Digital Image Chaotic Encryption

Digital Image Chaotic Encryption (DICE – A Partial-Symmetric Key Cipher for Digital Images)

Akhil Kaushik

Satvika

Assistant Professor, Computers Department T.I.T&S College Bhiwani, Haryana, India [email protected]

Assistant Professor, It Department T.I.T&S College Bhiwani, Haryana, India [email protected]

Krishan Gupta

Anant Kumar

M. Tech Scholar, Computers Department T.I.T&S College Bhiwani, Haryana, India [email protected]

Software Engineer Newgen Software Technologies Ltd Okhla, Delhi, India Email: [email protected]

Abstract- The swift growth of communication facilities and ever decreasing cost of computer hardware has brought tremendous possibilities of expansion for commercial and academic rationales. With widely incremented communiqué like Internet, not only the good guys, but also bad guys have advantage. The hackers or crackers can take advantage of network vulnerabilities and pose a big threat to network security personnel. The information can be transferred by means of textual data, digital images, videos, animations, etc and thus requires better defense. Especially, the images are more visual and descriptive than textual data; hence they act as a momentous way of communication in the modern world. Protection of the digital images during transmission becomes more serious concern when they are confidential war plans, top-secret weapon photographs, stealthy military data and surreptitious architectural designs of financial buildings, etc. Several mechanisms like cryptography, steganography, hash functions, digital signatures have been designed to provide the ultimate safety for secret data. When the data is in form of digital images; certain features of images like high redundancy, strong correlation between neighbouring pixels and abundance in information expression need some extra fortification while transmission. This paper proposes a new cryptographic cipher named Digital Image Chaotic Encryption (DICE) to convene the special requisites of secure image transfer. The strength of DICE lies in its partial-symmetric key nature i.e. even discovery of encryption key by hacker will not guarantee decoding of the original message. Keywords- DICE, Partial-Symmetric key algorithm, Block cipher, Digital watermarking.

I.

INTRODUCTION

Through the rapid development of computer and network technology, sending data in form of text, images, sounds and videos have become an important security issue. Now more and more data in image format needs to

be transferred over the network, as images are more descriptive in nature. If the digital images are highly confidential and contain covert data like medical images/ X-rays/ military maps, etc., then they need a solid mechanism for transmission of images. The special attributes of digital images like bulk data capacity, high redundancy and high correlation between neighboring pixels must be considered carefully while choosing them as media for electronic data transfer. Hence an inevitable and best solution to send images can be cryptography. Cryptography can be defined as the art of creating an unintelligible form from intelligent information[1][7]. Encryption of textual data is relevant mostly to onedimensional data, that’s why these techniques are incompetent for two-dimensional digital images. The established encryption algorithms like DES, AES, etc concentrate on changing two-dimensional data into onedimensional data and then applying encryption on it[8]. But this technique is less efficient to encrypt and decrypt images and because of these grounds, it is not preferred to use this primitive style of encryption on digital images. Now-a-days, research on image transfer has reached a high standard and lots of encryption mechanisms have been devised. These techniques can be broadly classified into cryptography, steganography and digital watermarking. Steganography is the art of hiding a secret message within a larger one in such a way that others cannot discern the presence or contents of the hidden message[2]. Digital watermark technology mainly embeds the watermark information into the image information, provides copyright protection to achieve the information property right for digital image. This technology may be feasible referring to information property right in entertainment industries, but it is not recommended for

sending confidential information such as maps of military battlefields. However, cryptography still stands out as the preferred way of image transfer. Encryption in digital images mostly works at pixel level, which is the lowest level of information in the image. But due to strong correlation between neighboring pixels, data of one pixel can be decoded easily if one of neighboring pixel becomes known. However, an image can also be interpreted as an ordered arrangement of image blocks instead of pixels. An accurate orientation of these image blocks makes us able to sense information from the image changing of which causes visual disruption[5]. Thus using block level encryption for images will help to overcome correlation between neighboring pixels, which is the chief nuisance in image encryption. The block size should be smaller for better transformation because only fewer pixels will keep their neighbor’s data[6]. In this paper, a new block crypotosystem for encoding images named: Digital Image Chaotic Encryption (DICE) has been proposed. II.

PROPOSED ENCRYPTION ALGORITHM

This new proposed color image encryption algorithm is a basically a block cipher which takes blocks of images as input and encrypts them simultaneously using a special mathematical set of functions known as key. This algorithm uses same key at both sending and receiving ends i.e. it uses symmetric key tactic. This new proposed algorithm has a special characteristic i.e. it operates at two levels of image data: block level and pixel level to add more chaos and confusion. Another imperative attribute of this algorithm is its partial dependence on the secret key for encryption; hence it is more rigid to break. It actually uses different keys at block level and pixel level; thus breaking one key will not be enough to decode whole image. Moreover, keys at both levels are altered for each pixel and block to give headache to the eavesdropper. The varying nature of key makes it tougher against traditional attacks like brute-force attacks[4]. Working at both block level and pixel level will also protect from high redundancy of color images. This algorithm is designed to provide ultimate security against unauthorized attacks.

Fig 1. Generation of Round-Key by Primary Key

III.

DICE ENCRYPTION PROCESS

In this algorithm, a diverse set of binary operations are performed on the block and pixel-level data like binary XOR, which gives output as 1 (or true) if both inputs are different i.e. 0 and 1. On similar inputs, it gives a false value i.e. 0. The following figure shows the truth tables of XOR and XNOR operation:

Fig 2: Truth tables of XOR and XNOR

The steps of encryption process for DICE algorithm are as follow: i. The given image is divided into number of blocks as a multiple of 6. ii. The blocks are shuffled row wise first. iii. Then blocks are jumbled column wise to ensure security. iv. After scrambling of blocks, a block of 3 pixels are selected and their red component’s value (RRR) out of RGB content is considered. v. Binary shift-left operation is performed on this 24-bit data ‘n’ number of times, where ‘n’ is a predefined value. vi. A 24-bit round primary key is randomly chosen from the original 40-bit key. As the name suggests, round key is different for every round of encryption. vii. This primary key is then XORed with Red component of that block. viii. Then the Green component (GGG) and Blue component (BBB) value of same block using other binary operations (XOR and XNOR). After that the primary key is applied on both components and it is ensured that resultant data is of 24 bits. ix. After encoding of one block, next block is selected and then each of Red, Green and Blue components is operated by primary key using different binary operations. x. Steps iv to ix is repeated till every block of image is encrypted. xi. After block level encryption is achieved, the next level of information i.e. pixel is taken into account. xii. A random secondary key is chosen from a preselected range of delimiters and is converted into an 8-bit number.

xiii.

xiv.

xv.

xvi.

This 8-bit number is divided into 8 individual bits and then appended in the 24-bit value of the pixel (RGB Component) at particular positions. The resultant 32-bit data is divided into 4-bit segments and these segments are changed into 3bit by XORing 1st and 4th bit. Hence, finally achieving 24-bit encrypted data. This 24-bit data is then converted back to RGB component of pixel and then a new pixel is formed, which is poles-apart from the plain image’s pixel. Steps xi to xv is repeated till every pixel of image is encoded to get the cipher image.

vi. vii. viii.

Reverse binary operations are performed on the block’s Red, Green and Blue component. Steps v and vi are repeated for every block till the end of image is reached. After block decryption, the blocks are descrambled first vertically and then horizontally to get the final output as plain image.

Fig 4: Decryption Procedure of DICE Algorithm

V.

Fig 3: Encryption Procedure of DICE Algorithm

IV.

DICE DECRYPTION PROCESS

Decryption process of DICE cipher is exactly reverse of the encryption method, because it is based on symmetric methodology. The steps of decryption are listed as follows: i. The pixels of encrypted image are read from the received file and their corresponding RGB value is measured in binary form. ii. The respective secondary key is read from the central database server. iii. Reverse binary operations are performed on the cipher text with help of the secondary key. iv. Steps ii to iii are performed for all pixels till last pixel of the encoded image is reverted back to its previous state. v. Then the corresponding primary key is selected from the central server for first block.

PERFORMANCE EVALUATION OF DICE

The most prominent feature of DICE algorithm is that it is not fully dependent on the encryption key as a supplementary secondary key is used to give an extra security edge. Moreover, DICE algorithm keeps changing the secret key from a preselected range of delimiters. This feature makes it nearly impossible for a cracker to deduce the key, with help of replay attacks. A chief predicament in image encrypting algorithms is the strong association between neighboring pixels due to smooth value changes in large areas of digital images[9]. DICE cryptosystem handles this concern pretty well and is designed to use dissimilar keys (i.e. secondary key at pixel level) to encrypt varied pixels and hence resolving the spatial problematic issue. Another predicament that arises in image encryption is high redundancy i.e. same color of encrypted pixels that have same color in original image[3][8]. This behavior will make it easier for the cryptanalyst to decode the given image. These problematic issues of digital images are not a concern in DICE algorithm as it encodes not only at chunk level but also at pixel level of image, hence achieving crucial security. Further, it does shuffling of blocks

before encryption, making it harder to crack than traditional image encryption ciphers. The above mentioned features of DICE cipher make it safe and secure than traditional image cryptosystems.

image. To testify this issue, an image with variety of colours is chosen as candidate for input to DICE.

A. Timing Analysis The DICE algorithm is consciously planned in such a way to accomplish major aim i.e. the speed of encryption and decryption of data[10]. Code optimization techniques are exercised to achieve greater encryption speed. During the performance analysis of DICE, the following results of encryption and decryption are obtained. Input Size (Pixels)

TABLE 1. Performance analysis of DICE Encryption Decryption Total Execution Time (sec) Time (sec) Time (sec)

240 * 240

0.75

0.75

1.5

480 * 480

3.0

3.0

6.0

720 * 720

5.5

5.5

11.0

960 * 960

15

15

30

B. Memory Analysis The following table shows memory prerequisite of DICE algorithm in comparison to other image encryption algorithms. Encryption Algorithms MIE

Fig 5: Input Image taken for DICE Algorithm

TABLE 2. Memory requirement of DICE Key Length Plaintext Size Cipher Text (Bits) (Bits) (Bits) -

9962344

9962200

VC

-

196610448

-*

DICE

24

32

32

From the above table it is clear that DICE memory requirements are negligible as compared to popular image encryption algorithms such as Mirror-like Image Encryption (MIE) and Visual Cryptography (VC). Visual cryptography exploits the human visual system to read the secret message from some overlapping shares[11].Thus it can be observed that DICE has greater encryption speed and lower memory requirements than recognised image encryption standards. VI.

CASE STUDY

Here is an example showing the screen shots for the plaintext, cipher text and encryption process of DICE algorithm applied on a 2-dimensional digital image. A major concern in image encryption is that if eavesdropper intercepts the message. Then, he may be able to see the ciphered image; and he maybe able to guess the colour and shapes of some objects in the

Fig 6: Output of DICE Algorithm

VII. CONCLUSION & FUTURE WORK In this paper, a new block cipher for two-dimensional digital images has been proposed. The algorithm is based on symmetric key approach and it has some special security feature of using an additional key to make it partially dependent on the encryption key. In this algorithm, we divide image into blocks and scramble

them to add confusion. Then these blocks are further encrypted by means of primary encryption key, followed by pixel level encryption using a secondary key. Hence, the encryption process involves three levels of security and thus making it tougher against unauthorized attacks. Experiments show that algorithm fully encrypts twodimensional digital images and original images are also reconstructed without any distortion. The maximum distortion that can be noticed is 1 out of 1000 pixels which is far better than other established image encryption standards. Moreover, same size of encrypted and decrypted image proves least distortion in the image. Performance analysis show that memory requirements of proposed cipher is quite low. Last but not the least, most vital facet of DICE algorithm is its ultimate security which is the main aim to design any cryptosystem. Finally, it can be concluded that DICE cryptosystem proves a competent technique for digital image encryption, and is a key for secure transfer of images. REFERENCES [1]

A. J. Menezes, P. C. Van Oorschot, and S. A. Vanstone Handbook of Applied Cryptography. New York: CRC Press, Inc., 1997. [2] B. Schneier, Applied Cryptography : Protocols, Algorithms, and Source Code in C, John Wiley & Sons, Inc., New York, USA, 1994. [3] C. Li, S. Li, M. Asim, J. Nunez, G. Alvarez, G. Chen, On the security defects of an image encryption scheme, Image and Vision Computing, p. 1371–1381, 2009. [4] L. Shujun and X. Zheng, "Cryptanalysis of a chaotic image encryption method," Inst. of Image Process. Xi'an Jiaotong Univ. Shaanxi, IEEE International Symposium, Vol. 2, p. 708-711, 2002. [5] M. Jakobsson, J.P. Stem, and M. Yung, "Scramble All, EncryptSmall, http://www.julienstern.org/files/scramble/scramble.html, visited in December 2002. [6] N. Pareek, V. Patidar, K. Sud, Image encryption using chaotic logistic map, Image and Vision Computing 24 (9), p. 926–934, 2006. [7] P.P Charles & P.L Shari, “Security in Computing: 4th edition”, Prentice-Hall, Inc, 2008. [8] P.P. Dang, P.M. Chau, Image encryption for secure Internet multimedia applications [J].IEEE Transactions on Consumer Electronics, 46(8), p.395- 403, 2000. [9] R. Lukac, N. Konstantinous Plantaniotis, "A cost-effective encryption scheme for color images," in Science Direct real time imaging, p.454-464, 2005. [10] W.Stallings, “Cryptography and network security principles and practice,” Fourth edition, Prentice hall, 2007. [11] Y.C. Hou, “Visual cryptography for color images”, Pattern Recognition 36, www.elsevier.com/locate/patcog, 1619-1629, 2003.