A Multi-Level Blocks Scrambling Based

Chaotic Image Cipher

Musheer Ahmad and Omar Farooq

Department of Computer Engineering,

ZH College of Engineering and Technology
A. M. U., Aligarh-202 002, India
Department of Electronics Engineering,
ZH College of Engineering and Technology
A. M. U., Aligarh-202 002, India

Abstract. In this paper, an image encryption scheme based on multi-

level blocks scrambling is proposed. The image is first decomposed into
non-overlapping blocks and scrambling of these blocks is done by using
2D Cat transform. Multi-level blocks scrambling (MLBS ) is implemented
by starting with a large block size and the size of blocks gets reduced it-
eratively at each level. The scrambling of blocks is performed at multiple
levels to get cummulative effect. At each level, the control parameters
of scrambling are randomly generated through 2D coupled Logistic map
to make the process of scrambling key dependent. The scrambled image
obtained after carrying out MLBS scrambling is encrypted using chaotic
sequence generated by one-dimensional Logistic map. The experimental
results show that the proposed encryption scheme can successfully en-
crypts/decrypts the images with same secret keys. The algorithm has
large key space and high sensitivity to a small change in secret keys. The
simulation analysis also demonstrates that the ciphered images have high
information entropy, very low correlation coefficients and uniform gray
level distribution.

Keywords: Information security, image encryption, image scrambling,

logistic map, information entropy.

1 Introduction
In recent years, the chaos based cryptographic algorithms have suggested some
new and efficient ways to develop secure image encryption techniques in order
to meet the demand for real-time image transmission over the communication
channels. Chaotic systems are considered good for practical image encryption
and transmission because they have important characteristics such as sensitiv-
ity to initial conditions and system parameters, pseudo-random property and
non-periodicity as the chaotic signals usually noise-like, etc. The characteristics
of chaotic signals make image cryptosystems excellent and robust against any
statistical attacks. Therefore, chaos based image encryption is given much atten-
tion in the research of information security. A lot of image encryption algorithms

S. Ranka et al. (Eds.): IC3 2010, Part I, CCIS 94, pp. 171–182, 2010.

c Springer-Verlag Berlin Heidelberg 2010
172 M. Ahmad and O. Farooq

based on chaotic systems have been proposed [1,2,3,4,5,6,7,8,9,10,11,12,13].

These image encryptions algorithms are based on chaotic maps like the Logistic
map [5,6,7], the Standard map [8], the Baker map [9,10], the Piecewise nonlinear
chaotic map [11] the Cat map [12,13], the Chen map [6,13] etc.
The most widely used methodology to encrypt an image is to use the concept
of scrambling the positions of pixels in the plain-image and changing the gray
values of pixels in the scrambled image. The purpose of image scrambling is to
transform a meaningful image into a meaningless or distorted image to enhance
the security [14]. This can be done by permuting the coordinates of image pixels
to achieve the visual effect of disorder [15]. A good scrambled image prevents
human visual or computer vision system from understanding the real meaning
of the original image. In the past decade, many scrambling schemes have been
proposed; these schemes are based on Cat transform [12,13,16,17,18,19], Lorenz
system [20,21], Logistic map [6,7], Chebyshev map in DCT domain [22] etc. In
these scrambling schemes, all the pixels of an image are permuted to achieve
image scrambling. The application of single iteration of these methods needs to
perform a lot of computations.
In this paper, an image encryption scheme is proposed which is based on
novel concept of multi-level blocks scrambling (MLBS ). MLBS scrambling is
performed by scrambling the blocks of image at multiple levels rather than scram-
bling all pixels. MLBS scrambling scheme reduces the number of computations
required to achieve good scrambling effect as compared to the conventional pix-
els scrambling. Two chaotic maps are used to perform MLBS scrambling: one
map is employed to carry out scrambling of blocks while the control parameters
of scrambling are randomly generated by other map. To enforce the security of
the proposed scheme, the encryption of scrambled image obtained after MLBS
phase is performed using a third chaotic map.
The paper is organized as follows: Section 2 describes the different chaotic
maps used. Section 3 discusses the proposed MLBS scrambling based image
encryption algorithm in detail. The experimental results are discussed in Section
4 followed by conclusions in Section 5.

2 Chaotic Maps
According to the chaos theory, chaotic maps are the nonlinear dynamical systems
whose state evolves with time and are highly sensitive to initial conditions. The
future dynamics of these systems are fully defined by their initial conditions. As a
result, the behavior of these systems appears random. In the proposed MLBS based
image encryption algorithm, the following three distinct chaotic maps are used:

2.1 Two-Dimensional Coupled Logistic Map

The two-dimensional coupled Logistic map [23] is described as follows:

xn+1 = μ1 xn (1 − xn ) + γ1 yn2
(1)
yn+1 = μ2 yn (1 − yn ) + γ2 (x2n + xn yn )
 A Multi-Level Blocks Scrambling Based Chaotic Image Cipher 173

Three quadratic coupling terms introduce strength to the complexity of 2D

Logistic map. This map is chaotic when 2.75 < μ1 ≤ 3.4, 2.7 < μ2 ≤ 3.45,
0.15 < γ1 ≤ 0.21, 0.13 < γ2 ≤ 0.15 and generate chaotic sequences x, y ∈ (0, 1).
To examine the statistical performance, the map is iterated for 16000 times with
initial conditions/parameters as: x0 = 0.0215, y0 = 0.5734, μ1 = 2.93, μ2 = 3.17,
γ1 = 0.197 and γ2 = 0.139. The statistical analysis of x and y sequences shows
that they have poor balance, autocorrelation and cross-correlation properties.
The mean values of the sequences x and y are 0.6456 and 0.6590 respectively.
To improve these statistical properties of the sequences generated by 2D map,
the following pretreatment is done.

xi = 106 xi − f loor(106 xi )
(2)
yi = 106 yi − f loor(106 yi )
The mean values after preprocessing of the sequences x and y are 0.5004 and
0.4984 respectively, which are closer to the ideal value 0.5. Now, the preprocessed
sequences have better balance distribution, auto as well as cross correlation
properties and they can be utilized in a cryptographic process.

2.2 2D Cat Transform

A 2D Cat map is first presented by V. I. Arnold in the research of ergodic theory.

Let the coordinates of positions of pixels in an image are P = {(x, y) | x, y = 1,
2, 3, . . ., N}. A 2D Cat map with two control parameters [12] is as follows:

x = (x + ay)mod(N )
(3)
y = (bx + (ab + 1)y)mod(N )
Where, a, b are control parameters which are positive integers. The Cat map
translates the pixel position (x, y) of plain-image of size N × N to a new position
(x, y) when applied once to plain-image. Cat transform permutes the organiza-
tion of pixels of plain-image by replacing the position of the image pixel points
with new coordinate. After several iterations, the correlation among the adjacent
pixels is disturbed completely and the image appears distorted and meaningless.
Due to the periodicity of the Cat transform it is possible to recover the original
image after iterating Cat transform many times [13]. To deal with this problem
of Cat transform, a block based image scrambling is performed at multiple levels
in which the control parameters a, b of transform are randomly generated by 2D
coupled Logistic map at each level of MLBS scrambling.

2.3 One-Dimensional Logistic Map

The one-dimensional Logistic map proposed by May [24] is one of the simplest
nonlinear chaotic discrete systems that exhibit chaotic behavior and is defined
by the following equation:

zn+1 = λzn (1 − zn ), n ≥ 0 (4)

174 M. Ahmad and O. Farooq

where zo is initial condition, λ is the system parameter and n is the number of

iterations. The research shows that the map is chaotic for 3.57 < λ < 4 and
zn+1 ∈ (0, 1) for all n. The sequence generated from Eq. 4 does not require any
pretreatment as the sequence has random-like behavior. The sequence generated
by Logistic map is used to improve the statistical properties of the scrambled
image obtained after MLBS scrambling phase.

3 Proposed Image Encryption Algorithm

Statistical analysis of images shows that the image data have strong correlations
among adjacent pixels. For image security and secrecy, one has to disturb this
correlation. In the proposed algorithm, the average correlation among the ad-
jacent pixels of plain-image is reduced by first scrambling larger image blocks
and then iteratively reduces the block size at each level to achieve the desired
scrambling result. At level-k, the blocks of size 2L-k+1 × 2L-k+1 of image ob-
tained after applying previous level scrambling are scrambled to get cumulative
effect, where L = log2 (N ) - 1 for a plain-image of size N × N and k varies
from 1 to L. The multi-level blocks scrambling (MLBS ) scheme makes use of 2D
Cat transform in which the control parameters are randomly generated through
two key dependent sequences obtained from 2D coupled Logistic map. The con-
trol parameters of Cat transform are the control parameters of scrambling in
MLBS phase. The sequences obtained from 2D Logistic map are first prepro-
cessed through Eq. 5 and then the control parameters a, b are evaluated. The
process of random generation of control parameters is as follows:
First the map given in Eq. 1 is iterated t = 500 times with initial conditions:
x0 = 0.0215, y0 = 0.5734, μ1 = 2.93, μ2 = 3.17, γ1 = 0.197 and γ2 = 0.139,
these 500 values of x and y are discarded. The map is again iterated to produce
next xi and yi values, where i = 1, 2, . . ., L. The control parameters ai and bi
of MLBS scrambling from xi and yi are calculated as:
⎫
Ψ (xi ) = 1014 (106 xi − f loor(106 xi )) ⎪⎪
⎪
Ψ (yi ) = 1014 (106 yi − f loor(106 yi )) ⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎬
φ(xi ) = (Ψ (xi )mod(83)) + 17
(5)
φ(yi ) = (Ψ (yi )mod(107)) + 19 ⎪ ⎪
⎪
⎪
⎪
⎪
⎪
⎪
ai = (Ψ (xi )mod(φ(yi )) + 1 ⎪
⎪
⎭
bi = (Ψ (yi )mod(φ(xi )) + 1
The control parameters ai , bi of scrambling are made sensitive to secret keys of
2D Logistic map. As a result, the MLBS scheme becomes sensitive to a small
change in secret keys. Thus, the attacker cannot make use of Cat transform’s pe-
riodicity to obtain the plain-image without secret keys. Moreover, the statistical
properties like gray value distribution, mean value of gray values, information
entropy etc of an image do not change after scrambling. The attacker can make
use of these properties of scrambled image to recover the plain-image. To further
 A Multi-Level Blocks Scrambling Based Chaotic Image Cipher 175

improve the security provided by the algorithm, these statistical properties of

scrambled image needs to be improved. The statistical properties are improved
by encrypting the scrambled image using chaotic key sequence generated from
1D Logistic map. The complete procedure of proposed image encryption scheme
is as follows:

P(x, y): Plain-image of size N × N, where x, y = 1, 2, 3, . . ., N.

x0 , y0 , z0 : Secret keys of algorithm.
μ1 , μ2 , γ1 , γ2 , λ : Initial parameters of chaotic maps.
C(x, y): Ciphered image of size N x N obtained as output of algorithm.

Step 1. [Initialization]: Set I(x, y) = P(x, y), k = 1, L = log2 (N ) - 1, m =

L.
Step 2. Repeat Steps 3 to 6 while k ≤ L.
Step 3. [Decomposition]: Decompose the whole image I(x, y) into 2m × 2m size
blocks.
Step 4. [Control Parameter]: Generate control parameters ak and bk for level-k
scrambling using 2D coupled logistic map and Eq. 5.
Step 5. [Level-k Scrambling]: Apply Cat transform to scramble the blocks of size
2m × 2m within the image I(x, y) using control parameters ak and bk . We get
Sk (x, y) scrambled image as output of level-k scrambling process.
Step 6. Set I(x, y) = Sk (x, y), k = k + 1 and m = m - 1.
Step 7. [Encryption Variable]: Iterate 1D logistic map to get discrete variable zi
and evaluate key K(i) as:

K(i) = (1014 zi )mod(256)

Step 8. [Binary Conversion]: Convert the decimal gray values (0 to 255) of scram-
bled image SL (x, y) and key K(i) to equivalent 8-bit binary numbers as:

SB(x, y) = DecimalToBinary(SL (x, y)) .

KB(i) = DecimalToBinary(K(i)) .

The function DecimalToBinary(x) converts the decimal number x to 8-bit bi-

nary number and (x)mod(y) returns the remainder whenever x is divided by y.
Step 9. [Encryption]: Encrypt the scrambled image SL (x, y) using the key K(i)
as:

CB(x, y) = SB(x, y) ⊕ KB(i) .

Where i = N × (x − 1) + y; and x, y = 1, 2, 3, . . ., N. The CB(x, y) is the

8-bit binary equivalent of gray value of the ciphered image C(x, y) with pixel
coordinate (x, y). The symbol ⊕ represents the exclusive-OR operation bit by
bit.
Step 10. [Repeat]: Steps 7 to 9 are repeated for N x N times to encrypt all the
pixels of the scrambled image.
176 M. Ahmad and O. Farooq

Step 11. [Decimal Conversion]: Convert all the binary numbers CB(x, y) to dec-
imal numbers to get resultant Ciphered image C(x, y) as:

C(x, y) = BinaryToDecimal(CB(x, y))

The function BinaryToDecimal(x) converts an 8-bit binary number x to its

equivalent decimal number.
The plain-image can be recovered successfully by applying the proposed al-
gorithm in reverse order. The block diagram of the proposed image encryption
algorithm is shown in Fig. 1.

Fig. 1. Proposed MLBS based image encryption algorithm

4 Experimental Results
The proposed encryption scheme is applied to various plain-images like Lena,
Baboon, Peppers, Cameraman, Barbara, Boat and Airplane of 256 × 256 size.
Among them, the plain-images of Lena, Baboon, Cameraman, Boat and their
respective histograms are shown in Fig. 2. As can be seen in this figure that the
gray value distributions of plain-images are not uniform. The initial conditions
and system parameters taken for experimentation are: x0 = 0.0215, y0 = 0.5734,
z0 = 0.3915, μ1 = 2.93, μ2 = 3.17, γ1 = 0.197, γ2 = 0.139 and λ = 3.9985. The
results of multi-level blocks scrambling (MLBS ) scheme are shown in Fig. 3.
 A Multi-Level Blocks Scrambling Based Chaotic Image Cipher 177

800
1000 1000
800
700 900 900
700
800 800
600
600
700 700
500
500 600 600
400 500 500
400

300 400 400

300
300 300
200 200
200 200
100 100 100 100

0 0 0 0

0 50 100 150 200 250 0 50 100 150 200 250 0 50 100 150 200 250 0 50 100 150 200 250

Fig. 2. Plain-images of Lena, Baboon, Cameraman, Boat and their histograms

The ciphered images of Lena, Baboon, Cameraman and Boat using proposed
encryption algorithm and their histograms are shown in Fig. 4 respectively. It is
clear from the Fig. 4 that the ciphered images are very much indistinguishable
and appears like a noise. Moreover, as seen in Fig. 4 that the distribution of
gray values of the ciphered images is fairly uniform and much different from
the histograms of the plain-images shown in Fig. 2. The statistical property of
the plain-images is changed absolutely in ciphered images in such a manner that
ciphered images have good balance property. To quantify the balance property of
the ciphered images, the mean value of gray values of plain-images and ciphered
images are evaluated and listed in Table 1. It is evident from the mean values
obtained that no matter how gray values of plain-images distributed, the mean
value of ciphered images comes out to about 127. This shows that the ciphered
image doesn’t provide any information regarding the distribution of gray values
to the attacker. Hence, the proposed algorithm can resist any type of histogram
based attacks and strengthen the security of ciphered images significantly.

4.1 Key Space of Cipher

Key space is the total number of different keys that can be used in the crypto-
graphic system. A cryptographic system should be sensitive to all secret keys.
There are total eight initial conditions of chaotic map used in the algorithm.
The initial conditions for x0 , y0 , and z0 used as secret keys of encryption and

Fig. 3. Scrambled images of Lena, Baboon, Cameraman and Boat using MLBS scheme
178 M. Ahmad and O. Farooq

600 600 600 600

500 500 500 500

400 400 400 400

300 300 300 300

200 200 200 200

100 100 100 100

0 0 0 0

0 50 100 150 200 250 0 50 100 150 200 250 0 50 100 150 200 250 0 50 100 150 200 250

Fig. 4. Ciphered images of Lena, Baboon, Cameraman, Boat using proposed scheme
and their histograms

decryption. In the proposed cryptosystem, all the variables are declared as type
double which has a 15-digit precision, the decimal fractions of the variables are
multiplied by 1014 , therefore the key space of proposed image cipher is (1014 )3
i.e. 1042 ≈ 2140 . The initial conditions to parameters μ1 , μ2 , γ1 , γ2 and λ also
increase the size of key space. Thus, the key space of the cipher is extensively
large enough to resist the exhaustive attack.

4.2 Sensitivity to Keys

A good cryptosystem should be sensitive to a small change in secret keys i.e. a
small change in secret keys during decryption process results into a completely
different decrypted image. Proposed encryption algorithm is sensitive to a tiny
change in the secret keys. Any change in the initial conditions: even of the order of
(10−14 ), a decrypted image is totally different from plain-image. As an example,
consider the ciphered image Lena shown in Fig. 4, if one of the initial conditions,
say x0 , y0 and z0 is changed a little (10−14 ) then the respective decrypted images
obtained are shown in Fig. 5. The images shown in Fig. 5 are totally different
from the plain-image shown in Fig. 2. Further, the decrypted images appear like

Table 1. Mean values of Original and Ciphered Images

Test Images Original Ciphered

Lena 124.09 127.48
Baboon 129.41 127.14
Peppers 123.10 127.56
Cameraman 128.72 127.01
Barbara 127.39 127.42
Boat 129.71 127.28
Airplane 129.17 127.59
 A Multi-Level Blocks Scrambling Based Chaotic Image Cipher 179

Fig. 5. Key sensitivity: decrypted image with x0 = 0.02150000000001; decrypted image

with y0 = 0.57340000000001 and decrypted image with zo = 0.39150000000001

a noise. Similar sensitivity is noticed for the case of wrong μ1 , μ2 , γ1 , γ2 and λ.

Hence, the proposed encryption scheme can be said to be highly sensitive to a
small change in secret keys.

4.3 Computational Analysis

Image scrambling is carried out by re-locating the pixels of the image from
one position to another. In most of the existing image scrambling algorithms
[6,7,12,13,16,17,18,19,20,21,22], many iterations of these algorithms are applied
to get good scrambling effect. In order to evaluate computation required to
scramble an image using Cat transform, we calculate the number of times the
Cat transform given in Eq. 3 is executed. If only single iteration of Cat transform
is applied to scramble all the pixels of an image of size 256 × 256, it needs to
perform 65536 computations of Cat transform. However, only 2×2 + 4×4 + 8×8
+ 16×16 + 32×32 + 64×64 + 128×128 = 21844 computations of Cat transform
are required to scramble the image of same size using proposed MLBS scheme.
This results in a saving of about 66.6% computations required to scramble an
image using the proposed scrambling scheme as compared to the pixel level
scrambling used in [12,13,16,17,18,19].

4.4 Correlation Coefficient Analysis

In order to evaluate the encryption quality of the proposed encryption algo-
rithm, the correlation coefficient is used. The correlation coefficients between
two vertically, horizontally and diagonally adjacent pixels of a ciphered image is
calculated as [9].
N N N
N i=1 (xi × yi ) − i=1 xi × i=1 yi
ρ=   N N N (6)
N
(N i=1 x2i − ( i=1 xi )2 ) × (N i=1 yi2 − ( i=1 yi )2 )

Where x and y are gray values of two adjacent pixels in an image. Correla-
tion coefficients of 1000 pairs of vertically, horizontally and diagonally adjacent
randomly selected pixels are evaluated. The average of 50 such correlation co-
efficients of adjacent pixels in plain and ciphered images in three directions are
listed in Table 2 and Table 3 respectively. The values of correlation coefficients
show that the two adjacent pixels in the plain-images are highly correlated to
180 M. Ahmad and O. Farooq

Table 2. Correlation Coefficient of two adjacent pixels in Original images

Test Images Vertical Horizontal Diagonal

Lena 0.95967 0.92479 0.90236
Baboon 0.79835 0.84413 0.75661
Peppers 0.95807 0.95513 0.91878
Cameraman 0.95899 0.93702 0.91097
Barbara 0.95012 0.92208 0.91518
Boat 0.94182 0.91856 0.87697
Airplane 0.92944 0.93696 0.88629

Table 3. Correlation Coefficient of two adjacent pixels in Ciphered images

Test Images Vertical Horizontal Diagonal

Lena 0.00095 0.00205 0.00188
Baboon 0.00088 0.00170 0.00129
Peppers 0.00118 0.00191 0.00031
Cameraman 0.00133 0.00046 0.00182
Barbara 0.00077 0.00085 0.00173
Boat 0.00026 0.00032 0.00115
Airplane 0.00372 0.00081 0.00324

each other and correlation coefficients are close to 1, whereas the values obtained
for the ciphered images are close to 0. This means that the proposed algorithm
highly de-correlate the adjacent pixels in the ciphered images.

4.5 Information Entropy Analysis

The entropy H of a symbol source S can be calculated by following equation [25].
255
H(S) = − p(si )log2 (p(si )) (7)
i=0

Where p(si ) represents the probability of symbol si and the entropy is expressed
in bits. If the source S emits 28 symbols with equal probability, i.e. S = { s0 ,
s1 , . . . , s255 }, then the result of entropy is H(S) = 8, which corresponds to a
true random source and represents the ideal value of entropy for message source
S. Information entropy of a ciphered image can show the distribution of gray
value. The more the distribution of gray value is uniform, the greater the infor-
mation entropy. If the information entropy of a ciphered image is significantly
less than the ideal value 8, then, there would be a possibility of predictability
which threatens the image security. The values of information entropy obtained
for plain-images and ciphered images are given in Table 4. The values obtained
for ciphered images are close to the ideal value 8. This implies that the infor-
mation leakage in the proposed encryption process is negligible and the image
encryption system is secure against the entropy attack.
 A Multi-Level Blocks Scrambling Based Chaotic Image Cipher 181

Table 4. Information Entropy of Original and Ciphered Images

Test Images Original Ciphered

Lena 7.4439 7.9969
Baboon 7.2649 7.9976
Peppers 7.5327 7.9973
Cameraman 7.0097 7.9969
Barbara 7.5482 7.9974
Boat 7.1124 7.9969
Airplane 6.7074 7.9975

5 Conclusion
In this paper, a new algorithm of encryption and decryption of images is pre-
sented. The algorithm is based on the concept of scrambling the pixel’s positions
and changing the gray values of the image pixels. To perform the scrambling of
the plain-image’s pixels, a multi-level blocks based scrambling scheme is sug-
gested. At each level, the image is first decomposed into blocks whose size de-
pends on the level and then the blocks are scrambled through 2D Cat transform.
The control parameters of scrambling are randomly generated using a 2D cou-
pled Logistic map to enforce the secrecy of the images. The encryption of the
scrambled image is done using chaotic sequence generated through 1D Logistic
map. All the simulation and experimental analysis show that the proposed image
encryption system has very large key space, high sensitivity to secret keys, better
diffusion of information in the ciphered images and low correlation coefficients.
Hence, the proposed image encryption algorithm has high level of security with
less computation and is more robust towards cryptanalysis.

References
1. Fridrich, J.: Symmetric Ciphers based on two-dimensional Chaotic maps. Interna-
tional Journal of Bifurcation and Chaos 8(6), 1259–1284 (1998)
2. Yen, J.C., Guo, J.I.: A New Image Encryption Algorithm and Its VLSI Architec-
ture. In: IEEE Workshop on Signal Processing Systems, pp. 430–437 (1999)
3. Yen, J.C., Guo, J.I.: A New Chaotic Key-based Design for Image Encryption and
Decryption. In: IEEE International Symposium on Circuits and Systems, vol. 4,
pp. 49–52 (2000)
4. Zhang, L., Liao, X., Wang, X.: An Image Encryption Approach based on Chaotic
maps. Chaos, Solitons and Fractals 24(3), 759–765 (2005)
5. Pisarchik, A.N., Flores-Carmona, N.J., Carpio-Valadez, M.: Encryption and De-
cryption of Images with Chaotic map lattices. CHAOS Journal, American Institute
of Physics 16(3), 033118-033118-6 (2006)
6. Dongming, C., Zhiliang, Z., Guangming, Y.: An Improved Image Encryption Al-
gorithm Based on Chaos. In: IEEE International Conference for Young Computer
Scientists, pp. 2792–2796 (2008)
7. Pareek, N.K., Patidar, V., Sud, K.K.: Image Encryption using Chaotic Logistic
map. Image and Vision Computing 24(9), 926–934 (2006)
182 M. Ahmad and O. Farooq

8. Lian, S., Sun, J., Wang, Z.: A Block Cipher based on a Suitable use of Chaotic
Standard map. Chaos Solitons and Fractals 26(1), 117–129 (2005)
9. Mao, Y., Lian, S., Chen, G.: A Novel Fast Image Encryption Scheme based on
3D Chaotic Baker maps. International Journal of Bifurcation and Chaos 14(10),
3616–3624 (2004)
10. Salleh, M., Ibrahim, S., Isnin, I.F.: Enhanced Chaotic Image Encryption Algorithm
based on Baker’s map. In: IEEE International Conference on Circuits and Systems,
vol. 2, pp. 508–511 (2003)
11. Behnia, S., Akhshani, A., Ahadpour, S., Mahmodi, H., Akhavand, A.: A Fast
Chaotic Encryption Scheme based on Piece-wise Nonlinear Chaotic maps. Physics
Letter A 366(4-5), 391–396 (2007)
12. Chen, G.Y., Mao, Y.B., Chui, C.K.: A Symmetric Image Encryption Scheme based
on 3D Chaotic Cat maps. Chaos Solitons and Fractals 21(3), 749–761 (2004)
13. Guan, Z.H., Huang, F., Guan, W.: Chaos-based Image Encryption Algorithm.
Physics Letters A 346(1-3), 153–157 (2005)
14. Hsu, C.T., Wu, J.L.: Hidden Digital watermarks in Image. IEEE Transactions on
Image Processing 8(1), 58–68 (1999)
15. Yang, Y.L., Cai, N., Ni, G.Q.: Digital Image Scrambling Technology Based on
the Symmetry of Arnold Transformation. Journal of Beijing Institute of Technol-
ogy 15(2), 216–220 (2006)
16. Liu, H., Zhu, Z., Jiang, H., Wang, B.: A Novel Image Encryption Algorithm Based
on Improved 3D Chaotic Cat Map. In: 9th IEEE International Conference for
Young Computer Scientists, pp. 3016–3021 (2008)
17. Shang, Z., Ren, H., Zhang, J.: A Block Location Scrambling Algorithm of Digital
Image Based on Arnold Transformation. In: 9th IEEE International Conference for
Young Computer Scientists, pp. 2942–2947 (2008)
18. Liehuang, Z., Wenzhou, L., Lejian, L., Hong, L.: A Novel Image Scrambling Algo-
rithm for Digital Watermarking Based on Chaotic Sequences. International Journal
of Computer Science and Network Security 6(8B), 125–130 (2006)
19. Fu-Yan, S., Shu-Tang, L., Zong-Wang, L.: Image Encryption using high-
dimensional Chaotic System. Chinese Physics 16(12), 3616–3623 (2007)
20. Jian-Liang, M., Hui-jing, P., Wang-qing, G.: New Color Image Encryption Algo-
rithm based on Chaotic Sequence Ranking. In: IEEE International Conference on
Intelligent Information Hiding and Multimedia Signal Processing, pp. 1348–1351
(2008)
21. Fu, C., Zhang, Z., Cao, Y.: An Improved Encryption Algorithm Based on Chaotic
Maps. In: Third IEEE International Conference on Natural Computation, pp. 189–
193 (2007)
22. Li, T., Zhou, S., Zeng, Z., Ou, Q.: A New Scrambling Method Based on Semi-
frequency Domain and Chaotic System. In: IEEE International Conference on
Natural Computation, pp. 607–610 (2005)
23. Wang, X.Y., Shi, Q.J.: New Type Crisis, Hysteresis and Fractal in Coupled Logistic
Map. Chinese Journal of Applied Mechanics, 501–506 (2005)
24. May, R.M.: Simple Mathematical Model with very Complicated Dynamics. Na-
ture 261, 459–467 (1976)
25. Tao, X., Liao, X.F., Tang, G.P.: A Novel Block Cryptosystem based on Iterating
a Chaotic map. Physics Letter A 349(1-4), 109–115 (2006)