International Journal of Ele ctrical and Ele ctronics Engineering Research (IJEEER) ISSN(P): 2250-155X; ISSN(E): 2278-943X Vol. 4, Issue 2, Apr 2014, 33-40 ÂŠ TJPRC Pvt. Ltd.

OPTIMUM OPERATIONAL COST ALGORITHM FOR CRYPTOGRAPHY OF MULTIMEDIA DATA ATUL S. J OSHI1 & P. R. DES HMUKH2 1

Department of Electronics & Teleco mmunication Engineering, Dr. Panjab rao Deshmu kh Co llege of Engineering, Amravati, Maharashtra, India 2

Department of Co mputer Science & Engineering, Dr. Panjabrao Deshmu kh Co llege of Engineering, Amravati, Maharashtra, India

ABSTRACT Fro m the last decade the world is experiencing the unprecedented explosion in the amount of multimed ia data. Hence security of this data is an important issue. The aspects of information security can address by cryptography. Cryptography is the practice of storing and communicating data in such a form that only whom it is intended for can read and process. In this paper we propose optimu m operational cost encoding technique which offers higher co mpression ratios and better security towards the cryptanalysis attacks during transmission of the multimedia data. It is simp le, low complexity algorith m and suitable for text, image as well as audio applications. At the transmitter side, source file is first converted into binary form. Source bit stream is decompose to the blocks of equal length, (Xi ). Pseudorandom generator is used to generate secured key, (Yi ). Fro m the pair of bits in the block, a sequence number is calculated i.e.

2i.

This number selects the particular bits in the key bit of the key. This bit is then complemented. For each chunk, the pseudorandom generator generates a new key which provide integrity protection, in addition to confidentiality. A selective encryption technique is employed in this algorith m. Hamming Distance in between the Yi & âˆ†Yi are taken care by the algorith m thereafter to form a codebook. At the receiver side, cipher ext racted fro m the channel is decompressed and decrypt jointly with the help of codebook and secret key. Deco mpression and Decryption of the compressed cipher is again integrated process in this algorithm. Entire bit stream is then obtained from the chunk of bits which thereafter converted to original source data. Due joint execution of encryption and compression processes t his algorithm provides high degree of security with better compression performance at very low operational cost.

KEYWORDS: Hamming Distance, Polygra m Substitution, Key, Encryption and Co mp ression INTRODUCTION Most of the data transported between source and destination is on the danger of being accessed or altered by the eavesdropper. Cryptography is the practice of storing and communicat ing data in secured form. The basic purpose of cryptography is to protect the data from unauthorized individuals who may exp loit it for their own benefit. Due to large size of mu ltimed ia data and because of real time constraint some algorith ms which are better for security may inc reases operational cost in execution [1]. In these circu mstances we propose this optimu m operational cost algorith m for cryptographic application.

www.tjprc.org

editor@tjprc.org

34

Atul S. Joshi & P. R. Deshmukh

The proposed Hamming distance based Polygram substitution cipher (HDPSC) algorith m is a joint paradig m of the selective encryption & compression. It is potentially possible to reduce the complexity of the compression and encryption in our algorithm because of the joint scheme for both these processes. According to the results made by survey, the joint compression and encryption algorithms reduces 40% of the memory storage size, increases execution speed up to 21%, increases compression ratio up to 25% and provides two level of security as compared to independent encryption techniques [2]. Selective encryption, encrypt only a subset of the data. Selective encryption reduces the amount of data to encrypt, at the same time it provides sufficient level of security. Overall co mputation cost reduction of data can be achieved using selective encryption techniques [3]. The basic philosophy of proposed algorithm is to transform the bit stream in to some intermediate form using Hamming distance technique & then compressed it using Polygram substitution with better efficiency. Hamming distance in between plaintext block and key block is calculated by the algorithm to form the codebook. Code words based on the Hamming distance are entered in the codebook. Codebook is a look up table used for coding and decoding. Polygram substitution cipher is type of substitution cipher in wh ich blocks of characters are substituted in groups. The proposed algorithm use Polygram substitution for indexing the code words by 3-bit pointer. Substitution of 3-bit pointer for 4-bit code, co mpressed the bit stream. Proposed algorithm uses dynamic approaches for encryption and compression of similar and dissimilar b it pairs. In similar bit pair approach first 2-b it are similar to next 2-bit in a block of 4-bit like 0000, 0101, 1010 and 1111. In dissimilar bit pair approach first pair of bit is unlike the next pair in a block of 4-b it. These approaches add second level of security in the proposed algorithm. Universal data co mpression technique has been used in the algorith m. In universal data compression techniques, data is compressed without statistical knowledge of the data.

LITERATURE REVIEW Intelligent dictionary based encoding algorithm develop a transformat ion yielding compression with security . The transformation used here in this method is basically depending upon visual perception.[4]. In RAC algorithm as the encoding is done on a bit-by-bit basis, and an interval part ition is associated to each bit, an independent decision is made as to whether the LPS or MPS subinterval co mes first [5]. Secure med ical image lossless compression scheme to transfer a med ical image fro m one place to another place or to store a medical image in a particular place in a secure manner discusses secure medical image lossless compression techniques uses Block Pixel Sort A lgorith m (BPSA) for solving the problems of co mpression and secure transmission [6]. Selectively encryption of the I-frames on the MPEG stream and application of DES on DC coefficients is discussed in the selective encryption algorith m [ 7]. Video encryption algorith m (VEA) wh ich uses a secret key to randomly change the signs of all DCT coefficients in an MPEG stream in [8] reduces the computational comp lexity by encrypting the sign bits of differential values of DC coefficients of I-frames and sign bits of differential values of motion vectors of B and P frames. The approach in [9] is a randomization of the arithmetic coder. This is achieved by randomly swapping the most probable symbol (MPS) and least probable symbol (LSP) intervals. Both total and selective encryptions are possible by choosing the layers or resolution levels to encrypt. In a syntax compliant encryption algorithm, encryption is inserted within the encoder. To achieve syntax co mpliance, selected compliant code words are randomly permuted with other compliant code words. The shift used for permutation is determined by the AES counter [10]. The JPEG Huffman coder terminates runs of zeros with code words in order to approach the entropy. Appended bits are added to these code words to fully specify the magnitudes and signs of nonzero Impact Factor(J CC): 5.9638

Index Copernicus Value(ICV): 3.0

35

Optimum Ope rational Cost Algorithm for Cryptography of Multimedia Data

coefficients, only these appended bits are encrypted using DES or IDEA [11]. The Encryption scheme in [12] is a JPEG2000 co mp liant algorith m wh ich iterat ively encrypts code block contribution to packets (CCP). The encryption process acts on CCP using stream ciphers or block ciphers. The chaotic map algorith m [13] is used to perform co mp ression and encryption simultaneously. This is can be applied for lossless data and with loss of image compression but not suitable for audio data encryption and compression. Literature survey of the related wo rk reveals the facts that most of the algorith ms that possess data security and data compression issues do not support text, image and audio data applications simultaneously. Also most of the algorith ms increase computational co mplexity. Hence effo rts have been made through this research work by designing innovative algorith m wh ich is low operational cost, combines data security and data compression issues without scarifying their individual performance and wo rk suitably with text, image and audio data.

PRPPOSED SCHEME The proposed algorithm is exp lained mathematically using two approaches, Dissimilar Bit Pair Approach and Similar Bit Pair Approach as presented below. Dissimilar Bit Pair Approach Let K represents chunks of 64 b its, XK represents decomposition of binary bit stream (M) into 64 bits, XKB represents 4-bit blocks of XK so that, XK ⊂ M XKB ⊂ XK Let Y represents pseudorandom key of 128 bits and YS represents randomly selected key of 64 bits fro m Y, so that YS € Y: YS ⊂ Y Let YSB represents decomposition of YS into the 4-b it blocks, hence YSB ⊂ YS Fro m the pair of b its in XKB , a sequence number n is calculated as n=

2i

The bit wh ich is to be map is represented by YSm and the value of n decides the mapping of YSm to YSB so that, n: YS m ↦ YSB Co mplement of YSm is represented by Changed form of YS is ∆YS & ∆YS contents both complemented and non-complemented bits. i. e ∆YS = {

, YS }

Hence Encryption, ES = ∆YS Let ESB represents 4-bit b lock of ES as

www.tjprc.org

editor@tjprc.org

36

Atul S. Joshi & P. R. Deshmukh

ESB ⊂ ES Let CH represents possible code words of which the selection is strictly depend on the value of Hamming distance of ‘2’, in between YSB and ESB CH : HD [([YSB]

, [ESB]

)] = 2

Polygram substitutions based codeword indexing is represented by Pi, Pi = {001, 010, 011, 100, 101, 110} Let CB is the codebook form by all possible code words CH , so that YSB : CH ↦ Pi Hence compressed cipher text, ESC fro m codebook is given by ESC = {Pi } Similar Bit Pair Approach In proposed HDPSC algorithm, pairs of the plaintext bits are mapped to the key bits in Two -on-One manner. However if the pair o f the plaintext bits are similar in a nibb le. i. e. If

=

This situation produces conflict with algorithm. Hence proposed scheme is design with separate approach called as similar bit approach to overcome this conflict. Th is difficulty strengthens the security of the algorithm. Since the generated cipher text is the combination of the encrypted bits either by encrypting similar b its or by encrypting dissimilar bits, it offers another flavor to the security of the plaintext bits. This approach is expressed mathematically as , = i.e

= 00

then ESB = aaaa & ESCB = 00 Both these approaches work together to achieve adaptability in proposed algorithm.

OPERATIONAL COST ANALYSIS Operational cost of the algorithm can be analyzed by evaluating the number of operations requires during encoding or decoding process. The encryption and the decryption costs are usually similar, and they are more important than the key-setup cost because one single key-setup can often be followed by thousands of encryption and decryption operations. However the actual complexity may vary & are h ighly depend on the particular architecture [13, 14]. In proposed algorith m we analy zed the co mputational comp lexity by evaluation of key set up cost and encoding cost that is encryption & compression cost as below

Impact Factor(J CC): 5.9638

Index Copernicus Value(ICV): 3.0

Optimum Ope rational Cost Algorithm for Cryptography of Multimedia Data

37

Plaintext Setup Cost (PS ) When multimedia input is provided to the algorithm; it is first converted to Binary bitstream. Let us consider that the single operation is require for conversion of the input to binary bit stream (Con). Another operation is require for the division of binary bit stream into 64 bits each (Di v). There after single operation is requires to form a chunks of 4 bits in each 64 b it frame (C4 ). Hence for 64 bit frame the number of C4 operations requires to be 64/4 = 16. Total Plaintext Setup Cost is the algebraic sum o f operations mentioned above. i.e. PS = 1(Con) + 1 (Di v) + 16(C4 ) = 18 Key Setup Cost (KS ) The Key setup cost includes all the computations prior to actual encryption process of the first bit of the plaintext .Let us assume that the single operation is requires for random key generation of 128 b its by using Pseudorandom Sequence generator for each 64 bits (KG ). Key bits are selected thereafter randomly equal in number with the plaintext bits (KB ). Then chunks of 4 bits are formed fro m total 64 KS bits that requires 16 operations (K4 ). Thus total Key setup cost is given by KS = 1 (KG ) + 1 (KB ) + 16 (K4 ) = 18 Encryption Cost (E) Selective Encryption technique is emp loyed in proposed algorithm. It starts with the finding of decimal value of the bits from every chunk of 4 b its (

2 i ). If this is consider as a single operation then for each chunk 2 operations

are require. Since each 64 bit frame has 16 chunks, total 16* 2 = 32 such operations are required. According to decimal value of bits, a single bit of K4 gets change. Hence total 32 KB bits gets change requires 32 operations (∆KB ). Hence total Encryption Cost is E = 32 (

2 i ) + 32(∆KB ) = 64

Compression Cost (C) In the proposed scheme compression requires Hamming Distance (HD) calculations and Polygram Substitution (PS). Cipher text (Ci ) available as input to compression process is the combination of K G & ∆KB bits. Comp ression procedure starts with the format ion of the codebook on the basis of Hamming d istance. Using 4 b it binary total 16 codes are possible. For each 4 bit chunk of the cipher text code (Ci4 ) total 6 codes are at a Hamming Distance of 2 from that particular code. These 6 codes are index using lesser bits. Each Ci4 is entering in the codebook against the 6 codes of Hamming distance 2 from part icular Ci4 and its 6 indices. For codebook format ion each Ci4 requires 6 operations for 6 codes and another 6 operations are requires for 6 indices. It means that each Ci 4 requires 12 operations. Since there are 16 Ci4 in 64 bits, the nu mber of operations requires being 12*16 operations (CO ). The algorithm co mpares the bit of cipher text with the bits of KG , chunk by chunk. Since in 64 b its there are 16 chunks, 16 co mparisons are requires (CP ). Co mparison is to find out Hamming d istance code which exist in codebook fro m each 4 bit chunk of K G by noticing the changed bits. For 16 chunks of KG , requires 16 such operations (HdC ). Each (Hd C) is there after replace by the corresponding Codebook Index of 3 bits results into compressed code further requires 16 operations ( CI ). Thus total Co mpression Cost is given by

www.tjprc.org

editor@tjprc.org

38

Atul S. Joshi & P. R. Deshmukh

C = 12*16 (CO) +16(C P) +16 (HdC ) + 16 (Ci) = 240 Total Computational Cost of the algorithm is CC = PS + KS + E + C = 18 + 18 + 64 + 240 = 340 operati ons. However the Average value of the Computational Cost is less than 340 operations for maximu m input frame size of 64 b its. The above cost is calculated by considering first 2 bits of the C 4 are not similar with the next two b its of the plaintext. When they are similar algorith m work with different approach results into lesser value of Encryption and Co mpression cost as explained below Encryption Cost for Si milar Bits (ES) In this case algorithms need not to calculate

2 i . Also âˆ†KB operations are not required. There is Direct

Assignment of the alphabets like aaaa depend upon the 2 bit similarity for C 4 irrespective of the KB. Hence for 16 chunks, 16 such operations are reqires (DA ). Hence ES = 16(DA ) = 16 Average value of the Encryption Cost (AE) = E + ES / 2= 64 + 16 / 2 Thus, AE = 40 Compression Cost for Similar Bits (CS) When the first 2 bits of C4 is similar with next 2 bits, steps taken by the algorithm is comparatively simple results into very less computational cost. In this case there is no need of codebook format ion and further operations. However encrypted DA is replace by 2 bits depend on the form of DA .If 16 such DA are consider in 64 bits, then 16 replacement operations are requires(R2 ). Hence CS = 16 (R2 ) = 16 Average value of the Compression Cost (AC) = C + CS / 2 = 240+ 16 / 2 = 128 Thus, AC = 128 The average value of co mputational cost of the algorithm is ACC = PS + KS + AE + AC = 18 + 18 + 40 + 128 = 204 operations.

Impact Factor(J CC): 5.9638

Index Copernicus Value(ICV): 3.0

39

Optimum Ope rational Cost Algorithm for Cryptography of Multimedia Data

The another approach handled by the algorithm for Similar bits not only reduce the average computational cost but also it provide another level of security as the cipher text is the mixing of the code provided by both these approaches. Because of this fact it is too much difficu lt to guess the plaintext by the adversary.

RESULTS & DISCUSSIONS The result & co mparision is based on the Key set up cost & Encryption cost.

Figure 1: Encrypti on Cost

Figure 2: Key Setup Cost Fro m Figure 1, it is clear that Encryption cost of the proposed HDPS algorith m is high as compare to other algorith ms. But this high value of the encryption cost is compensated by the Key Setup cost as shown in fig ure 2. Hence overall average operational cost is very less without scarifying the compression performance as well as security.

CONCLUSIONS Hamming Distance Polygram Substitution (HDPS) algorith m based on joint selective encryption and compression approach, proves superior in operational cost performance over other standard algorithms.

REFERENCES 1.

M. Droogenbroeck and R. Benedett, “Techniques for a selective encryption of uncompressed and compressed images,” Proc. ACIVS, Ghent, Belgium, Sep. 2002.

2.

K. John Singh and R. Manimegalai, “A Survey on joint co mpression and encryption techniques for video data,” Journal of Computer Science, 8 (5): 731-736, 2012.

www.tjprc.org

editor@tjprc.org

40

Atul S. Joshi & P. R. Deshmukh

3.

W. Puech and J.M. Rodrigues , “Analysis and cryptanalysis of a selective encryption method for JPEG Images,” Journal of Computer applications and research, Vol. 2, 2013.

4.

Dr. V. K. Gov indan and B.S.Shajee Mohan, “An intelligent dictionary based encoding algorith m for text data compression for high speed data transmission over internet,” International Conference on Intelligent Signal Processing and Robotics, Feb2004.

5.

Marco Grangetto and Enrico Magh, “Multimedia selective encryption by means of randomized arith metic coding,” IEEE transactions on multimedia, Nov 2005.

6.

S. Manimu rugan and K.Porku maran, “Fast and Efficient Secure Medical Image Co mp ression Schemes,” European Journal of Scientific Research, Vo l.56 No.2, 2011.

7.

L. Tang, “Methods for encrypting and decrypting MPEG video data efficiently,” Proceedings of the 4th ACM International Multimedia Conference and Exhibition, pp. 219–229, Boston, Mass, USA, November 1996.

8.

C. Sh i and B. Bhargava, “A fast MPEG video encryption algorithm,” 6th ACM International Conference on Multimedia, pp. 81–88, Bristol, UK, September 1998.

9.

M. Grangetto, E. Magli, and G. Olmo , “Multimed ia selective encryption by means of randomized arith metic coding,” IEEE Transactions on Multimedia, Vol. 8, no. 5, pp. 905– 917, 2006.

10. C. Bergeron and C. Lamy-Bergot, “Co mp liant selective encryption for H.264/A VC video streams,” 7 th IEEE Workshop on Multimedia Signal Processing (MMSP ’05) , pp. 1–4, Shanghai, China, October 2005. 11. M. Van Droogenbroeck and R. Benedett, “Techniques for a selective encryption of uncompressed and compressed images,” Advanced Concepts for Intelligent Vision Systems (ACIVS ’02) , pp. 90–97, Ghent, Belgiu m, September 2002. 12. Y. Wu and R. H. Deng, “Co mp liant encryption of JPEG2000 code streams,” International Conference on Image Processing, vol. 5, pp. 3439–3442, Singapore, October 2004. 13. Wong, K.W, “Performing comp ression and encryption simu ltaneously using chaotic map,” in city University of Hong Kong, China, 2008. 14. Y. Sadourni and V. Conan, “A proposal for supporting selective encryption in JPSEC,” IEEE Transactions on Consumer Electronics, volume 49, nu mber 4, pp. 846– 849, Nov. 2003.

AUTHOR’ S DETAILS Prof.

Atul

Joshi

is

currently

working

as

a

Associate

Professor

in

Department

of

Electronics & Teleco mmunication Engineering, at Dr.Panjabrao Deshmukh College of Eng ineering, A mravati (India). Dr. Prashant Deshmukh is currently working as Head of CMPS & IT Depart ment, Engineering & Technology, Amravati (India). He has completed his Ph.D. in the faculty of Electronics Engineering fro m SGBA U A mravati University, Amravati (India). His areas of interest are Dig ital Signal Processing, VLSI Design and Embedded Systems.

Impact Factor(J CC): 5.9638

Index Copernicus Value(ICV): 3.0