Medical image encryption algorithm based on DNA and spiral transformation

doi:10.21203/rs.3.rs-4722290/v1

Download PDF

Research Article

Medical image encryption algorithm based on DNA and spiral transformation

https://doi.org/10.21203/rs.3.rs-4722290/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Medical image encryption plays a key role in healthcare and privacy protection as an important application in the field of image encryption. In this paper, a new medical image encryption method is proposed based on spiral transform and DNA operations. The algorithm makes use of Chaotic sequences generated by a 5-dimensional hyperchaotic system to perform double helical transform, DNA encoding, arithmetic and decoding operations on medical images respectively. The emulation experiments show that the algorithm has good encryption performance and can resist various attacks, especially in the anti-clip attack with good results.

DNA dynamic coding and computing

Spiral rotation

Medical image

Hyperchaotic system

With the fast development of Internet technology, digital images have become an essential way of information storage and transmission. In medicine, medical digital images are rich in sources and diverse in forms, has become an important auxiliary tool for the prevention, diagnosis and treatment of diseases. which can help doctors to understand the treatment and diagnosis of diseases, so as to provide patients with more excellent and advanced treatment methods. Medical images include physiological and pathological information about the human body, and they often also involve the privacy of patients, and there is a certain need for confidentiality. However, malice hackers tend to seek to steal medical images through illegal tactics, selling the information or engaging in privacy extortion for financial gain. Therefore, in order to avoid information leakage and privacy exposure, effective protection measures should be built into medical images so as to ward off criminal elements' malicious attacks and extortion.

Currently, the most common methods to protect the content of medical images are medical image watermarking, medical image steganography and medical image encryption, In this, medical image cryptography is the most powerful means of securing the images because it can convert the original medical image into an unreadable one by means of a key. It has many medical image encryption methods to improve the security of the image. And in improving the security of medical image cryptography it mainly depends on these three phases of the encryption process which are key, scrambling and diffusion phases. In the case of key, the key is mainly created by introducing chaotic mapping to improve the security of the key. Whereas in the case of disarray, it mainly changes the pixel position without changing the pixel value, the diffusion phase mainly changes the pixel value.

In recent years, researchers have proposed many medical image encryption techniques, for example, For example, Ammar[1] proposed an algorithm is based on combining a seven-dimensional hyperchaotic regime with Pascal's matrix, but it has a simpler encryption process. Li et al[2] proposed a medical image cryptography algorithm based on hash algorithm, DNA encoding and four-dimensional hyper chaotic system, but its key is relatively easy to be cracked. Xiong et al[3] proposed a medical image cryptography algorithm based on improved chaotic system, block rotation and DNA computation to design a fully chaotic, strong plaintext management, but the algorithm is simpler in the process of DNA operations. Gao et al[4] proposed a hybrid multi-image encryption algorithm that can encrypt different types and sizes at the same time, and DNA operations are also used in the process. But the encryption time of this algorithm is too long. Yasmine et al[5] proposed a grey scale image encryption model based on dynamic DNA coding, hash function, and Arnold transform, but the complexity of its algorithm is too high.

After learning from the above experiences and based on the existing work and progress, in order to improve the safety of algorithm and to decrypt it correctly even if it is attacked by external attack during transmission, this paper proposes a medical image encryption algorithm based on helical transform and dynamic DNA, and the main innovations of this paper are as follows:

1. Spiral transformation is based on chunking, which can choose the appropriate sub-chunk size according to the size of medical images, thus enhancing the encryption effect and accelerating the encryption efficiency;
2. In the diffusion process, eight DNA coding rules, three DNA arithmetic rules and one random DNA decoding rule are introduced, which can complete the full diffusion operation on the scrambled image.
3. The encrypted medical image has better performance against clip attack and can protect the central region from loss;

After experimental simulation and performance evaluation analysis, it is evident that the algorithm exhibits a high level of comprehensiveness compared to current industry advanced encryption algorithms in the industry. The encryption process is straightforward and suitable for practical scenarios.

The rest is organized as shown below. Section II presents the theoretical theory of the algorithm. Section III describes in detail the encryption algorithm's structure and steps. Section IV shows the experimental emulation outcomes. Section V is the specific performance analysis. Section VI summarizes the work and expresses the future outlook.

The main focus is to introduce some basics of the algorithm of the latter.

2.1 Chaotic sequence

The Lorentzian hyperchaotic system is a typical chaotic system evolved from the Lorentzian chaotic system[6], which is formulated as follows:

$$\:\begin{array}{c}\left\{\begin{array}{c}\frac{dx}{dt}=a\left(y-x\right)+\omega\:\\\:\begin{array}{c}\frac{dx}{dt}=cx-y-xz\\\:\frac{dx}{dt}=xy-bz\\\:\frac{dx}{dt}=-yz+r\omega\:\end{array}\end{array}\right.\#\left(1\right)\end{array}$$

where x, y, z and ω are state variables; a, b, c and r are control parameters[6]. The state of the system is in a hyperchaotic state when a = 10, b = 8/3, c = 28, and r is in the range [-1.52, -0.06]. The Lorentzian hyperchaotic system can produce multiple positive Lyapunov exponents and its behavior is also more unpredictable than normal chaotic systems, thus playing an important role in the field of image security.

2.2 Spiral Transform

In the context of classical encryption algorithms, the most prevalent structure is the permutation-diffusion structure[7]. This paper introduces a disruption method based on the spiral transform, which effectively scrambles pixel data, thereby reducing correlations between pixels. Firstly, according to the size of the medical image, we need to select the appropriate size of the sub-block, and perform the intra-block helical transform on these sub-blocks first, and then perform the inter-block disruption on the image which is divided into multiple sub-blocks. In this process, the intra-block disarray is divided into four cases because of the different starting directions, which are transformations starting from the upper left corner, upper right corner, lower left corner, and lower right corner, respectively.

After the intra-block and inter-block disarrangement of the image, the disarrangement process of the algorithm is completed, and the specific operation process is shown in Fig. 1

2.3 DNA operation

After completing the permutation process, only the pixel positions are altered, and the pixel values remain unchanged[8]. This incomplete encryption process is vulnerable to cracking, necessitating a subsequent diffusion operation to alter the pixel values. This improves the safety of the encryption algorithm. In this paper, the algorithm primarily employs DNA rules for diffusion during this stage.

2.3.1 DNA coding and decoding rules

DNA molecule is mainly composed of four alkaline deoxyribonucleotides A、T、G、C[9]. Between them, A, T, G, C because of the principle of complementary pairing of the bases two by two, will be in accordance with the four bases in accordance with the binary coding rules for complementary coding, there are also eight kinds of DNA coding rules, as shown in Table 1.

Table 1

DNA coding rules
kinds	A	G	C	T
rule1	00	01	10	11
rule2	00	10	01	11
rule3	11	01	10	00
rule4	11	10	01	00
rule5	10	11	00	01
rule6	01	11	00	10
rule7	10	00	11	01
rule8	01	00	11	10

2.3.2 DNA operation rules

After DNA encoding, it is also necessary to carry out DNA arithmetic on the encoded image, mainly there are three kinds of arithmetic rules, namely DNA addition, DNA subtraction and DNA XNOR, the specific rules are shown in Tables 2, 3 and 4.

Table 2

DNA addition rules
་	A	G	C	T
A	A	G	C	T
G	G	C	T	A
C	C	T	A	G
T	T	A	G	C

Table 3

DNA subtraction rules
-	A	G	C	T
A	A	G	C	T
G	G	A	T	C
C	C	T	A	G
T	T	C	G	A

Table 4

DNA XNOR rules
XNOR	A	G	C	T
A	T	C	G	A
G	C	T	A	G
C	G	A	T	C
T	A	G	C	T

In the process of DNA arithmetic on the encoded medical images, different arithmetic rules are selected according to different sub-blocks until all sub-blocks have completed the arithmetic.

After the completion of the operation, it is necessary to carry out DNA decoding, the specific method is similar to the encoding, the same rules as those used in encoding can be decoded, the pixel value of the decoded image is binary, and it is also necessary to convert it to decimal in order to obtain the final encrypted result.

2.3.3 DNA inverse operation rules

Here is mainly used in the decryption process, because the inverse operation is needed in the decryption, it is different from the rules used in the forward operation, which need to be explained, the inverse operation of addition, subtraction and XNOR are shown in Tables 5, 6 and 7 respectively.

Table 5

Addition inverse rules
inverse +	A	G	C	T
A	A	T	C	G
G	G	A	T	C
C	C	G	A	T
T	T	C	G	A

Table 6

subtraction inverse rules
inverse -	A	G	C	T
A	A	G	C	T
G	G	A	T	C
C	C	T	A	G
T	T	C	G	A

Table 7

XNOR inverse rules
inverse XNOR	A	G	C	T
A	T	C	G	A
G	C	T	A	G
C	G	A	T	C
T	A	G	C	T

The part mainly describes the medical image encryption arithmetic proposed in this paper, through which medical images can be transmitted between the sender and the receiver in the form of encrypted snowflake diagrams, which can effectively avoid the direct leakage of the information and also improve the safety of the image transmission process. The way implementation framework is depicted in Fig. 2.

3.1 Key Generation

In there, we use the method related to the plaintext image to generate the key[7], input a M*N plaintext image I, calculate the average of the grey value and the standard deviation sd of the image, and use its row and column numbers M and N to generate the four keys K1, K2, K3 and K4, the specific formulas are as follows:$\:\begin{array}{c}K1=floor(1+mod(\left(average+sd)/\text{128,10}\right))\#\left(2\right)\end{array}$

$$\:\begin{array}{c}K2=floor(1+mod(\left(average+sd+M)/\text{128,10}\right))\#\left(3\right)\end{array}$$

$$\:\begin{array}{c}K3=floor(1+mod(\left(average+sd+N)/\text{128,10}\right))\#\left(4\right)\end{array}$$

$$\:\begin{array}{c}K4=floor(1+mod(\left(average+sd+M+N)/\text{256,4}\right))\#\left(5\right)\end{array}$$

3.2 Generating Sequences and Processing

The above key associated with the plaintext image is used as the initial value of the chaotic system to generate the chaotic sequence Y, which is used in the subsequent encryption algorithm, in which a total of three chaotic sequences need to be generated, z1, z2 and z3, where z1 is used to select the starting direction during the helical transformation, z2 is used to select the coding rules when DNA encoding is carried out for the chaotic image after the helical transformation, z3 is used to select the operation rules when DNA operations are carried out after the completion of the encoding. z3 is used to select the operation rule when DNA operation is performed after the completion of encoding. Their calculation formulae are as follows:

$$\:\begin{array}{c}z1=floor\left(mod\left(Y*{10}^{10}\right),4\right)\#\left(6\right)\end{array}$$

$$\:\begin{array}{c}z2=floor\left(mod\left(Y*{10}^{10}\right),8\right)\#\left(7\right)\end{array}$$

$$\:\begin{array}{c}z3=floor\left(mod\left(Y*{10}^{10}\right),3\right)\#\left(8\right)\end{array}$$

After calculating the above three sequences there is still a need to eliminate transient effects, so the number o of sub-blocks of the plaintext image is chosen as the starting position of the sequence, ending with o + o-1, so that the length chosen for each sequence is exactly the number of sub-blocks of the plaintext image.

3.3 Encryption process

After finishing of the preliminary work, we start to accomplish the encryption process with the structure of permutation-diffusion. The process is as follows:

Step 1: Preprocessing: input the plaintext image I, obtain its row and column numbers M, N, and then select the appropriate size of sub-blocks according to the row and column numbers to carry out the subsequent encryption operations;

Step 2: Generate chaotic sequences: Generate initial values x1, x2, x3 and x4 from the average and standard deviation sd of the grey values of the original image I and the number of rows and columns M and N. After that, put them into the Lorenz super chaotic system to generate the sequence Y, and then after eliminating the transient effect and setting the values, the sequences z1, z2 and z3 are generated, which are used for the subsequent encryption operations;

Step 3: Spiral Transformation: Here, the original image I is firstly divided into sub-blocks of appropriate size, and then one by one, the spiral transformation is carried out within the sub-blocks, and it is worth mentioning that the sequence z1 is used to limit the starting point of the spiral transformation in this process. After the process is completed, the helically transformed sub-blocks are stored in the block1 array, and inter-block helical transformations are performed on them;

Step 4: DNA encoding: in this process, it is also divided into sub-blocks for encoding. Specifically, you need to use the sequence z2 to select which encoding rule to use for each sub-block, and after the selection is complete, each element in the sub-block is first converted to binary and then DNA encoded;

Step 5: DNA arithmetic: after the DNA encoding is completed, DNA arithmetic is required, which also requires selecting the corresponding arithmetic mode for the elements in the sub-block through the sequence z3;

Step 6: DNA decoding: when decoding, a random DNA decoding rule is used for decoding;

Step 7: the decoded pixel values are converted from binary to decimal to get the encrypted image.

Algorithm 1

Encryption process

Input: original image I with size of M*N.
Output: encrypt image
1: imread(I) 2: [M N] = size(I) 3: z1 = floor(mod((Y*10^10),4))
4: for i = 1:8:M
5: for j = 1:8:N
6: block2 = spiral_transformation(blocksize,z1 + 1)
7: end
8: end
9: block3 = spiral_transformation(M/8,1)
10: scrambling_I = block3
11: z2 = floor(mod(Y*10^10,8))
12: bianma_DNA = DNA_bianma_rule(scrambling_I,z2 + 1)
13: z3 = floor(mod((Y*10^10),3))
14: yunsuan_DNA = DNA_yunsuan(bianma_DNA,z3)
15: jiema_DNA = DNA_jiema(yunsuan_DNA) 16: encrypt image = = bin2dec(jiema_DNA)
17: end

3.4 Decryption process

Since the operations performed by the encryption process are all reversible, decryption is the inverse process of encryption and is not described in detail.

This section presents the simulation results of medical image encryption using the proposed encryption scheme. All experiments were conducted using Matlab 2017a software on a computer equipped with an Intel(R) Core (TM) i7-9750H CPU @ 2.60GHz processor and running Windows 10 operating system.

This chapter tests the performance of the algorithm in the following aspects:

5.1 Statistical Feature Analysis

Since the strong correlation between pixels within an image makes it easy for hackers to crack through this vulnerability, the encryption algorithm must be able to eliminate this property between pixels. The next part of this paper describes and analyses the statistical performance of the algorithm.

5.1.1 Histogram Analysis

Digital images are saved in the form of a matrix of pixels in the computer[10], so we need to analyze the spread of pixels in the encrypted image, here histogram is used, which visually helps us to observe the distribution of pixels to conclude whether the statistical attack is mitigated or not. Figure 3 shows the histogram distribution of the image encrypted by the algorithm of this paper.

Figure 3 Histogram analysis plot ((a), (e) original image, (b), (f) histogram of original image, (c), (g) encrypted image, (d), (h) histogram of encrypted image)

5.1.2 Correlation analysis

The pixels of the image are highly correlated with each other, if this property is not reduced in the image after the encryption is completed, it can be easily cracked, so it is necessary to detect the image encrypted by this algorithm, the correlation is calculated by the formula such as (9), (10), (11), and (12)[3]:

$$\:\begin{array}{c}r=\frac{cov\left(p,q\right)}{\begin{array}{cc}\sqrt{D\left(p\right)}&\:\sqrt{D\left(q\right)}\end{array}}\#\left(9\right)\end{array}$$

$$\:\begin{array}{c}cov\left(p,q\right)=\frac{1}{L}\sum\:_{i=1}^{L}\left(p\left(i\right)-E\left(p\right)\right)\left(q\left(i\right)-E\left(q\right)\right)\#\left(10\right)\end{array}$$

$$\:\begin{array}{c}D\left(p\right)=\frac{1}{L}\sum\:_{i=1}^{L}{\left(p\left(i\right)-E\left(p\right)\right)}^{2}\#\left(11\right)\end{array}$$

$$\:\begin{array}{c}E\left(p\right)=\frac{1}{L}\sum\:_{1}^{L}\left(p\left(i\right)\right)\#\left(12\right)\end{array}$$

where p and q denote the vector of selected pixels, N denotes the number of selected pixel couples, and E(p), D(p), and cov(p, q) denote the mean, variance, and covariance of the pixels. r values close to 1 denote high correlation, and those close to 0 denote low correlation[11].

The correlation coefficients for the test image are presented in Table 8. Additionally, the correlation between pixel pairs is visualized through pixel scatter plots. Specifically, correlations in horizontal, vertical, and diagonal directions are considered for both plaintext and ciphertext images. These results are depicted in Fig. 4 and summarized in Table 8.

Table 8

Correlation coefficients for selected directions
kinds
direction	plain image			Ciphered images
direction	P1	P2	image 3	C1	C2	C3
horizontal	0.8870	0.9982	0.9983	0.0193	0.0399	0.0475
vertical	0.8700	0.9983	0.9987	0.0020	0.0141	0.0273
diagonal	0.9282	0.9972	0.9975	-0.0073	0.0147	0.0126

5.1.3 Information Entropy Analysis

Information entropy is an index reflecting the distribution of pixel intensity values[12], which can be used to measure the degree of disorder of an image, and is calculated by the formula (13), where pi is the odds of pixel i. For an 8-bit grey scale image, the closer it is to 8 after encryption, the better the pixel intensity value distribution is, and the better the ability to resist statistical attacks. Table 9 shows the it of the image before and after encryption. It can be know that encryption using this algorithm can effectively conceal message and defend against attacks.

$$\:\begin{array}{c}IE=\sum\:_{i=0}^{255}\left({p}_{i}*{log}_{2}{p}_{i}\right)\#\left(13\right)\end{array}$$

Table 9

Information entropy before and after encryption
image	plain	ciphered
1	2.2044	7.9998
2	7.8651	7.9998
3	7.5360	7.9998
4	7.4298	7.9998
5	7.5167	7.9997
6	7.6642	7.9998
7	7.4791	7.9997

5.2 Security Feature Analysis

When employing encryption algorithms for image processing, it is crucial to assess both the security of the encryption algorithm itself and its sensitivity to different types of images. This ensures effective encryption and broad applicability of the algorithm. Such considerations are essential to prevent images from being easily compromised in the event of an attacker launching checkpoint attacks.

5.2.1 Plaintext Sensitivity Analysis

Plaintext sensitivity is defined to using the identical key to encrypt two slightly different images by an encryption algorithm and then comparing the discrepancy between them[13]. If the discrepancy between the ciphertext images is higher, it means that the plaintext sensitivity is higher and more resistant to differential attacks. Equations (14), (15) and (16) define the quota of image difference level.

$$\:\begin{array}{c}sign\left(i,j\right)=\left\{\begin{array}{cc}0&\:P1\left(i,j\right)=P2\left(i,j\right)\\\:1&\:P1\left(i,j\right)\ne\:P2\left(i,j\right)\end{array}\right.\#\left(14\right)\end{array}$$

$$\:\begin{array}{c}NPCR\left(P1,P2\right)=\frac{1}{XY}\sum\:_{i=1}^{X}\sum\:_{j=1}^{Y}\left|sign\left(i,j\right)\right|*100\%\#\left(15\right)\end{array}$$

$$\:\begin{array}{c}UACI\left(P1,P2\right)=\frac{1}{XY}\sum\:_{i=1}^{X}\sum\:_{j=1}^{Y}\frac{\left|P1\left(i,j\right)-P2\left(i,j\right)\right|}{255-0}*100\%\#\left(16\right)\end{array}$$

Among them sign (i, j) is the sign function and the theory values of NPCR and UACI of encrypted image are 99.6094% and 33.4635% separately. The results of NPCR and UACI values of two encrypted images are shown in Table 10, which indicates that the presented algorithm has excellent plaintext sensitivity.

Table 10

NPCR and UACI for plaintext sensitivity test
image	NPCR	UACI
1	0.9962	0.3276
2	0.9961	0.3367
3	0.9961	0.3388
4	0.9960	0.3338
5	0.9962	0.3361
6	0.9961	0.3344

5.2.2 Key Space Analysis

The key space is the set containing all the keys. The key length is defined as the logarithmic number of elements in the key space (with a base of 2)[14]. In order to prevent various attacks, algorithms need a large enough key space to ensure a long key length. Encryption algorithms with key lengths longer than 100 bits are generally considered secure enough.

In this paper, the algorithm employs 8 parameters in its process: x0, y0, z0, w0, a, b, c, r. These parameters operate independently of each other. According to the floating-point standard definition, the computational precision of 64-bit double-precision data is 10^–15. Therefore, the key space value of the algorithm is calculated as (10¹⁵)⁸ = 10¹²⁰≈2³⁹⁹. These calculations confirm that the algorithm possesses a sufficiently vast key space to defend against exhaustive attacks.

5.2.3 Key Sensitivity Analysis

The key sensitivity test is crucial for assessing the reliability of an encryption scheme. It serves as an important metric because even a slight alteration to the encryption key can result in significantly different decryption outcomes, potentially leading to complete inconsistency between the decrypted image and the original[15]. For instance, taking image 1 in this paper as an example, the key sensitivity analysis can be conducted by varying parameters a, b, c, and r. Each of these parameters is modified individually, and the decryption results are observed as follows:

Figure 5 Key sensitivity test diagram ((a) is the original image, (b) is the normal decrypted image, (c), (d), (e), (f) are the parameter a is changed to 8, parameter b is changed to 3, parameter c is changed to 30 and parameter r is changed to 0, respectively)

From Fig. 5, it can be seen that the modification of the parameters will have an impact on the key, which will lead to an impact on the decryption result as well, resulting in a wrong decryption result, i.e., the decryption cannot be achieved. Therefore, the algorithm in this paper has excellent key sensitivity.

5.3 Robustness Analysis

Currently, digital images are recorded in the form of digital signals stored and delivered over a network, and for images in specific fields such as medical images, which have a high degree of privacy and value, they are susceptible to attacks that can lead to data loss or damage during transmission. Therefore, we must know whether we can decrypt the original image when the ciphertext image is attacked[16].

5.3.1 Analysis of Clipping Attack

The clip attack involves replacing a portion of the ciphertext image with an all-black block to simulate information loss during image transmission. Different proportions of replacement correspond to varying degrees of severity in the clip attack. In this paper, the encrypted image is segmented according to ratios of one-half, one-quarter, one-eighth, and one-sixteenth, respectively. The decrypted images after segmentation are illustrated in Fig. 6:

Figure 6 Schematic diagram of clipping attack ((a)-(d) are the encrypted image clipped in one-half, one-quarter, one-eighth and one-sixteenth respectively, and (e)-(h) are its decrypted image respectively)

In medical images, the central region stores important information and is easy to be attacked during transmission, while in this paper, due to the characteristics of the spiral transform, the central region of the medical image is transformed to a position that is not easy to be affected by the clip attack in the process of the disruption, so it can be decrypted when subjected to the clip attack, and there is no loss of the central region of the image after the decryption.

5.3.2 Noise Attack Analysis

Noise attack is a common attack during the transmission of information, in this paper use different pretzel noise parameters to simulate it and the experimental results are shown in Fig. 7.

Figure 7 Schematic diagram of the noise attack ((a) is the original encrypted image, (b), (c), (d) are pretzel noise with parameters of 0.001, 0.005, 0.01, (e)-(h) are its decrypted images, respectively)

It can be seen that after adding pretzel noise with different parameters to the image, the original image can be recovered almost without distortion, which shows that our algorithm can effectively resist the noise attack.

In this paper, a medical image encryption album based on spiral transform and dynamic DNA is presented, and its feasibility is verified by experimental simulation. The encryption algorithm takes permutation-diffusion as the basic framework and uses the Lorenz hyperchaotic system to produce chaotic sequences for encryption. It is mainly divided into two parts: the chaotic sequence is used to perform double helix transformation on the chunked image, and then it is converted to binary, and DNA encoding, DNA operation and decoding operations are performed respectively, so as to complete the process of disruption and diffusion. Analysis of performance shows that the encryption scheme is an accurate medical image encryption algorithm. It exhibits resilience against various common attack methods including differential attacks, statistical attacks, and exhaustive attacks, highlighting its high robustness and security. Especially in clip attack, due to the characteristics of the algorithm itself, the image can protect its central region from loss and be completely restored when it is subjected to clip attack. Therefore, this encryption method meets the requirements for the security and confidentiality of medical images and can be effectively utilized in the field of medical images.

However, there are some problems with our algorithm, firstly, due to the chunking operation of the image, it needs to have certain requirements on the size of the image. Secondly, because the dynamic DNA encoding and binary conversion process takes a long time, The algorithm is long time consuming and cannot be applied to the real-time transmission of a large number of large-scale images. In our future work, we will try to make a trade-off between the safety and effort of the encryption arithmetic, so that the algorithm can have both better encryption and higher encryption efficiency as much as possible.

Author Contribution

Z Z completed the main experimental content and manuscript, XG Z helped to find relevant literature and draw the flowchart, and JN L helped to draw the table. All authors reviewed the manuscript.B Z provided the main ideas for the literature and pressed for progress.

Ammar Ali Neamah(2023) An image encryption scheme based on a seven-dimensional hyperchaotic system and Pascal's matrix. Journal of King Saud University - Computer and Information Sciences, 35(3): 238-248. https://doi.org/10.1016/j.jksuci.2023.02.014
Li M, et al (2022) Medical image encryption algorithm based on hyper-chaotic system and DNA coding. Cognitive Computation and Systems. 4(4): 378–390. https://doi.org/10.1049/ccs2.12070
Xiong J, et al (2023) Fully chaotic medical image encryption scheme based on dynamic DNA and block rotation. Physica Scripta, 98(7). https://doi.org/10.1088/1402-4896/acde1e
Gao XY, Mou J, Li B, et al (2023) Multi-image hybrid encryption algorithm based on pixel substitution and gene theory. Fractals. 31(6). https://doi.org/10.1142/S0218348X23401114
Afify YM, Sharkawy NH, Gad W, et al (2024) A new dynamic DNA-coding model for gray-scale image encryption. Complex & Intelligent Systems. 10: 745–761. https://doi.org/10.1007/s40747-023-01187-0
Zhang L, Song X, et al (2024) Reversibly selective encryption for medical images based on coupled chaotic maps and steganography. Complex & Intelligent Systems. 10: 2187–2213. https://doi.org/10.1007/s40747-023-01258-2
Xu JJ, Zhao B (2023) Designing an Image Encryption Algorithm Based on Hyperchaotic System and DCT. International Journal of Bifurcation and Chaos. 33(2). https://doi.org/10.1142/S0218127423500219
Wang X, Chen Y (2021) A New Chaotic Image Encryption Algorithm Based on L-Shaped Method of Dynamic Block. Sensing and Imaging. 22(31). https://doi.org/10.1007/s11220-021-00357-z
Wu Y, Zhang L, Berretti S, Wan S (2023) Medical Image Encryption by Content-Aware DNA Computing for Secure Healthcare. IEEE Transactions on Industrial Informatics. 19(2): 2089-2098. https://doi.org/10.1109/TII.2022.3194590
Wang Y, Li XW, Wang QH (2021) Integral Imaging Based Optical Image Encryption Using CA-DNA Algorithm. IEEE Photonics Journal, 13(2): 1-12. https://doi.org/10.1109/JPHOT.2021.3068161
Iakovidou ND, Christodoulakis M, Papathanasiou ES, Papacostas SS, Mitsis GD (2015) Introducing weighted approaches to study network brain dynamics from EEG epilepsy measurements: The EigenBrain algorithm. 2015 IEEE 15th International Conference on Bioinformatics and Bioengineering (BIBE), Belgrade, Serbia. 1-5. https://doi.org/10.1109/BIBE.2015.7367685
Zhao, CF, Ren, HP (2020) Image encryption based on hyper-chaotic multi-attractors. Nonlinear Dynamics. 100: 679–698. https://doi.org/10.1007/s11071-020-05526-5
Arab, A, Rostami, MJ & Ghavami, B (2019) An image encryption method based on chaos system and AES algorithm. The Journal of Supercomputing. 75: 6663–6682. https://doi.org/10.1007/s11227-019-02878-7
Li M, et al (2022) Medical image encryption algorithm based on hyper-chaotic system and DNA coding. Cognitive Computation and Systems. 4(4): 378–390. https://doi.org/10.1049/ccs2.12070
Xie HW, Gao YJ, Zhang H (2023) An image encryption algorithm based on novel block scrambling scheme and Josephus sequence generator. Multimedia Tools and Applications. 82: 16431–16453. https://doi.org/10.1007/s11042-022-14139-6
Mohamed AG, et al (2021) New DNA Coded Fuzzy Based (DNAFZ) S-Boxes: Application to Robust Image Encryption Using Hyper Chaotic Maps. IEEE Access. 9: 14284-14305. https://doi.org/10.1109/ACCESS.2021.3052161

No competing interests reported.

Download PDF

Version 1

posted

You are reading this latest preprint version

Medical image encryption algorithm based on DNA and spiral transformation

Status:

Version 1

Abstract

Figures

1. Introduction

2 Preparation

2.1 Chaotic sequence

2.2 Spiral Transform

2.3 DNA operation

2.3.1 DNA coding and decoding rules

2.3.2 DNA operation rules

2.3.3 DNA inverse operation rules

3 Encryption Algorithm for Medical Images

3.1 Key Generation

3.2 Generating Sequences and Processing

3.3 Encryption process

3.4 Decryption process

4 Simulation Experiments

5 Analysis of Experimental Results

5.1 Statistical Feature Analysis

5.1.1 Histogram Analysis

5.1.2 Correlation analysis

5.1.3 Information Entropy Analysis

5.2 Security Feature Analysis

5.2.1 Plaintext Sensitivity Analysis

5.2.2 Key Space Analysis

5.2.3 Key Sensitivity Analysis

5.3 Robustness Analysis

5.3.1 Analysis of Clipping Attack

5.3.2 Noise Attack Analysis

6 Conclusion

Declarations

Author Contribution

References

Additional Declarations

Status:

Version 1