The Application of Chaotic Mapping Based On Finite Field Operation in Digital Image Encryption
The Application of Chaotic Mapping Based On Finite Field Operation in Digital Image Encryption
Abstract—A new digital image encryption method is encryption algorithm according to the initial sensitivity of
proposed by introducing a high-dimensional chaotic mapping chaotic system[12]. Singh combined the two chaotic mappings
and combining matrix operation in finite field. The original to make the encryption process of digital images more robust
image is mapped into two dimensional matrices by using chaotic against attack, and he proved that the encryption performance
mapping, and then the original image matrix and the previously of this method had been improved significantly through
generated two matrices are successively subjected to a various anti-attack performance tests[13]. Chong used the
multiplication operation in finite field. The experimental results Logistic 3D chaotic system as the scrambling model of image
show that the proposed encryption method can obtain more encryption, and adopted the registration mechanism to encrypt
uniform and detailed image scrambling effect than the
the decryption key, which made the digital image encryption
traditional method, has stronger anti-jamming performance and
faster execution efficiency, and has better performance in image
method extremely safe[14]. Zhu proposed a parallel image
secure storage and communication Value. encryption strategy based on hyperchaotic system optimization
sequences combined with ciphertext interlaced diffusion, its
Key words—digital image; image encryption; chaotic mapping; ciphertext interlacing duffusion ability was strengthened[15].
finite field operation In view of the successful application of chaotic theory in
I. INTRODUCTION the field of digital image encryption, this paper introduces a
new 3D chaotic mapping to construct a digital image
Compared with other types of data, digital images possess encryption algorithm based on this mapping to obtain a better
such advantages as carrying more information and not image encryption effect.
substantially limited by the identification subject. The image
data transmitted through network is exploding. In order to II. T MAPPING
transmit it more securely and conveniently, researchers have In the field of digital image encryption, Arnold chaotic
been studying various image encryption technologies, and a lot mapping, Logistic chaotic mapping and Lorenz chaotic
of successful usages have been achieved. With time going on, mapping are relatively successful in application[16]. In order to
these methods have been gradually deciphered, its safety and further enhance the anti-attack ability of the subsequent image
reliability are greatly reduced[1], which leads to more complex encryption algorithm, a new 3D chaotic mapping is proposed
encryption methods entering the field of image encryption[2,3]. by Prof. TANG[17], which is called T mapping in this paper.
In recent years, the encryption method based on chaotic The Kinetic equation is shown as (1):
theory and chaotic system has been widely used in digital
image information encryption[4-8]. Chaotic theory has its
dx
dt = −ax + by
inherent advantages for the encryption of digital image
information. On the one hand, it is difficult to describe the
chaotic mapping with definitive rules. On the other hand, the
dy
chaotic system is strongly sensitive to the initial value, which = cx − xz − dy (1)
makes it difficult for decoders to construct decryption model dt
and accurately confirm decryption key, thereby its encryption dz
security and reliability have been greatly enhanced[9,10]. dt = xy − e( x + z )
To the digital image encryption method based on the
chaotic mapping, there are many successful cases. On the basis According to the proof of literature [17], T mapping has all
of the basic chaotic mapping, Elabady adopted a dynamic the common characteristics of chaotic systems: certainty,
update mechanism which further enhances the scrambling boundedness, extreme sensitivity to the initial value, long-term
degree of encryption process and upgrade the implementation unpredictability, the largest plus Lyapunov exponent, the
speed of encryption algorithm[11]. Banerjee constructed the continuous spectrum in certain frequency ranges, and
chaotic encryption flow framework of digital image ergodicity, etc. If the parameters in equation (1) are
information and scrambled the original image information with “ a ∈ [10,30 ] , b ∈ ( 22,70] , c ∈ (13.8,40] , d ∈ [0,0.5] ,
chaotic mapping, which increased the reliability of the whole e ∈ (0.9,1.3] ∪ (1.5,5) ”, the system is in a chaotic state, and
Then, Y and Z in equation (2) are converted into 2D image In equation (4), E is the final encrypted image.
matrixes. The mathematical processing is shown in equation
C. Decryption Process Design
(3).
In fact, decryption is the inversion of encryption. The
YM × N = Y decryption process is as follows:
(3)
Z M ×N = Z Step 1: Initial state is set according to T chaotic mapping;
In equation (3), M and N represent the width and height Step 2: Chaotic sequences Y and Z are generated according
of the original image. to T mapping and the encrypted image matrix E;
B. Finite Field Processing Step 3: Transform Y and Z into image matrix YM ×N and
Traditional image encryption methods often adopt the Z M ×N .
operation of XOR, SOR or modular addition, but they are
weak in security. To this end, the method proposed in [15] can Step 4: Calculate original image O according to equation
enhance the diffusion capacity of ciphertext interlacing to a (5).
certain extent. However, due to the use of XOR or modulo −1 ' −1
addition in this method, and its non-linear operation is weak, it O = (E ⊙ Z M × N ) ⊙ (YM × N ) (5)
still lacks the ability of attacking plaintext and differential
attack, so the encryption result is still not very ideal, the IV. EXPERIMENTAL RESULTS AND ANALYSIS
encryption algorithm based on the finite field multiplication is In order to verify the effectiveness of this method, two
proposed in this paper. By encrypting the pixels, the grayscale images are selected from the international common
correlation and non-linearity between plaintext and ciphertext test library, one is Lena image and the other is Landscape
can be improved to improve the anti-attack ability of the image. Computer hardware configuration: Core Duo, clocked
encryption result. at 2.0GHz CPU, 4G size of memory. Computer software
configuration: windows 7.0 operating system, the compiler
In order to enhance the complexity and validity of the
environment for Matlab2014b. The experimental results are
whole encryption process, a finite field matrix operation is shown in figure 1 and figure 2.
further performed on the basis of T mapping. In this paper, we
( )
introduce the operations of group Z q• ,⊙ ,Where the finite
From the visual point of view, figure 1 and figure 2 In the above equation, x and y are the values of two
qualitatively prove the effectiveness of the encryption method adjacent pixels in the image. cov( x, y ) and D (x) are
in this paper. After the original image is encrypted, a very defined as follows:
good uniform scrambling effect is shown. The detailed
information of the original image can not be seen completely, m×n
∑ (x
1
and the decrypted image has no difference from the original cov( x, y ) = i − E ( x))( y i − E ( y ))
m×n i =1
one. The quantitative analysis is to be carried on from the (7)
following five aspects. m×n
∑ (x
1 2
D ( x) = i − E ( x))
A. Key Space and Execution Time Analysis m× n i =1
The encryption method in this paper uses five parameters m×n
∑x
1
of the chaotic system as the original key and is represented by Where E ( x) = i .
a double precision real number of 15 decimal places. Therefore, m×n i =1
the key space can be 1015×1015×1015×1015×1015=1075≈ The correlation of neighborhood pixels in this paper
2240, which is equivalent to the key length of 240 bits. In the involves three kinds of pixel correlation. They are the adjacent
existing hardware conditions, it is fully able to cope with the pixel correlation in the horizontal, vertical, and diagonal
ability to resist the exhaustive attack. direction respectively.
For Lena images and Landscape images, the encryption Using the above equation, the correlation coefficients of
method proposed in this paper obtains the ideal encryption the Lena image and the Landscape image before and after
effect and anti-attack performance. Next, the execution time of encryption are calculated, which are shown in table 2.
this method is tested. The results are shown in table 1.
Table 2 The table of correlation coefficients in three
Table 1 Time analysis of encryption and decryption directions before and after encryption
∑p
test library with "(a, b, c, d, e) = (22, 65, 32, 0.3, 1.4)" as the
Let p i = P{S = s i } , if i = 1 , then the entropy of “S” is encryption key, then the encrypted image is encrypted with a
i =1 slightly different key decryption(the decryption key changes
defined as: only one of the initial variables 10-10 at a time). Figure 5
n respective shows the decrypted Cameraman image using the
H (S ) = − ∑p
i =1
i ln p i (8 ) correct and wrong key.
(a)Before encryption (b)After encryption Table 3 The mean square error between the decrypted
image and the original image
Figure 3 Evaluation of gray scale histogram of Lena image This paper
Decryption key Literature [17]
algorithm
The correct key 0.00 0.00
δa ≈ 10−10 9408.26 9317.62
δb ≈ 10−10 9352.67 9412.38
Error
δc ≈ 10−10 9418.46 9358.29
value
δd ≈ 10−10 9423.35 9317.42
δe ≈ 10−10 9339.24 9436.82
For the Cameraman image, the first column of table 3
shows the mean square error between the original image and
the decrypted image of the correct key and five sets of error
(c)Before encryption (d) After encryption keys. The results show that the correct key can achieve
complete and accurate decryption. And the image decrypted
Figure 4 Evaluation of gray scale histogram of Landscape image
with a small error decryption key will have a large difference
from the original image. This reflects the sensitivity of the
It can be seen from figure 3 and figure 4 that the histogram
algorithm to the key. The third column in table 3 respectively
distribution of the two images encrypted with this method is
shows the mean square error between the decrypted image
very uniform, which proves that the encryption security is
obtained from the literature[17] and the original image. In
high.
contrast, the algorithm is more sensitive to the initial keys a, c,
and d, while the literature[17] is more sensitive to the initial
keys b and e. This is due to the difference in the characteristics [5] LIU Lepeng, ZHANG Xuefeng, Image encryption algorithm based on
of the two chaotic systems. chaos and bit operations. Journal of Computer Applications, 2013,
33(4): 1070-1073
V. CONCLUSION [6] YANG T, Research on VANET privacy protecting based on proxy
signature. Beijing: Peking University, 2013.
The encryption algorithm proposed in this paper is a [7] ZHOU X, A novel chaotic system and its circuit simulation. Acta
scrambling one combined with finite field operation. In this Physica Sinica, 2012, 61(3): 030504
paper, the chaotic mapping is T mapping. With this method, [8] Yong WANG, etc., A chaos-based image encryption algorithm with
the original image is scrambled into two chaotic matrices, then variable control parameters. Chaos Solitons Fractals, 2009
the original image and the two chaotic matrices are [9] H.J. Shiu, K.L. Ng, J.F. Fang, R.C.T. Lee, C.H. Huang, Data hiding
successively subjected to a multiplication operation in finite methods based upon DNA sequences, Information Sciences, 2010, 180
field, the encrypted image is achieved finally. In the simulation (11): 2196–2208.
experiment, Lena image, Landscape image and Cameraman [10] SUN Jinguang, WANG Jie, MENG Xiangfu, Improved Image
Encryption Algorithm Based on Double Scrambling of Fibonacci
image are tested, analyzed and compared from five aspects: Transforms. Computer Science, 2012, 39(11): 249-253
key space and execution time, correlation, information entropy, [11] Elabady N.F, Abdalkader H.M, Moussa M.I, Sabbeh S.F, Image
histogram evaluation and key sensitivity analysis. The encryption based on new one-dimensional chaotic map. ICET
experimental results show that the proposed method possesses 2014-2nd International Conference on Engineering and Technology,
better encryption effect, stronger anti-attack performance and 2015, 20: 851-892.
excellent execution efficiency. It is suitable for different types [12] S.Banerjee, etc., Synchronization of spatiotemporal semiconductor
of images, and has certain practical value in the field of high lasers and its application in color image encryption. Optics
Communications, 2009,284(9):2278-2291.
security requirements such as confidential storage and
transmission etc. [13] Singh H, Yadav A.K, Vashisth S, Singh K, Double phase-image
encryption using gyrator transforms, and structured phase mask in the
frequency plane. Optics and Lasers in Engineering, 2015, 67: 145-156.
REFERENCES
[14] Chong Fu, Zhiliang Zhu, A chaotic image encryption scheme based on
[1] ZHAO Wenbo, TIAN Xiaoping, WU Cheng-mao, Image encryption circular bit shift method. The 9th International Conference for Young
based on low density parity check coding and chaotic system. Journal Computer Scientists, 2013: 3057–306l.
of Computer Applications, 2012, 32(7): 2018-2021
[15] Zhu Congxu, Hu Yuping, Sun Kehui, New Image Encryption
[2] YUE Honghui, LI Tao, SHI Lei, Improved image encryption algorithm Algorithm Based on Hyperchaotic System and Ciphertext Diffusion in
based on Henon hyperchaotic system. Journal of Computer Crisscross Pattern. Journal of Electronics & Information Technology,
Applications, 2011, 31(7): 1909-1906 2012, 34(7):1735-1743
[3] LIU Shihua, WANG Deshi, CHEN Long, Analysis of the Ambiguity [16] Song Wei, Hou Jianjun, Li Zhaohong, Huang Liang, A novel zero-bit
Characteristic of Digital Synthesis Signals with Chaotic Frequency watermarking algorithm based on Logistic chaotic system and singular
Modulation. ACTA ELECTRONICA SINICA, 2007, 35(9): value decomposition. ACTA PHYSICA SINICA, 2009, 58(7): 4449-08
1784-1788
[17] Tang LiangRui, Li Jing, Fan Bing, Zhai Mingyue, A new
[4] ZHAO Feng, Wu Chengmao, Image Encryption Algorithm Combined three-Dimensional chaotic system and its circuit simulation. Acta
Self-Encoded Theory with Super-Chaotic Mapping n. Journal of Physica Sinica, 2009, 58(02): 0785-09
Computer-Aided Design & Computer Graphics, 2016, 28(1): 119-128