www.nature.

com/scientificreports

OPEN A hybrid encryption framework

leveraging quantum and
classical cryptography for secure
transmission of medical images in
IoT-based telemedicine networks
Arslan Shafique1, Syed Ali Atif Naqvi1, Ali Raza1, Masoud Ghalaii2,
Panagiotis Papanastasiou3, Julie McCann4, Qammer H. Abbasi1 & Muhammad Ali Imran1
In the era of the Internet of Things (IoT), the transmission of medical reports in the form of scan
images for collaborative diagnosis is vital for any telemedicine network. In this context, ensuring
secure transmission and communication is necessary to protect medical data to maintain privacy. To
address such privacy concerns and secure medical images against cyberattacks, this research presents
a robust hybrid encryption framework that integrates quantum, and classical cryptographic methods.
The proposed framework not only secure medical data against cyber threats but also protects the
secret security keys. Initially, a Quantum Key Distribution (QKD) is employed to generate a shared
key, which is then used to secure the symmetric keys via One-Time Pad (OTP) encryption. Next, bit-
planes are extracted from each color component. The rows and columns of the extracted bit-planes
are scrambled using random sequences which are generated by a 6D hyperchaotic Chen system and
the Ikeda map. To further increase confusion in the original data, multiple-step pixel scrambling
operations such as pixel shuffling, pixel value shuffling, and rotational and flipping operations are
implemented. After the confusion phase, a combination of affine transformations with non-linear
functions, Discrete Cosine Transform (DCT) with complex modulation, Discrete Wavelet Transform
(DWT) with random phase modulation, bilinear transformation, and nonlinear polynomial mapping
are employed to create diffusion in the scrambled components. These multiple encryption operations
aim to maximize randomness in the final ciphertext image. Additionally, to reduce computational
complexity, only the Most Significant Bit-Planes (MSBs) are encrypted, as they contain more than
94% of the plaintext information. Several experimental results and analyses are conducted to assess
the proposed encryption framework, including entropy analysis, key sensitivity analysis, correlation
analysis lossless analysis, and histogram analysis. Furthermore, the framework is tested against
various cyberattacks such as brute-force attacks, clipping attacks, and noise attacks on the ciphertext
images, to demonstrate its resilience against such threats.

The evolution of the Internet from digitization to intellectualization is transforming daily life and industry1. As
the transfer of digital data, specifically digital images, occurs in huge amounts between Internet of Things (IoT)
devices, it is necessary to secure them from eavesdropper access. Developing a robust technology to secure
digital images is not enough because IoT devices such as smartphones and tablets are very resource-constrained,
including limited power and limited storage2. As digital images contain a huge amount of data with a high
correlation between image pixels, while developing algorithms for securing digital data, it should be kept in
mind that they should be capitalised with such IoT devices having limited resources.
To protect image information from unauthorised access, three main technologies, such as encryption
methods and steganographic techniques, and techniques for securing encryption secret keys have been
frequently used in the past several years3–5. Image encryption includes several approaches such as frequency

1School of Electronic and Nanoscale Engineering, University of Glasgow, Glasgow G12 8QQ, UK. 2School of
Electronic and Electrical Engineering, University of Leeds, Leeds LS2 9JT, UK. 3Department of Computer Science,
University of York, York YO10 5DD, UK. 4Department of Computing, Imperial College, London SW7 2AZ, UK.
email: [email protected]

Scientific Reports | (2024) 14:31054 | https://doi.org/10.1038/s41598-024-82256-3 1

Content courtesy of Springer Nature, terms of use apply. Rights reserved
www.nature.com/scientificreports/

domain encryption6, spatial domain encryption, compressed sensing7, and encryption based on optical and
quantum computing8,9. Spatial domain encryption modifies pixel values directly, while transform-based
encryption uses transformations to alter pixel values in both the spatial and transform domains. Compressive
sensing incorporates both compression and encryption to enhance the processing time while also increasing the
chances of losing data due to decryption. A new image encryption approach that enhances the use of encryption
icing with optical properties10. Quantum image encryption, on the other hand, employs quantum encoders
to secure digital images from cyber threats11. It is worth noting that encryption methods based on quantum
phenomena are rather superior to those classical methods, which are logistic in nature. Most of the encryption
techniques have a positive impact depending on the purpose of use12–15. For example, the advanced encryption
standard (AES)16 is very effective when enhanced privacy is needed over digital pictures, but when it comes
to IoT devices, it is not suitable as it has multiple encryption rounds and also includes very high-resolution
encrypted images that take large storage space. However, with the growth in quantum computing, even the most
secure algorithms like AES could become vulnerable.
To overcome the vulnerabilities related to the high latency and weak security, quantum operations such as
quantum scrambling, and quantum substitution can be utilized. The scrambling process rearranges positions of
the image pixel using random numbers generated by methods such as Blum Blum Shub17, chaotic systems7,11,18,
ChaCha2019, and Random Number Generation Uniform (RANDU)20. Among them, chaotic systems are
very reliable due to their robust properties, such as sensitivity to initial conditions, aperiodicity, topological
mixing, and ergodicity. However, pixel permutation alone does not provide strong security for images, as it only
rearranges pixel positions without altering the pixel values themselves21. Combining scrambling with diffusion
provides stronger encryption by reducing the image pixel correlations. An effective encryption should use both
permutation and diffusion for strong protection. Traditional encryption methods such as AES, and DES are
unsuitable for real-time applications like video conferencing due to limited efficiency and security. However,
quantum image encryption provides enhanced security and efficiency by leveraging quantum mechanics.
In this research, a robust and time-efficient hybrid quantum encryption framework is proposed. The
proposed encryption framework utilizes QKD for key management, where a shared key is generated through
QKD and then encrypted using a One-time Pad (OTP). This combination ensures that the secret keys are highly
secure which makes it virtually impossible for eavesdroppers to recover the original secret keys. Moreover, the
encryption process begins by extracting the R, G, and B components from the plaintext image. Bit-planes are
then extracted from each color component of the input image. As most of the plaintext information present
in the Most Significant Bit Planes (MSBPs)22, therefore, only these bit-planes are selected for encryption to
reduce the computational complexity of the proposed encryption framework. The remaining, unencrypted bit-
planes are simply combined to form pre-ciphered images for each color component. For the encryption of the
MSBPs, random sequences are generated using specific initial conditions and control parameters of the chaotic
maps such as 6D hyper Chen chaotic map(6DHCCM), and Ikeda map that act as secret keys. These sequences
are employed to perform three types of scrambling: pixel scrambling, pixel value permutation, and rotation
and flipping. Additionally, the process of generating random sequences is extended to enhance the diffusion
process using XOR operation. Further diffusion is achieved by applying a variety of operations, including Affine
transformations with nonlinear functions, discrete cosine transform (DCT), wavelet transform with random
phase modulation, bilinear transformation with complex weights, and nonlinear polynomial mapping to
produce the final encrypted image. The contributions of the paper are as follows:

• A novel hybrid encryption framework is proposed that combines quantum and classical cryptographic tech-
niques developed to secure the transmission of medical images in IoT-based telemedicine networks.
• In the proposed work, QKD is combined with OTP encryption to generate shared keys and secure the sym-
metric secret keys, which enhances the key management process.
• To reduce the computational complexity, bit-planes are extracted from the color components of the medical
images and scramble only such bit-planes that contain the majority of the image information (over 94%) us-
ing random sequences generated by a 6D hyperchaotic Chen system and the Ikeda map.
• Instead of using a single scrambling step, three-step pixel scrambling operations such as pixel shuffling, pixel
value shuffling, and rotational and flipping operations are applied to create randomness in the plaintext image.
• In the diffusion phase, affine transformations with non-linear functions, DCT with complex modulation,
DWT with random phase modulation, bilinear transformations, and nonlinear polynomial mapping are
combined to maximize randomness in the encrypted images.The remainder of the paper is structured as fol-
lows: Section 2 offers a brief summary of existing works including their advantages and vulnerabilities. Also,
this section provides the advantages, vulnerabilities, and possible solutions to overcome such vulnerabilities.
Section 3 delivers a comprehensive explanation of the foundational knowledge necessary for the development
of the proposed encryption framework. Section 4 presents a step-by-step explanation of the proposed encryp-
tion framework. In Section 5, the experimental results and analysis of the proposed encryption framework
are discussed. Section 6 presents a concise discussion of the entire proposed work, along with an overview of
the advantages of utilizing multiple encryption layers within the framework. Finally, Section 7 concludes the
entire proposed research, highlights the limitations of the proposed work, and also provides a few recommen-
dations for future research to address these limitations.

Related work
The development of image encryption techniques has seen substantial progress, incorporating contemporary
technologies to enhance data security. This section explores the existing encryption schemes and investigates
their vulnerabilities. For instance, In23, Hu et al. proposed a quantum image encryption method using a qubit-
level scrambling23,24. This approach enhances operational effectiveness by the use of hyperchaotic states and

Scientific Reports | (2024) 14:31054 | https://doi.org/10.1038/s41598-024-82256-3 2

Content courtesy of Springer Nature, terms of use apply. Rights reserved
www.nature.com/scientificreports/

sometimes controlling periodic windows, while the diffusion and scrambling methods aid image encryption
further. Quite, the scheme is not effective if the parameters of the chaotic maps and quantum model are not
efficiently selected.
In25, He et al. proposed a quantum image encryption algorithm, that enhances quantum image preparation
by optimizing quantum circuits to reduce auxiliary qubits. The algorithm also deals with non-random key
issues with a modified logistic map26and enhances security with row-column-based permutation27. However,
vulnerabilities may arise if the circuit optimization or permutation methods are exploited, and the encryption
could be weakened if the improved logistic map lacks adequate randomness or security. In28, Mohamed et al.
proposed a quantum-based framework for the encryption of digital images. They perform pixel diffusion and
pixel permutation by using quantum operations like Hadamard gates29and quantum-controlled gates30and
perform key generation using a hyperchaotic system, bloomschaotic sequences, and circular shits for
permutation. This scheme poses a risk of attacks owing to the hyperchaotic being predictable. In31, Gao et al.
proposed quantum DNA-based cryptographic techniques. To round off the pixel transformation images of the
new frame, alternation of colours of particular pixels within the images is achieved using Hilbert scrambling30.
Their scheme is vulnerable to attacks due to weak management of the inverse operations and the key matrix.
In32, Zhang et al. presented quantum image cryptosystems that include superposition states in reversible pixel
scramblers. It also incorporates a pixel reordering feature in which a feature-orientated order of the image is
processed with sequential cries. In33, Wen et al presented an encryption scheme based on chaos-based block
scrambling and confusion-diffusion operations. It encrypts the digital images using multiple cryptographic
operations, such as block-wise scrambling, and rotation. It also satisfies cryptographic criteria for confusion,
diffusion, and avalanche effects.
In34, Rehman et al. proposed an encryption framework for the protection of color image schemes dependent
on OTP keys and chaos theory with a rotor machine concept. The image pixel rows and columns are transformed
and rotated to create new configurations for substitution. Vulnerabilities arise in their scheme due to the
predictable patterns of the rotor rotation and chaotic sequences. Moreover, the one-time keys and chaotic maps
are not securely managed or sufficiently randomized. In35, Kumar et al. introduced an encryption scheme that
converts digital images into a 2D difference matrix using a median edge detector. The image is then encoded
with a bit-plane representation to reduce the size and enable the embedding more secret data. In36, Zhu et al.
performed cryptanalysis of encryption frameworks using bit plane extraction to find the key streams for bit-level
permutation and XOR diffusion. This independence allows attackers to recover the permutation and diffusion
keys through chosen plaintext attacks using just two specific plaintext images and their ciphertexts, exposing a
significant vulnerability in the encryption system. In37, Sing et al. presented a chaotic system based encryption
scheme in which key is initialized with confusion- diffusion operations. While the method offers strong security
and high speed, it is not enough to resist cyberattacks because of the exploitable patterns in the object detection
process. A summary of the existing works is provided in Table 1.

Preliminaries
In the proposed research, several key components are used, and it is essential to understand their importance
and roles within the framework. The following subsections provides a brief overview of these critical encryption
elements and explain why they are integral to the proposed encryption approach.

6D Hyperchaotic Chen system

The 6D Hyperchaotic Chen system (6DHCCS)38 is a higher-dimensional extension of the classic Chen system
which is designed to exhibit complex and extremely chaotic behavior. The additional dimensions in this
system increase its complexity and unpredictability which makes it suitable for secure communications and
cryptography applications. Mathematically, 6DHCCS can be expressed using Equation 1.

Existing schemes Application domain Real-World performance Robustness against attacks disadvantages Potential solutions
Chaotic maps
Hu et al.23 Quantum Image Encryption Enhanced security Vulnerable to poor parameters Improve parameter selection
vulnerabilities
Circuit and key randomness
He et al.25 OCPBP Quantum Encryption Optimized BRQI prep Weak circuit optimization Better optimization techniques
issues
Mohamed et al.28 Quantum Cellular Automata Effective diffusion Predictable hyperchaotic system Predictable chaos Use less predictable chaos
Gao et al.31 Quantum DNA Code Uses Hilbert scrambling Weak key management Weak key matrix Improve key management
Zhang et al.32 Quantum State Superposition Large key space Vulnerable to attacks Weak state transmission Enhance state security
Chaos-Based Block
Wen et al.33 Satisfies criteria Not detailed Not addressed Further robustness evaluation
Permutation
Rehman et al.34 Color Image Encryption Transforms pixel rows Predictable rotor patterns Weak key management Secure key management
Kumar et al.35 Bit-Plane Representation Effective embedding Not detailed Limited robustness details Enhance robustness evaluation
Zhu et al.36 Bit Plane Extraction Key streams analyzed Vulnerable to attacks Exploitable key recovery Improve key management
Singh et al.37 Chaos-Based for Smart Cities Fast object detection Exploitable detection patterns Patterns in detection Better detection security

Table 1. Summary of Image Encryption Schemes.

Scientific Reports | (2024) 14:31054 | https://doi.org/10.1038/s41598-024-82256-3 3

Content courtesy of Springer Nature, terms of use apply. Rights reserved
www.nature.com/scientificreports/

ϖ̇1 = σ1 (ϖ2 − ϖ1 )
ϖ̇2 = ρ1 ϖ 1 − ϖ2 − ϖ 1 ϖ 3
ϖ̇3 = −β1 ϖ3 + ϖ1 ϖ2
(1)
ϖ̇4 = σ2 (ϖ5 − ϖ4 )
ϖ̇5 = ρ2 ϖ 4 − ϖ5 − ϖ 4 ϖ 6
ϖ̇6 = −β2 ϖ6 + ϖ4 ϖ5

where ϖ1 , ϖ2 , ϖ3 , ϖ4 , ϖ5 , ϖ6 are the state variables, and σ1 , ρ1 , β1 , σ2 , ρ2 , β2 are the system parameters.
In 6D, visualization can be challenging, but projections of the attractor into lower-dimensional spaces (such as
3D) reveal intricate and complex trajectories as shown in Figure 1. By analyzing various 3D projections of the
system-such as ϖ1 − ϖ2 − ϖ3 , ϖ4 − ϖ5 − ϖ6 , and combinations involving other dimensions, it can be seen
that each projection reveals unique patterns and interactions among the state variables, which can be utilized
to generate diverse and unpredictable pseudorandom sequences for encrypting image pixels. Moreover, 3D
projections of the 6D hyperchaotic Chen system highlight its complex trajectories and high sensitivity, which
enable it to disrupt predictable image information when used in pixel scrambling.

The values of the state variables ϖ1 , ϖ2 , ϖ3 , ϖ4 , ϖ5 , ϖ6 are obtained through numerical integration of the 6D
Hyperchaotic Chen System’s differential equations, and these values are computed over time and are not fixed
but rather evolve according to the system dynamics. However, the values of σ1 , ρ1 , β1 , σ2 , ρ2 , β2 are 70, 150, 20,
70, 150, and 20, respectively.

Ikeda map
The Ikeda map is a discrete dynamical system that models the behavior of light in an optical cavity with a
nonlinear medium39. Equation 2 is the mathematical representation of Ikeda map.
Ξn+1 = 1 + u(xn cos(θn ) − ξn sin(θn ))
(2)
ξn+1 = u(Ξn sin(θn ) + ξn cos(θn ))

where: Ξn and ξn are the coordinates of the point at the nth iteration. u controls the behavior of the map, and
θn represents the phase shift at the n-th iteration as given in Equation 3.
p
θn = k − (3)
1 + x2n + yn2

Fig. 1. 6D Hyperchaotic Chen system attractors.

Scientific Reports | (2024) 14:31054 | https://doi.org/10.1038/s41598-024-82256-3 4

Content courtesy of Springer Nature, terms of use apply. Rights reserved
www.nature.com/scientificreports/

where k and p control the nonlinearity and feedback in the system. Figure 2(a-f) illustrates the six attractors
of Ikeda map attractors under different initial conditions that demonstrate its chaotic behavior and sensitivity,
which produce non-repeating sequences for further scrambling and diffusion of image data. Together, these
chaotic systems can be used to create strong randomness in the encrypted image to enhance security by
concealing any identifiable structure and making it difficult for attackers to retrieve the original image.

Proposed encryption process

The proposed encryption process includes four main phases: (i) Generating a shared key using QKD with OTP
encryption, (ii) extracting bit-planes from each color component and encrypting them through scrambling
operations using random sequences generated by the 6DHCCS and the Ikeda map, (iii) applying confusion
operations through multiple permutation techniques, and (iv) achieving diffusion in the scrambled image. A
detailed overview of the entire encryption framework is illustrated in Figure 3. Each step is explained in detail
in the following subsections.
The proposed encryption process consists of four key phases: (i) Generating of the shared key using QKD
which is further based on OTP encryption chaotic systems, (ii) bit-planes extraction from each color component,
and encrypting using scrambling operations applied based on the random sequences generated using 6DHCCS,
and Ikeda map, (iii) carrying out confusion operations using multiple permutation operations, and (iv) creating
diffusion in the scrambled image. The details of each step are provided in the next subsections.

QKD based symmetric key sharing

Our proposed encryption framework begins with the implementation of QKD for generating secret keys.
QKD allows multiple parties to generate shared secret cryptographic keys, which can be used for secure
communication. In the proposed research, the BB84 algorithm, which is proposed by Charles Bennett and Gilles
Brassard40, is used to secure the actual symmetric keys that are used in the generation of the chaotic sequences.
Alice randomly prepares a sequence of quantum bits (qubits), where each qubit can be in
one of four possible states, based on two different bases: Rectilinear (+) and Diagonal (×). The
rectilinear (+) and the diagonal (×) basis states are |1⟩ (bit value 1), |1⟩ (bit value 1) and
|+⟩ = √12 (|0⟩ + |1⟩) (bit value 0), |−⟩ = √12 (|0⟩ − |1⟩) (bit value 1), respectively.
Now Alice prepares a sequence of 8 qubits. For this example, bit sequence (b) is (1, 0, 1, 1, 0, 1, 0, 0), and
basis choice (bB ) are (+, ×, +, ×, +, ×, +, ×). Here, the Rectilinear basis + is represented as |0⟩ and |1⟩, and
the Diagonal basis × is represented as |+⟩ and |−⟩.

Fig. 2. 3D Ikeda map attractors demonstrating chaos and sensitivity.

Scientific Reports | (2024) 14:31054 | https://doi.org/10.1038/s41598-024-82256-3 5

Content courtesy of Springer Nature, terms of use apply. Rights reserved
www.nature.com/scientificreports/

Figure 3. Propped encryption framework.

Bits Qubits Measured results

1. Qubit 1 Measured in × basis, Bob gets 1 (correct basis)
2. Qubit 2 Measured in + basis, Bob gets0 (correct basis)
3. Qubit 3 Measured in × basis, Bob gets nothing (incorrect basis)
4. Qubit 4 Measured in + basis, Bob gets0 (correct basis)
5. Qubit 5 Measured in × basis, Bob gets 0 (correct basis)
6. Qubit 6 Measured in + basis, Bob gets 1 (correct basis)
7. Qubit 7 Measured in × basis, Bob gets 1 (correct basis)
8. Qubit 8 Measured in × basis, Bob gets nothing (incorrect basis)

Table 2. Bob’s basis choice and measurement.

Bob randomly chooses his measurement basis. For instance, Bob’s basis (bB ′ ) are (×, +, +, ×, × +, ×, + ).
Now Bob will measure the qubits and obtain the following results as mentioned in Table 2
Now Alice and Bob publicly compare their chosen bases and retain only those bits where their bases match.

• Alice’s Bases: (+, ×, +, ×, +, ×, +, ×)

• Bob’s Bases: (×, +, ×, +, ×, +, ×, +)

The retained bits will be 1,0,0,0,1,1. Therefore, Alice and Bob both have the shared key of 1,0,0,0,1,1. No, the
following procedure is used to encrypt the symmetric key.

• Define the symmetric key: Assume α = 3.49. Convert α to binary, which is αbinary = 11.011111.Suppose
a scaling factor of 100 is used (two decimal places), 3.49 × 100 = 349 → - 349 in binary is 1010111012 .
Now pad the binary of 100011 to match the length of 101011101:
1 0 1 0 1 1 1 0 1
0 0 0 1 0 0 0 1 1
1 0 1 1 1 1 1 1 0

This merged output (M) is sent over the public channel. The receiver, knowing both the shared key and the
merged output, performs XOR again to recover the original symmetric key α. This ensures that even if the merged
output is intercepted by the eavesdropper, without the shared key, it is impossible to recover the symmetric key.

Scientific Reports | (2024) 14:31054 | https://doi.org/10.1038/s41598-024-82256-3 6

Content courtesy of Springer Nature, terms of use apply. Rights reserved
www.nature.com/scientificreports/

Algorithm 1. Bit-plane extraction based permutation.

Algorithm 2. Confusion phase.

Scientific Reports | (2024) 14:31054 | https://doi.org/10.1038/s41598-024-82256-3 7

Content courtesy of Springer Nature, terms of use apply. Rights reserved
www.nature.com/scientificreports/

Justification of secrete key security

The security of the process relies on the properties of the XOR operation and the fact that the shared key is
known only to the legitimate parties (Alice and Bob).
The XOR operation has the property that if you know the result C = A ⊕ B and one of the operands (either
A), one can uniquely determine the other operand (B = C ⊕ A). However, if you only know C without knowing
either A or B , it is computationally infeasible to determine either A or B without additional information.
In the proposed setup, the shared key K is known only to Alice and Bob, and it is never shared publicly. Even
if an attacker (Eve) intercepts the merged output, without knowledge of the shared key K , the receiver cannot
recover the symmetric key α.

Mathematical proof of security

Theorem 1 Prove that an attacker who intercepts the merged output M = α ⊕ K cannot recover α without
knowing K .

• Given:
• Symmetric key: α
• Shared key: K
• Merged output: M = α ⊕ K
• To prove: Knowing M alone does not allow an attacker to recover α.

Case 1: Eve has only the merged output M = α ⊕ K . To recover α, Eve needs to compute:α = M ⊕ K .
Challenge for Eve: Eve knows M , but M depends on both α and K . K is unknown to Eve. Therefore, the
XOR operation is such that every possible key K could correspond to a different symmetric key α. This means
that without knowing K , Eve cannot uniquely determine α.
Case 2: Randomness of K : Assume K is a random string of bits, generated independently of α. Since K
is random and known only to Alice and Bob, the merged output M can be any random binary sequence of the
same length as α. Therefore, for every possible value of M , there is an equally likely K that could correspond
to any possible α. Thus, from M alone, the probability distribution of possible values of α is uniform over all
possible values, providing no information about α.
Case 3: Conditional Entropy: The conditional entropy H(α|M ) represents the uncertainty about α given
M . Since K is random and unknown to Eve, the conditional entropy remains high: H(α|M ) = H(α). This
means that knowing M does not reduce the uncertainty about α.

Image encryption
In the proposed encryption scheme following major components are used to provide robust security to the
digital images.

• Bit-plane extraction-based cryptography

• Random numbers generated using chaos
• Confusion phase: Multiple scrambling operations
• Diffusion phase: XOR operation, Affine transformations with nonlinear functions, discrete cosine transform
(DCT), wavelet transform with Random Phase Modulation, bilinear transformation with complex weights,
and nonlinear polynomial mapping

Bit-plane extraction based cryptography

Let I be the input color image, which can be represented as a 3D matrix I(x, y, c) where x and y are the spatial
coordinates, and c denotes the color channel (Red, Green, Blue) as given in Equation 4.
{
R(x, y) if c = 1
I(x, y, c) = G(x, y) if c = 2 (4)
B(x, y) if c = 3

For each color component C(x, y) ∈ {R(x, y), G(x, y), B(x, y)}, extract the k-th bit-plane PC,k (x, y) using
Equation 5.
⌊ ⌋
C(x, y)
PC,k (x, y) = mod 2(5)
2k

where k ranges from 0 to 7, representing each bit-plane. Apply encryption functions as given in Equations 6-10
to each bit-plane extracted from the R, G, and B components. Suppose, there are n rounds of encryption with a
round key Ki for each round i.

Now, use a substitution function S (S-box) for non-linear substitution suing Equation 6. The S-box used in this
research are given in41–43.

Scientific Reports | (2024) 14:31054 | https://doi.org/10.1038/s41598-024-82256-3 8

Content courtesy of Springer Nature, terms of use apply. Rights reserved
www.nature.com/scientificreports/

S(PC,k (x, y)) = S-box(PC,k (x, y))(6)

The permutation layer, which scrambles the bits, and the key mixing, which XORs the result with the round key,
are applied according to Equation 7.
{ ′
PC,k (x, y) = P (S(PC,k (x, y)))
(7)
′′
PC,k (x, y) = PC,k
′
(x, y) ⊕ Ki

Repeat substitution, permutation, and key mixing for n rounds. Let PC,k
n
(x, y) be the pre-encrypted bit-plane
after n rounds as given in Equation 8.
( ( n ))
⊕
n
PC,k (x, y) = Pn Sn (PC,k
i−1
(x, y) ⊕ Ki ) (8)
i=1

where Sn and Pn represent the substitution and permutation operations in the n-th round.

Combine the encrypted bit-planes to create the pre-ciphertext image by reconstructing each pixel value for each
color component from the encrypted bit-planes using Equation 9.
7
∑
Cenc (x, y) = n
PC,k (x, y) · 2k (9)
k=0

where C ∈ {R, G, B} denotes the color component. Combine the encrypted color components to form the
pre-encrypted image IP-enc as given in Equation 10. The process to generate P − enc is given in Algorithm 1.
{
Renc (x, y) if c = 1
Penc (x, y, c) = Genc (x, y) if c = 2 (10)
Benc (x, y) if c = 3

Random numbers generation using chaos

To generate a random image using a chaotic map and convert it to a 2D image, the following steps are followed:

• Choose initial values and control parameters: Select initial values x0 and control parameters a for the
chaotic map.
• Iterate Equations 1, and 2 for M × N times to generate a sequence of values.
• Scale the Values: Since the chaotic map produces values in the range (0, 1), multiply each value by a large
integer L to amplify it as given in Equation 11.

yn = xn · L(11)

• Truncate the decimal part of yn to convert it into an integer value.

• Adjust to the desired range [0, 255]: Take the modulo 256 of each value to ensure it fits within the 8-bit range
according to Equation 12.

zn = mod(yn , 256)(12)

• Generate the image: Reshape the 1D sequence of values into a 2D matrix with size M × N , where M × N
is the total number of color components in the image.

Confusion phase
In the confusion phase of the proposed encryption framework, multiple image shuffling techniques such as pixel
shuffling, pixel value permutation, rotation and flipping, and scrambling by image transformation are applied
to Penc .

Pixel shuffling
It involves rearranging the pixels of an image according to some permutation rules. The primary goal is to
disrupt the spatial structure of the image without altering the pixel values. In the proposed encryption frame
work, given an image Penc of size M × N , let each pixel be denoted as Penc (i, j), where i ∈ {1, 2, . . . , M } and
j ∈ {1, 2, . . . , N }. The two types of pixel shuffling such as row and column scrambling are used. For the row
scrambling, a permutation function is defined as πr : {1, 2, . . . , M } → {1, 2, . . . , M } that randomly shuffles
the rows of the Penc . The new image Pshuf f led_rows after row shuffling is given by Equation 13.
Pshuf f led_rows (πr (i), j) = Penc (i, j),
(13)
∀i ∈ {1, 2, . . . , M }, ∀j ∈ {1, 2, . . . , N }

Scientific Reports | (2024) 14:31054 | https://doi.org/10.1038/s41598-024-82256-3 9

Content courtesy of Springer Nature, terms of use apply. Rights reserved
www.nature.com/scientificreports/

For the column scrambling, a permutation function is defined as πc : {1, 2, . . . , N } → {1, 2, . . . , N } that
randomly shuffles the columns of Pshuf f led_rows (πr (i), j). The new image Pshuf f led after both row and
column shuffling is given by Equation 14.
Pshuf f led (πr (i), πc (j)) = Penc (i, j),
(14)
∀i ∈ {1, 2, . . . , M }, ∀j ∈ {1, 2, . . . , N }

Pixel value permutation

It changes the order of pixel values in the image without changing their spatial locations. This technique disrupts
the color distribution while maintaining the structural content. In this step, there are also two techniques of
permutation such as flattening the image, and a permutation of pixel values is applied. First, flatten the image
(Pshuf f led (πr (i), πc (j))) into a one-dimensional vector p = [p1 , p2 , . . . , pM N ], where pk = Penc (ik , jk )
for k = 1, 2, . . . , M N . Whereas, in the permutation of pixel values, first define a permutation function
πv : {1, 2, . . . , M N } → {1, 2, . . . , M N } that rearranges the pixel values. The new pixel vector p′ after
permutation is p′ (πv (k)) = p(k), ∀k ∈ {1, 2, . . . , M N }. After that, reshape the permuted vector p′ back
to the original image dimensions M × N to obtain the permuted image Pperm .

Rotation and flipping

To rotate an imagePshuf f led (πr (i), πc (j)) byθ degrees around its center, each pixelPenc (i, j) is mapped to a
new location(i′ , j ′ ) using the rotation matrix which is given in Equation15.
[ ′] [ ][ ] [ ]
i cos θ − sin θ i − ic ic
= sin θ cos θ j − jc + jc (15)
j′
M +1 N +1
where(ic , jc ) is the center of the image, calculated as:ic = 2
, jc = 2 .

To flip the image, both horizontal and vertical flipping techniques are employed according to the Equations 16,
and 17, respectively.
Pf lip_horizontal (i, j) =Pperm (i, N − j + 1),
(16)
∀i ∈ {1, 2, . . . , M }, ∀j ∈ {1, 2, . . . , N }
Pf lip_vertical (i, j) =Pf lip_horizontal (M − i + 1, j),
(17)
∀i ∈ {1, 2, . . . , M }, ∀j ∈ {1, 2, . . . , N }

Algorithm 2 provides the pseudocode used to generate the confusion depicted in the image.

Diffusion phase
The objective of this phase is to disperse the influence of each pixel in Is throughout the entire image for enhancing
the overall security. In this research, this is achieved using a different complex mathematical operation. The
detail of such operations is provided in the next subsections.

Affine transformation with nonlinear functions

Affine transformation is integrated with nonlinear functions to enhance diffusion. There are two major
transformations in this process: (i) Affine transformation and (ii) nonlinear function application. Mathematical
representations of these transformations are given in Equations 18, and 19, respectively.
A(i, j) =a · Pflip_vertical (i, j) + b (18)
⌊ [ ( )
( ) (x + y) · π
A(x, y) = a · Pflip_vertical (x, y) + b · sin
N ·M
( )p ]⌋ (19)
(x − y)2
+ √
N ·M

Where a and b are constants, and p is a polynomial degree.

Discrete Cosine Transform (DCT) with complex modulation

Apply the 2D DCT to the A(x, y) using Equation 20.

∑∑
N −1 M −1 [ ] [ ]
π(2i + 1) π(2j + 1)
D(x, y) = A(x, y) · cos · cos (20)
2N 2M
x=0 y=0

Scientific Reports | (2024) 14:31054 | https://doi.org/10.1038/s41598-024-82256-3 10

Content courtesy of Springer Nature, terms of use apply. Rights reserved
www.nature.com/scientificreports/

After applying 2d DCT, use a complex modulation function to the DCT coefficients to get a second pre-ciphertext
image (D′ (x, y)) using Equation 21.
( [ ] ( ))
π π2
D′ (x, y) = D(x, y) · exp · sin (21)
N +M N · M2
2

Now, apply the inverse 2D DCT using Equation 22 to obtain the third ciphertext image (D′′ (u, v)).

2 ∑∑ ′
N −1 M −1

D (x, y) =
′′
D (x, y)
NM
](22)
x=0 y=0
[ ] [
π(2x + 1) π(2y + 1)
· cos · cos
2N 2M

Wavelet transform with random phase modulation

In this research, the wavelet transform is used to decompose D′′ (x, y) into different frequency sub-bands (LL,
LH, HL, and HH) which consist of both high- and low-frequency components. The 2D wavelet transform is
implemented using Equation 23.
∑
W(j, k, a, b) = D′′ (x, y) · ψj,k (x − a, y − b)(23)
x,y

Where W(j, k, a, b), and ψj,k (x − a, y − b) are the wavelet coefficients, and wavelet function, respectively. The
wavelet function is shifted by (a, b) and scaled according to indices j and k. The random phase modulation is
applied to the wavelet coefficients to for the modification of phase information of the image components using
Equation 24.
[ ( )
π · (a + b)
W (j, k, a, b) = W(j, k, a, b) ·
′
exp i ·
M ·N
( ( ) ( )) ](24)
2π · j · k a π · (a2 + b2 )
· sin + · cos
N b M ·N

Where a and b are constants. The inverse wavelet transform is applied using Equation 25.
∑∑
C(x, y) = W′ (j, k, a, b) · ψj,k (x − a, y − b)(25)
j,k a,b

Bilinear transformation
The purpose of bilinear transformations is to create non-linear relationships between image pixels in a way
that is not easily reversible without exact replica of the secrete keys and the knowledge of the transformation
parameters. Furthermore, adding a bilinear transformation to the encryption process enhances the security of
the enciphered images. The bilinear transformation for C(x, y) is given in Equation 26.

∑∑
N −1 M −1 [ ]
(i + k) · (j + l)
B(i, j) = C(x, y) · (26)
N ·M
x=0 y=0

To increase further complexity, a complex weighting function is applied to the transformed pixel values using
Equation 27.
 

 1
C′ (i, j) = B(i, j) ·  ( )
1 + exp − √(i+j)
N ·M (27)
( 2 2
)]⌋
π · (i + j )
+ cos
N ·M

Nonlinear polynomial mapping

In the last step of the proposed encryption framework, a nonlinear polynomial mapping is applied to change the
pixel values based on higher-order polynomials. This introduces non-linearity and makes the transformation
difficult to reverse without knowing the exact polynomial used. A nonlinear polynomial mapping is applied to
each pixel value to produce the final ciphertext image (Cf (I, j)) using Equation 28.

Scientific Reports | (2024) 14:31054 | https://doi.org/10.1038/s41598-024-82256-3 11

Content courtesy of Springer Nature, terms of use apply. Rights reserved
www.nature.com/scientificreports/

⌊( ) [ ( )
π · (i · j)
3 i2 + j 2
Cf (i, j) = C(i, j) + · exp
N +M N ·M
( )]⌋ (28)
2π · i · j
+ sin
N ·M

The pseudocode of the proposed diffusion phase is given in Algorithm 3. The entire encryption framework
provides a significantly robust non-linear relationship between the original, and the final ciphertext image. The
generated ciphertext images are displayed in Figure 4, where all plaintext information is entirely concealed,
and the original image is unrecognizable. This demonstrates the effectiveness of the proposed framework in
encrypting the plaintext image by disrupting pixel correlations.

Experimental results and analysis

To test and analyze the proposed encryption framework for securing digital images, four test images are used:
X-ray, Tumor, Ultrasound, and Electrocardiogram (ECG) signal images. The entire encryption framework
is implemented in MATLAB 2018 on a system with the following hardware specifications: 8 GB RAM, 512
GB SSD, 11th generation Intel Core i5 processor at 2.4 GHz, running Windows 11. The effectiveness of the
encryption framework is assessed through various statistical security analyses, including entropy, histogram,
lossless analysis, and key sensitivity. The statistical values for the encrypted images are averaged across the R, G,
and B components. The framework is also tested against multiple cyberattacks including brute force, noise, and
cropping attacks to show its resilience to these threats.

Fig. 4. (a,f,k, p) Plaintext images, (b-d, g-i, l-n, q-s) their corresponding R, G, B components, and (e, j, o, k)
corresponding decrypted versions.

Scientific Reports | (2024) 14:31054 | https://doi.org/10.1038/s41598-024-82256-3 12

Content courtesy of Springer Nature, terms of use apply. Rights reserved
www.nature.com/scientificreports/

Algorithm test
The proposed encryption framework is tested on different plaintext color images including X-ray and tumor
images with a resolution of 256 × 256, as well as ultrasound and ECG signal images with a resolution of
512 × 512 having varying sizes such as 256 × 256, and 512 × 512. The quality of the encrypted images is then
assessed visually. As shown in Figure 4, the plaintext information is completely concealed within the ciphertext
images, with no visible patterns of the original data. This shows that the proposed encryption framework is
effective in protecting plaintext information.

Entropy
Entropy measures the randomness in an image, with the ideal value varying based on the bit depth of the image.
The maximum entropy value for an 8-bit image is 8, while for a binary (2-bit) image, it is 2. Mathematically, it
can be calculated using Equation 29.

∑
M

EI = p(il )logb p(il )(29)

L=0

Where, EI is the entropy of any image. logb (p(il )) is the logarithm of the probability p(tk ) with a base b. The
base b depends on the entropy
∑calculation. For example, if b = 2, it is a binary logarithm, which is commonly
N
used when dealing with bits. K=0 is the summation symbol which indicates that the sum is calculated over all
possible pixel values L in the image ranges form 0 to M . For robust encryption, it is essential that the entropy
value is as high as possible. A higher entropy value indicates greater randomness in an image, which contributes
to stronger encryption. The relationship between high entropy and robust encryption is expressed in Equation
30.
Entropy ∝ Strong Encryption(30)

Table 3 presents entropy values for both the proposed and existing methods. The results show that the entropy
values of the encrypted images from the proposed framework are closer to 8, indicating a higher level of
randomness in the ciphertext compared to existing methods.

In addition to the global entropy calculation, the encrypted and plaintext images are divided into blocks of
various sizes containing 16, 64, 256, and 1024 pixels in each block. The 3D entropy distribution plots for these
blocks are displayed in Figure 5. Even with a block size of 4 × 4, the entropy remains above approximately 3.58,
indicating that no pattern of even 16 pixels repeats in the encrypted image, with each pixel appearing random.
This results in a high global entropy, as shown in Table 3. In contrast, the 3D block-wise entropy distribution for
the plaintext image in Figure 5(a-d) shows entropy dropping to zero for 4x4 and 8x8 blocks, revealing smooth
patterns in the plaintext image. Furthermore, the global entropy for 100 plaintext and corresponding ciphertext
images is displayed in the polynomial surface fit plots in Figure 5 (i, j). It can be observed that the global entropy
of the plaintext images falls within the range of approximately [6.7981, 6.79983], while for the corresponding
ciphertext images, it ranges from roughly [7.9994, 7.9995].

Correlation analysis
Correlation evaluates how adjacent pixels relate to each other in different orientations. In general, plaintext
images exhibit high pixel correlation because of the similarity between neighboring pixels. However, in image
encryption, it is crucial to break this pixel correlation to protect the plaintext information securely. Reduced
correlation indicates a more robust encryption, as it signifies that the encryption process has effectively broken
the inherent relationships between pixels, enhancing the security of the image. The pixel correlation coefficients
for horizontally (H), vertically (V), and diagonally (D) adjacent pixels are calculated using Equations 31, 32, and
33 to assess the effectiveness of the proposed encryption framework.
∑W −1
(I(i, j) − I¯i )(I(i, j + 1) − I¯i )
Corrhorizontal =
j=1
(31)
(W − 1)σi2
∑H−1
(I(i, j) − I¯j )(I(i + 1, j) − I¯j )
Corrvertical = i=1
(32)
(H − 1)σj2

Image size Images 32 28 44 45 46 Proposed

Xray 7.9967 7.9919 7.9936 7.9955 7.9949 7.9993
256 × 256 × 3
Tumor 7.9946 7.9966 7.9919 7.9977 7.9991 7.9994
Ultrasound 7.9961 7.9979 7.9987 7.9949 7.9966 7.9992
512 × 512 × 3
ECG signal 7.9916 7.9910 7.9933 7.9969 7.9982 7.9991

Table 3. Entropy analysis.

Scientific Reports | (2024) 14:31054 | https://doi.org/10.1038/s41598-024-82256-3 13

Content courtesy of Springer Nature, terms of use apply. Rights reserved
www.nature.com/scientificreports/

Algorithm 3. Diffusion phase.

∑H−1 ∑W −1 ¯ ¯
(I(i, j) − I)(I(i + 1, j + 1) − I)
Corrdiagonal =
i=1 j=1 (33)
(N − 1)σ 2

In Equations 31-33, I(i, j) represents the pixel value at row i and column j . W and H denote the width and
height of the image, respectively; I¯i is the mean pixel value for row i or column j ; and σi2 is the variance of pixel
values in row i. Table 4 displays various correlation values, showing that the ciphertext images exhibit minimal
correlations in all directions compared to existing encryption methods, with coefficients approaching zero. This
indicates that the proposed encryption method is more effective at randomizing pixel values and disrupting the
patterns present in the plaintext information.

A visual analysis using scatter plots shows the statistical correlation analysis. Figure 6 compares the pixel
correlations of the plaintext and ciphertext images. The closely clustered points in Figures (a-c, g-i) indicate a
high correlation among plaintext pixels, while the more scattered points in Figures (d-f, j-l) demonstrate very
little correlation in the ciphertext pixels.

Key sensitivity analysis

For any encryption framework, it is crucial that the secret keys used are highly sensitive. Sensitivity
to secret keys means that even a tiny change in the keys makes it impossible to successfully decrypt
the plaintext information. In this work, a total of fifteen secret keys (ϖ1 , ϖ2 , · · · , ϖ6 , σ1 , ρ1 , β1
,σ2 , ρ2 , β2 , Ξ, ξ, u, k, p) are used. To perform the key sensitivity analysis, a minute change of ∆ = 10−15

Scientific Reports | (2024) 14:31054 | https://doi.org/10.1038/s41598-024-82256-3 14

Content courtesy of Springer Nature, terms of use apply. Rights reserved
www.nature.com/scientificreports/

Fig. 5. Polynomial surface fit plots for entropy values of 100 original and corresponding encrypted images.

Image size Images Directions 32 28 44 45 46 Proposed

H 0.0023 0.0015 0.0014 0.0016 0.0028 0.0001
Xray V 0.0016 0.0019 0.0022 0.0019 0.0029 0.0002
D 0.0011 0.0031 0.0026 −0.0016 0.0019 −0.0003
256 × 256 × 3
H 0.0014 0.0013 −0.0023 −0.0015 0.0014 −0.0013
Tumor V 0.0011 0.0021 −0.0016 0.0031 −0.0020 0.0001
D 0.0011 0.0013 −0.0011 0.0018 −0.0011 −0.0007
H 0.0021 −0.0022 −0.0013 0.0022 −0.0016 −0.0002
Ultrsound V 0.0011 0.0019 −0.0017 0.0031 0.0010 0.0003
D 0.0016 0.0021 −0.0033 0.0017 0.0033 0.0006
512 × 512 × 3
H 0.0023 −0.0014 −0.0013 −0.0017 0.0032 −0.0013
ECG signal V 0.0016 −0.0011 0.0031 0.0008 0.0031 0.0005
D 0.0034 0.0064 0.0046 0.0044 −0.001 −0.0003

Table 4. Correlation analysis.

is introduced to each secret key. The modified keys are ϖ1′ = ϖ1 + ∆, ϖ2′ = ϖ2 + ∆, . . . , ϖ6′ =
ϖ6 + ∆, σ1′ = σ1 + ∆, ρ′1 = ρ1 + ∆, β1′ = β1 + ∆, σ2′ = σ2 + ∆, ρ′2 = ρ2 + ∆,
β2′ = β2 + ∆, Ξ′ = Ξ + ∆, ξ ′ = ξ + ∆, u′ = u + ∆, k ′ = k + ∆, p′ = p + ∆. These modified keys are then
used to decrypt the plaintext image from the ciphertext image.
The resulting decrypted images, which are recovered using the modified secret keys, are shown in Figure
7. It can be observed in Figure 7(c) that the decrypted images contain no recognizable information from the
original plaintext image, which indicates that even a tiny modification to the secret keys renders decryption
completely ineffective. This demonstrates that the secret keys used in the proposed encryption framework are

Scientific Reports | (2024) 14:31054 | https://doi.org/10.1038/s41598-024-82256-3 15

Content courtesy of Springer Nature, terms of use apply. Rights reserved
www.nature.com/scientificreports/

Fig. 6. Correlation in multi directions (a-c) Correlation in Tumor image, (d-f) Correlation in encrypted
Tumor image, (g-i) (a-c) Correlation in Xray image, (j-l) Correlation in encrypted Xray image.

highly sensitive. To successfully decrypt the plaintext information, the exact keys must be used. Figure 7 (d)
shows the decrypted image using the correct keys, where it is evident that the image is successfully recovered.

Key space analysis

Keyspace analysis refers to the size of the secret keys used in an encryption framework and is also related to brute
force attacks. In a brute force attack, an adversary attempts every possible combination of secret keys to breach
the encryption. To determine the size of the secret keys in the proposed encryption framework, first assess key
sensitivity as explained in Section 5.4. In this proposed framework, the sensitivity of each key is at least 10−15 ,
meaning the size of each secret key is 10+15 . Therefore, the total key space for the fifteen keys is 1015×15 , which

Scientific Reports | (2024) 14:31054 | https://doi.org/10.1038/s41598-024-82256-3 16

Content courtesy of Springer Nature, terms of use apply. Rights reserved
www.nature.com/scientificreports/

Fig. 7. Key sensitivity analysis.

is approximately 2747.43 . According to Alvarez’s47criteria for key space, any encryption algorithm with a key
space equal to or greater than 2100 is considered secure against brute force attacks.

Noise and cropping attack analysis

To make decryption fail, attackers often attempt to modify the ciphertext image by adding external noise to it or
cropping a portion of the ciphertext image.
In the ciphertext image, 10% salt and 10% pepper noise are added randomly to alter each pixel’s value. To
evaluate the proposed encryption framework, first, random noise is added to the ciphertext image pixels. For
each pixel at position (i, j) in the image I , the noisy image I ′ will be calculated using Equation 34.
{
0 with probability 0.1 (pepper noise)
I (i, j) =
′
255 with probability 0.1 (salt noise) (34)
I(i, j) with probability 1 − 0.2 = 0.8 (no change)

After adding noise to the ciphertext image pixels, the plaintext image is recovered from the noisy version of
the ciphertext. The visualized results are displayed in Figure 8. The decrypted image, shown in Figure 8(e),
demonstrates that the information in the plaintext image is still visible, although some noise remains in the
recovered image. This indicates that while the exact pixel values are not fully restored, the content of the plaintext
image is still easily recognizable. This result shows that the proposed encryption framework is effective in
resisting noise attacks.

To perform a cropping attack analysis on the encryption framework, 20% portion of the ciphertext image is
cropped and then attempt to decrypt the plaintext image from this cropped version. The decrypted image is
shown in Figure 8(f). Based on your cropping attack analysis, the results indicate that the proposed encryption
framework is effective at resisting cropping attacks. Despite removing 20% of the ciphertext image, the
decryption process is able to produce an output where the plaintext image remains largely recognizable. This
suggests that the encryption framework is resilient to partial data loss, as significant portions of the original
content are still recoverable from the cropped ciphertext. The ability of the framework to maintain the integrity
of the plaintext image under such conditions shows its robustness and effectiveness in handling attacks that
involve manipulating or removing portions of the ciphertext.
Several noise and cropping curves are generated for different images to demonstrate how much plaintext data
can be recovered after adding external salt and pepper noise at varying percentages from 10% to 50%, with a step
size of 2%. Likewise, for the cropping attack analysis, the image is cropped from 10% to 50% in increments of
2%. Figure 9 illustrates that even with the addition of 50% salt and pepper noise to the ciphertext, the proposed
framework is capable of decrypting more than 90% of the plaintext information. Similarly, after cropping 50% of
the ciphertext image, over 90% of the plaintext information can still be recovered.

Histogram analysis
Histogram analysis is a technique used to assess the security and effectiveness of an encryption algorithm
by analyzing the pixel distribution in the image48. For an unencrypted image, the histogram shows different
patterns corresponding to the visual content of the image. For an encrypted image, the histogram should be
uniform and distinctly different from that of the plaintext image. This uniformity indicates that the encryption
algorithm has effectively encrypted the original pixel values, which makes it difficult for an eavesdropper to
extract any meaningful information.
In the proposed research, several histograms are generated for the different color components of the images.
Figure 10 displays the histograms of the unencrypted color components, while Figure 11 shows those of
the encrypted color components. The histograms of the encrypted color components show a uniform pixel
distribution and differ significantly from those of the plaintext color components, indicating that the proposed
encryption framework effectively resists histogram attacks.

Scientific Reports | (2024) 14:31054 | https://doi.org/10.1038/s41598-024-82256-3 17

Content courtesy of Springer Nature, terms of use apply. Rights reserved
www.nature.com/scientificreports/

Fig. 8. Noise and cropping attack analysis.

Fig. 9. Noise and cropping attack analysis for different medical images.

Scientific Reports | (2024) 14:31054 | https://doi.org/10.1038/s41598-024-82256-3 18

Content courtesy of Springer Nature, terms of use apply. Rights reserved
www.nature.com/scientificreports/

Fig. 10. Histograms of the plaintext color images.

Histogram variance analysis

Histogram variance analysis is used to statistically evaluate the effectiveness of an encryption algorithm by
examining the distribution of pixel intensities in the encrypted image. Mathematically, histogram variance can
be calculated using Equation 35.

2
∑
L−1

σ = p(i) · (i − µ)2 (35)

i=0

Where, σ 2 is the histogram variance, L is the number of possible intensity levels, p(i) is the probability of
occurrence
∑L−1 of intensity level i, and µ is the mean intensity of the histogram which can be calculated as
µ = i=0 i · p(i). The histogram variance σ 2 quantifies how pixel intensities deviate from the mean intensity
value. For a strong encryption, the histogram variance should be low which indicates a uniform distribution of
pixel values. Table 5 shows that the proposed encryption framework achieves lower histogram variance values
than existing schemes, indicating its superior performance.

Lossless analysis
Lossless analysis determines whether any plaintext information is lost after decrypting the plaintext image from
the ciphertext image. To assess whether the proposed encryption method is lossless or lossy, two well-known
metrics are used: Peak Signal-to-Noise ratio (PSNR) and Mean Squared Error (MSE). PSNR measures the

Scientific Reports | (2024) 14:31054 | https://doi.org/10.1038/s41598-024-82256-3 19

Content courtesy of Springer Nature, terms of use apply. Rights reserved
www.nature.com/scientificreports/

Fig. 11. Histogram of the encrypted color components.

Image size Images 32 28 44 45 46 Proposed

Xray 260.36 261.46 261.30 260.15 259.16 258.31
256 × 256 × 3
Tumor 261.03 262.34 265.66 270.16 266.03 257.01
Ultrasound 266.31 261.09 270.16 267.78 260.33 259.46
512 × 512 × 3
ECG signal 260.19 265.19 270.19 259.66 26011 259.88

Table 5. variance analysis.

similarity between the plaintext image and the decrypted image, while MSE calculates the difference between
these images. Mathematically, PSNR and MSE can be computed using Equations 36 and 37, respectively.
( )
MAX2
PSNR =10 · log10 (36)
MSE

1 ∑∑
M −1 N −1

MSE = (I(i, j) − K(i, j))2 (37)

MN
i=0 j=0

Scientific Reports | (2024) 14:31054 | https://doi.org/10.1038/s41598-024-82256-3 20

Content courtesy of Springer Nature, terms of use apply. Rights reserved
www.nature.com/scientificreports/

PSNR
Image size Images 32 28 44 45 46 Proposed
Xray 50.5 59.12 63.32 71.65 50.3 ∞
256 × 256 × 3
Tumor 55.61 66.15 70.64 84.64 59.16 ∞
Ultrasound 59.61 61.76 68.97 74.68 68.31 ∞
512 × 512 × 3
ECG signal 60.19 66.97 70.64 73.15 69.11 ∞
MSE
Image size Images 32 28 44 45 46 Proposed
Xray 10.69 15.19 16.59 5.16 21.34 0
256 × 256 × 3
Tumor 11.65 16.19 18.33 3.59 22.66 0
Ultrasound 16.11 19.49 20.97 1.69 24.09 0
512 × 512 × 3
ECG signal 19.19 18.16 20.19 4.98 24.67 0

Table 6. Lossless analysis.

Image size Images 32 28 44 45 46 Proposed

Xray 0.014 0.220 0.362 0.0001 0.79 0.002
256 × 256 × 3
Tumor 0.016 0.259 0.038 0.0002 0.89 0.003
Ultrasound 0.030 0.430 0.636 0.002 1.68 0.003
512 × 512 × 3
ECG signal 0.035 0.510 0.729 0.004 1.59 0.005

Table 7. Computational time analysis.

Where, I(i, j) and K(i, j) represent the pixel values at position (i, j) in the plaintext and decrypted images,
respectively, and MAX is the maximum possible pixel value of the image (e.g., 255 for an 8-bit grayscale image).
Table 6presents the PSNR and MSE values for both the proposed and existing encryption schemes. The MSE
values for the encryption framework proposed in44 are nearly zero for each corresponding plaintext color image.
In contrast, for the proposed encryption framework, all PSNR values are infinite and all MSE values are zero,
indicating no difference between the decrypted image and the original plaintext image. This demonstrates that
the proposed encryption algorithm is lossless.

Computational time analysis

Apart from statistical security analysis and conducting cyberattacks on the ciphertext images to assess the
effectiveness of the proposed encryption framework, it is also essential to evaluate the encryption scheme’s
computational complexity. For real-time applications and low-memory devices, such as IoT devices, the
encryption framework must minimize the time required to encrypt plaintext images. To determine the
computational complexity of the proposed encryption framework, MATLAB commands ‘tic‘ and ‘toc‘ are used
to measure the elapsed time between two points in a script.
The ‘tic‘ command records the current time when it is executed, capturing the precise time in seconds since a
fixed reference point in the past. In contrast, the ‘toc‘ command calculates the time elapsed since the most recent
‘tic‘ command was executed. Mathematically, if Tstart and Tend represent the times at which ‘tic‘ and ‘toc‘ are
executed, respectively, then the elapsed time ∆T is given by Equation 38.
∆T = Tend − Tstart (38)

Table 7presents the computational complexity analysis for the proposed and existing encryption algorithms.
It can be observed that the encryption algorithm proposed in45 performs slightly better than the proposed
encryption framework. However, the proposed encryption framework is still suitable for real-time and IoT
applications that require fast processing, as it can encrypt plaintext images of sizes 256 × 256 and 512 × 512
in under one second. Moreover, apart from the encryption scheme in45, the proposed encryption framework
outperforms other comparable encryption schemes in terms of computational complexity.

Discussion
The proposed work consists of two main components: (i) securing secret keys from potential attackers using
Quantum Key Distribution (QKD), and (ii) protecting plaintext medical images through the integration of
multiple encryption techniques. The QKD is utilized alongside One-Time Pad (OTP) encryption to protect the
secret keys, and the proposed research mathematically demonstrates that these keys are fully secured against

Scientific Reports | (2024) 14:31054 | https://doi.org/10.1038/s41598-024-82256-3 21

Content courtesy of Springer Nature, terms of use apply. Rights reserved
www.nature.com/scientificreports/

cyberattacks. Additionally, the integration of multiple encryption techniques aims to maximize randomness in
the plaintext data to enhance protection against cyber threats.
For instance, using a single encryption method, such as a permutation operation that makes the encryption
vulnerable to various attacks. When only a permutation function is applied to a plaintext image, the process
remains insecure, as it only rearranges pixel positions without altering their values. This means that key patterns
and structural information within the original image are largely preserved which makes it easier for attackers to
reconstruct the image. Moreover, a ciphertext-only attack (COA) becomes possible when only a permutation is
employed. In such cases, an attacker can take advantage of the statistical characteristics of the image, including
patterns in color distribution or variations in intensity, to partially or completely reconstruct the original image.
One more significant weakness of a permutation-only encryption method is its vulnerability to known-plaintext
attacks (KPA). When an attacker has access to both the encrypted image and its corresponding plaintext image,
they can analyze the pixel rearrangements to determine the permutation pattern. This ability to exploit statistical
features makes the encryption scheme highly susceptible to reconstruction efforts. Furthermore, histogram
analysis also presents a significant threat to permutation-only encryption. Since permutation does not change
pixel values, the histogram of the encrypted image remains identical to that of the original image. This similarity
enables attackers to analyze the histogram which provides statistical clues about the structure and content of the
original image. Finally, the lack of diffusion in permutation-only encryption schemes makes them particularly
weak. According to Claude Shannon49, an encryption scheme can be considered secure only if it incorporates
both confusion and diffusion operations. Therefore, the proposed encryption framework effectively integrates
multiple encryption methods to create both confusion and diffusion in the plaintext data. This ensures strong
security with low latency, as demonstrated by statistical security analyses, computational evaluations, and
comparisons of the proposed encryption scheme’s strength with existing encryption methods presented in
Section 5.

Conclusion
This research presented a new hybrid encryption framework that combines quantum and classical cryptographic
techniques for the secure transmission of medical images in IoT-based telemedicine networks. By leveraging
QKD for the generation of secure shared keys and integrating classical cryptographic methods, the proposed
framework addressed several security concerns such as weak key management, and weak security of the digital
data. The proposed framework employed multiple encryption techniques, such as pixel shuffling, bit-plane
extraction, and various transformation techniques for the enhancement of the confusion and diffusion between
the plaintext image pixels. Experimental results demonstrated the robustness of the proposed encryption
framework. The proposed encryption framework achieved impressive values in entropy, correlation, and key
space, which are 7.999, 0.0001, and 2741.67 , respectively. Moreover, the proposed framework is also tested against
different cyber threats, including noise, cropping, and brute-force attack. It is found that after the addition of
the external noise and cropping various portions of the ciphertext image, the proposed framework remains
capable of decrypting over 90% of the plaintext image. This also showed its capability to protect sensitive medical
data against unauthorized access and tampering. Additionally, the proposed encryption framework is well-
suited for real-time and IoT applications where fast processing speed is required. This is demonstrated by the
computational complexity analysis, where it is observed that the proposed encryption framework can encrypt
plaintext color images of sizes 256 × 256 or 512 × 512 within one second.
The proposed encryption framework can be further enhanced by incorporating visual cryptography using
DWT. Additionally, to further reduce computational time, Artificial Intelligence (AI) will be integrated to
identify and focus on the Region of Interest (ROI) for encryption.

Data availability
The datasets used and/or analysed during the current study available from the corresponding author on reason-
able request.

Received: 3 October 2024; Accepted: 3 December 2024

References
1. Kim, S., Choi, B. & Lew, Y. K. Where is the age of digitalization heading? the meaning, characteristics, and implications of
contemporary digital transformation. Sustainability 13, 8909 (2021).
2. Liu, F. et al. Gearing resource-poor mobile devices with powerful clouds: architectures, challenges, and applications. IEEE Wireless
communications 20, 14–22 (2013).
3. Shafique, A., Mehmood, A. & Elhadef, M. Survey of security protocols and vulnerabilities in unmanned aerial vehicles. IEEE Access
9, 46927–46948 (2021).
4. Douglas, M., Bailey, K., Leeney, M. & Curran, K. An overview of steganography techniques applied to the protection of biometric
data. Multimedia Tools and Applications 77, 17333–17373 (2018).
5. Kaur, M., Singh, D., Kumar, V., Gupta, B. B. & Abd El-Latif, A. A. Secure and energy efficient-based e-health care framework for
green internet of things. IEEE Transactions on Green Communications and Networking 5, 1223–1231 (2021).
6. Shafique, A. & Ahmed, F. Image encryption using dynamic s-box substitution in the wavelet domain. Wireless Personal
Communications 115, 2243–2268 (2020).
7. Brahim, A. H., Pacha, A. A. & Said, N. H. Image encryption based on compressive sensing and chaos systems. Optics & Laser
Technology 132, 106489 (2020).
8. Wang, J., Chen, J., Wang, F. & Ni, R. Optical image encryption scheme based on quantum s-box and meaningful ciphertext
generation algorithm. Optics Communications 525, 128834 (2022).
9. Cherbal, S., Zier, A., Hebal, S., Louail, L. & Annane, B. Security in internet of things: a review on approaches based on blockchain,
machine learning, cryptography, and quantum computing. The Journal of Supercomputing 80, 3738–3816 (2024).

Scientific Reports | (2024) 14:31054 | https://doi.org/10.1038/s41598-024-82256-3 22

Content courtesy of Springer Nature, terms of use apply. Rights reserved
www.nature.com/scientificreports/

10. Shafique, A. A noise-tolerant cryptosystem based on the decomposition of bit-planes and the analysis of chaotic gauss iterated
map. Neural Computing and Applications 34, 16805–16828 (2022).
11. Mehmood, A., Shafique, A., Alawida, M. & Khan, A. N. Advances and vulnerabilities in modern cryptographic techniques: A
comprehensive survey on cybersecurity in the domain of machine/deep learning and quantum techniques. IEEE Access 12, 27530–
27555 (2024).
12. Sonko, S., Ibekwe, K. I., Ilojianya, V. I., Etukudoh, E. A. & Fabuyide, A. Quantum cryptography and us digital security: a
comprehensive review: investigating the potential of quantum technologies in creating unbreakable encryption and their future in
national security. Computer Science & IT Research Journal 5, 390–414 (2024).
13. Mehmood, A. et al. A time-efficient and noise-resistant cryptosystem based on discrete wavelet transform and chaos theory: An
application in image encryption. Journal of Information Security and Applications 78, 103590 (2023).
14. Chamola, V., Jolfaei, A., Chanana, V., Parashari, P. & Hassija, V. Information security in the post quantum era for 5g and beyond
networks: Threats to existing cryptography, and post-quantum cryptography. Computer Communications 176, 99–118 (2021).
15. Abd El-Latif, A. A., Abd-El-Atty, B., Mazurczyk, W., Fung, C. & Venegas-Andraca, S. E. Secure data encryption based on quantum
walks for 5g internet of things scenario. IEEE Transactions on Network and Service Management 17, 118–131 (2020).
16. Rijmen, V. & Daemen, J. Advanced encryption standard. Proceedings of federal information processing standards publications,
national institute of standards and technology 19, 22 (2001).
17. Yu, S. et al. Development of modified blum-blum-shub pseudorandom sequence generator and its use in education. Measurement
Science Review 22, 143–151 (2022).
18. Zhou, S. et al. A novel image encryption cryptosystem based on true random numbers and chaotic systems. Multimedia Systems
1–18 (2022).
19. Muhammed, R. K. et al. Comparative analysis of aes, blowfish, twofish, salsa20, and chacha20 for image encryption. Kurdistan
Journal of Applied Research 9, 52–65 (2024).
20. Yang, C., Taralova, I., El Assad, S. & Loiseau, J.-J. Image encryption based on fractional chaotic pseudo-random number generator
and dna encryption method. Nonlinear Dynamics 109, 2103–2127 (2022).
21. Zia, U. et al. Survey on image encryption techniques using chaotic maps in spatial, transform and spatiotemporal domains.
International Journal of Information Security 21, 917–935 (2022).
22. Shafique, A. & Shahid, J. Novel image encryption cryptosystem based on binary bit planes extraction and multiple chaotic maps.
The European Physical Journal Plus 133, 331 (2018).
23. Hu, M., Li, J. & Di, X. Quantum image encryption scheme based on 2d s ine 2-l ogistic chaotic map. Nonlinear Dynamics 111,
2815–2839 (2023).
24. Zhang, W. et al. An image encryption scheme using self-adaptive selective permutation and inter-intra-block feedback diffusion.
Signal Processing 151, 130–143 (2018).
25. He, J., Zhu, H. & Zhou, X. Quantum image encryption algorithm via optimized quantum circuit and parity bit-plane permutation.
Journal of Information Security and Applications 81, 103698 (2024).
26. Li, R., Liu, Q. & Liu, L. Novel image encryption algorithm based on improved logistic map. IET Image Processing 13, 125–134
(2019).
27. Iqbal, N. & Hanif, M. An efficient grayscale image encryption scheme based on variable length row-column swapping operations.
Multimedia Tools and Applications 80, 36305–36339 (2021).
28. Mohamed, N.A.E.-S., El-Sayed, H. & Youssif, A. Mixed multi-chaos quantum image encryption scheme based on quantum cellular
automata (qca). Fractal and Fractional 7, 734 (2023).
29. Qu, G. et al. Optical color image encryption based on hadamard single-pixel imaging and arnold transformation. Optics and Lasers
in Engineering 137, 106392 (2021).
30. Hu, W.-W., Zhou, R.-G., Jiang, S., Liu, X. & Luo, J. Quantum image encryption algorithm based on generalized arnold transform
and logistic map. CCF Transactions on High Performance Computing 2, 228–253 (2020).
31. Gao, J., Wang, Y., Song, Z. & Wang, S. Quantum image encryption based on quantum dna codec and pixel-level scrambling.
Entropy 25, 865 (2023).
32. Zhang, J. et al. Quantum image encryption based on quantum image decomposition. International journal of theoretical physics 60,
2930–2942 (2021).
33. Wen, H., Lin, Y., Kang, S., Zhang, X. & Zou, K. Secure image encryption algorithm using chaos-based block permutation and
weighted bit planes chain diffusion. IScience 27 (2024).
34. Rehman, A. U. et al. A color image encryption algorithm based on one time key, chaos theory, and concept of rotor machine. IEEE
Access 8, 172275–172295 (2020).
35. Kumar, R., Sharma, A. K. et al. Bit-plane based reversible data hiding in encrypted images using multi-level blocking with quad-
tree. IEEE Transactions on Multimedia (2023).
36. Zhu, S. & Zhu, C. Security analysis and improvement of an image encryption cryptosystem based on bit plane extraction and multi
chaos. Entropy 23, 505 (2021).
37. Singh, K. N., Singh, O. P., Baranwal, N. & Singh, A. K. An efficient chaos-based image encryption algorithm using real-time object
detection for smart city applications. Sustainable Energy Technologies and Assessments 53, 102566 (2022).
38. Wu, X., Wang, D., Kurths, J. & Kan, H. A novel lossless color image encryption scheme using 2d dwt and 6d hyperchaotic system.
Information Sciences 349, 137–153 (2016).
39. Valli, D. & Ganesan, K. Chaos based video encryption using maps and ikeda time delay system. The European Physical Journal Plus
132, 1–18 (2017).
40. Bennett, C. H., Brassard, G. & Ekert, A. K. Quantum cryptography. Scientific American 267, 50–57 (1992).
41. Özkaynak, F. Construction of robust substitution boxes based on chaotic systems. Neural Computing and Applications 31, 3317–
3326 (2019).
42. Shafique, A. A new algorithm for the construction of substitution box by using chaotic map. The European Physical Journal Plus
135, 194 (2020).
43. Hussain, I., Shah, T., Gondal, M. A., Khan, W. A. & Mahmood, H. A group theoretic approach to construct cryptographically
strong substitution boxes. Neural Computing and Applications 23, 97–104 (2013).
44. Zhou, N.-R., Hu, L.-L., Huang, Z.-W., Wang, M.-M. & Luo, G.-S. Novel multiple color images encryption and decryption scheme
based on a bit-level extension algorithm. Expert Systems with Applications 238, 122052 (2024).
45. Alexan, W., Elkandoz, M., Mashaly, M., Azab, E. & Aboshousha, A. Color image encryption through chaos and kaa map. Ieee Access
11, 11541–11554 (2023).
46. Yasser, I., Khalil, A. T., Mohamed, M. A., Samra, A. S. & Khalifa, F. A robust chaos-based technique for medical image encryption.
IEEE Access 10, 244–257 (2021).
47. Alvarez, G. & Li, S. Some basic cryptographic requirements for chaos-based cryptosystems. International journal of bifurcation and
chaos 16, 2129–2151 (2006).
48. Shafique, A., Mehmood, A., Alawida, M., Khan, A. N. & Khan, A. U. R. A novel machine learning technique for selecting suitable
image encryption algorithms for iot applications. Wireless communications and mobile computing 2022, 5108331 (2022).
49. Shannon, C. E. A mathematical theory of communication. The Bell system technical journal 27, 379–423 (1948).

Scientific Reports | (2024) 14:31054 | https://doi.org/10.1038/s41598-024-82256-3 23

Content courtesy of Springer Nature, terms of use apply. Rights reserved
www.nature.com/scientificreports/

Acknowledgements
This work has been supported by the CHEDDAR: Communications Hub for Empowering Distributed ClouD
Computing Applications and Research funded by the UK EPSRC under grant numbers EP/Y037421/1 and EP/
X040518/1.

Author contributions
A.S: Conceptualisation, Methodology, Writing - Original Draft; S.A.A.N: Data analysis; A.R: Data curation;
M.G: Review and editing; P.P: Theoretical analysis; Q.H.A: Supervision & funding acquisition; M.A.I: Supervi-
sion & funding acquisition; J.M: Project administration. All authors reviewed the manuscript.

Declarations

Competing interests
The authors declare no competing interests.

Additional information
Correspondence and requests for materials should be addressed to A.S.
Reprints and permissions information is available at www.nature.com/reprints.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and
institutional affiliations.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which
permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give
appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and
indicate if changes were made. The images or other third party material in this article are included in the article’s
Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included
in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or
exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy
of this licence, visit http://creativecommons.org/licenses/by/4.0/.

© Crown 2024

Scientific Reports | (2024) 14:31054 | https://doi.org/10.1038/s41598-024-82256-3 24

Content courtesy of Springer Nature, terms of use apply. Rights reserved
Terms and Conditions
Springer Nature journal content, brought to you courtesy of Springer Nature Customer Service Center GmbH (“Springer Nature”).
Springer Nature supports a reasonable amount of sharing of research papers by authors, subscribers and authorised users (“Users”), for small-
scale personal, non-commercial use provided that all copyright, trade and service marks and other proprietary notices are maintained. By
accessing, sharing, receiving or otherwise using the Springer Nature journal content you agree to these terms of use (“Terms”). For these
purposes, Springer Nature considers academic use (by researchers and students) to be non-commercial.
These Terms are supplementary and will apply in addition to any applicable website terms and conditions, a relevant site licence or a personal
subscription. These Terms will prevail over any conflict or ambiguity with regards to the relevant terms, a site licence or a personal subscription
(to the extent of the conflict or ambiguity only). For Creative Commons-licensed articles, the terms of the Creative Commons license used will
apply.
We collect and use personal data to provide access to the Springer Nature journal content. We may also use these personal data internally within
ResearchGate and Springer Nature and as agreed share it, in an anonymised way, for purposes of tracking, analysis and reporting. We will not
otherwise disclose your personal data outside the ResearchGate or the Springer Nature group of companies unless we have your permission as
detailed in the Privacy Policy.
While Users may use the Springer Nature journal content for small scale, personal non-commercial use, it is important to note that Users may
not:

1. use such content for the purpose of providing other users with access on a regular or large scale basis or as a means to circumvent access
control;
2. use such content where to do so would be considered a criminal or statutory offence in any jurisdiction, or gives rise to civil liability, or is
otherwise unlawful;
3. falsely or misleadingly imply or suggest endorsement, approval , sponsorship, or association unless explicitly agreed to by Springer Nature in
writing;
4. use bots or other automated methods to access the content or redirect messages
5. override any security feature or exclusionary protocol; or
6. share the content in order to create substitute for Springer Nature products or services or a systematic database of Springer Nature journal
content.
In line with the restriction against commercial use, Springer Nature does not permit the creation of a product or service that creates revenue,
royalties, rent or income from our content or its inclusion as part of a paid for service or for other commercial gain. Springer Nature journal
content cannot be used for inter-library loans and librarians may not upload Springer Nature journal content on a large scale into their, or any
other, institutional repository.
These terms of use are reviewed regularly and may be amended at any time. Springer Nature is not obligated to publish any information or
content on this website and may remove it or features or functionality at our sole discretion, at any time with or without notice. Springer Nature
may revoke this licence to you at any time and remove access to any copies of the Springer Nature journal content which have been saved.
To the fullest extent permitted by law, Springer Nature makes no warranties, representations or guarantees to Users, either express or implied
with respect to the Springer nature journal content and all parties disclaim and waive any implied warranties or warranties imposed by law,
including merchantability or fitness for any particular purpose.
Please note that these rights do not automatically extend to content, data or other material published by Springer Nature that may be licensed
from third parties.
If you would like to use or distribute our Springer Nature journal content to a wider audience or on a regular basis or in any other manner not
expressly permitted by these Terms, please contact Springer Nature at

[email protected]