng, H.W. “A Steganographic method for digital Oct-Dec., 2006. images using side match”. Pattern Recognition Letters, 25, pp. 1431 – [21] Nedal M.S. Kafari, Hani Y. Suleiman, “Bit-4 of Frequency Domain 1437, 2004. DCT Steganography Technique”, First national on networked digital[8] Sayuthi Jaafar, Azizah A Manaf, Akram M Zeki, “Steganography technology, p.p. 286-291, 2009. Technique using Modulus Arithmetic”, 9th International Symposium [22] Dr. Ekta Walia, Payal Jain and Navdeep, “An Analysis of LSB & on Signal Processing and Its Applications, pp. 1 – 4, April DCT based Steganography”, Global Journal of Computer Science and 2007. Technology, Vol. 1, pp. 4-8, April, 2010.[9] W. Bender, D. Gruhl, N. Morimoto, and A. Lu, “Techniques for Data [23] Zhiping Zhou and Maomao Hui, “Steganalysis for Markov Hiding”, I.B.M. Systems Journal, 35(3-4): pp. 313-336, 1996. Feature of Difference Array in DCT Domain”, Proceedings of Sixth[10] N. Nikolaidis, and I. Pitas, “Robust Image Watermarking in the International Conference on Fuzzy Systems and Knowledge Spatial Domain”, Signal Processing, 66(3), pp. 385-403, 1998 Discovery, pp. 581 - 584 , Aug. 2009.[11] T. Morkel, J. Eloff, and M. Olivier, “An overview of image [24] L. Davidson, and P. Goutam, “Locating secret message in images”, In steganography”, In Proceedings of the Fifth Annual Information ACM SIGKDD international conference on Knowledge discovery Security South Africa Conference (ISSA2005). and data mining, (Seattle, Washington, Aug.22-25. ACM 1-58113- 888-1, 2004.[12] J. Fridrich, M. Goljan, “ Steganalysis of JPEG Images: Breaking the F5 Algorithm”, Publisher: Springer Berlin, Heidelberg, Lecture Notes [25] Piyush Marwaha, Paresh Marwaha, “Visual cryptographic in Computer Science, vol. 2578, pp 310-323, 2003. steganography in images”, Proceedings of international conference[13] M. A. Bani Younes, A. Jantan, “A New Steganography Approach for on Computing and Networking Technologies, p.p. 1-6, 2010. Image Encryption Exchange by Using the Least Significant Bit [26] Wai Wai Zin, Than Naing Soe, “Implementation and Analysis of Insertion”, IJCSNS, International Journal of Computer Science and Three Steganographic Approaches”, Proceedings of 3rd international Network Security, vol. 8 No. 6, June 2008. conference on computer research and development, vol. 2, p.p. 456-460, 2011.[14] J. Rodrigues, J. Rios, and W. Puech “SSB-4 System of Steganography using bit 4”, In International Workshop on Image Analysis for Multimedia WIAMIS, May, 2005. www.ijceee.org Page 43
Clipping is a handy way to collect and organize the most important slides from a presentation. You can keep your great finds in clipboards organized around topics.