Hindawi

Complexity
Volume 2022, Article ID 3676265, 20 pages
https://doi.org/10.1155/2022/3676265

Research Article
Fast and Robust Image Encryption Scheme Based on Quantum
Logistic Map and Hyperchaotic System

Nehal Abd El-Salam Mohamed ,1 Aliaa Youssif,2 and Hala Abdel-Galil El-Sayed 3

1
College of Information Technology, Misr University for Science & Technology (MUST),
6th of October City 77, Egypt
2
College of Computing and Information Technology, Arab Academy for Science, Technology and Maritime Transport,
Smart Village 12577, Egypt
3
College of Computers and Artificial Intelligence, Helwan University, Ain Helwan (Helwan University Building),
Helwan 11795, Egypt

Correspondence should be addressed to Nehal Abd El-Salam Mohamed; [email protected]

Received 7 October 2021; Revised 9 February 2022; Accepted 14 February 2022; Published 29 March 2022

Academic Editor: Ahmed A. Abd El-Latif

Copyright © 2022 Nehal Abd El-Salam Mohamed et al. This is an open access article distributed under the Creative Commons
Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is
properly cited.
Topic of quantum chaos has begun to draw increasing attention in recent years. So, to ensure the security of digital image, an
image encryption algorithm based on combining a hyperchaotic system and quantum 3D logistic map is proposed. This algorithm
is applied in four stages. Initially, the key generator builds upon the foundation of mean for any row or column of the edges of the
plain image. Its output value is used to yield initial conditions and parameters of the proposed image encryption scheme. Next, it
diffuses the plain image by the random sequences generated by 3D hyperchaotic system, and the diffusion process is realized by
implementing XOR operation. Then, the diffused image and chaotic sequences are produced by the 3D quantum chaotic logistic
map, expressed as a quantum superposition state using density matrix which is a representation of the state of a quantum system,
and finally the resulting quantum image is then confused and diffused simultaneously by a unitary matrix generated by logistic
chaos using XNOR operation to obtain the final cipher image. Because of the dependence on the plain image, the algorithm can
frustrate the chosen-plaintext and known-plaintext attacks. Simulation results and theoretical analysis verify that the presented
scheme has high safety performance, a good encryption effect, and a large key space. The method can effectively resist exhaustive,
statistical, and differential attacks. Moreover, the encryption time of the proposed method is satisfactory, and the method can be
efficiently used in practice for the secure transmission of image information.

1. Introduction daily needs of people where digital information is being

applied in all the fields in the society.
In today’s era [1–6], with the fast development of electronic Images originated in some scenarios such as any social
technology and the scale of the communication network, a media servers, business, personal privacy, healthcare or
lot happens over a time of one minute. Along with this rapid military systems, organizations, banks, and other private
development of Internet and multimedia, usage of digital sectors contain private information which is placed and
media has increased tremendously in past decades. In this maintained in very big databases, since it can be transmitted,
period of digital data technology, today, we are in the sphere shared, and stored on the Internet, so if this information is
of digitally advanced era, where most of the private data and stolen or an unauthorized person accesses it, this may cause
secure digital information is being exchanged by the help of a serious damage and serious consequences to any orga-
electronic media such as television, smartphones, personal nization [7–9].
computers, tablets, facsimiles, satellites, and so forth to all With the widespread application of a digital image,
corners of the world over just one minute to facilitate the providing digital image information security in the
2 Complexity

transmission channel has become an increasingly serious techniques can be exploited to process these images. Due to
issue to be urgently solved because the data can be inter- the promising prospect of quantum image encryption, more
cepted, cracked, or destroyed [10, 11]. Hence, the security of and more researchers devoted their attention to developing
the important and valuable image information has become a quantum image representation models and designing image
hot recent topic of the field of information security. encryption algorithms.
Image encryption [5, 12, 13] is one of the possible ef- For example, Li et al. [3] proposed an efficient chaos-
fective solutions used to protect these images from this based image encryption scheme, which uses the imitating
threat where it is extensively recognized as a useful technique jigsaw method containing revolving and shifting operations
for secure transmission and its objective is to accomplish and shows good performance in both security and speed. Liu
privacy and integrity of data. It converts images into noise- et al. [7] proposed a quantum image encryption algorithm
like encrypted images with key by disrupting pixel positions based on bit-plane permutation and sine logistic map which
or changing pixel values and decryption will reveal the has good performance in the aspect of security and the
original message or information by utilizing same key uti- computational complexity is superior to its classical coun-
lized for encryption. terpart. Dong et al. [9] proposed a self-adaptive image en-
To satisfy the emerging demand, a lot of useful image cryption algorithm based on the quantum logistic map,
encryption algorithms based on optical transformation, which can achieve secure communications and frustrate the
DNA sequence operations, wave motion, Brownian motion, chosen-plaintext and known-plaintext attacks. In [15], an
cellular automata, compressive sensing, and chaotic system innovative quantum color image encryption method focused
[6, 14] were developed in literature to secure these digital on the Lucas series-based substitution box is suggested to
images. enhance the competence of encryption. This cryptosystem
Since the chaos theory was first proposed by Lorenz, has more excellent key space and significant confidentiality.
many chaotic phenomena were found in many fields, such as In [19], an image encryption algorithm based on 3D DNA
physics, astronomy, chemistry, biology, and medicine. In level permutation and substitution scheme is proposed,
1998, Fridrich firstly proposed a chaos-based image en- where the proposed encryption scheme has large key space
cryption algorithm composed of two stages: permutation and high key sensitivity and may resist some typical attacks,
and diffusion. After that, many scholars have designed and it may effectively secure the secret image information.
numerous efficient algorithms for chaotic systems and El-Latif et al. [20] presented a new method for constructing
chaos-based image encryption to be applied for the secure substitution boxes (S-boxes) based on cascaded quantum-
communications [3, 4, 8, 11]. inspired quantum walks and chaos inducement, which will
Chaotic systems [2, 4] have many noteworthy features offer gains in many cryptographic applications where the
which satisfy the requirements of image encryption, such as performance of the proposed S-box scheme is investigated
random-like behaviors, high sensitivity to initial conditions via established S-box evaluation criterion and outcomes
and control parameters where the wrong initial condition suggest that the constructed S-box has significant qualities
will lead to nonchaotic behavior, nonperiodicity, and er- for viable applications information security. In [21], a new
godicity, and low cost in the computer operation system and method for the encryption by utilizing quantum chaotic
microprocessor [8–10]. Therefore, these systems can be maps and continuous chaotic dynamical systems is designed
rapidly applied to cryptographic systems which achieve which contributes to achieving the security of data with the
superior performance with respect to the trade-offs between minimum time of encryption. Sridevi and Philominathan
the security and efficiency. However, the appearance of [22] presented a quantum encryption technique which is
quantum computing brought a great challenge to classic built by adopting Haar Integer Wavelet Transform (HIWT),
encryption methods [15, 16]. RC6 (Rivest Cipher) block cipher, and DNA (deoxy-
Additionally, with the advancement in technology in the ribonucleic acid) sequences. In addition, a Unified Chaotic
modern era of computer world, brute-force attack [4, 6] will Logistic Tent Map (ULTM) has been employed in the
be quite easily performed in quantum computers which are permutation phase to produce the pseudorandom sequence
based on quantum information theory. This vulnerability for shuffling the RGB planes of the quantum represented
gives potential danger to idealized security required at na- source image in spatial and transform domains. This
tional security and protected innovation level. To beat this cryptosystem has confirmed the significant immune level of
threat, it is necessary to study novel and safer cryptosystem the quantum cryptosystem. In [23], an enhanced quantum
to meet the current safety requirements in image encryption, scheme is proposed for generalized novel enhanced quan-
and, therefore, quantum encryption can be applied in the tum image representation which has good visual effects and
image encryption process as it gives us a secure encryption high security. Wen et al. [24] proposed an image crypto-
method. system adopting a quantum chaotic map and the certain
Quantum computation [7, 17, 18] has shown great security-enhanced mechanisms where the cryptosystem has
potential for improving information processing speed and excellent performance and can resist various cryptographic
enhancing communication security. Combining quantum attacks. Moreover, the feasibility and effectiveness of the
computing and image encryption is a secure and effective image cryptosystem are verified on the Internet of Things
approach to design the encryption algorithms. The essence secure communication experimental platform. It proves that
task of quantum image encryption is to store the images into the proposed image cryptosystem is a preferred and
quantum computers, and then quantum encryption promising secure communication technology solution.
Complexity 3

After conducting a detailed analysis of the breaking ⎪

⎧ x_ � a(y − x),
methods, it was found that some chaos-based image en- ⎪
⎨
⎪ y_ � cx − xz − y, (1)
cryption schemes have security vulnerabilities, which are as ⎪
⎩
follows: (1) key dependence and fixed key; (2) one cycle of z_ � xy − bz,
permutation-diffusion architecture; (3) low-dimensional
systems used for image encryption; (4) single chaotic system where a, b, and c are constant parameters of the system.
still used for encryption operation, which leads to inability to Typically, when a � 10, b � 8/3, and c � 28, the system is in a
resist brute-force attack; and (5) low sensitivity to all the chaotic state.
chaotic secret keys. In 1999, Chen [27] discovered another chaotic system
To overcome these security shortcomings and design with more complex dynamic behaviors than Lorenz system
secure and effective image encryption, an image encryption when studying chaotic feedback control. Chen’s hyper-
algorithm based on integrating a hyperchaotic system and chaotic system is defined as follows:
quantum 3D logistic map is presented in this paper. The ⎪
⎧ x_ 1 � a x2 − x1 􏼁,
⎪
⎪
essence goals of the proposed scheme are listed as follows: ⎪
⎪
⎨ x_ 2 � −x1 x3 + dx1 + cx2 − x4 ,
(i) First, it can fight against the chosen-/known- ⎪ (2)
⎪
⎪ x_ 3 � x1 x2 − bx3 ,
plaintext attacks due to the use of symmetric key ⎪
⎪
⎩
image cryptosystem based on original image. x_ 4 � x1 + k,
(ii) Second, the generated key cryptosystem based on where a, b, c, d, and k are the system parameters. In this
the plain image is used to determine the number of system, when the values of the parameters (a, b, c, d, k) � (36,
cycles of composite chaotic algorithms. 3, 28, −16, −0.7 <k <0.7), the system is hyperchaotic in a very
(iii) Third, multidimensional chaotic maps like hyper- wide parameter range in this case and has many more in-
chaos and 3D quantum logistic map are used which teresting complex dynamical behaviors than those of Lorenz
have more chaotic attractors, so the high-dimen- system. The hyperchaos attractors of this system are shown
sional chaotic system has stronger randomness, in Figure 1, while the corresponding bifurcation diagram of
better confidentiality, greater amount of informa- state x with respect to k is given in Figure 2.
tion, and higher communication efficiency, pro- Its Lyapunov exponents are λ1 � 1.552,
viding sufficiently large key space and having high λ2 � 0.023, λ3 � 0, λ4 � −12.573; Lyapunov exponents for
security. this system are depicted in Figure 3. As the hyperchaos has
(iv) Fourth, two different chaotic systems (quantum four positive Lyapunov exponents, the prediction time of a
logistic map and hyperchaotic Chen’s system) are hyperchaotic system is shorter than that of a chaotic system
combined, which have the advantage of excellent [28]; as a result, it is safer than chaos in security algorithm.
random sequence to expand the key space, enhance
the performance of resisting brute-force attack, and
achieve better encryption effect and high level of 2.2. 3D Quantum Logistic Chaotic Map. Quantum chaotic
security. systems are the quantized of classical chaotic system, such as
(v) Fifth, high sensitivity with respect to all secret keys quantum logistic map which [1, 29, 30] is constructed by the
is achieved, which leads to creating a completely classical logistic system and that is a perfect example of
different cipher image when applied to the same complex chaotic maps which arises from nonlinear dy-
plain image whenever flipping one bit in a key. namical equations. Classical chaotic maps have a small range
for key space as they suffer from low control parameters
Based on the above literature, it is evident that, for which in turn lead to a limited chaotic range, whereas the
generating excellent encryption effects and producing a chaotic maps with higher dimensional as the used one in the
highly secure encryption scheme, it is needed to design a proposed scheme can be lead to increase the key space range,
combination of hyper- and multidimensional chaotic sys- have excessive complexity, high degree of randomness, and
tems through density matrix which describes the quantum high sensitivity to initial conditions and control parameters.
state of a system. Therefore, quantum logistic system is suitable as seed system
in encryption algorithm.
Based on the classical logistic map and the effect of
2. Preliminary Knowledge quantum correlations on a dissipative system [31], the
proposed quantum logistic map was applied to image en-
2.1. Chen’s Hyperchaotic System. In order to improve the cryption, which can be defined as follows:
security and efficiency performance, many image encryption
methods based on three-dimensional chaotic systems, 􏼌􏼌 􏼌􏼌2
xn+1 � r􏼐xn − 􏼌􏼌xn 􏼌􏼌 􏼑 − ryn ,
hyperchaos, and even spatiotemporal chaos have been
presented in recent years [25]. yn+1 � −yn e−2β + e−β r􏼂 2 − xn − x∗n 􏼁yn − xn z∗n − x∗n zn 􏼃, (3)
In 1963, Lorenz [26] found the first chaotic attractor in a
three-dimensional autonomous system: zn+1 � −zn e−2β + e−β r􏼂2 1 − x∗n 􏼁zn − 2xn yn − xn 􏼃,
4 Complexity

80 40
60 30
z 40 20
20
10
0
40 y 0
40
20 20
0 -10
0
y -20 -20

x
-20
-40 -40
-30

-40
-30 -20 -10 0 10 20 30
x
(a) (b)
70 70

60 60

50 50

40 40
z z
30 30

20 20

10 10

0 0
-40 -30 -20 -10 0 10 20 30 40 -30 -20 -10 0 10 20 30 40
y x
(c) (d)

Figure 1: x − yx − zFigure 1Hyperchaos attractors of Chen’s chaotic system: (a) Distribution in the direction of x-y-z. (b) Plane graph of x-
y. (c) Plane graph of y-z. (d) Plane graph of x-z.

10

x 0

–10

–5 0 5
k
Figure 2: Bifurcation diagram of Chen’s chaotic system with parameter k.

where β is dissipation parameter and c represents control 3. The Proposed Image Encryption and
parameter. However, the initial conditions (xn , yn , zn ) are Decryption Scheme
set as real numbers to meet the requirement of communi-
cation. Figure 3 shows the phase diagram of quantum lo- 3.1. Image Encryption Process. This section presents the details
gistic map, and its bifurcation diagram is displayed in of the design of the proposed method based on the adopted
Figure 4. fundamental Fridrich’s permutation-diffusion model,
Complexity 5

×10-5 ×10-3
0

y z

0
0 0.5 1 0 0.5 1
x x
×10-3
0

0 ×10-5
y
Figure 3: Phase diagram of quantum logistic map.

1
xn

0
3 4 5 6 7 8 9 10
β
Figure 4: Bifurcation diagram of quantum logistic map.

hyperchaotic system, and a 3D quantum logistic mapping. The mean to make a number of iterations for both
proposed algorithm is designed in the context of sensitive Chen’s hyperchaotic system and the quantum
information of digital color and gray images. Consider a color logistic map in order to modify the initial seeds and
image I with size W × H, where W and H represent the image’s control parameters for them.
rows and columns, respectively. The R, G, and B components of
(ii) Iterate continuously Chen’s hyperchaotic system
I are denoted as R, G, and B, respectively. The proposed
W × H times to generate a random sequence of
framework consists of four main phases, and the details of these
integers EI whose values range from [0. . .255],
phases are presented as follows:
where the length of sequence EI , that is, n, will be
(i) The first phase is key extraction from a plain image equal to the number of pixels in the image. Then
through computing the mean of any of the four split it into three chaotic sequences ERi , EGi , EBi
edges of the plain image and then utilizing that which are computed using the following equations:
6 Complexity

ERi (i, j) � unit8􏼐round􏼐mod􏼐(abs(x) − floor(abs(x))) ∗ 10(14) , 256􏼑􏼑􏼑,

EGi (i, j) � unit8􏼐round􏼐mod􏼐(abs(y) − floor(abs(y))) ∗ 10(14) , 256􏼑􏼑􏼑, (4)
EBi (i, j) � unit8􏼐round􏼐mod􏼐(abs(z) − floor(abs(z))) ∗ 10(14) , 256􏼑􏼑􏼑,

where i � 1, 2, . . . , W × H. (iv) Perform quantum logistic map to produce a

(iii) Diffuse three components of the plain image by the chaotic sequence Qi ; after that separate it into three
random sequences generated by 3D hyperchaotic channels QRi , QGi , QBi which can be calculated as
system to obtain their corresponding cipher se- follows:
quences. The diffusion process is performed by
implementing XOR operation as follows:
CRi � PRi ⊕ ERi ,
CGi � PGi ⊕ EGi , (5)
CBi � PBi ⊕ EBi .

QRi � mod floor ε1 ∗ xi+1 + ε2 􏼁, 256􏼁, i � 1, 2, . . . , W × H,

QGi � mod floor ε1 ∗ yi+1 + ε2 􏼁, 256􏼁, i � 1, 2, . . . , W × H, (6)
QBi � mod floor ε1 ∗ zi+1 + ε2 􏼁, 256􏼁, i � 1, 2, . . . , W × H,

SCRi � CRi ⊕ H,
where (ε1 , ε2 ) are two large prime numbers and
(xi+1 , yi+1 , and zi+1 ) are random sequences SCGi � CGi ⊕ H,
which are generated by 3D quantum logistic
map (3). SCBi � CBi ⊕ H,
(8)
(v) Generate density matrix H using the following
SQBi � QBi ⊕ H,
equations:
a 2 SQBi � QBi ⊕ H,
H11 � (p +(1 − p)) ∗ 􏼠cos 􏼒 􏼓 􏼡,
2 SQBi � QBi ⊕ H,
a a
H12 � (1 − p) ∗ 􏼒sin􏼒 􏼓 ∗ ∗ cos􏼒 􏼓􏼓, where operator ⊕ denotes bitwise exclusive NOR.
2 2
(7) (vii) Finally, the final cipher channels FCRi , FCGi , FCBi
a a are obtained by applying XNOR function on both a
H21 � (1 − p) ∗ 􏼒sin􏼒 􏼓 ∗ cos􏼒 􏼓􏼓,
2 2 unitary matrix generated by logistic chaos
(SQRi , SQGi , SQBi ) and the diffused components
a 2 (SCRi , SCGi , SCBi ) generated density matrix to
H22 � (1 − p) ∗ 􏼠sin 􏼒 􏼓 􏼡,
2 confuse and diffuse pixels simultaneously, which
can be expressed as follows:
where p is probability and a is angle.
(vi) The diffused layers (CRi , CGi , CBi ) and chaotic se- FCRi � SCRi ⊕ SQRi ,
quences are produced by the 3D chaotic logistic FCGi � SCGi ⊕ SQGi , (9)
map (QRi , QGi , QBi ) which are expressed as a
quantum superposition state, using the XNOR FCBi � SCBi ⊕ SQBi .
function as follows:
Complexity 7

(viii) Combine (FCRi , FCGi , and FCBi ) into a chaotic matrix quantitative performance of the newly resulted image en-
FCi with transpose rows and columns of the border of cryption algorithm could be measured through different
the image to get the final cipher image CW×H . evaluation parameters, including statistical, differential,
sensitivity, and key space metrics. Each of these measures is
The sketch of the proposed encryption scheme is
discussed in detail in the accompanying subsections.
exhibited in Figure 5 with a succinct explanation of each
phase presented herewith while the specific implementation
process of the proposed image encryption scheme is pre-
4.1. Key Space Analysis. The key space of a cryptosystem is the
sented in Algorithm 1.
very important factor on security when brute-force attack is
happening. For high-security cryptosystem, it should be highly
3.2. Decryption Method. The architecture of the proposed sensitive to a tiny change in the cryptographic keys and the key
decryption algorithm is shown in Figure 6, which is applied space is suggested to be much larger than 2100 to resist ex-
on a cipher image to produce a plain image. haustive attack effectively [33–36]. Moreover, the keys should
be easy to establish and exchange for practical communication.
4. Experimental Results and Numerical Analysis The key space is the total number of different keys that can be
used in the encryption/decryption procedure. According to the
Due to the absence of a practical and functional quantum algorithm structure, the secret key format should consist of the
computer, the experimental results are performed with following: (1) The parameters of Chen’s hyperchaotic system
MATLAB R2017b platform on a classical computer to verify the are a, b, c, d, and k and each of the original variables (x1, x2, x3,
security and effectiveness of the proposed quantum image x4) has 2 decimal places; there exist 102 possible values for each
encryption algorithm. The operation system used is Windows value. This contributes to 6 possible guesses of value. This
10 Professional operating system with the specific configuration applies to (a, b, c, d, k, x1, x2, x3, x4) as well. Thus, there are 254
being i7-8550U applied as the central processing unit (CPU) possible values of (a, b, c, d, k, x1, x2, x3, x4). (2) The initial values
and the random-access memory (RAM) adopted is 8 GB. of hyperchaotic system (x1, x2, x3) are obtained by iterating
For simulation, the control parameters and initial values system; each has 14 decimal places with the range between 0
of Chen’s hyperchaotic system, given in (2), are set as a � 36, and 1, and there exist 1014 possible values for each value. This
b � 3, c � 28, and d � −16, and x0 � 0.3, y0 � −0.4, z0 � 1.2, and contributes to 246.5 possible guesses of value. This applies to (x1,
q � 1, we carry out the encryption scheme. The keys for this x2, x3) as well. Thus, there are 2139.5 possible values of (x1, x2,
proposed cryptosystem include the iteration times of Chen’s x3). (3) Parameters β and r are used in the quantum logistic
hyperchaotic system and quantum logistic chaotic map M, chaotic map, where β consists of 4 decimal places; there exist
where the discarded number M is set according to the mean 104 possible values for each value. This contributes to 212
of plain image. For color images, the encryption key is the possible guesses of its value and r consists of 2 decimal places,
same in RGB channels. and there exist 102 possible values for each value. This con-
To demonstrate the practical benefits of the proposed tributes to 26 possible guesses of its value. Thus, there are 218
image encryption scheme, a number of experiments were possible values of β and r. (4) Each initial value of quantum
performed based on the USC-SIPI (the University of Southern map consists of 12 decimal places with the range between 0 and
California Signal and Image Processing Institute) Image Da- 1; there exist 1012 possible values for each value. This con-
tabase [32]. This database is divided into four groups of images: tributes to 240 possible guesses of value. This applies to (x0, y0,
Textures (64 images), Aerials (38 images), Miscellaneous (39 z0) as well. Thus, there are 2120 possible values of x0, y0, and z0.
images), and Sequences (69 images). Each group contains (5) Two large prime numbers are of 8 decimal places with the
images of various sizes m × m, m � 256, 512, 1024. Different range between 0 and 1; there exist 108 possible values for each
sample images (gray and color) are chosen as test images from value. This contributes to 226 possible guesses of value. This
the USC-SIPI “Miscellaneous” dataset and the simulation re- applies to (ε1 , ε2 ) as well. Thus, there are 252 possible values of
sults of these encryption and decryption images are presented (ε1 , ε2 ). (6) Density matrix has probability p and an angle a,
in Figure 7, where the plain images of “Aerial,” “Boat,” “Male,” where p has only one decimal place with the range between 0
“Airplane,” “Lena,” and “Baboon” are shown in Figures 7(a)– and 1, and a has 2 decimal places; thus there exist 101 possible
7(f), their corresponding cipher images are shown in values for p; this contributes to 23 possible guesses of value,
Figures 7(g)–7(l), and the recovery images from decryption whereas there exist 102 possible values for a; this contributes to
process with correct secret keys are shown in Figures 7(m)–7(r) 26 possible guesses of value. This contributes to 210 possible
which are identical to the original images, and their detailed guesses of value. Thus, there are 29 possible values of (p, a).
information is listed in Table 1. Consequently, the overall key space of the proposed
As illustrated in Figures 7(g)–7(l) that the proposed image encryption scheme is
encryption scheme can encrypt different size images, besides
that it destroys the obvious pattern of the plain image and TOTAL KEY SPACE � 254 ∗ 218 ∗ 2139.5 ∗ 2120 ∗ 252 ∗ 29
makes the ciphered image display a space filling with a noise-
like pattern which makes the ciphered image seem random � 254+139.5+120+52+20
to the intruder. Therefore, the proposed encryption algo-
rithm has good encryption and decryption effect; it can � 2392.5 .
attain the image data security and appearance security. The (10)
8 Complexity

Plaintext Image
1
12 8 2 Calculate mean (M) for any
4
row or column (1...12)
3

Iterated continuously for M times

PiB
PiG
PRi

Chen’s Hyperchaotic System Quantum Logistic System

QiB
QiG
EiB QiR
EiG
ERi Hermitian Matrix (H)
(Density Matrix)

CiB G
Ci H
CRi

SCiB
SCiG SQiB
SCRi SQiG
SQiR
FCiB G
4 FCi
Ciphertext Image 9 5 3 FCRi
1
2

Figure 5: Block diagram of the proposed image encryption algorithm.

Input: Plain Image P of size W × H, initial conditions and control parameters for hyperchaotic system (3D Chen’s system), and seeds
for the chaotic generator.
Output: Cipher Image C of size W × H
Step 1: Plain image P is resized to a dimension of ((W − 2) × (H − 2)) pixels and is stored as P2 , and compute the mean M of any of
the edges of the plain image P.
Step 2: Iterate both Chen’s hyperchaotic system (equation (2)) and quantum logistic map (equation (3)) M times according to the
computed mean M.
Step 3: Generate three chaotic sequences ERi , EGi , EBi by using a hyperchaotic system with given parameters and initial state values as
secret keys.
Step 4: Separate each of the color pixel Pi ∈ P2 of the resized image P2 into its three grayscale components of PRi , PGi , PBi , then apply
XOR function between three components PRi , PGi , PBi of the resized image P2 and three chaotic sequences ERi , EGi , EBi produced by
chen’s hyperchaotic system. The result is considered as diffused R, G, and B components, which are CRi , CGi , CBi .
Step 5: Quantum logistic map is initiated and utilized to generate a chaotic keystream sequence Qi , after that split it into R, G, and B
components QRi , QGi , QBi .
Step 6: Generate Density matrix which is described as Hermitian matrix HW−2×H−2 .
Step 7: Employ Density matrix on the diffusion components (CRi , CGi , CBi ), as well as the output of quantum logistic map (QRi , QGi , QBi )
using XNOR function to put each of them in a superposition environment.
Step 8: The three components of the cipher image FCRi , FCGi , FCBi are generated by XNORing the output of applying density matrix
on the diffused components (SCRi , SCGi , SCBi ), and quantum logistic map (SQRi , SQGi , SQBi ).
Step 9: Take transpose of the edges of the plain image P in order to increase the randomness within the plain image by shuffling the
pixels.
Step 10: Recombine the cipher image FC with the shuffled edges of the plain image P to obtain the final cipher image C.

ALGORITHM 1: Image encryption method.

As a result, the key space is reasonably large enough for 4.2. Key Sensitivity Analysis. To resist violent attacks, a
the cryptosystem to withstand exhaustive attacks and even password system should be highly sensitive. Hence, key
quantum computer attacks. Table 2 shows the key space sensitivity [37–40] is an important index to measure the
comparison of similar recent algorithms. Obviously, the strength of encryption algorithm. The key sensitivity of an
proposed encryption algorithm has larger key compared to image cryptosystem can be evaluated in two aspects: First,
the existing works [4, 15, 24, 35, 37], which is sufficiently the cipher image will be completely different when
large to resist all presently known brute-force attacks. encrypting the same plain image with slightly different keys,
Complexity 9

Ciphertext Image
4 Calculate mean (M) for any
9 5 3
1 row or column (1...12)
2

Iterated continuously for M times

FCBi G
FCi
FCRi
Quantum Logistic System Chen’s Hyperchaotic System

QiBi G
Qi
QRi EBi G
Ei
Hermitian Matrix (H) ERi
(Density Matrix)

SCBi G SQBi
SCi SQGi R
SCRi
SQi H

CBi
CGi
CRi

1 PBi G
12 8 Pi
2 PRi
Plaintext Image 4
3

Figure 6: Block diagram of the proposed image decryption algorithm.

Input: Cipher Image C of size W × H

Output: Decrypted Image P of size W × H
Steps: Inverse steps of image encryption routine are carried out in the reverse order using the same encryption keys.

ALGORITHM 2: Image decryption method.

which is measured by the change rate t of the cipher image. 4.3. Statistical Attack Analysis. To verify the security per-
Second, a small change in the decryption key makes a huge formance of the proposed algorithm, the statistical analyses
difference to the result, and the original image will not be including histogram, correlation, and entropy analysis are
decrypted correctly, indicating that the algorithm has a high demonstrated in this subsection.
sensitivity. The Lena color image with size 512 × 512 is
utilized to verify the sensitivity of the suggested image
encryption scheme. During the test process, one of the keys 4.3.1. Histogram Statistical Analysis. Histogram statistical
has undergone a tiny change, while other keys were kept analysis is a kind of statistical attack, and the histogram can
untouched. Suppose that K1 and K2 are the two keys that are characterize the image. It has been widely used in image
slightly different from each other, which gives encrypted retrieval, classification, and other fields [41–47]. Image
outputs of E1 and E2, respectively, where K1 is the correct key histogram is probability density function of discrete gray
and K2 is the wrong one. In the proposed cryptosystem, the level, plotted by gray level on horizontal axis and the cor-
control parameters of Chen’s hyperchaos system are set as responding frequency on the vertical. The more uniform the
a � 36, b � 3, c � 28, and d � −16, and the initial values of the histogram distribution for the encrypted image, the stronger
system are x0 � 0.3, y0 � −0.4, z0 � 1.2, and q � 1, which are the ability of antistatistical analysis. Therefore, the elimi-
denoted as K1, to obtain encrypted image E1. Another nation of correlation among pixels was necessary, and pixels
encrypted image E2 is generated with a tiny change in only x0 of the encrypted image had to be distributed evenly to
(x0 � 0.4, y0 � −0.4, z0 � 1.2, and q � 1), which are denoted as prevent the opponent from extracting any useful informa-
K2. As shown in Figures 8(b) and 8(c), the image encrypted tion from the fluctuating histogram. In addition, comparing
using K1 is completely different from the image encrypted cipher image histogram with the original image histogram,
using K2. From the result, as shown in Figures 8(e) and 8(f ), there is a significant difference.
it is clear that decryption of the encrypted image is possible We have analyzed the histograms of two original images
only when we use the same key. Therefore, it can be seen that as well as their encryptions using the proposed approach.
only a subtle difference in the secret key can have a huge The histogram of the original grayscale image of “Boat” with
effect which guarantees the security against brute-force dimensions 512 × 512 pixels and the histogram of its cipher
attacks and known plain-text attacks. image are shown in Figure 9, while Figure 10 illustrates the
10 Complexity

(a) (b) (c) (d) (e)

(f ) (g) (h) (i) (j)

(k) (l) (m) (n) (o)

(p) (q) (r)

Figure 7: Encryption and decryption results: ((a)–(f )) plain images of “Aerial,” “Boat,” “Male,” “Airplane,” “Lena,” and “Baboon”; ((g)–(l))
the corresponding encrypted images; and ((m)–(r)) decrypted images.

Table 1: Selected test images.

Image Aerial Boat Male Airplane Lena Baboon
Size 256 × 256 512 × 512 1024 × 1024 256 × 256 512 × 512 1024 × 1024
Type Grayscale Grayscale Grayscale Color Color Color

Table 2: Key space comparative analysis. the gray value and RGB component histogram of cipher
Encryption scheme Key space
images are very uniform and flat distribution, which indi-
cates that the attack based on histogram analysis is difficult
Ref. [4] 2256
Ref. [15] 2125
as attackers cannot use a statistical attack to obtain any
Ref. [24] 1015×3 + 2256 useful information by analyzing the histogram of the
Ref. [35] 2186 encrypted image. Thus, the proposed scheme is strong
Ref. [37] 2364 enough to withstand statistical attacks.
Proposed algorithm 2392.5 Consequently, it is concluded that the proposed image
encryption scheme can achieve good performance and meet
the requirements of image encryption.
histograms of the R, G, and B channels of the color plain Furthermore, for quantity analyses of the image histo-
image “Lena” alongside its encrypted counterparts with the gram, a metric called variance of the histogram (var) is
size 512 × 512, respectively. measured to evaluate and guarantee the uniformity of pixels
Clearly, it can be seen from Figures 9 and 10 that the values of the encrypted images. The higher the uniformity of
histograms of the original images have obvious peaks, and ciphered images, the lower the value of variances of
Complexity 11

(a) (b) (c)

(d) (e) (f )

Figure 8: Key sensitivity analysis. (a) Plain image, (b) correctly encrypted image (E1), (c) incorrectly encrypted image (E2), (d) difference of
E1 and E2, (e) incorrectly decrypted image, and (f ) correctly decrypted image.

histogram [48]. The variance of histogram can be computed Therefore, attackers could try to infer adjacent pixel values
as follows: based on probability theory. Conversely, in order to resist
the statistical attack and achieve better security of the
1 n n 1 2
var(H) � 􏽘 􏽘 􏼐h − hj 􏼑 , (11) encrypted image, an excellent image encryption algorithm
n2 i�1 j�1 2 i should be able to break high correlation between adjacent
pixels of the plain image and produce a very small corre-
where H � 􏼈h1 , h2 , . . . , h256 􏼉 is a one-dimensional array of lation value near the optimal value of zero.
the histogram values; hi and hj are considered as the Normally, three different types of correlation are per-
numbers of pixels where gray values are equal to i and j, formed to ensure the strength of the encrypted image: the
respectively. Tables 3 and 4 display the values of histogram horizontal, the vertical, and the diagonal correlation [54]. To
variance for the experimented grayscale and color images, evaluate the proposed encryption scheme, 3000 pairs of
respectively, and illustrate that the variance of images after adjacent pixels are selected randomly in the three different
encryption is greatly reduced when compared with the adjacent directions in both original and encrypted images of
variance of those images prior to encryption. the different sample images to calculate the correlation
The simulation results indicate that the difference in coefficient. Then, the correlation coefficient rxy of each pair,
variance value shows that the histogram depends on the defined in (12), is calculated as follows:
plain image; in addition, the proposed algorithm can
strongly withstand statistical analysis attack as it is efficient ⎪
⎧ cov(x, y)
⎪
⎪ rxy � 􏽰����� 􏽰����� ,
to prevent attackers from obtaining any useful statistical ⎪
⎪ D(x) × D(y)
⎪
⎪
information to decrypt the cipher image. ⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪ 1 N
4.3.2. Correlation Coefficient Analysis. It is known that some ⎪
⎪ cov(x, y) � 􏽘 xi − E(x)􏼁 yi − E(y)􏼁,
⎪
⎪ N i�1
algorithm was broken by using correlation analysis between the ⎪
⎪
⎨
adjacent pixels. So, correlation coefficient analysis [49–52] is ⎪ (12)
performed to evaluate the statistical relationship between image ⎪
⎪
⎪
⎪ 1 N
pixels, and its value is in the range of [−1, 1]. This type of ⎪
⎪ D(x) �
2
􏽘 xi − E(x)􏼁 ,
⎪
⎪
analysis visually shows the distribution between the neigh- ⎪
⎪ N i�1
⎪
⎪
borhood pixels of both the original and encrypted images. ⎪
⎪
⎪
⎪
Due to the intrinsic features of the digital image [53], ⎪
⎪
⎪
⎪
there is a strong correlation between the adjacent pixels, ⎪
⎪ 1 N
⎪
⎩ E(x) � N 􏽘 xi ,
namely, the gray value of one pixel of the plaintext image is i�1
very close to the gray value of the surrounding pixels.
12 Complexity

4000

3500

3000

2500

2000

1500

1000

500

0 50 100 150 200 250

(a)
2500

2000

1500

1000

500

0 50 100 150 200 250

(b)

Figure 9: Histogram distribution analysis of plain and encrypted grayscale image. (a) Plain grayscale “Boat” image; (b) encrypted grayscale
“Boat” image.

where xi and yi are the grayscale pixel values of the ith pair close to 0, which means that the plain image has strong
of the selected adjacent pixels in the tested image, N is the relationships, but weakness exists in the encrypted image.
total number of the randomized chosen samples, cov(x, y) Therefore, these results show that the proposed image en-
is the covariance of x and y, and E(x) and D(x) represent cryption scheme has a good performance in fighting against
the mean value and the variance of vector x, respectively. attacks based on statistical properties of the images.
Figures 11 and 12 show the correlation distribution
between neighborhood pixels in the three directions of the
grayscale “Boat” image and color “Lena” image with size 4.3.3. Information Entropy Analysis. Information entropy
512 × 512 before and after image encryption. It is obvious [55–57] is the most important criterion to evaluate the ef-
that the correlation of adjacent pixel pairs of the plain image ficiency of an image encryption algorithm. In information
is distributed intensively, but those of encrypted image are theory, the entropy parameter is considered as the standard
scattered randomly which looks very uniform, and the to test randomness. For a digital image, information entropy
correlation is greatly reduced. (IE) is one of the outstanding criteria that is usually utilized
Numerically, Table 5 demonstrates values of correlation to evaluate the degree of disorder or randomness of each
coefficient parameter for the proposed technique in different gray value in the encrypted image and measure the amount
test images with diverse sizes. According to the quantitative of information hidden in an image. The color-level distri-
results, it can be concluded that the correlation degrees bution values in an image can also be determined via entropy
between adjacent pixels in the plain images are close to 1, analysis. Ideally, in the case of 8-bit grayscale image, a robust
while those of the encrypted images are very small and are encryption scheme has an entropy value of 8; otherwise, it
Complexity 13

Original Histogram R Encrypted Histogram R

5000 1150
4000 1100
3000 1050
2000 1000
1000 950
0 900
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Original Histogram G Encrypted Histogram G
4000 1150
3000 1100
1050
2000
1000
1000 950
0 900
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Original Histogram B Encrypted Histogram B
6000 1150
1100
4000
1050
1000
2000
950
0 900
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Figure 10: RGB color histogram of original “Lena” image and its corresponding cipher image.

Table 3: Variance of histogram for encrypted grayscale images.

Variance value
Image name
Plain image Cipher image
Aerial 51062 780.8235
Boat 1541901.8039 9791.7098
Male 11393958.6980 138136.8627

Table 4: Variance of histogram for encrypted color images.

Variance value
Image name Plain image Cipher image
R G B R G B
Airplane 165621.8980 163801.6941 274155.3333 783.1607 765.4039 796.7137
Lena 1021383.0980 457505.9372 1382757.2627 8929.4431 8930.2823 9570.1019
Baboon 6346579.1843 10106060.6980 5938608.9490 133133.6470 135706.9490 136423.2078

causes a plausibility of consistency which undermines its Besides, to verify the randomness, local Shannon entropy
security. The closer the value is to 8, the greater the un- should be applied. It can be calculated by the following
certainty is and the stronger the randomness of image is, operations: ① divide the image into noninterlocked K
which leads to better-secured encryption where the less blocks containing a certain fixed number of pixels; ②
visual information can be obtained from the image. The compute Shannon entropy IE(Ki ) using the former
most famous entropy formula is Shannon’s entropy equa- equation (12); ③ calculate the sample mean of global
tion, calculated in terms of the probability of each available Shannon entropy over all these K image blocks as local
data value, which can be defined as follows: Shannon entropy.
Table 6 presents the simulation results of information
255
entropy and local Shannon entropy values, where K � 16, on
IE � − 􏽘 P(i)log2 P(i), (13) some standard original images and their respective
i�0
encrypted images, which were encrypted by the proposed
where P(i) denotes the probability of occurrence of gray image encryption algorithm. It can be seen that the results
level i in an image, that is, the proportion of the number reveal that the entropy values of each cipher image are very
of pixels with gray value i to all pixels in an image. close to the ideal value of 8, while the information entropy of
14 Complexity

Horizontal Horizontal
250 300
200
150 200

100
100
50
0 0
0 50 100 150 200 250 0 50 100 150 200 250 300
Vertical Vertical
250 300
200
150 200

100
100
50
0 0
0 50 100 150 200 250 0 50 100 150 200 250 300
Diagonal Diagonal
250 300
200
150 200

100
100
50
0 0
0 50 100 150 200 250 0 50 100 150 200 250 300
Figure 11: Adjacent pixel correlation test for plain grayscale “Boat” image (a) and the corresponding encrypted image (b) for horizontal,
vertical, and diagonal directions.

300 250 250

Pixel Gray Value on Location (x+1, y)
Pixel Gray Value on Location (x+1, y)

Pixel Gray Value on Location (x+1, y)

250 200 200

200 150 150

150 100 100

100 50 50

50 0 0
60 80 100 120 140 160 180 200 220 240 260 0 50 100 150 200 250 0 50 100 150 200 250
Pixel Gray Value on Location (x, y) Pixel Gray Value on Location (x, y) Pixel Gray Value on Location (x, y)

(a) (b) (c)

300 300 300
Pixel Gray Value on Location (x+1, y)
Pixel Gray Value on Location (x+1, y)
Pixel Gray Value on Location (x+1, y)

250 250 250

200 200 200

150 150 150

100 100
100

50 50
50

0 0
0 0 50 100 150 200 250 300
0 50 100 150 200 250 300 0 50 100 150 200 250 300
Pixel Gray Value on Location (x, y) Pixel Gray Value on Location (x, y)
Pixel Gray Value on Location (x, y)

(d) (e) (f )

Figure 12: Correlation distribution of color “Lena” image: (a)–(c) show RGB layers of plain image; (d)–(f ) show RGB layers of cipher image
for vertical, horizontal, and diagonal directions, respectively.

each plain image is much less than the ideal one. This result encrypted images have a good randomness. As a conclusion,
makes obtaining image information by analyzing this in- the proposed scheme is safe against the perspective of in-
formation difficult for attackers. This indicates that the formation entropy attack.
Complexity 15

Table 5: Correlation coefficients results.

Plain image Cipher image
Image name
Vertical Horizontal Diagonal Vertical Horizontal Diagonal
Aerial 0.8602 0.9050 0.8213 0.0062 −0.0017 0.0031
Boat 0.9713 0.9381 0.9222 −0.0012 0.0007 0.0004
Male 0.9813 0.9774 0.9671 0.0001 0.0002 0.0023
Airplane 0.9174 0.9314 0.8643 0.0061 0.0109 0.0012
Lena 0.9902 0.9804 0.9695 0.0040 −0.0003 −0.0012
Baboon 0.9765 0.9877 0.9671 0.0023 0.0012 0.0001

Table 7 presents the values of information entropy of the 􏽐M N

i�1 􏽐j�1 D(i, j) (15)
proposed scheme as compared with the values which NPCR � × 100%,
resulted from other recent schemes. It can be seen that the M×N
information entropy of the different cipher images is very M N 􏼌􏼌
􏼌 􏼌􏼌
􏼌
close to 8 bits and the proposed algorithm has greater su- UACI �
1 ⎛⎝􏽘 􏽘 􏼌C1 (i, j) − C2 (i, j)􏼌⎞ ⎠ × 100%.
periority or in the same range. M × N i�1 j�1 255
(16)
4.4. Differential Attack Analysis. Differential attacks are Taking the images that are listed in Table 1 as examples
another effective and commonly used cryptanalysis tech- and experimenting on them for 100 times, the theoretical
nique. A differential attack is attempted to learn the key and ideal values of the NPCR and UACI for a gray image are
figure out the encryption scheme by tracing differences. An 99.6094% and 33.4635%, respectively. Table 8 lists the test
Assailant may make a trivial change in the plain image, results of NPCR and UACI of grayscale encrypted images,
encrypt two plain images, and then carry out cryptanalysis whereas the theoretical values of NPCR and UACI for
by tracing the meaningful relationship between two cipher different color images in three channels are shown in Table 9.
images. According to the principles of cryptography, the It can be observed from the former tables that the proposed
encryption algorithm should be sufficiently sensitive to the image encryption algorithm can achieve better perfor-
changes of plaintext image or secrete key in order to keep mances against differential attacks, since the values of NPCR
high security, such that a minor change in the plaintext and UACI are close to their theoretical values. Thus, the
image or the initial key parameters causes a significant system has guaranteed that the designed system is applicable
change in the ciphertext image [59–64]; then differential for real-time communication.
analysis may become useless. The high sensitivity of the
system shows that the generated algorithm is sturdy against
any probable attack, since it would indicate no meaningful
relationship between the plain image and the cipher image. 4.5. Known-Plaintext Attack and Chosen-Plaintext Attack
In this test, the number of pixels changing rate (NPCR) and Analysis. A cryptosystem is supposed to be secure if it
unified average changing intensity (UACI) become two resists all known types of cryptographic attacks. In
widely used security analyses in the image encryption cryptanalysis, the fundamental assumption enunciated by
community for differential attacks. The tests signify the Kerckhoffs’s principle is that encryption and decryption
chance of occurrence of the attack and its sensitivity towards algorithms are known or transparent in a cryptosystem
the source image by changing the value. [65–68]. Therefore, the security of the cryptosystem de-
Considering C1 and C2 as the two cipher images obtained pends on the key rather than the encryption algorithm
itself. In the cryptanalysis, there are four traditional
from encrypting two one-pixel different images with M × N
cryptanalysis attacks: (1) ciphertext-only attack, (2) known-
size or encrypting same plain image with two secret keys of
plaintext attack, (3) chosen-plaintext attack, and (4) cho-
only 1-bit difference, introduce a bipolar array, D, with the
sen-ciphertext attack. Among these attacks, chosen-
sizes similar to images C1 and C2 as follows:
plaintext attack is the most threatening attack. Therefore, it
0, C1 (i, j) � C2 (i, j), is claimed that the cryptosystem can resist the other three
D(i, j) � 􏼨 (14) types of attacks if it can resist the chosen-plaintext attack.
1, C1 (i, j) ≠ C2 (i, j).
In order to assess the resistance of encryption algorithms
The NPCR reflects the change rate of the gray value of against the main attacks, two tests are generally used,
different pixels at the same position between two corre- namely, the known-plaintext attack (KPA) and chosen-
sponding encrypted images which are obtained by two plaintext attack (CPA). In known-plaintext attack and
original images with one-bit difference. In other words, chosen-plaintext attack, the attackers usually choose spe-
NPCR helps us to understand the effect of change of single cial plaintext and make minor changes to observe the
pixel over an image, while the UACI reflects the average changes of ciphertext. Or they choose some plaintext with
change of the gray value within the two paired cipher images linear relationship to observe the characteristics of ci-
(C1 and C2). Then, the formula used to calculate UACI and phertext. By using this method, they can obtain secret key.
NPCR is shown in the two following equations: By using this method, they can obtain secret key.
16 Complexity

Table 6: Results of information entropy and local Shannon entropy.

Information entropy Local Shannon entropy
Image name Dimension
Plain image Cipher image Plain image Cipher image
Aerial 256 × 256 3.3556 7.9970 3.2893 7.9556
Boat 512 × 512 3.3153 7.9993 3.1037 7.9880
Male 1024 × 1024 3.3540 7.9998 3.2127 7.9971
Airplane 256 × 256 × 3 6.6906 7.9987 6.1280 7.9834
Lena 512 × 512 × 3 7.7495 7.9997 7.3136 7.9959
Baboon 1024 × 1024 × 3 7.7208 7.9999 7.4281 7.9990

Table 7: Information entropy comparison.

Information entropy
Image name Dimension
Proposed Scheme Ref. [6] Ref. [39] Ref. [58] Ref. [59]
Aerial 256×256 7.9970 — — 7.9024 —
Boat 512 × 512 7.9993 — 7.9993 — —
Male 1024 × 1024 7.9998 — 7.9998 — —
Airplane 512 × 512 × 3 7.9997 — — — 7.9994
Lena 512 × 512 × 3 7.9997 7.9988 — — 7.9994
Baboon 512 × 512 × 3 7.9997 — — — 7.9993

Table 8: NPCR and UACI results for cipher grayscale images.

Image name Image size NPCR (%) UACI (%)
Aerial 256 × 256 99.6458 33.5243
Boat 512 × 512 99.6517 33.4416
Male 1024 × 1024 99.6300 33.4864

Table 9: NPCR and UACI scores on color images of different sizes.

NPCR (%) UACI (%)
Image name Dimension
Red Green Blue Red Green Blue
Airplane 256 × 256 × 3 99.6892 99.6380 99.64113 33.3572 33.5681 33.6369
Lena 512 × 512 × 3 99.6394 99.6417 99.6574 33.4980 33.4593 33.5460
Baboon 1024 × 1024 × 3 99.6241 99.6168 99.6238 33.4607 33.4702 33.4707

In the presented encryption scheme, the mean (M) value other cipher images by using the same keys. By observing the
of the plaintext image is computed to generate the number of resulting encrypted images, we can find that it is impossible to
preiterations, which is related chaotic sequences generation, extract any information from the encrypted images. Therefore,
and the initial value of diffusion process. In other words, the the proposed encryption scheme is sufficiently robust to resist
generated random sequences are related to the plaintext, and all forms of potential attacks.
the chaotic systems are sensitive to the initial value. Con-
sequently, the keystream used in the proposed algorithm has
a high connection with the plain image, which means that a 4.6. Time Complexity Analysis. Apart from security analysis
small change in the plaintext image produces a completely of the image encryption scheme, performance analysis is also
different key, as detailed in the “Key Sensitivity Analysis” an important aspect to evaluate the encryption/decryption
section. That means the attacker cannot extract any useful time and time complexity of the algorithm [69, 70]. A good
information by encrypting certain selected images because encryption algorithm needs to have a fast encryption time
the encrypted image is only relevant to the selected image, and low computation complexity.
which implies the excellent performance in withstanding the The encryption/decryption time can be calculated
known-plaintext attack and chosen-plaintext attack. manually where it is mainly analyzed into six parts as fol-
Besides, to test the ability of defending this kind of attack, lows: (a) mean for any row column of R, G, and B channels,
both plain images with “pure white” and “pure black” images, so its complexity is O(1); (b) the cyclic process N times for
their encrypted images, and the corresponding histograms are Chen’s hyperchaotic system quantum logistic chaotic map,
derived, which are shown in Figure 13. From the results, it can so it has complexity of O(n); (c) the generation of three
be seen that the pixels in the cipher image are uniformly chaotic sequences (ER , EG , EG ), which are produced by
distributed with random noise, and the attacker cannot decrypt Chen’s hyperchaotic system with length M × N and hence
Complexity 17

2500

2000

1500

1000

500

0
0 50 100 150 200 250
(a) (b) (c)
2500

2000

1500

1000

500

0
0 50 100 150 200 250
(d) (e) (f )

Figure 13: Simulation result of cryptanalysis tests: (a) all white image; (b) cipher image of (a); (c) histogram of (b); (d) all black image; (e)
cipher image of (d); (f ) histogram of (e).

Table 10: Running time analysis.

Image name Image size Encryption/decryption time (s)
Grayscale Aerial 256 × 256 0.4844
Color Airplane 256 × 256 0.8125
Grayscale Boat 512 × 512 1.4375
Color Lena 512 × 512 3.2969
Grayscale Male 1024 × 1024 5.7500
Color Baboon 1024 × 1024 13.6250

Table 11: Speed performance analysis (seconds).

Encryption scheme Encryption time (s) Processor speed RAM Platform
Ref. [71] 3.45 3 GHz 4 GB Python 3.6
Ref. [22] 13.90 — — MATLAB R2016b
Ref. [72] 1.67 — — MATLAB
Ref. [73] 9.36 3.60 GHz 32 GB MATLAB R2019b
Proposed algorithm 1.11 1.80 GHz 8 GB MATLAB R2017b

the time cost is (3 × M × N); (c) XOR operation having time presented in Table 10. As a result, the proposed scheme reflects
complexity of O(1); (d) time cost of chaotic map sequences the efficiency to be used in practical cases.
and the generation of random matric being Taking the 256 × 256 “Lena” image as an example,
O(n2 ) � max􏼈O(1), O(n2 )􏼉; (e) the computational cost of comparative analyses of the execution time among different
density matrix being O(1); and (f ) the computational cost of encryption algorithms are illustrated in Table 11. It is ob-
XNOR operation being O(1). From the above analyses, the served that the proposed algorithm runs faster than the
total time cost of the proposed scheme is O(n2 ), so that the referenced algorithms [71–73]. In addition, it has less
time consumption of proposed scheme hinges on t repre- computational complexity.
senting the number of code loops.
It can be calculated by using the in-built operations of the 5. Conclusion
software used for implementation. Here, the elapsed time was
measured by the tic and toc functions of MATLAB. The Complex nonlinearity was preserved by choosing suitable
running speed of the proposed encryption scheme for a chaotic maps. By choosing a high-dimensional chaotic
number of standard images with diverse sizes (M × M) is system, the key space is increased. This study employed a
18 Complexity

chaotic quantum logistic map, combining with both con- [6] M. Khan and H. M. Waseem, “A novel image encryption
fusion and diffusion operations, to propose a new symmetric scheme based on quantum dynamical spinning and rota-
image encryption algorithm. This algorithm is based on tions,” PLoS ONE, vol. 13, no. 11, Article ID e0206460, 2018.
Chen’s hyperchaotic system to diffuse image pixels. Among [7] X. Liu, D. Xiao, and C. Liu, “Quantum image encryption
them, the keystreams extracted are different for the same algorithm based on bit-plane permutation and sine logistic
secret key associated with the plain image, which are true map,” Quantum Information Processing, vol. 19, no. 8,
pp. 1–23, 2020.
random numbers generated from noise arrays. Thus, the
[8] M. Ge and R. Ye, “A novel image encryption scheme based on
presented approach can achieve high resistance to the
3D bit matrix and chaotic map with Markov properties,”
known-plaintext attack and chosen-plaintext attack as well Egyptian Informatics Journal, vol. 20, no. 1, pp. 45–54, 2019.
as high level of sensitivity where the randomness of the [9] Y. Dong, X. Huang, Q. Mei, and Y. Gan, “Self-adaptive image
random sequence displayed better behavior. At last, to encryption algorithm based on quantum logistic map,” Se-
confuse the relationship between original and encrypted curity and Communication Networks, vol. 2021, Article ID
images, the transpose process is applied to rows and columns 66749448, 12 pages, 2021.
of image. Through the results of extensive experiments and [10] H. Liu, B. Zhao, and L. Huang, “Quantum image encryption
corresponding security analysis, it can be found that the scheme using arnold transform and S-box scrambling,” En-
salient features of the proposed symmetric image encryption tropy, vol. 21, no. 4, p. 343, 2019.
algorithm can be summarized as follows: (a) large enough [11] Y. Luo, R. Zhou, J. Liu, Y. Cao, and X. Ding, “A parallel image
key space to resist brute-force attacks, (b) high level of encryption algorithm based on the piecewise linear chaotic
security and being quite worthy of being called a good se- map and hyper-chaotic map,” Nonlinear Dynamics, vol. 93,
curity system, (c) less computational complexity, and (d) no. 3, pp. 1165–1181, 2018.
[12] Y. Pourasad, R. Ranjbarzadeh, and A. Mardani, “A new al-
being suitable for applications like wireless communications
gorithm for digital image encryption based on chaos theory,”
due to its fast implementation. An actual implementation of
Entropy, vol. 23, no. 3, p. 341, 2021.
different kinds of operations in the scrambling stage to [13] Z. Tang, Y. Yang, S. Xu, C. Yu, and X. Zhang, “Image en-
increase the security without affecting drastically the pro- cryption with double spiral scans and chaotic maps,” Security
cessing time is concerned and more detailed analysis on the and Communication Networks, vol. 2019, Article ID 8694678,
chaotic or hyperchaotic dynamical systems deserves further 15 pages, 2019.
investigation in the near future. [14] X. Chai, Z. Gan, K. Yang, Y. Chen, and X. Liu, “An image
encryption algorithm based on the memristive hyperchaotic
system, cellular automata and DNA sequence operations,”
Data Availability Signal Processing: Image Communication, vol. 52, pp. 6–19,
The data that support the findings of this study are openly 2017.
[15] K. K. Butt, G. Li, F. Masood, and S. Khan, “A digital image
available in [USC-SIPI Image Database] at [http://sipi.usc.
confidentiality scheme based on pseudo-quantum chaos and
edu/ database/], reference number [32].
Lucas sequence,” Entropy, vol. 22, no. 11, p. 1276, 2020.
[16] Y. Liu, B. Zhou, Z. Li, J. Deng, and Z. Cai, “An image en-
Conflicts of Interest cryption method based on quantum fourier transformation,”
International Journal of Intelligence Science, vol. 8, no. 3,
The authors declare that they have no conflicts of interest. pp. 75–87, 2018.
[17] X. Liu, D. Xiao, and C. Liu, “Double quantum image en-
cryption based on arnold transform and qubit random ro-
References tation,” Entropy, vol. 20, no. 11, pp. 1–16, 2018.
[18] N. Zhou, X. Yan, H. Liang, X. Tao, and G. Li, “Multi-image
[1] B. Sinha, S. Kumar, and C. Pradhan, “Comparative analysis of
encryption scheme based on quantum 3D Arnold transform
color image encryption using 3D chaotic maps,” in Pro-
and scaled Zhongtang chaotic system,” Quantum Information
ceedings of the International Conference on Communication
and Signal Processing (ICCSP), pp. 332–335, Melmaruvathur, Processing, vol. 17, no. 12, pp. 1–36, 2018.
India, April 2016. [19] C. Zhu, Z. Gan, Y. Lu, and X. Chai, “An image encryption
[2] J. Xu, P. Li, F. Yang, and H. Yan, “High intensity image algorithm based on 3-D DNA level permutation and sub-
encryption scheme based on quantum logistic chaotic map stitution scheme,” Multimedia Tools and Applications Inter-
and complex hyperchaotic system,” IEEE Access, vol. 7, national Journal, vol. 17, no. 12, pp. 7227–7258, 2019.
pp. 167904–167918, 2019. [20] A. A. El-Latif, B. A. El-Atty, M. Amin, and A. M. Iliyasu,
[3] Z. Li, C. Peng, W. Tan, and L. Li, “An effective chaos-based “Quantum-inspired cascaded discrete-time quantum walks
image encryption scheme using imitating jigsaw method,” with induced chaotic dynamics and cryptographic applica-
Complexity, vol. 2021, Article ID 88249115, 18 pages, 2021. tions,” Scientific Reports, vol. 10, p. 1, 2020.
[4] X. Wang, N. Guan, H. Zhao, S. Wang, and Y. Zhang, “A new [21] A. Alghafis, N. Munir, M. Khan, and I. Hussain, “An en-
image encryption scheme based on coupling map lattices with cryption scheme based on discrete quantum map and con-
mixed multi-chaos,” Scientific Reports, vol. 10, no. 1, 2020. tinuous chaotic system,” International Journal of Theoretical
[5] R. K. Singh, B. Kumar, D. K. Shaw, and D. A. Khan, “Level by Physics, vol. 59, no. 4, pp. 1227–1240, 2020.
level image compression-encryption algorithm based on [22] R. Sridevi and P. Philominathan, “Quantum colour image
quantum chaos map,” Journal of King Saud University - encryption algorithm based on DNA and unified logistic tent
Computer and Information Sciences, vol. 33, no. 7, pp. 844– map,” Information Sciences Letters, vol. 9, no. 3, pp. 219–231,
851, 2021. 2020.
Complexity 19

[23] W.-W. Hu, R.-G. Zhou, S. Jiang, X. Liu, and J. Luo, “Quantum self-invertible matrix,” Computational Intelligence and Neu-
image encryption algorithm based on generalized Arnold roscience, vol. 2019, no. 12, 12 pages, Article ID 9524080, 2019.
transform and Logistic map,” CCF Transactions on High [42] C. Li, F. Zhao, C. Liu, L. Lei, and J. Zhang, “A hyperchaotic
Performance Computing, vol. 2, no. 3, pp. 228–253, 2020. color image encryption algorithm and security analysis,”
[24] H. Wen, C. Zhang, P. Chen et al., “A quantum chaotic image Security and Communication Networks, vol. 2019, Article ID
cryptosystem and its application in IoT secure communica- 8132547, 8 pages, 2019.
tion,” IEEE Access, vol. 9, pp. 20481–20492, 2021. [43] H. Fan, K. Zhou, E. Zhang, W. Wen, and M. Li, “Subdata
[25] X. Wu, Y. Li, and J. Kurths, “A new color image encryption image encryption scheme based on compressive sensing and
scheme using CML and a fractional-order chaotic system,” vector quantization,” Neural Computing & Applications,
PLoS ONE, vol. 10, no. 3, Article ID e0119660, 2015. vol. 32, no. 16, pp. 12771–12787, 2020.
[26] J. Lü and G. Chen, “A new chaotic attractor coined,” Inter- [44] X. Xue, H. Jin, D. Zhou, and C. Zhou, “Medical image
national Journal of Bifurcation and Chaos, vol. 12, no. 3, protection algorithm based on deoxyribonucleic acid chain of
pp. 659–661, 2002. dynamic length,” Frontiers in Genetics, vol. 12, pp. 1–18,
[27] R. Zhang, L. Yu, D. Jiang et al., “A novel plaintext-related Article ID 654663, 2021.
color image encryption scheme based on cellular neural [45] S. Zhou, P. He, and N. Kasabov, “A dynamic DNA color
network and chen’s chaotic system,” Symmetry, vol. 13, no. 3, image encryption method based on SHA-512,” Entropy,
p. 393, 2021. vol. 22, no. 10, p. 1091, 2020.
[28] A. Z. Mahmoud, On some new approaches for multimedia [46] H. Zhu, X. Zhang, H. Yu, C. Zhao, and Z. Zhu, “A novel image
content encryption, Ph.D. dissertation, Dept. Comp. Science, encryption scheme using the composite discrete chaotic
Menoufia Univ., Al Minufya, Egypt, 2015. system,” Entropy, vol. 18, no. 8, p. 276, 2016.
[29] X. Liu, D. Xiao, and Y. Xiang, “Quantum image encryption [47] F. Yang, J. Mou, J. Liu, C. Ma, and H. Yan, “Characteristic
using intra and inter bit permutation based on logistic map,” analysis of the fractional-order hyperchaotic complex system
IEEE Access, vol. 7, pp. 6937–6946, 2019. and its image encryption application,” Signal Processing,
[30] G. Ye, K. Jiao, C. Pan, and X. Huang, “An effective framework vol. 169, pp. 1–19, Article ID 107373, 2020.
for chaotic image encryption based on 3D logistic map,” [48] N. Tsafack, A. M. Iliyasu, N. J. De Dieu et al., “A memristive
Security and Communication Networks, vol. 2018, no. 11, RLC oscillator dynamics applied to image encryption,”
11 pages, Article ID 8402578, 2018. Journal of Information Security and Applications, vol. 61,
[31] Y. He, Y.-Q. Zhang, X. He, and X.-Y. Wang, “A new image
Article ID 102944, 2021.
encryption algorithm based on the OF-LSTMS and chaotic
[49] Z. Deng and S. Zhong, “A digital image encryption algorithm
sequences,” Scientific Reports, vol. 11, no. 1, pp. 1–22, 2021.
based on chaotic mapping,” Journal of Algorithms & Com-
[32] SIPIUSC, “The USC-SIPI image database,” 2021, http://sipi.
putational Technology, vol. 13, pp. 1–11, 2019.
usc.edu/database/.
[50] J. Zeng and C. Wang, “A novel hyperchaotic image encryption
[33] Y. Zhou, C. Li, W. Li, H. Li, W. Feng, and K. Qian, “Image
system based on particle swarm optimization algorithm and
encryption algorithm with circle index table scrambling and
cellular automata,” Security and Communication Networks,
partition diffusion,” Nonlinear Dynamics, vol. 103, no. 2,
vol. 2021, Article ID 6675565, 15 pages, 2021.
pp. 2043–2061, 2021.
[51] L. Ding and Q. Ding, “A novel image encryption scheme
[34] X. Yan, X. Wang, and Y. Xian, “Chaotic image encryption
based on 2D fractional chaotic map, DWT and 4D hyper-
algorithm based on arithmetic sequence scrambling model
and DNA encoding operation,” Multimedia Tools and Ap- chaos,” Electronics, vol. 9, no. 8, p. 1280, 2020.
[52] I. Yasser, M. A. Mohamed, A. S. Samra, and F. Khalifa, “A
plications, vol. 80, no. 7, pp. 10949–10983, 2021.
[35] M. Liu and G. Ye, “A new DNA coding and hyperchaotic chaotic-based encryption/decryption framework for secure
system based asymmetric image encryption algorithm,” multimedia communications,” Entropy, vol. 22, no. 11,
Mathematical Biosciences and Engineering, vol. 18, no. 4, p. 1253, 2020.
pp. 3887–3906, 2021. [53] H. Liu, B. Zhao, J. Zou, L. Huang, and Y. Liu, “A lightweight
[36] C. Fu, J.-j. Chen, H. Zou, W.-h. Meng, Y.-f. Zhan, and image encryption algorithm based on message passing and
Y.-w. Yu, “A chaos-based digital image encryption scheme chaotic map,” Security and Communication Networks,
with an improved diffusion strategy,” Optics Express, vol. 20, vol. 2020, pp. 1–12, 2020.
no. 3, pp. 2363–2378, 2012. [54] N. Sanam, A. Ali, T. Shah, and G. Farooq, “Non-associative
[37] Y. Wan, S. Gu, and B. Du, “A new image encryption algorithm algebra redesigning block cipher with color image encryp-
based on composite chaos and hyperchaos combined with tion,” Computers, Materials & Continua, vol. 67, no. 1,
DNA coding,” Entropy, vol. 22, no. 2, pp. 1–19, 2020. pp. 1–21, 2021.
[38] K. Jiao, G. Ye, Y. Dong, X. Huang, and J. He, “Image en- [55] F. Naz, I. A. Shoukat, R. Ashraf, U. Iqbal, and A. Rauf, “An
cryption scheme based on a generalized arnold map and RSA ASCII based effective and multi-operation image encryption
algorithm,” Security and Communication Networks, vol. 2020, method,” Multimedia Tools and Applications, vol. 79, no. 31-
pp. 1–14, 2020. 32, pp. 22107–22129, 2020.
[39] J. Ge, “ALCencryption: A secure and efficient algorithm for [56] D. W. Ahmed, T. M. Jawad, and L. M. Jawad, “An effective
medical image encryption,” Computer Modeling in Engi- color image encryption scheme based on double piecewise
neering and Sciences, vol. 125, no. 3, pp. 1083–1100, 2020. linear chaotic map method and RC4 algorithm,” Journal of
[40] S. Zhu, C. Zhu, and W. Wang, “A new image encryption Engineering Science & Technology, vol. 16, no. 2, pp. 1319–
algorithm based on chaos and secure hash SHA-256,” En- 1341, 2021.
tropy, vol. 20, no. 9, p. 716, 2018. [57] S. Zhu and C. Zhu, “Security analysis and improvement of an
[41] X. Zhang, L. Wang, Y. Niu, G. Cui, and S. Geng, “Image image encryption cryptosystem based on bit plane extraction
encryption algorithm based on the H-fractal and dynamic and multi chaos,” Entropy, vol. 23, no. 5, p. 505, 2021.
20 Complexity

[58] Y. Chen, C. Tang, and Z. Yi, “A novel image encryption

scheme based on PWLCM and standard map,” Complexity,
vol. 2020, Article ID 3026972, 23 pages, 2020.
[59] F. Masood, J. Ahmad, S. A. Shah, S. S. Jamal, and I. Hussain,
“A novel hybrid secure image encryption based on julia set of
fractals and 3D Lorenz chaotic map,” Entropy, vol. 22, no. 3,
p. 274, 2020.
[60] Z. Hua, Z. Zhu, S. Yi, Z. Zhang, and H. Huang, “Cross-plane
colour image encryption using a two-dimensional logistic tent
modular map,” Information Sciences, vol. 546, pp. 1063–1083,
2021.
[61] U. Erkan, A. Toktas, S. Enginoglu, E. Akbacak, and
D. N. H. Thanh, “An image encryption scheme based on
chaotic logarithmic map and key generation using deep
CNN,” Multimedia Tools and Applications, vol. 81, no. 78,
pp. 7365–7391, 2022.
[62] C. Xu, J. Sun, and C. Wang, “A novel image encryption al-
gorithm based on bit-plane matrix rotation and hyper chaotic
systems,” Multimedia Tools and Applications, vol. 79, no. 9-10,
pp. 5573–5593, 2020.
[63] G. Ye, K. Jiao, X. Huang, B.-M. Goi, and W.-S. Yap, “An image
encryption scheme based on public key cryptosystem and
quantum logistic map,” Scientific Reports, vol. 10, no. 1, p. 19,
2020.
[64] M. A. A.-J. A. Mizher, R. Sulaiman, A. M. A. Abdalla, and
M. A. A. Mizher, “A simple flexible cryptosystem for meshed
3D objects and images,” Journal of King Saud University -
Computer and Information Sciences, vol. 33, no. 6, pp. 844–
851, 2019.
[65] Y. Luo, X. Ouyang, J. Liu, and L. Cao, “An image encryption
method based on elliptic curve elgamal encryption and
chaotic systems,” IEEE Access, vol. 7, pp. 38507–38522, 2019.
[66] H.-Y. Gu, W.-Q. Yan, and J.-H. Zhang, “A novel image en-
cryption scheme based on hyperchaotic cellular automaton,”
Journal of Computers, vol. 31, no. 6, pp. 155–168, 2020.
[67] Y. Dong, X. Huang, and G. Ye, “Visually meaningful image
encryption scheme based on DWTand schur decomposition,”
Security and Communication Networks, vol. 2021, Article ID
6677325, 16 pages, 2021.
[68] L. M. Heucheun Yepdia, A. Tiedeu, and G. Kom, “A robust
and fast image encryption scheme based on a mixing tech-
nique,” Security and Communication Networks, vol. 2021,
Article ID 6615708, 17 pages, 2021.
[69] B. Mondal, P. K. Behera, and S. Gangopadhyay, “A secure
image encryption scheme based on a novel 2D sine-cosine
cross-chaotic (SC3) map,” Journal of Real-Time Image Pro-
cessing, vol. 18, no. 1, pp. 1–18, 2021.
[70] R. I. Abdelfattah, H. Mohamed, and M. E. Nasr, “Secure image
encryption scheme based on DNA and new multi chaotic
map,” Journal of Physics: Conference Series, vol. 1447, no. 1,
pp. 1–11, Article ID 012053, 2020.
[71] X. Hu, L. Wei, W. Chen, Q. Chen, and Y. Guo, “Color image
encryption algorithm based on dynamic chaos and matrix
convolution,” IEEE Access, vol. 8, pp. 12452–12466, 2020.
[72] X. Wang and Y. Su, “Color image encryption based on chaotic
compressed sensing and two-dimensional fractional Fourier
transform,” Scientific Reports, vol. 10, pp. 1–19, Article ID
18556, 2020.
[73] D. Zhang, L. Chen, and T. Li, “Hyper-chaotic color image
encryption based on transformed zigzag diffusion and RNA
operation,” Entropy, vol. 23, no. 3, p. 361, 2021.