Multimedia Tools and Applications

https://doi.org/10.1007/s11042-018-6989-y

A New Approach in Adaptive Encryption Algorithm

for Color Images Based on DNA Sequence Operation
and Hyper-Chaotic System

Hamid Reza Amani 1 & Mahdi Yaghoobi 1

Received: 7 November 2017 / Revised: 14 October 2018 / Accepted: 28 November 2018

# Springer Science+Business Media, LLC, part of Springer Nature 2019

Abstract
The security requirements of digital images led to the development of effective encryption
techniques. In this paper, the proposed approach includes the use of the DNA sequence and
hyper-chaotic dynamics in adaptive encryption of color images. This method consists of two
main steps: In the first step, the pixels logical arrangement is upset by using Arnold’s cat map.
In the second step, the scrambled image is encrypted by using a combination of three methods,
including the Chen’s hyper-chaotic system, DNA sequence, and proposed adaptive method,
which play a significant role in the complexity of the proposed algorithm. The simulation and
experimental results and security analysis indicated that the proposed algorithm not only
produces a significant encryption effect, but is also capable of effectively resisting statistical
and brute-force attacks.

Keywords Image Encryption . DNA sequence . Hyper Chaotic System . Adaptive Encryption

1 Introduction

In recent years, it became imperative to secure digital images due to their large number
everywhere. Many applications, such as military image database, secret video conferencing,
medical imaging system, personal photo album et cetera require a safe, fast, and reliable
method of transferring and storing digital images. Over the past decade, multiple encryption
algorithms have been developed based on various principles in literature works. Among many
methods of storage, image encryption methods are one of the most effective and common

* Mahdi Yaghoobi
[email protected]
Hamid Reza Amani
[email protected]

1
Electrical Engineering Department, Mashhad Branch, Islamic Azad University, Mashhad, Iran
 Multimedia Tools and Applications

methods to protect image information. Due to the special format of image storage, and
especially the large volume of video data, classic block encryption algorithms are not any
longer considered effective, such as the Data Encryption Standard (DES), International Data
Encryption Standard (IDES), and the Advanced Encryption Standard (AES) etc. Many
encryption methods have already been proposed with the purpose of reducing redundancy
of image content through specific operations, such as chaos-based encryption [20, 39, 44] and
DNA-based encryption operations [12]. The chaos system process includes a variety of
characteristics, such as high sensitivity to initial conditions, certainty, ergodicity, and so on.
Chaotic sequences are created by chaotic mappings that are pseudo-random sequences. Such
structures are very complex and difficult to analyze and predict [8, 13, 21, 22, 29]. The general
encryption based on chaotic mapping can be divided into two steps: permutation and diffusion.
In practice, researchers often combine permutation and diffusion for high computational
security. Lian et al. [23] proposed an image encryption method with a standardized chaos
mapping. Key space is large and of high sensitivity. However, the security is not high enough.
The hyper-chaotic systems show more resistance to hacker attacks because of their complexity
and large key space compared to chaotic systems [10]. In addition, an effective encryption
algorithm must have a large enough key space to resist hackers’ blind and intense attacks [51].
Some research works focused on a hyper-chaos-based image encryption algorithm [48]. A
chaotic image encryption algorithm has a large key space, high sensitivity to initial conditions,
and high security. With DNA calculations, DNA encryption was first introduced and emerged
in the encryption field. DNA is used as a carrier of information and modern biological
technology is used in the implementation tool [47]. DNA encoding and computing is com-
posed of kernel in these cryptographic algorithms which are characterized as complementary
rules for bases, DNA addition, subtraction, and XOR operations. However, some cryptograph-
ic techniques are vulnerable due to security defects. Image encryption techniques using chaotic
systems and DNA sequences provide extremely chaotic image, a few of which are discussed
below. Clelland et al. [3] hid the famous statement "June 6 invasion: Normandy" in DNA.
They provide a new coding method instead of traditional binary decoding. Nucleotides are
used as quadruple codes, and each letter is identified with three nucleotides. Ailenberg and
Rotstein [1] proposed an improved method called Huffman for encoding in DNA that encrypts
an image successfully. Shyam et al. [36] proposed a new encryption scheme based on DNA
calculations; they encode the information and image using the natural DNA sequences, and
encrypt them using the logical XOR operation. For instance R. Guesmi, M.A.B.Farah, A.
Kachouri, M. Samet [14] have come up with a new image encryption algorithm according to
DNA masking, SHA-2, and Lorenz system. The results of the implementation suggest that this
algorithm enhances the efficiency of encoding. It betters large key space and high key
sensitivity and it is resistant to statistical and exhaustive attacks. Ping Zhen, Geng Zhao,
Lequan min, Xin-Jin [53] presented a secure image encryption plan, which was based on
logistic and spatiotemporal chaotic systems coupled with DNA encoding technique. Experi-
mental findings confirm that the plan is resistant to brute-force, statistical, and differential
attacks. Ye Liu, J. Wang, J. Fan, and L. Gong [34] presented a new algorithm according to
Lorenz and Chen’s chaotic systems. Dynamic S-boxes are mixed with chaotic system.
Experimental findings confirm that the plan has good performance and is resistant to several
attacks. Yicong Zhou, Long bao and A.L.P.chen [55] developed a new 1D chaotic system for
image encryption. Many 1D chaotic maps are mixed with larger chaotic ranges to improve the
level of security. The outcomes of the implementation confirm that the plan has is resistance to
Multimedia Tools and Applications

different attacks. Jun Peng, Shangzhu Jin and Qi Han [18] introduced a new 192-bit key-based
hybrid of chaotic maps and DNA encoding. Experiment findings confirm that the image is
highly chaotic and shows resistance to several security analyses. Xiaoling H and Guodong Ye
[50] developed a new algorithm with the help of 4D hyper chaotic system and DNA sequence.
Hyper-chaotic system is applied to produce pseudo-random sequence, transformed in to
biological DNA sequence image block diffusion. The findings of implementation confirm
the system to be resistant to known-plaintext and chosen-plaintext attacks. Recently, An image
cryptographic scheme was lately presented by Ting Hu et al. [15]. By employing high-
dimensional chaotic systems and cycle sequencing was used for DNA. Using DNA encoding
rule in this scheme, pixels of original image are encoded randomly and controlled by key
stream given by hyper-chaotic Chen system. Based on the above analyzes, considering the
properties of hyper-chaotic systems and DNA sequences, a new encryption method is pro-
posed with a more complex structure than other algorithms. This produces a good encoding
effect in order to overcome the abovementioned weaknesses by DNA sequencing and hyper-
chaotic dynamics are used in adaptive encryption of color images. The present article is
organized as follows. Section 2 is an introduction to basic theories of the proposed algorithm.
Section 3 includes the design of the proposed image encryption scheme. Section 4 covers the
simulation results. Section 5 focuses on analysis and Section 6 is the conclusion.

2 Preliminary materials

2.1 Hyper-chaotic system

A system being hyper-chaotic is important in encryption as the key length and sensitivity to the
initial values drastically increase, reducing the risk of attack and exposing of the image .In the
proposed encryption scheme, hyper-chaotic sequences generated from Chen’s hyper-chaotic
system, were used for image encryption. Chen’s hyper-chaotic system is described as Eq. (1):

8
>
> x ¼ aðx2 −x1 Þ
<
y ¼ −dx1 þ cx2 −x1 x3 −x4
ð1Þ
>
> z ¼ x1 x2 −bx3
:
q ¼ x1 þ k

In Eq. (1), a, b, c, d, k are the system parameters, when a = 36, b = 3, c = 28, d = 16, and
−0.7≤k≤0.7, the Chen’s hyper-chaotic system is in the chaotic state and able to generate four
chaotic sequences. According to a = 36, b = 3, c = 28, d = 16, and k = 0.2, Lyapunov
exponents were obtained as follows: λ1=1.552, λ2=0.023, λ3=0, λ4=-12.573 [11] (Gao et al.,
2006). The hyper-chaotic attractors are presented in Fig. 1 since the hyper-chaotic system
includes two positive Lyapunov exponents; the prediction time is often shorter than that of a
chaotic system. Therefore, it is often considered more effective for designing security algo-
rithms [49] (Yanchuk and Kapitaniak, 2001). Here, the four-order Runge–Kutta method was
employed to solve the equations and obtain the sequences x, y, z, and q. In the implementation
process, the decimal part was reserved and the integer part of the element was removed from
hyper-chaotic sequences in order to achieve better effects. Hence, four sequences with better
 Multimedia Tools and Applications

randomness were obtained. Since the numerical sequence generated by the system has better
randomness, Chen’s hyper-chaotic system compared to low dimensional chaotic system is
more suitable for encryption; in addition, the wide range of choices for the initial conditions
and control parameters result in a larger key space.

Fig. 1 The attractors of Chen's hyper-chaos system, (a) (x-y) plan, (b) (x-z) plan
Multimedia Tools and Applications

2.2 DNA sequence encryption

2.2.1 DNA encoding and decoding for image

A DNA sequence consists of four nucleic acid bases, including A (adenine), C (cytosine), G
(guanine), and T (thymine) where A and T are complementary [41], and so are G and C. Since
0 and 1 are considered complementary in the binary system, 00 and 11 and 01 and 10 are
complementary as well. By using the four bases A, C, G, and T to encode 00, 01, 10, and 11,
there must be 24 types of coding schemes. However, there are only 8 types to satisfy the
Watson–Crick complement rule (Watson and Crick, 1953), which are listed in Table 1.
Here, the DNA code was used to encode the color image. A color image can be divided into
three channels: Red, Green, and Blue channel. For an 8-bit single channel image, each pixel is
a DNA sequence with a length of 4 (the binary sequence is 8 in length). For example, if the
first pixel value of the Red channel image is 173, it is converted into a binary sequence
[10101101]. By using the above DNA encoding Rule 1 to encode, the DNA sequence can be
obtained [CCTG]. By using the DNA encoding Rule 1 to decode the above DNA sequence, a
binary sequence can be obtained [10101101]. However, if DNA encoding Rule 2 is used to
decode the same DNA sequence, another binary sequence is obtained [01011110]. Accord-
ingly, it can be considered a simple method of encryption.

2.2.2 XOR algebraic operation for DNA sequences

The rapid development of DNA computing have resulted in the development of biological and
algebraic operations based on DNA sequence [9, 19] such as XOR operation. XOR operation
for DNA sequences is carried according to traditional XOR in the binary system. For the eight
types of DNA encoding schemes, there are eight types of DNA XOR rules.
In the current study, the XOR operation was used to fusion the original and key image. For
example, if there are two DNA sequences (AGCT and CTGA), one type of XOR operation is
necessary (Table 2) to XOR them and there will be also one sequence [CCTT] as the result.
The XOR operation is reflexive. Thus, the sequence can be also obtained [AGCT] by sequence
[CATT] and XOR sequence [CTGA] through XOR. According to Table 2, any base in any row
or column is unique, in other words, the results of XOR operation is unique as well. Here, this
XOR operation rule was used to scramble the pixel values of the original image.

2.3 Scrambling the pixels of the original image by Arnold cat's map

Due to the high similarity between the pixels of an image, which is the properties of each
image, usually before encryption, a conversion is performed on the location of the pixels of the
image so that the adjacent pixels are less similar. In other words, the order of the pixels of the

Table 1 Eight kinds of schemes encoding and decoding map rule of DNA sequence

1 2 3 4 5 6 7 8

A 00 00 01 01 10 10 11 11
T 11 11 10 10 01 01 00 00
G 01 10 00 11 00 11 01 10
C 10 01 11 00 11 00 10 01
 Multimedia Tools and Applications

Table 2 One type of XOR opera-

tion for DNA sequences XOR A G C T

A A G C T
G G A T C
C C T A G
T T C G A

image is mixed by a way this is called shuffling. In 1960, Russian mathematician Vladimir
Arnold introduced the most general form of two-dimensional chaotic map, i.e. image of a cat,
called Arnold cat's map [31].

x
If there is a pixel for an N*N matrix, according to Arnold's mapping:
y
 0  
x x
0 ¼A ðmodðN ÞÞ;
y  y  ð2Þ
1 t
A¼
q qt þ 1

The control parameters of the matrix A are t and q that must be selected so that the determinant
of the matrix is equal to one. These two parameters are part of the decryption key. The control
parameters t and q were selected to be 40 and 8 respectively in the Arnold mapping.

3 The proposed image encryption scheme

Our proposed approach is using the DNA sequence and hyper-chaotic dynamics in compar-
ative cryptography of color images. This method consists of two main steps. In the first step,
the logical ordering of pixels is eliminated by Arnold's chaotic cat map; and in the second step,
the scrambled image in encrypted with the combination of three important models using Chen
hyper-chaotic system, DNA sequence, and the proposed comparative method - with great
efficiency in the complexity of our algorithm. Pseudo random numbers are the numbers
generated by a particular pattern and can be re-generated. Pseudo-random numbers are used
to find new places of image pixels and encoding an image.

3.1 Key image generation

In modern encryption, one-dimensional or multi-dimensional chaotic functions and mappings

are used to generate pseudo-random sequences to find new pixel locations and also to xor with
color values of pixels, which are generated with a completely randomized appearance sensitive
to key changes, but a pattern is followed for their production that helps us in decryption. The
key is a string of information that allows authorized persons to decrypt encrypted information.
The key image was derived from the Chen’s hyper-chaotic system and Arnold cat map. The
Chen’s system was also used to produce four 256*256 matrices.

3.2 Image encryption

Color image has overall three layers - R (red), G (green), and B (blue). The results of experiments
indicate that, although mixing the three G, R, B layers of the image into one layer prior to
Multimedia Tools and Applications

encryption measurement would bring about faster calculation speed, and simplify the method, the
layer mixed by R, G, and B may be distorted. Hence, here, firstly, we separate R, G, B layers of the
color image, respectively, and encrypt them. Finally, we merge the three encrypted layers.
Encryption plan is shown as follows in Fig. 2.
Step 1:

There are many ways to disrupt the location of pixels. In this article, we use the Arnold's cat
map method which is addressed in section 2.3 completely. So that, Arnold cat is repeated on
the image five times, the entire pixels sequence is mixed and the image is vague. Fig. 3 shows
the original image of Lena and Baboon before and after Arnold Cat map (Shuffling).
Step 2:

We define the three R, G, and B layers of the color image pixels displaced by Arnold's cat map(step
1), in the form of three color matrices, where any value of the pixel is in the range from 0 to 255.
Here, the size of plaintext is defined as m*n, i.e. the size of each layer color matrixes is m*n,
as well. We will obtain three tow-dimensional matrixes, and respectively mark them as .Trans-
form Á bR, Á bG, Á bB into binary tow-dimensional matrices, and respectively mark Á bR, Á bG, A
. After that, we use the DNA encoding rules in this article in conducting DNA coding, and
́
respectively marked DNAAbR ; DN AAbG ; DN AAbB (according to section 2.1.1 Rule 1 (Table 1)).

Step 3:

Here, one of the four variables derived from the Chen’s system is removed for the production
of the key system of the above-mentioned hyper-chaotic system (Section 2.1), which consisted
of four variables. This process is repeated as many times as the number of pixels in the image.
The index of variable that should be removed is obtained according to the following formula:

Index ¼ modððx þ y þ z þ qÞ; 4Þ;

ð3Þ
Index ¼ Index þ 1

DNAA bR
Separate Layer R
Decimal Color Binary Color DNA
Color Image Shuffled Image
the Layers Layer G DNAA bG
Matrices Matrices Encoding
Layer B DNAA bB

Chen’s
Binary Random DNA DNA
Hyperchaoc Adapve Swiching DNA Matrices of Z
Matrices Encoding Subtracon
System

DNAT R
DNA DNAT G
Decoded Image Decimal Matrices Binary Matrices
Encoding
DNAT B

Fig. 2 Block diagram of the chaotic encryption algorithm

 Multimedia Tools and Applications

Fig. 3 Original and Shuffling image

For all pixels, the sum of variables was calculated and the remainder (divided by four) was
reserved in the variable index. The index values were zero, one, two, and three. Then, the
resulting index was increased by one for achieving the following values, one, two, three, and
four. The four variables of the Chen system, which are x, y, z, q, are equivalent to numbers 1,
2, 3, and 4, respectively. The index was used to specify the variable to be discarded. For
example, if the index is equal to two, then the y variable is discarded and other variables take
new indexes. This is repeated as many times as the number of pixels in the image, and in each
step one of the four variables is discarded and the other three ones remain. This increases the
complexity and security of the proposed algorithm. It should be noted that the values obtained
were decimal, negative, and small. For example
, if index = 1, the variable X is removed from the Chen’s hyper-chaotic system, and the
other three variables are changed from the decimal form to positive ones ranging between zero
and 255 and called adaptive in the ours algorithm and this methods makes the more
complexity of the algorithm:

modðfloorðabsðsyði; jÞÞÞ*ð10^14Þ; 256Þ;

modðfloorðabsðszði; jÞÞÞ*ð10^ 14Þ; 256Þ;
ð4Þ
modðfloorðabsðsqði; jÞÞÞ*ð10^14Þ; 256Þ;
Multimedia Tools and Applications

Where the floor statement is used to round off the decimal number and the abs is the absolute
value of the number. After normalizing the value, we let alone the elements in the random
matricesZR, ZG, ZB until the size of each random matrix is m × n. This means we need to
produce 3×m×n values to create three matricesZR, ZG, ZB.

Step 4:

In this step, the variables that were attained in the step before are first transformed into binary
numbers and are encoded into a DNA sequence matrix with the help of Table 1. This paper
encodes matrix ZR , ZG , ZB to DNA matrices, and, respectively, mark the obtained matrixes
with DNAZR , DNAZG , DNAZB .

Step 5:

The three matrices derived from the preceding step, as DNA code, are Xorred, respectively by
red, green, and blue colors, transformed into DNA in step one, accordingly as is shown in
Tables 2, 3, 4, and 5 and show withDNATR, DNATG, DNATB.
DNAT R ¼ DNAzR ⊕DNAAbR
DNAT G ¼ DNAzG ⊕DNAAbG
DNAT B ¼ DNAzB ⊕DNAAbB

Step 6:

Convert the DNA matrices DNATR , DNATG , DNATB to binary matrices based on Rule 1
(Table 1), and then convert every element of them to decimal numbers, the three decimal
matrices of cipher text the pixel value of each element ranging from 0 – 255 are shaped now.
Ultimately, we get the encrypted image by mixing the three encrypted matrices (Fig. 4(b)).

4 Simulation result and analysis

In this paper, the standard gray 256 × 256 "Lena & Baboon" images (Fig. 4(a)) were the input
images for the proposed algorithm. Using the Matlab software to simulate the experiments and
parameters, the Chen’s hyper-chaotic system and Arnold cat’s map were set as follows:
x0 ¼ 0:225; y0 ¼ −0:587; z0 ¼ 0:98; q0 0:564; k ¼ 0:2; q ¼ 8; t ¼ 40

Table 3 Correlation coefficients of two adjacent pixels in the plain and cipher images

Image Plain image Cipher image

Horizontal Vertical Diagonal Horizontal Vertical Diagonal

Lena 0.9784 0.9758 0.9750 0.0054 0.0034 0.0053

 Multimedia Tools and Applications

Table 4 The information entropy

of encrypted image Model R G B

Original image 7.4150 7.6613 7.2148

Encrypted image 7.9971 7.9973 7.9974

The encrypted image is shown in Fig. 4(b), and the decoded image is shown in Fig. 4(c) and
(d), which are decoded image with different image key. From the visual point of view of the
image, there is no connection between the original and encrypted image. This indicates that the
proposed algorithm can produce a good encryption effect.

5 The security analysis

An effective encryption algorithm must resist a variety of known attacks, including cryptan-
alytic, statistical, and brute-force attacks. And, the encryption algorithm sensitive to secret keys
and key space must be enough large to resist great attacks [2, 7, 24, 30, 33, 38]. In this section,
the proposed scheme is analyzed in terms of security.

5.1 Secret key’s space analysis

A good encryption scheme should have a large key space to resist the brute-force attack. In this
paper, the initial values of the Chen’s hyper-chaotic system and the Arnold cat’s mapping
control parameters can be considered as keys (x0 , y0 , z0 , q0,a,b,c,d,k,q,t) . There are thus
seven keys.
From an encryption point of view, the size of the key space should be greater than 2100 to
provide a high level of security. If the accurate is 10 1 4 , the key space is
1014×1014×1014×1014×1014×1014×1014×1014×1014×1014×1014=10154. Therefore, the key space
is almost equal to 2514, which is much bigger than 2100 [27]. The algorithm will be thus
resistant to all brute-force attacks.

5.2 Secret key’s sensitivity analysis

We have used the Lena and Baboon color image as the input image for the proposed design. In
MATLAB software for simulation and testing, we have arranged the parameters of the Chen
chaotic system as well as Arnold cat map as follow: x0=0.225, y0=-0.587, z0=0.98,
q0=0.564,k=0.2,q=8,t=4. The encrypted images are shown in Fig. 4b, and the decoded images
in Fig. 4c and the decoded images under a different image key are shown in 4d. From the visual
point of view, there is no connection between the original image and the encrypted image.
Incorrect values for decoding Lena image: We only change the z0=0.980001 values and the rest
is the same. . Incorrect values for decoding Baboon image: We only change the y0=-0.5870001

Table 5 The NPCR% and UACI%

of ciphered image with one bit R G B

NPCR % 99.6428 99.6328 99.6773

UACI % 34.3935 34.6547 34.4841
Multimedia Tools and Applications

Fig. 4 Experimental results: (a) original image (b) encrypted image, (c) decrypted image, (d) decrypted image
under a wrong key
 Multimedia Tools and Applications

values and the rest is the same. With these interpretations (description) The Chen’s system is
sensitive to system parameters and initial values. This means that if the initial values are slightly
different, there will be no connection between the decoded and original image (Fig. 4(d)).

5.3 Resist statistical attack

5.3.1 The gray histogram analysis

The histogram reflects the frequency of emergence of pixels with different gray level in the
digital image. Cipher images of a good encryption method should have uniform histograms.
Fig. 5 is the gray histogram of the encrypted image on three levels of red, green, and blue.

Fig. 5 Histograms of the original image and the encrypted image

Multimedia Tools and Applications

Fig. 5 (continued)

Accordingly, it is evident that the values of the original gray pixels are concentrated in some
regions. But the value of the gray pixel after the encryption is more scattered and uniform over
the entire pixel value space. It is, in fact, clear that the two images are less similar. It is difficult
to use the statistics of the gray pixel value to recover the original image. As a result, the
proposed algorithm is of high capability to resist statistical attacks.

5.3.2 Correlation coefficient analysis

An effective algorithm must be able to greatly reduce the high correlation between neighboring
pixels. Figure 6 shows the horizontal, vertical, and diagonal relationship between neighboring
elements in the color image before and after encryption, which indicates a large decrease in the
 Multimedia Tools and Applications

Fig. 6 Correlation of two horizontally adjacent pixels in the original image and the encrypted image

correlation. To test the correlation between two neighboring pixels in an image, 2000
neighboring pixels (in horizontal, vertical and diagonal directions) are selected from the image
completely randomly. Then the each pair’s correlation coefficient is calculated using the
following relationships:
1 N
E ðxÞ ¼ ∑ xi
N i¼1
1 N 
DðxÞ ¼ ∑ ðxi −E ðxÞÞ2
N i¼1
1 N ð5Þ
covðx; yÞ ¼ ∑ ðxi −E ðxÞÞðyi −EðyÞÞ
N i¼1
covðx; yÞ
rx;y ¼ pffiffiffiffiffiffiffiffiffiffipffiffiffiffiffiffiffiffiffiffi
DðyÞ DðxÞ

Where x and y are the gray value of two neighboring pixels in the image, cov (x, y) is the
covariance, and E(x) is the mean. Figure 6 show the correlation between two horizontal,
vertical, and diagonal pixels in the original image and show the correlation between two
horizontal, vertical, and diagonal pixels in the encrypted image, respectively. The correlation
coefficient is listed in Table 3. According to Fig. 6 the correlation of the neighboring pixel in
the encrypted image decreased and Fig. 7 Correlation of two horizontally adjacent pixels in the
original image and the encrypted image Lena for three colors RGB. According to the results of
Table 3, it is clear that the correlation coefficient of the encrypted pixels’ correlation coefficient
of the encrypted image is close to zero. In other words, the proposed image encryption
algorithm is of high capability to resist statistical attacks.

5.4 Information entropy

Entropy is a prominent type of randomness and is defined to describe the degree of uncertainty
in the system [28, 35]. It can also be used to quantify the uncertainty of information provided
Multimedia Tools and Applications

Fig. 7 Correlation of two horizontally adjacent pixels in the original image and the encrypted image Lena for
three colors
 Multimedia Tools and Applications

by an image. Information entropy can measure the distribution of gray values in the image.
The results indicated that information with higher entropy is more uniform than the gray value
distribution. Information entropy is defined as:
n
H ðxÞ ¼ − ∑ pðxi Þlog2 pðxi Þ ð6Þ
i¼1

Where, xi is value of the L-level gray of the image. P (xi) is the possibility of xi. According
L
to ∑ P ðxiÞ ¼ 1, the information entropy value is 8 for an ideal random image; An effective
i¼0
encryption algorithm should target the information entropy of 8. In the present study, infor-
mation entropy of about 8 was obtained, the values of which are shown in Table 4 for the
pepper standard image, as well as red, green, and blue channels. As it is evident, the proposed
algorithm is very effective.

5.5 Attack analyses

5.5.1 Differential attack

According to the crypto logical principles, an effective encryption algorithm needs to be enough
sensitive to the plaintext. The sensitivity can be quantified as Number of Pixels Change Rate
(NPCR) and Unified Average Changing Intensity (UACI) [45]. NPCR is used to find meaningful
relationships between the original and encrypted image. And, UACI is used to measure the average
of two different encrypted images. NPCR and UACI are defined as following respectively,

∑M
i¼1 ∑ j¼1 Dði; jÞ
N
NPCR ¼  100%
h M N i ð7Þ
∑M
i¼1 ∑ j¼1 j
N
pði; jÞ−qði; jÞj
UACI ¼  100%
255  M  N

0 Pði; jÞ ¼ qði; jÞ
In (1), D (i,j)= Where P(i, j), q(i, j) represent the pixel value of the
1 Pði; jÞ≠qði; jÞ
two encrypted images at location (i, j). M and N represent the number of rows and columns in
the original image, respectively.
Moreover, the ideal values of NPCR and UACI is calculated through the following formula,

NPCRE ¼ ð1−2−n Þ  100%;

1
1 ∑2 −1 iði þ 1Þ ð1 þ 2−n Þ  100% ð8Þ
n
¼
UACI E ¼ 2n i¼1 n  100% 3
2 2 −1

Table 6 Comparison entropy of

the proposed image encryption Model R G B
method with recent state-of-the-art
encryption algorithms (the test im- Ours 7.9971 7.9973 7.9974
age is Lena) REF[42] 7.9971 7.9969 7.9962
REF[54] 7.9987 7.9983 7.9987
Multimedia Tools and Applications

Table 7 Comparison correlation

coefficient analysis of the proposed Model R G
image encryption method with re-
cent state-of-the-art encryption al- Horizontal 0.0064 0.0018 0.0099
gorithms (the test image is Lena) Ours Vertical 0.0053 0.0047 0.0043
Diagonal 0.0061 0.0027 0.0035
Horizontal 0.0054 0.0059 0.0012
REF[42] Vertical 0.0062 0.0016 0.0022
Diagonal 0.0017 0.0029 0.0026
Horizontal 0.0681 0.0471 0.0044
REF[54] Vertical 0. 0422 0.0119 0.1271
Diagonal 0.0497 0.0703 0.1043

Where n is the number of bits that represents the different bit planes of an image. A gray scale
image is 8 bits per pixel, n = 8. In the experiment, five groups of Lena images were encrypted.
In each group, one image is the original one and the other images are the original image with
only one randomly changed pixel value. The results are listed in Table 5. It can be concluded
that every value is near to the ideal values. Therefore, the proposed encryption algorithm is
very sensitive to small changes in the plain image and highly resistance to differential attacks.

6 Comparison results and discussion

6.1 Comparison results

In this section the performance of the proposed algorithm in entropy analysis and correlation
coefficient is compared with other image encryption methods for all three RGB parts stated in
references [42, 54]. The numerical results are shown in Tables 6 and 7, respectively, as is seen.
To illustrate the better performance of the algorithm compared to the other encryption methods
in Table 8, Key space, Entropy, NPCR%, UACI% are compared with several other references.
The results indicated that the UACI% and Key space in the proposed algorithm had better
values than other references. In addition, entropy, NPCR% and correlation coefficients
between pixels, although in some references indicated a better number than our proposed

Table 8 Performances of the pro-

posed scheme with recent state-of- Methods Key space Entropy NPCR% UACI%
the-art encryption algorithms
Ours 2514 7.9983 99.53 34.26
Ref[37] 2349 7.997 99.5804 33.4782
Ref[17] 2212 7.9998 99.122 33.1298
Ref[26] 2370 7.9896 99.6521 33.3438
Ref[52] 2100 7.9972 99.6521 33. 4152
Ref[43] 2239 7.998 99.6063 34.1951
Ref[46] 2233 7.9968 99.5506 33.4055
Ref[25] 2299 7.9989 99.6033 33.4655
Ref[40] 2186 7.9989 0.0004 0.0005
Ref[16] 266 7.9975 99.6077 33.457
Ref[5] 2128 7.997 99.57 33.42
Ref[6] 2269 7.9974 99.59 33.42
Ref[32] 2505 7.99881 99.60917 33.46566
Ref[4] 2268 7.9993 99.62 33.45
 Multimedia Tools and Applications

algorithm, but with the closeness of the values of these indices to the ideal values in the
proposed algorithm and the complexity of the proposed algorithm to other algorithms it can be
concluded that the proposed algorithm had better performance than other algorithms.

6.2 Conclusions

Given national security and economic interests, this paper emphasized the importance of
encryption work. The present study proposed a new approach to adaptive encryption algorithm
for color images based on DNA Sequence operation and hyper-chaotic dynamics. According
to the above, the conclusion is that the position of pixels can be changed by the Arnold cat's
mapping and gray value of the color image pixels by using a combination of three methods, the
chen hyper-chaotic system, the DNA sequence, and the adaptive method.
The combination of these three methods makes the more complexity of the algorithm and
the rise of the security. Through the test and security analysis, the proposed algorithm proved
to produce good encryption effects, a larger secret key space, and a high sensitivity to the
secret key. In addition, the proposed algorithm can also resist known attacks, such as
cryptanalytic, statistical, and brute-force attacks. All of these characteristics indicate that the
proposed algorithm is very effective in encoding digital color images.

References

1. Ailenberg M, Rotstein OD (2009) An improved Huffman coding method for archiving text, images, and
music characters in DNA. BioTechniques 47:747–754
2. Baum EB (1996) DNA sequences useful for computation. Proc 2nd DIMACS Workshop DNA Based
Comput:122–127
3. Cell CT, Risca V, Bancroft C (1999) Hiding messages in DNA. Microdots Nature 399:533–534
4. Chai X, Chen Y, Broyde L (2017) A novel chaos-based image encryption algorithm using DNA sequence
operations. Opt Lasers Eng 88:197–213
5. Chai X, Gan Z, Yang K, Chen Y, Liu X (2017) An image encryption algorithm based on the memristive
hyperchaotic system. Cellular automata and dna sequence operations. Signal Processing Image.
Communication 52:6–19
6. Chai X, Chen Y, Broyde L (2017) A novel chaos-based image encryption algorithm using dna sequence
operations. Opt Lasers Eng 88:197–213
7. Dong EZ, Chen ZQ, Yuan ZZ, Chen ZP (2008) A chaotic image encryption algorithm with the key mixing
proportion factor. International Conference on Information Management, Innovation Management. Ind Eng:
169–174
8. Fu C, Zhu ZL (2008) A chaotic image encryption scheme based on circular bit shift Method. 9th
International Conference for Young Computer Scientists 3057–3061
9. Gaborit P, King OD (2005) Linear constructions for DNA codes. Theor Comput Sci:33499–33113
10. GAO TG, Chen ZQ (2008) A new image encryption algorithm based on hyper-chaos. Phys Lett A 372:394–
C400
11. GAO TG, Chen ZQ, Yuan ZY, Chen G (2006) Hyper chaos generated from Chen’s system. Mod Phys C 17:
471
12. Gehani A, LaBean TH, Reif JH (2000) DNA based cryptography. DIMACS:54233249
13. Guan ZH, Huang F, Guan W (2005) Chaos-based image encryption algorithm. Phys Lett A 346:153–157
14. Guesmi R, Farah MAB, Kachouri A, Samet M (2016) A novel chaos based image encryption using DNA
sequence operation and secures hash algorithm sha-2. Nonlinear Dynamics 83(3):1123–1136
15. Hu T, Liu Y, Gong L-H (2016) An image encryption scheme combining chaos with cycle operation for
DNA sequences. Nonlinear Dyn. https://doi.org/10.1007/s11071-016-30246
16. Hu T, Liu Y, Gong LH, Ouyang CJ (2016) An image encryption scheme combining chaos with cycle
operation for dna sequences. Nonlinear Dynamics:1–16
17. Huang X, Ye G (2015) An image encryption algorithm based on hyperchaos and dna sequence. Multimed
Tools Appl 72(1):57–70
Multimedia Tools and Applications

18. Jin X, Tian Y, Song C, Wei G, Li X, Zhao G, Wang H (2015) An invertible and anti-chosen plaintext attack
image encryption method based on DNA encoding and chaotic mapping. Chinese Automation Congress
(CAC) 1159–1164
19. King OD, Gaborit P (2007) Binary templates for comma-free DNA codes. Discret Appl Math 155:831–839
20. Lian SG (2008) Multimedia Content Encryption: Techniques and Applications. ISBN: 1420065270,
Auerbach Publication Taylor & Francis Group, Boca Raton
21. Lian SG (2009) A block cipher based on chaotic neural networks. Neurocomputing 72:1296–C1301
22. Lian SG (2009) Efficient image or video encryption based on spatiotemporal Chaos system. Chaos Solitons
Fractals (Elsevier) 40(15):2509–2510
23. Lian SG, Sun JS, Wang ZQ (2005) A block cipher based on a suitable use of the chaotic standard map.
Chaos, Solitons Fractals 26:117–129
24. Lian SG, Sun JS, Wang ZQ (2005) Security analysis of a chaos-based image encryption algorithm. Phys A:
Stat Theor Phys 351(2–4):645–661
25. Liu L, Zhang Q, Wei X (2015) A rgb image encryption algorithm based on dna encoding and chaos map.
Comput Electr Eng 38(5):1240–1248
26. Liu Y, Wang J, Fan J, Gong L (2016) Image encryption algorithm based on chaotic system and dynamic s-
boxes composed of DNA sequences. Multimed Tools Appl 75(8):4363–4382
27. Mohammad SS, Benyamin N, Reza MM, Sattar M (2015) A novel color image encryption algorithm based
on spatial permutation and quantum chaotic map. Nonlinear Dyn 81:511–529
28. Norouzi B, Mirzakuchaki (2014) A fast color image encryption algorithm based on hyper-chaotic systems.
Nonlinear Dyn 78(2):995–1015
29. Pareek NK, Patidar V, Sud KK (2006) Image encryption using chaotic logistic map. Image Vis Comput
24(9):926–934
30. Peng J, Jin SZ, Liu YG et al (2008) A novel scheme for image encryption based on piecewise linear chaotic
map. IEEE Conf Cybernet Int Syst 1012–1016
31. Peterson G (1997) Arnold’s cat map
32. Rostami MJ, Shahba A, Saryazdi S, Nezamabadi-Pour H (2017) A novel parallel image encryption with
chaotic windows based on logistic map. Comput Electric Eng
33. Sabery MK, Yaghoobi M (2008) A new approach for image encryption using chaotic logistic map, in: 2008
International Conference on Advanced Computer Theory and. Engineering:585–590
34. Saranya MR, Mohan AK, Anusudha K (2015) Algorithm for enhanced image security using DNA and
genetic algorithm. Sign Proc, Inform, Commun Energy Syst (SPICES), IEEE Int Conf 1–5
35. Shannon CE (1949) Communication theory of security systems. Bell Syst Tech J 28:656715
36. Shyam M, Kiran N, Maheswaran V (2007) A novel encryption scheme based on DNA computing.
HIPC2007
37. Som S, Kotal A, Chatterjee A, Dey S (2015) A color image encryption based on DNA coding and chaotic
sequences. International Conference on Emerging Trends and Applications in Computer. Science:108–114
38. Wang L, Ye Q, Xiao YQ et al (2008) An image encryption scheme based on cross Chaotic map. Cong
Image Sign Proc:22–26
39. Wang Y, Wong K, Liao X, Xiang T, Chen G (2009) A chaos based image encryption Algorithm with
variable control parameters. Chaos, Solitons Fractals:411773–411783
40. Wang XY, Zhang YQ, Bao XM (2015) A novel chaotic image encryption scheme using DNA sequence
operation. Opt Lasers Eng 73:53–61
41. Watson JD, Crick FHC (1953) A structure for deoxyribose nucleic acid. Nature 171(4356):737–738
42. Wei X, Guoa L, Zhanga Q, Zhanga J, Lianb S (2012) A novel color image encryption algorithm based on
DNA sequence operation and hyper-chaotic system. Syst Software:85290–85299
43. Wei X, Guo L, Zhang Q, Zhang J, Lian S (2015) A novel color image encryption algorithm based on dna
sequence operation and hyper-chaotic system. J Syst Softw 85(2):290–299
44. Wong K, Kwok B, Law W (2008) A fast image encryption scheme based on chaotic standard map. Phys
Lett A 372:2645–2652
45. Wu Y, Noonan JP, Agaian S (2011) NPCR and UACI randomness tests for image encryption. Cyber J:
Multidiscip J Sci Technol J Sel Areas Telecommun (JSAT):31–38
46. Wu X, Kan H, Kurths J (2015) A new color image encryption scheme based on DNA sequences and
multiple improved 1d chaotic maps. Appl Soft Comput 37(C):24–39
47. Xiao GZ, Lu MX, Qin L, Lai XJ (2006) New field of cryptography: DNA cryptography. Chin Sci Bull
51(12):1413–1420
48. Xue XL, Zhang Q (2010) An image fusion encryption algorithm based on DNA sequence and multi-chaotic
maps. Comput Theor Nanosci 7(2):397–403
49. Yanchuk S, Kapitaniak T (2001) Chaos-hyperchaos transition in coupled Rossler systems. Phys Rev E 64:
056235
 Multimedia Tools and Applications

50. Zhang Y (2015) Cryptanalysis of a novel image fusion encryption algorithm based on DNA sequence
operation and hyper-chaotic system. Optik – Int J Light Electron Opt 126(2):223–229
51. Zhang X, Chen WB (2008) A new chaotic algorithm for image encryption. ICALIP:889–892
52. Zhang Q, Guo L, Wei X (2010) Image encryption using DNA addition combining with chaotic maps. Math
Comput Model 52(1112):20282035
53. Zhen P, Zhao G, Min L, Jin X (2016) Chaos-based image encryption scheme combining DNA coding and
entropy. Multimed Tools Appl 75(11):6303–6319
54. Zheng W, Wang F-Y, Wang K (2017) An ACP-based Approach to Color Image Encryption Using DNA
Sequence Operation and Hyper-chaotic System. IEEE International Conference on Systems, Man, and
Cybernetics (SMC). 978-1-5386-1645-1
55. Zhou Y, Bao L, Chen CP (2014) A new 1D chaotic system for image encryption. Signal Process 97:172–
182

Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and
institutional affiliations.

Hamid Reza Amani has passed a bachelor’s degree in Electrical Engineering at Islamic Azad University. He is
graduating with a master’s degree in Control Engineering from Islamic Azad University of Mashhad.

Mahdi Yaghoobi Faculty member, Associate professor, electrical engineering department, Azad university of
MASHHAD, Iran. His area of research is chaos application, predictive control, fuzzy control and evolutionary
algorithm.