Received 30 May 2023, accepted 15 June 2023, date of publication 20 June 2023, date of current version 27 June 2023.

Digital Object Identifier 10.1109/ACCESS.2023.3287858

Multimedia Security Using Encryption: A Survey

KHALID M. HOSNY 1 , (Senior Member, IEEE), MOHAMED A. ZAKI 1 , NABIL A. LASHIN 1,

MOSTAFA M. FOUDA 2 , (Senior Member, IEEE), AND HANAA M. HAMZA 1

1 Department of Information Technology, Zagazig University, Zagazig 44519, Egypt
2 Department of Electrical and Computer Engineering, Idaho State University, Pocatello, ID 83209, USA
Corresponding author: Khalid M. Hosny (k_hosny@yahoo.com)

ABSTRACT Considering the current dependency on digital technology in modern society, the protection
of multimedia is highly important. Encryption is vital in modern digital communication, ensuring data
confidentiality, authentication, integrity, and non-repudiation. Multimedia encryption-based security tech-
niques are becoming increasingly important as they allow for the secure sharing of multimedia content on
digital platforms. This survey aims to review the state of secure and privacy-preserving encryption schemes
applicable to digital multimedia, such as digital images, digital video, and digital audio. An extensive
analysis of the existing cryptography schemes and multimedia encryption algorithms will be conducted
to give an extensive overview of the current state of security encryption schemes specifically designed
for digital multimedia technology. The survey results will be used to understand better the effectiveness
and reliability of secure multimedia encryption schemes and contribute to developing efficient and secure
encryption schemes in the future.

INDEX TERMS Audio, encryption, image, multimedia, video.

I. INTRODUCTION them, resulting in growth in the branch of image encryp-

A. MULTIMEDIA tion [2], [3], [4], [5], [6], [7], [8], [9].
The impact of multimedia on daily life is undeniable. Multi- A digital video is a form of digital media that captures
media has become integral to our lives, from communicating and preserves moving visual content. As technology advances
to consuming entertainment [1]. It has changed the way we and video becomes more prevalent in entertainment, edu-
interact with each other, the way we learn, and the way we cation, and surveillance industries, the necessity to protect
work. It can create a more immersive experience for the user, this form of digital media’s privacy and security is growing.
allowing them to interact meaningfully with the content. For Various encryption techniques are employed to protect video
example, multimedia can create interactive learning experi- content from combating the risk of unauthorized access and
ences, allowing students to explore a topic more engagingly. breaches, resulting in growth in the branch of video encryp-
It can also create engaging advertisements, allowing compa- tion [10], [11], [12], [13], [14], [15], which aims to safeguard
nies to reach a wider audience and increase sales. video files and ensure the confidentiality of information.
Digital images are digital media that can capture and Digital audio is a type of digital media that can cap-
store visual information. Images are increasingly used to ture and store audio information. As the use of audio
transmit sensitive information across networks in various becomes more prevalent across industries such as music,
fields, such as medicine, defense, online banking, and podcasting, and voice communication, there is a grow-
telecommunications. There is a significant risk of this con- ing concern for the security and confidentiality of this
fidential information being stolen by unauthorized parties. type of digital media. Encryption techniques are employed
The goal is to safeguard these images through encryption to protect audio content from unauthorized access and
algorithms so that adversaries cannot access or understand breaches, resulting in growth in the branch of audio encryp-
tion [16], [17], [18], [19], [20], [21], [22], [23], which focuses
The associate editor coordinating the review of this manuscript and on ensuring the security of audio files and the information
approving it for publication was Gustavo Callico . they contain. The general diagram for multimedia protection

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 License.

VOLUME 11, 2023 For more information, see https://creativecommons.org/licenses/by-nc-nd/4.0/ 63027
 K. M. Hosny et al.: Multimedia Security Using Encryption: A Survey

using encryption is displayed in Figure 1. Also, Figure 2 that securely distributes a symmetric key to two or more
displays different applications for multimedia Encryption. entities. User authentication verifies that someone trying to
access an application or service is genuine.
B. ENCRYPTION Because multimedia types have essential features, such as
Encryption is a branch of mathematics that concerns mul- numerous redundancies and a potent correlation, traditional
timedia data transformation. It is an important tool for algorithms such as AES and DES could not guarantee high-
providing data storage and transmission security. It converts quality encryption. The AES and DES are also unsuitable
the raw image, video, or audio into an unintelligible data form for encrypting multimedia content in real-time. Therefore,
using a secret key. Encryption methods can be categorized many suitable encryption algorithms are proposed to protect
into three types, as displayed in Figure 3. multimedia data. These algorithms can be categorized into
Keyless methods: The encryption transformation function chaotic-based, deoxyribonucleic acid (DNA)-based, elliptic
in these methods does not use any keys. The hash function and curve-based, and metaheuristics-based methods.
pseudorandom number generator are examples of the keyless Chaotic maps are a valuable tool in the realm of security
algorithm. The hash function transforms a variable amount of due to their chaotic nature. The chaotic behavior of these
text into a short, fixed-length value known as a hash code. The maps means that even the slightest modification in the initial
pseudorandom number generator produces a deterministic conditions can result in substantial changes in the output. This
sequence of bits or numbers. The sequence will repeat after property makes it extremely difficult for an attacker to predict
a certain sequence length, although the sequence looks to or replicate the encryption key, thus making the encryption
lack any distinct pattern. The sequence has the appearance process more secure. Additionally, chaotic-based encryption
of being a truly random sequence. methods possess several advantages, such as sensitivity to
Single-key methods: The transformation function in these initial conditions, ergodicity, and the requirement of a large
methods uses a single key. The key may be identified for a key space [24]. These factors combined make chaotic maps
single person to protect stored data, which is only accessible a widely used tool in various security implementations. They
to that person. Also, the key may be known to two or more are a potential solution to enhance security in image, video,
persons to protect communication traffic. The single-key and audio encryption applications.
encryption methods are also referred to as symmetric encryp- Various sectors have recently shown much interest in DNA
tion methods. The symmetric encryption algorithms take the technology, including information science and healthcare.
input data and a single secret key, then convert the data into DNA contains a genetic code that can store and convert
an unintelligible form. Also, the corresponding decryption data into different genetic codes. DNA technology has
algorithms take the protected data and the same secret key recently been used to create a simulation environment for
to recover the original data. The symmetric algorithms have biological studies. This idea has been used in the case of
two forms, block cipher and stream cipher. A block cipher encryption, where DNA is used for storing and encrypting
works on data in a series of blocks. A stream cipher works multimedia data [25]. The main advantages of using DNA
with data as a series of bits. in encryption are its ability to store large amounts of data,
Two-key methods: The transformation function in these parallel processing capabilities, and low power consump-
methods uses two different keys. The keys are related and tion. These features make DNA a promising technology for
referred to as public and private keys. The public key is known encryption [26].
to many people, but the corresponding private key is only Metaheuristic approaches are popular when a high level of
known to a single person. The two- key methods are also optimization is required. Recently, there has been an increase
referred to as asymmetric encryption methods. The asymmet- in the utilization of these techniques in the realm of encryp-
ric encryption methods can protect the data in two manners. tion [27]. There are two main ways in which metaheuristic
The first is taking the input data and a public key and then approaches are used in encryption. The first is to gener-
encrypting the data by converting it to an unintelligible form. ate multiple encrypted data and select the most optimized
The corresponding decryption algorithm can recover the orig- one. The second is to optimize the initial parameters of
inal data from unintelligible data by the corresponding private chaotic maps to create efficient encryption keys. Researchers
key known only to a single person. The other manner is taking have implemented various encryption methods based on
the input data and a private key and then encrypting the data metaheuristic approaches, considering different aspects and
by converting it to an unintelligible form. The corresponding perspectives.
decryption algorithm can recover the original data from unin- Cellular automata have been widely used in cryptogra-
telligible data by the corresponding public key that may be phy to generate pseudorandom numbers. These models are
known to many persons. There are different applications for famous for their complexity, contributing to their high robust-
asymmetric such as digital signature, key exchange, and user ness and efficiency. They utilize specific rules to generate
authentication. In the digital signature, a value is calculated random sequences, making them suitable for encryption.
with an asymmetric algorithm and associated with a data Cellular automata’s parallelism property and straightforward
object so that the data receiver can use the signature to assure hardware structure are two key benefits for image encryption,
the integrity and origin of data. Key exchange is the process making it a very important option for encryption meth-

63028 VOLUME 11, 2023

K. M. Hosny et al.: Multimedia Security Using Encryption: A Survey

FIGURE 1. Multimedia Encryption Method.

FIGURE 2. Multimedia Encryption Applications.

ods. [28]. These properties of cellular automata make them requirements for maintaining a secure multimedia system,
a popular choice for many encryption techniques. which includes a threat model for detecting and prioritizing
Many papers reviewed multimedia security, whether an potential risks. They describe the risks to multimedia security
image, video, audio, or all. The authors in [29] reviewed and their impacts on users using textual descriptions and
multimedia encryption schemes based on chaotic maps. The an overview of the current state-of-the-art video-encryption
paper did not classify the encryption schemas into any cat- schemes. The authors also examine the performance param-
egories. Also, the authors did not mention the difference eters of these encryption schemes. Although, the paper did
between their and other surveys. Furthermore, they did not not use standard metrics such as entropy, Number of Pixel
outline the future research direction in the field. The authors Change Rates, Unified Average Changing Intensity, and other
in [30] surveyed multimedia cryptography. Their survey did metrics to evaluate the schemas against various attacks. Also,
not mention the last advancement schemas in the field. It is the paper did not state the future directions in the field. The
only stated the fundamental concepts and approaches in mul- authors in [32] surveyed medical image security approaches.
timedia encryption. The authors in [31] present the design They presented encryption schemas, watermarking schemas,

VOLUME 11, 2023 63029

 K. M. Hosny et al.: Multimedia Security Using Encryption: A Survey

FIGURE 3. Encryption Method Types.

and steganography schemas. The authors did not mention the The rest of the paper’s outline: The Common attacks and
existing surveys that include the three security schemas and related performance evaluation metrics are discussed in
how their survey differs from other surveys. The drawbacks Section II. Section III discusses various existing encryption
of the existing surveys can be summarized as follows: methods for images. The existing encryption methods for
• Some papers only introduce the main concepts and videos are presented in section IV. Section V discusses var-
approaches used in multimedia security and do not men- ious existing encryption methods for audio. In Section VI,
tion the recent schemas the paper presents a discussion and a comparison of different
• Some papers did not use the standard evaluation met- existing multimedia encryption methods. Finally, the paper
rics in their evaluation of the schemas against different concludes in Section VII.
attacks
• Some authors did not mention the existing surveys and
II. COMMON ATTACKS AND RELATED PERFORMANCE
the difference between their work and related work.
EVALUATION METRICS
• Some papers did not introduce future ideas to improve
Evaluation criteria are used to measure the effectiveness of
multimedia data security.
multimedia encryption. Attackers can attempt various secu-
In this study, we thoroughly examined utilizing encryption rity breaches to compromise the encryption method and
methods based on the types of multimedia for securing mul- discover the key. Attackers frequently employ cryptanalysis
timedia, as displayed in Figure 4. The key features of this to examine encryption techniques [33], [34].
survey include the following:
1) A thorough examination of the literature on encryption • Differential threat: This method is employed to evaluate
methods based on different types of multimedia the vulnerability of encryption to slight variations in
2) Examination of the common attacks and related metrics plain multimedia. An adversary slightly alters the orig-
used to evaluate the performance of different types of inal multimedia and then applies the same encryption
multimedia encryption techniques method to both the modified and unmodified multimedia
3) Comparison of various existing approaches to deter- to find a relationship between the newly created original
mining their security parameters multimedia and the ciphered one.
4) Outline of future directions for research in the field of • Threat based on statistical analysis: This threat compares
multimedia encryption to inspire further study. the encrypted and original multimedia statistics. The

63030 VOLUME 11, 2023

K. M. Hosny et al.: Multimedia Security Using Encryption: A Survey

FIGURE 4. Multimedia Encryption Classification.

tools needed to perform this study are the histogram and Multiple methods for image encryption have been suggested,
correlation coefficient. including position permutation-based, position substitution-
• Brute-force threat: This method involves attempting all based, and value transformation-based techniques. However,
possible combinations of keys until the correct one is the survey focuses on organizing these techniques based on
found to break the secret key utilized in the encryption the domain in which they are applied. Table 2 summarizes all
process. encryption methods for the image.
• Ciphertext-only threat: In this attack, the cryptanalysts
only have access to a limited set of ciphertexts and A. SPATIAL DOMAIN
attempt to decrypt them to uncover the secret key or the This paper [47] presents a new chaotic image encryption
original plaintext. algorithm that uses dynamic DNA coding and RNA comput-
• Noise threat: In this attack, the adversary adds noise ing to improve the chaotic system’s complexity and ensure the
to the encrypted multimedia to corrupt the plain mul- cryptographic algorithm’s security. The proposed algorithm
timedia’s usable information, preventing the intended uses a four-dimensional hyperchaotic architecture to produce
recipient from being able to get back the original multi- a plaintext-related key sequence, which is then utilized to
media after decrypting it. dynamically DNA encode the plaintext image. The encoded
It is crucial to keep the plaintext and secret key statistics image is then converted through RNA encoding and the
hidden. The security and quality of multimedia encryption replacement of amino acids. The modular operation of RNA
can be examined to determine its robustness. Statistical anal- makes the encryption procedure simpler, making it more
ysis, differential analysis, key analysis, etc., are all types of suitable for image encryption. Lastly, a refined generator for
security analysis. The quality analysis measures the quality of replacement sequences produces pseudo-random numbers to
the decrypted multimedia using metrics such as peak signal- be used in replacement tasks to create the last encrypted
to-noise ratio, mean square error, etc. The possible threats image. The results prove that the method is more secure than
to multimedia and related metrics are displayed in Figure 5. conventional DNA operations.
Additionally, a summary of all these measures is presented in In this paper [48], key-generation models for chaos-based
Table 1. image encryption have been developed by combining conven-
tional key-generation processes with meta-heuristic methods.
III. ENCRYPTION OF IMAGE By using two meta-heuristic approaches, the architecture
In this survey, we look at ways to classify image encryption encrypts data using two keys. The transposition process is
techniques according to the domain in which they are used. done using the first key. The use of the genetic algorithm

VOLUME 11, 2023 63031

 K. M. Hosny et al.: Multimedia Security Using Encryption: A Survey

FIGURE 5. Evaluation metrics.

results in the creation of this key-string. The transposition 3 altered and enhanced 1D chaotic maps to secure the trans-
technique gives the image a confused appearance. The image mission of images on the internet. The method employs a
is scrambled here by performing a transposition operation key image to start the maps and as a cover during the diffu-
on the image’s pixels. The diffusion operation is carried out sion stage through XOR operation. Additionally, ExtraParam
across the encrypted result produced in the previous round obtained from the source image is employed in ciphering step
using a different key. Differential evolution theory has been to increase the chaotic maps’ sensitivity to bit changes during
used to the results of the logistic map generation to get this initialization. The algorithm features two main components,
second key. A chaotic map of the same size as the image confusion and diffusion, and has undergone rigorous testing
is created using the key string obtained using this LM-DE and cryptanalysis with successful outcomes.
approach. This map is employed in a diffusion procedure to The authors in [51] present a new encryption method
encrypt the image using a substitution cipher. This substi- that uses a novel two-dimension hyperchaotic map called
tution procedure is used twice to assure distinct outcomes, Sine ICMIC double cross-map. Combined bit-level grayscale
even for a single bit change in the source picture, to pro- matrices are produced from the color images. The method
vide complete proof of security. The suggested plan offers also performs random substitution, bit cyclic shifting, and
robust cryptography regarding information randomness, key XOR to improve encryption security. The technique also
sensitivity, resistance to noise, and various statistical and employs bidirectional diffusion to strengthen resistance to
differential assaults. differential attacks, scanning designs to lessen correlation
The paper [49] proposes a novel two-phase encryption between neighboring pixels, and selecting diffusion under a
design utilizing chaos theory for secure and efficient pro- quantized chaotic sequence to boost randomness. The results
tection of privacy images. The proposed algorithm combines stated that the method could encrypt three images simultane-
chaotic logistics and Arnold’s Cat mapping to encrypt gray ously.
images. The algorithm was tested on nine standard gray In [52], the authors introduce a new approach that builds
images and was found to be more efficient and secure when a 1D discrete chaotic structure that can be controlled by
compared to other similar algorithms. The researchers evalu- adjusting the parameters to achieve uniform distribution. The
ated the proposed algorithm using various image evaluation proposed method employs Marotto’s theorem, demonstrating
criteria. They also examined the algorithm’s resistance to that a 1D discrete structure can be chaotic in the sense of Li-
attacks on the original and encrypted image. These eval- Yorke. A chaos criterion theorem is presented to determine
uations and comparisons with similar algorithms showed the existence of chaos in the discrete system. An image
that the proposed algorithm performed better and was more encryption scheme is proposed based on this uniformly dis-
secure against attacks. The researchers suggested that future tributed chaotic system, which uses the chaotic sequence to
research could focus on developing better algorithms or intro- scramble, and XOR transforms the image. The plain text
ducing new evaluation criteria to enhance the algorithm’s determines DNA encoding and decoding operations. The
performance and protection. proposed method offers an efficient and secure way to encrypt
The authors in [50] present an innovative approach images and can be applied in various fields, such as military
for encrypting images that incorporates a combination of and financial applications.

63032 VOLUME 11, 2023

K. M. Hosny et al.: Multimedia Security Using Encryption: A Survey

TABLE 1. Evaluation metrics.

VOLUME 11, 2023 63033

 K. M. Hosny et al.: Multimedia Security Using Encryption: A Survey

TABLE 1. (Continued.) Evaluation metrics.

This paper [53] uses the Fibonacci Q-matrix with a 4D arrangement of the pixels in the sub-images and blocks is
hyperchaotic Chen system of fractional order to suggest a altered during the scrambling process to create a scrambled
novel technique for encrypting color images. Three steps image. Then the produced image is diffused using the logis-
in the encryption algorithm are presented. The three color tic map during the diffusion stage to obtain the encrypted
components of the input image are separated in the first image. According to the experimental findings, the suggested
stage, then confusion and diffusion procedures are carried out technique is highly effective and secure at encrypting color
separately for each component. In the second phase, the 4D images.
hyperchaotic Chen fractional-order system produces random The paper [56] introduces a new technique for protecting
integers to permute the pixel positions. In the final phase, grayscale images that utilize a Fibonacci Q-matrix and a
small 2 × 2 blocks of the permuted image are separated, and 6D hyperchaotic system. The proposed algorithm aims to
the Fibonacci Q-matrix is utilized in the diffusion process increase the encrypted image’s security level and resist dif-
for every block. The proposed method can resist various ferent attacks. The ciphering procedure includes two main
threats and is perceptive to key and pixel distribution changes, steps: scrambling and diffusion. In the scrambling step, the
leading to a highly secure encryption method. positions of the pixels in the original image are randomly
This paper [54] combines a 2D logistic-sine map and rearranged using a 6D hyperchaotic system. The permuted
FrMGMs that are fractional multichannel Gegenbauer image is diffused in the diffusion step using the Fibonacci Q-
moments to create a novel color image encryption technique. matrix. The approach is evaluated under various attacks. The
The three basic processes in the encryption process are: results show that the suggested procedure has an excellent
changing the values of picture pixels by utilizing the two- level of security and outperforms existing image ciphering
dimension logistic-sine map to produce an image that is procedures.
scrambled, using the logistic map and FrMGMs to generate a The paper [57] introduces a novel chaos-based medical
key, and using the key to diffuse the scrambled image to create image ciphering approach. This method aims to speed up
the ciphered image. The suggested approach can withstand the entire encryption process, which is crucial for protecting
common security threats and has a big keyspace and high key medical images received across IoT devices for healthcare.
sensitivity. According to the authors, future research might Four steps make up the encryption procedure: image divid-
use this technique to encrypt real-time video and medical ing, image scrambling, key creation, and diffusion. Using
pictures. a novel image-splitting method, the original image is split
The paper [55] uses the ideas of block scrambling and into blocks and sub-blocks as part of the algorithm’s initial
chaos to provide a novel method for encrypting color images. stage. The method then uses a zigzag pattern, rotation, and
The two primary components of the encryption technique random permutation between blocks to rearrange the pixels
suggested in the paper are scrambling and diffusion. Red, in the image blocks to scramble them. This phase eliminates
Green, and Blue, the three components that make up the pixel correlation and raises the encrypted image’s security
simple image, are separated into sub-images and blocks. The level. The logistic map is used to create a key in the third

63034 VOLUME 11, 2023

K. M. Hosny et al.: Multimedia Security Using Encryption: A Survey

stage, and the source image determines the starting param- strong encryption effect and can withstand common threats.
eter of the map. By doing this, the encryption procedure is The key contributions of the study are improving the security
made resistant to differential attacks. The algorithm’s last step and sensitivity of the suggested technique by using an NCA
includes diffusing the scrambled image by altering the pixel map in CML, a one-time pad encryption strategy, a DNA-
values with the secret key. The findings demonstrate that the level merge-shuffling process, and a pixel-level diffusion
suggested algorithm is highly secure. process.
This paper [58] utilizes the ideas of ribonucleic acid In this paper [61], The authors suggest an approach for
(RNA) and deoxyribonucleic acid (DNA) to introduce a image ciphering based on hyper-chaos to enhance the security
novel two-phase image encryption technique that enhances of existing low-dimensional chaotic-map-based encryption
the security of the encryption process. The first phase of techniques. The proposed approach creates a key stream
the suggested technique employs DNA rules, the DNA XOR linked to the source image using a 5D multi-wing hyper-
operator, and a chaotic map function to cipher the original chaotic system, which guarantees that different plain images
image. The second phase takes the encrypted image as its will produce entirely different key streams, making the cho-
input and encrypts it using the RNA codon truth table and an sen plaintext and ciphertext attacks useless. The authors
encryption key. The use of RNA is motivated by the fact that, further distorted the image using a bit- and pixel-level per-
although being a single-strand structure, it achieves strong mutation. The algorithm can increase the security of the
encryption quality that is competitive with DNA. When added cryptosystem by combining these two permutations. Finally,
to double-strand DNA, it improves encryption quality while the authors modify the image’s pixels using a diffusion
preserving algorithm efficiency. process. According to the authors’ theoretical research and
This paper [59] suggests a new color image ciphering numerical simulations, the suggested approach is secure and
procedure that utilizes a hash function in a chaotic sys- dependable for image encryption. The proposed technique is
tem’s parameter initialization. The hash function that is considered an improvement over traditional low-dimensional
used in the suggested approach is SHA-256. The procedure chaotic-map-based algorithms as it uses a more complex
consists of three steps: first, an image’s one-dimensional hyper-chaotic system and employs pixel-level and bit-level
vector of the three color channels is ordered and then sorted permutations to enhance security further.
using the Piecewise Linear Chaotic Map’s chaotic sequence. This paper [62] presents a new asymmetric image encryp-
The permuted array is then divided into three pieces, each tion algorithm that aims to improve the security of image
representing a color channel, and is individually permuted transmission among multiple people by using a small number
once again using Lorenz’s chaotic system. Finally, a chaotic of keys and a simple key transmission mode. The suggested
approach is used to individually encode each channel’s pixel approach first compresses the original image, then utilizes an
into Deoxyribonucleic Acid (DNA) bases. The XOR function updated 4D cat map and elliptic curve ElGamal encryption to
with DNA complementary rules is performed to substitute encrypt it, then diffuses it. The approach employs asymmetric
each pixel. The operation is performed many times with cryptography to increase the security of the encryption sys-
various DNA rules. The DNA rule’s choice and the opera- tem. It uses a chaos-based permutation and diffusion system,
tion’s continuation depend on Chen’s chaotic sequence. The including 3D Lorenz and 4D cat maps. Since it is secure, the
suggested approach has the benefit of resisting various noises algorithm can withstand attacks like differential and statisti-
and losing a portion of the ciphered image, and it achieves cal threats.
outstanding encryption results in just one round. This paper [63] presents a Cellular Automata-based image
This paper [60] proposes a unique one-time key, spa- ciphering technique to address the issue of the incapac-
tiotemporal chaos, and DNA sequence operations-based ity to create large sequences of states and a finite number
color image encryption technique. Four steps comprise the of reversal rules. The suggested approach includes chaos
algorithm: The NCA map-based CML first generates key mapping, hyper-chaotic mapping, confusion and diffusion
streams. The plain-image and secret keys are combined using processes, and a key image produced using non-uniform cel-
the SHA-256 hash function to update the system param- lular automata. The suggested technique uses a Chen hyper-
eters and initial conditions. Second, using DNA encoding chaotic function, a Logistic map, and a three-dimensional
rules, the basic image’s red, green, and blue components are Arnold mapping for encryption. The proposed scheme’s
split into three DNA matrices. Row-wise and column-wise primary contributions include robust key image generation
permutations are carried out on the concatenated DNA matri- employing hyper-chaotic functions and non-uniform CA.
ces. Thirdly, three equal blocks of the shuffled DNA matrix The paper [64] presents an asymmetric image encryption
are created and subjected to XOR, subtraction, and addi- paradigm that creates initial values, permutation sequences,
tion operations. Eventually, using DNA decoding methods, and diffusion sequences using a 2D discrete-time map and
the produced matrices are converted into numbers matrices. the source image’s hash value. The system encrypts the red,
Implementing a diffusion mechanism with the key streams green, and blue components independently using exclusive or
improves the system’s security. Experimental results and (XOR) operations after circularly shifting the image’s pixels
security assessments show that the suggested technique has a by rows and columns. The plain image’s hash value deter-

VOLUME 11, 2023 63035

 K. M. Hosny et al.: Multimedia Security Using Encryption: A Survey

mines the encryption keys, while the Hénon map’s iteration frequency domain is used for the first time in the suggested
count, beginning values, and control parameters determine encryption scheme.
the decryption keys. This paper [69] introduces a new lossless encryption
In [65], the authors suggested a chaos-based color image method for color images that uses a 6D hyperchaotic sys-
ciphering technique that uses a bijective function to cipher tem and a 2D discrete wavelet transform (DWT). The 2D
images. A sequence of random numbers is employed to dif- DWT splits the image into four sub-bands, then the algorithm
fuse the original image for random rounds, then separates the encrypts it by permuting the sub-bands with a key stream and
three color channels of the image and splits each channel into decreasing their size by a consistent amount. The image is
blocks of equal size. A Chen system with changeable condi- then reconstructed using the 2D inverse DWT, and additional
tions generates an 8 × 8 S-box that corresponds one-to-one key streams are employed to alter the pixel values further. The
to each block to create a bijection between the block set and key streams are created utilizing the hyperchaotic system and
the S-box set. Each block is replaced with its matched S-box the plain image to provide resistance against various assaults.
to create the ciphered image. Through numerical simulation Through experimental findings and security analysis, the
and security analysis, the suggested technique is workable suggested algorithm is demonstrated to have good security,
and appropriate for image encryption. speed, and resistance to attacks.
This paper [70] provides a study of optical image cipher-
ing in the Fresnel transform (FrT) domain utilizing a phase
mask based on the devil’s vortex toroidal lens (DVTL) in
B. TRANSFORM DOMAIN the frequency plane and a random phase mask (RPM) in
The paper [66] provides an approach for digital image cipher- the input plane. The suggested encryption technique is a
ing that employs the global bit scrambling (GBS) and the lensless optical system that has been digitally constructed
integer wavelet transform (IWT) techniques. Instead of the using MATLAB. This strategy resolves the optical setup’s
usual pixel scrambling technique, the proposed method uses axis alignment issue and increases the proposed scheme’s
a key-dependent bit scrambling mechanism to increase the security. The approach’s effectiveness is confirmed by cal-
security of the encryption process. IWT produces a faster, culating mean-squared error, and the scheme’s sensitivity
more memory-efficient encryption technique that can be used has been investigated for many parameters. The suggested
for lossless data compression applications. The proposed approach has been proven reliable against noise and occlusion
method is assessed with the help of a collection of standard attacks in the Fresnel domain.
images. The outcomes demonstrate its ability to withstand This paper [71] discusses a novel approach for encrypt-
intrusion attempts while offering high unpredictability in the ing digital images to protect secret data from unauthorized
encrypted image. access. In the suggested method, the Discrete Cosine Trans-
This paper [67] develops a strong and safe cryptosystem form Integer, Discrete Wavelet Transform (DWT), and
enabling color images to be distributed and shared on open Wavelet Transform (IWT) are used in combination with a
cloud servers. The random-multiresolution singular value cyclic shift and the 2-D chaotic Baker map (DCT). The use
decomposition and discrete trinion Fourier transform are the of transform domains allows for correlated data patterns to
foundation for the proposed cryptosystem. First, a trinion be transformed, resulting in a large degree of randomization
matrix is used to represent the color image. A block-wise upon returning to the spatial domain. The DWT domain
discrete trinion Fourier transform creates the cipher image, is the most secure and noise-immune, providing maximum
mapping from the trinion number domain to the real number randomness when using different keys.
domain and chaotic sequence-based scrambling. The research
concludes that the trinion matrix is useful in the realm of digi-
tal picture security, and further work will focus on developing IV. ENCRYPTION OF VIDEO
a paradigm for multiple-color images based on the theory of Various encryption methods for video protection are intro-
trinion duced. These methods can be classified into different cate-
The paper [68] develops an image ciphering approach gories. In this review, we focus on classifying the encryption
based on dynamic deoxyribonucleic acid coding and chaotic methods into two categories that most researchers are devel-
operations using hyper digital chaos in the frequency domain. oping their methods based on them. The first category is full
The proposed encryption algorithm uses the amplitude and encryption methods, and the other one is selective encryp-
phase components in the frequency domain to diffuse and tion methods. The entire video data is directly encrypted
scramble the image. The algorithm is evaluated using dif- in full encryption using an encryption technique, as dis-
ferent parameters. According to the findings, the encrypted played in Figure 6 (B). The full encryption procedures offer
image performs very well against statistical attacks, meaning high-security encryption, but its complexity is high. They
the original image’s statistical characteristics are lost. The are employed in important fields, including medicine and
security against differential assaults is additionally improved the military. With selective encryption, the video content is
by the encryption algorithm, making the ciphered image slightly damaged during the encryption operation but is still
more sensitive to plain-text. The chaotic DNA method in the partially understandable, as shown in Figure 6 (c). They are

63036 VOLUME 11, 2023

K. M. Hosny et al.: Multimedia Security Using Encryption: A Survey

TABLE 2. Image encryption methods.

VOLUME 11, 2023 63037

 K. M. Hosny et al.: Multimedia Security Using Encryption: A Survey

TABLE 2. (Continued.) Image encryption methods.

63038 VOLUME 11, 2023

K. M. Hosny et al.: Multimedia Security Using Encryption: A Survey

TABLE 2. (Continued.) Image encryption methods.

VOLUME 11, 2023 63039

 K. M. Hosny et al.: Multimedia Security Using Encryption: A Survey

TABLE 2. (Continued.) Image encryption methods.

employed in systems that demand less complexity time. The the methods that used selective encryption to provide video
first section discusses the methods that used full encryption security. Also, Table 3 summarizes all encryption methods for
techniques to protect the videos. The last section discusses video encryption.

63040 VOLUME 11, 2023

K. M. Hosny et al.: Multimedia Security Using Encryption: A Survey

FIGURE 6. Video Encryption Methods.

A. FULL ENCRYPTION This paper [75] provides a novel chaotic map-based

In [72], the authors suggested a ciphering method employ- video encryption technique independent of compression.
ing channel scrambling and key generation from a chaotic The method involves three main steps: generating chaotic
map to protect colored videos. The approach requires divid- sequences using a logistic-sine-coupling map, shuffling
ing the video frames into three color channels, each into individual frames using a sort index-based permuta-
blocks. The blocks are encrypted and scrambled with a tion technique, and performing block-level diffusion to
key created using a logistic map. The suggested strategy scramble the frames. The authors plan to investigate
is assessed for efficiency and suitability for use with IoT compression-dependent video encryption methods in future
devices utilizing a variety of security mechanisms, encryption work.
quality, and time complexity. The test results demonstrate The authors in [76] want to overcome the shortcomings
the suggested method’s high security and encryption effec- of the Advanced Encryption Standard (AES) for secure
tiveness, making it a promising option for protecting colored video storage and transmission; an Improved AES (IAES)
videos. algorithm is presented. The outcomes demonstrate that, com-
This paper [73] presents a technique for protecting privacy pared to the conventional AES, the suggested algorithm
in surveillance videos using object detection with YOLOv3 decreases the complexity time for the ciphering process and
and block scrambling-based encryption. The faces in the decryption and increases the ciphered video’s entropy. The
video are detected, then encrypted with a key taken by a proposed technique enhances diffusion and confusion in the
logistic map. The detected faces’ edges are protected to pre- ciphered video while reducing the run time by removing the
vent critical information from leaking. The encryption entails sub-byte and shift-row tasks and replacing them with a mix-
dividing the ROI into blocks and sub-blocks, eliminating row task. The Henon chaotic map is employed in the key
correlations between ROI pixels using a zigzag pattern and generation process, and the SHA-3 creates the chaotic attrac-
block permutation, then encrypting each ROI with a different tor’s beginning conditions. Overall, the results show that the
key. The proposed method is reversible, and its performance innovative security solution overcomes the shortcomings of
was evaluated using various analyses. From the findings, the the traditional AES and offers a trade-off between safety
suggested paradigm effectively detects and protects people’s levels and speed to protect the video content.
faces and can withstand attacks. This paper [77] suggests an effective hyperchaotic DNA
The authors in [74] suggested a lightweight stream cipher coding-based ciphering module combined with a sparse
method for video ciphering that employs the ChaCha20 representation-based spatial compression of video data. The
algorithm and a hybrid-chaotic map. The suggested technique suggested solution begins with the compression stage. Each
offers compression, effective computation, and high-level frame is encoded as a sparse matrix with a very small
security, the three major video transmission criteria. Two number of important sparse coefficients and most matrix
chaotic maps are employed in the approach for the key elements being zero. This sparse representation is accom-
generation process. With the assistance of the ChaCha20, plished using the matching pursuit technique and a trained
a symmetric approach is used in ciphering and deciphering. over-complete dictionary. The pursuit algorithm breaks the
The experimental findings demonstrate that the suggested signal into a regularized form, which gives compression an
method shrinks the size of the video by separating the vari- advantage. The encoded sparse coefficients are subjected to
ations between sequence frames using a dynamic reference the encryption stage, which involves designing and using a
frame and then encrypting the resulting video twice. novel 5D hyper-chaotic-DNA encryption module. The 5D

VOLUME 11, 2023 63041

 K. M. Hosny et al.: Multimedia Security Using Encryption: A Survey

hyper-chaotic-DNA coding-based encryption improves secu- preselected frame. In the second method, the XORing step
rity against various attacks. In video signals, the adjacent in VCMTH employs a key produced utilizing a chaotic
pixels are too associated. To further increase security, a global map. For both algorithms, the impact of block size and a
bit scrambling method decreases the correlation between Feistel structure on security is investigated. The findings
neighboring pixels in a cipher image. The suggested method of the proposed systems reveal that they reflect different
uses a 5D hyper-chaotic system, which is more chaotic and trade-offs between security and computing efficiency. Both
shows stable chaotic behavior. systems produce high-quality decrypted videos according to
This paper [78] uses a hyperchaotic structure to propose high PSNR values. However, VCMAB is slower but more
a video ciphering scheme. In the proposed approach, the resistant to known assaults than VCMTH. Additionally, the
video is separated into a frame and an audio stream, and suggested techniques have big keyspace that can withstand
then each component is ciphered using different chaotic exhaustive search assaults and are sensitive to small changes
encryption algorithms. The color image comprises three in the decryption key.
channels YCbCr, and the amount of information they carry The authors in [81] suggest a secure cryptosystem for video
differs. The Y channel encrypts more information using ciphering that uses chaos-based techniques to provide three
Arnold transforms and a DNA encoding algorithm. The levels of security. The initial parameters for the encryption
Lorenz hyperchaotic map is used to cipher the Cb and Cr system are produced using a mix of logistic and tent maps,
channels, which carry less information. The other stream and the encryption procedure entails one round of permuta-
is encrypted using the Logistic chaotic map. The proposed tion and diffusion. The video frames are randomly selected
scheme is designed to be incompatible with video encod- and then encrypted using permutation and diffusion. This
ing. The security tests and analysis reveal that the system method provides three levels of security: random frame selec-
is extremely resistant to statistical, power, and correlation tion, permutation, and diffusion. Multiple logistic and tent
assaults. maps increase sensitivity during the key generation process.
This paper [75] suggests new chaotic-based multi-media The proposed procedure is effective and resistant to various
encryption systems, which use 2D modification models assaults.
for safe data transfer. For confusion and diffusion rounds, In [82], the authors suggested two encryption approaches
a perturbation-based data encryption technique is suggested. based on chaos for video protection. A 12D chaotic map is
For media encryption, a hybrid chaotic structure incorpo- employed for the first method, and the Ikeda delay differential
rating various maps is used. The chaotic maps produce the equation (DDE) is employed for the other. These encryption
control parameters for permutation and diffusion structures. techniques use a substitution box (S-box) and a technique
The suggested schemes maintain a high level of encryption known as cipher block chaining (CBC) to diffuse the relation-
quality and have additional benefits like reduced residual ship between the pixels of the plain video and the encrypted
clarity and key sensitivity. Several different multimedia kinds video. Detailed analysis is carried out in this paper to demon-
have been evaluated in the article. The proposed pipeline is strate the security and uniqueness of the proposed schemes.
comprehensive and can encode all media formats, such as The major objectives of the suggested systems are to make
text, images, speech, and video. the key generation process more complex and to improve the
This paper [79] presents a new, cost-effective High- encryption process’s security. A larger key space in the 12D
Efficiency Video Coding (HEVC) encryption algorithm that chaotic map-based scheme can withstand statistical and brute
combines modified Mandelbrot sets, Arnold chaotic maps, force attacks.
and DNA encoding sequences. The suggested cryptosystem In contrast, the Ikeda DDE-based technique gener-
is made to withstand several multimedia attacks and improve ates keys using the chaotic dynamics of the system.
the security features of compressed HEVC frames for effec- It strengthens the encryption by creating a different
tive streaming. The cryptosystem can encrypt HEVC frames S-box for every frame. Both systems are suited for
of any size; its wide key space makes it resistant to brute-force real-time applications in terms of security and simplicity of
attacks. When tested and evaluated using different security implementation.
analysis methodologies, the proposed approach performs bet- In [83], the authors proposed a design for a real-time video
ter for secure video streaming. encryption system that uses the AES algorithm. The system
In the paper [80], VCMAB and VCMTH are two novel was implemented using a Xilinx FPGA and the Xilinx System
encryption algorithms suggested for secure video transmis- Generator for DSP. A video from a CMOS camera served as
sion. Both techniques involve a phase of substitution and the system’s input, while a DVI display screen showed the
permutation to induce confusion and diffusion. The chaotic encrypted video. The suggested architecture outperformed
maps are employed in the keystream production. The MPEG- previous comparable efforts in performance and hardware
2 standard is used to compress the video file before it is consumption. The device could encrypt video with a reso-
encrypted, but the methods can also be used with other lution of 720 × 480 at 60 frames per second. For camera
compression methods. VCMAB encrypts the video frames transfer and real-time encryption/decryption processing, the
by shuffling them using a chaotic map, and the scrambled system architecture needed 71% and 29% of the total time
frames are then obscured by XORing them with an encrypted level, respectively.

63042 VOLUME 11, 2023

K. M. Hosny et al.: Multimedia Security Using Encryption: A Survey

This paper [84] utilizes Discrete Cosine Transform (DCT) quantization parameter (QP) on inter-macroblock non-zero
and chaotic Arnold map Transform (AMT) for protecting coefficients and the impact of encrypting the sign of T1s are
the Multiview Video Coding/High-Efficiency Video Coding examined in the suggested encryption strategy. The suggested
MVC/HEVC data. The suggested framework has two phases: encryption system provides minimal computation cost, negli-
AMT, initial random phase mask (RPM1), and DCT are used gible bitrate increment, and effective resistance to brute-force
to shuffle, multiply, and decompose the video frames in the assaults, according to experimental data presented in the
first step. The final encrypted video frames are produced in study. In addition, a framework for implementing the sug-
the second stage, which involves shuffling the frames utiliz- gested SE strategy in multimedia social networks is presented
ing the AMT, modulating with RPM2, and decomposing with in the research.
IDCT. The use of AMT as a pre-processing shuffle method In this paper [88], the video under the HEVC standard is
ensures noise resistance. protected in the paper by suggesting three encryption strate-
gies. The proposed strategies are employed with insignificant
parsing overhead through the encoding and decoding opera-
B. SELECTIVE ENCRYPTION tions. The first technique, Sign Bin Encryption, manipulates
The paper [85] suggests a selective encryption (SE) proce- the signs of each CTU’s nonzero AC coefficient in the lumi-
dure for H.264 video that balances a high security and low nance channel. It randomizes the signs of Motion Vector
encryption ratio. The technique offers scalable encryption Displacement (MVD) Delta Quantization Parameter (dQP)
granularity by including an adjustable encryption parame- while avoiding extreme drift. The second method, Transform
ter to accommodate various security levels. The suggested Skip Bin Encryption, modifies each CTU’s transform skip
method makes precise key byte selection for encryption pos- flag array based on an encryption key. This technique will
sible by using a block weight model to examine the effects cause a minor decrease in the quality of the decrypted video
of each block on the motion compensation of other blocks. compared to the original video. The third technique is Suffix
The results of the experiments demonstrate how effective the Bin Encryption, which manipulates the fixed-length syntax
suggested method is in encrypting video files with sensitive of the transformed coefficients and motion vector differences
information. to preserve the video compression ratio and format compli-
In this paper () [86], the sensitive data of the HEVC ance, accomplished by adjusting the coefficient suffix and
standard is encrypted using a new technique. The suggested the motion vectors’ absolute horizontal and vertical values’
approach is maintaining HEVC constant bitrate and format suffixes. The paper concludes that, regardless of the texture
compliance while having a low complexity overhead and of the target movie, each approach can conceal the area of
quick encoding time. The Motion Vector Difference (MVD) interest.
and Discrete Cosine Transform (DCT) coefficients’ sign bits This paper [89] proposed a technique for HEVC-compliant
are encrypted using the low complexity chaotic logistic map video that combines selective encryption and data embed-
(CLM) in the suggested approach, known as the CLM-based ding. The encryption approach involves manipulating the
HEVC selective encryption (SE), during the entropy stage of Transform Skip Bin and Sign Bin elements in the HEVC cod-
the video encoding procedure. The suggested method’s secu- ing scheme to achieve security and data protection. One group
rity analysis, which is included in the paper, demonstrates that of elements in the HEVC coding structure is used to conceal
it is strong and superior to most attacks. the video material perceptually, and the other is used to embed
This paper [87] presents a selective encryption technique data. The method divides the data extraction and decryption
for video data that addresses the trade-off between com- operations, producing low parsing overhead. The results of
putation and security. The technique uses chaotic-based the experiments demonstrate that the suggested method can
substitution boxes (S-boxes) to permute a subset of the secure video content, enable data extraction in both encrypted
video data while maintaining compression efficiency and for- and decrypted forms, and perform joint selective encryption
mat compliance. The discrete cosine transforms coefficients, and data embedding with a minimal bitrate overhead.
trailing one sign, and non-zero transform coefficients were This paper [90] for the H.264 Scalable Video Coding
chosen to be secured in the video data. The remaining data (SVC) standard suggests an SE approach. The approach is
is subsequently encrypted using AES, integrated with the known as sufficient encryption (SE), and it seeks to keep
selected S-box. The results of the experiments demonstrate the bitstream’s compression efficiency and decoder format
that the proposed technique reduces computation without compliance while maintaining its confidentiality. Selected
compromising security, compression efficiency, or format codewords of the context-Adaptive Variable-Length Coding
compliance. (CAVLC) and Context-Adaptive Binary Arithmetic Cod-
This paper [85] protects the H.264/AVC video using ing (CABAC) entropy coders are subjected to encryption
a proposed selective encryption technique. The sign of to achieve SE. Sequences with various spatial resolutions
trailing ones (T1s), the sign of motion vector difference are used to test the suggested methodology. The out-
(MVD), intra-macroblock non-zero DCT coefficients, and comes demonstrate that it has several benefits over previous
the intra-prediction modes (IPMs) are ciphered to secure the approaches, including low computational latency, no bit-rate
motion information texture of the video. The impact of the escalation, and format compliance of the bitstream at the

VOLUME 11, 2023 63043

 K. M. Hosny et al.: Multimedia Security Using Encryption: A Survey

TABLE 3. Video encryption methods.

63044 VOLUME 11, 2023

K. M. Hosny et al.: Multimedia Security Using Encryption: A Survey

TABLE 3. (Continued.) Video encryption methods.

VOLUME 11, 2023 63045

 K. M. Hosny et al.: Multimedia Security Using Encryption: A Survey

TABLE 3. (Continued.) Video encryption methods.

63046 VOLUME 11, 2023

K. M. Hosny et al.: Multimedia Security Using Encryption: A Survey

TABLE 3. (Continued.) Video encryption methods.

VOLUME 11, 2023 63047

 K. M. Hosny et al.: Multimedia Security Using Encryption: A Survey

TABLE 3. (Continued.) Video encryption methods.

decoder. From the security analysis, the technique is reli- executed. The parameters of the chaotic maps are generated
able for commercial, real-time applications. The suggested by the suggested technique using a 128-bit shared key and
approach is appropriate for video delivery to customers who a 128-bit beginning vector. According to the article, future
have signed up for various video qualities on end systems research might expand this encryption method to cover other
ranging from modest portable devices to those with high multimedia applications like images and videos and enhance
spatial resolutions and frame rates. the suggested audio cryptosystem’s time efficiency.
The paper [91] presents research on H.264/AVC video This paper [18] suggests an audio ciphering procedure
security and privacy. The H.264/AVC video is encrypted called AEA-NCS that utilizes a nested chaotic system called
using the RC4 algorithm, a quicker stream cipher than block 2D-LNIC. The 2D-LNIC creates the keystream for the
cipher algorithms, which is covered in the first section of encryption technique by utilizing a logistic map with an
the research, which is the main topic of this work. The infinite collapse map. ESC-50 is used to test the algorithm,
compression and decompression time cycle is preserved since and the findings demonstrate that it performs well and con-
the encryption is done directly on the c H.264/AVC video siderably lowers the relationship of audio content in adjacent
encryption, data embedding, and data extraction are all part times.
of the research that is being proposed. This paper [93] employs a chaotic Henon and a tent map
The proposed selective video encryption algorithm [92] to produce a sequence of pseudo-random numbers as the
utilizes two selection layers to enhance the efficiency of secret key for a new audio ciphering technique. Both the
ciphering. The first layer examines the inter-frame depen- receiver and the sender generate the secret key. The random
dencies between neighboring frames at the macroblock number sequence and the audio data are combined using the
(MB) level using motion reference ratio (MRR), after which XOR technique to produce the encrypted audio file during
it chooses which MBs should be encrypted. The second the encryption stage. The same random number sequence
layer at the bitstream level employs bit sensitivity to deter- decodes the encrypted audio file in the decryption phase.
mine and encrypt the most important bits for the quality Instead of invitation links, the encrypted audio file can be
of the reconstructed video. The technique leaves the bit- used as an audio password in various e-learning activities.
stream’s compression ratio unchanged, which employs AES The encrypted audio file is designed to withstand statistical
encryption. The experimental findings demonstrate that the and brute force attacks with a uniform spectrogram, his-
suggested technique, which has a 50% lower computational togram, and a big keyspace. The values of the correlation
complexity than the current SEH264 algorithm, increases coefficients show no dependence between the original and
encryption efficiency and provides both perceptual and cryp- ciphered audio data values.
tographic security. Moreover, the entropy is high. Strong encryption and a high
degree of noise are indicated, respectively, by negative SNR
V. ENCRYPTION OF AUDIO and low PSNR values. The high value of NSCR demonstrates
Recently, various algorithms utilizing different techniques the method’s resistance against differential attacks.
have been proposed for audio encryption. Table 4 sum- This paper [96] introduces a new audio ciphering algorithm
marizes all encryption methods for audio encryption. This that combines substitution and permutation networks (SPN)
paper [21] provides a novel method for permuting and scram- for secure audio communication. The algorithm employs an
bling audio samples using a double DNA addition operation 8 × 8 S-box generated using a Mobius transformation for
and two chaotic maps. The Sine-Cosine map and the Logis- substitution and a chaotic map generated using the Hénon
tic Sine-Cosine map are employed. The suggested approach chaotic map for the permutation. Multiple tests of the sug-
consists of three steps: first, the Sine-Cosine map and a gested technique using audio files of various sizes reveal that
128-bit shared key are used in permutation; second, double it has an effective level of complexity and is appropriate for
DNA addition encoding is used to scramble the permuted audio encryption. Confusion and diffusion are achieved in
audio samples; and third, dynamic diffusion using either the the encryption process using a robust algebraic structure, the
Logistic Sine Cosine map or the Sine-Cosine chaotic map is AES S-box for substitution, and a chaotic map for the permu-

63048 VOLUME 11, 2023

K. M. Hosny et al.: Multimedia Security Using Encryption: A Survey

TABLE 4. Audio encryption methods.

VOLUME 11, 2023 63049

 K. M. Hosny et al.: Multimedia Security Using Encryption: A Survey

TABLE 4. (Continued.) Audio encryption methods.

TABLE 5. Comparison between image encryption methods.

tation. Additionally, the large keyspace almost prevents brute coding. The SHA-256 hash value of the audio is employed to
force attacks. The suggested technique also demonstrates the construct the chaotic sequences’ initial values created by the
ability to make statistical attacks ineffective. According to the PWLCM algorithm, resulting in an unpredictable, chaotic tra-
authors, future research might use small S-boxes to expand jectory. Next, a cyclical shift is applied to the audio’s binary to
the technology in an Android application. produce global scrambling. The key DNA matrix created by
This paper [17] suggests a new method for encrypting the chaotic sequence and the DNA matrix created by dynamic
audio recordings that combines chaotic systems and DNA coding is then XORed to create the encrypted audio.

63050 VOLUME 11, 2023

K. M. Hosny et al.: Multimedia Security Using Encryption: A Survey

TABLE 6. Comparison between video encryption methods.

This paper [97] suggests a multi-tier security system called This paper [99] offers a cosine number transform-based
segmentation ECC de-segmentation (SEED) for Securing audio encryption technique (CNT). The number of times
audio signal (SAIL) that combines elliptic curve encryption the CNT is applied to each block of samples in the non-
(ECC) and discrete wavelet transform (DWT) to increase the compressed digital audio signal is controlled by a key. The
security level in digital audio transmission for insecure public modified block substitutes the source block before the next
networks. The audio is divided into 8-bit chunks, compressed block is processed to diffuse the ciphered data to all processed
using DWT, encrypted using ECC, then de-segmented to blocks. The authors conduct computer experiments and report
create the encrypted audio signal. The authors claim that this that the approach satisfies secret-key cryptography’s funda-
is the first time ECC has been used for audio encryption, and mental security requirements. The proposed approach is also
the choice of ECC is motivated by its resistance to attack by noted for its simplicity, flexibility, fidelity, and computational
quantum computers. The system is also created to be com- efficiency.
putationally quick and simple for real-time applications like The paper [100] presents a proposed speech encryption
VoIP and live audio streaming. Additionally, the authors con- algorithm that uses three-dimensional chaotic maps. The
ducted several statistical tests to demonstrate that the system three primary components of the suggested algorithm are
offers greater protection and is resistant to statistical attacks, the generation of keys, the sample substitution, and the
making it appropriate for multi-channel audio processing. sample permutation procedure. The Lorenz and Rossler
This paper [98] proposes a novel block cipher and chaotic chaotic systems generate the keystreams for substitution
map-based audio file encryption scheme. The algorithm uses and permutation. The algorithm is designed to use chaotic
a three-stage procedure that entails permutation, XOR addi- cryptography by employing three dimension-chaotic maps
tion, and substitution to encrypt and decrypt audio data in in key generation. It aims to benefit from chaotic map
blocks of 625 bytes. A new algorithm built on the Chebyshev characteristics in speech encryption, such as the ability
polynomial produces the key. The algorithm is designed to to obtain complex global behavior, low residual intelli-
encrypt audio files without affecting the structure of the bit gibility, key sensitivity, large key space, and effectively
file, and it can also be used to encrypt photos and videos. The breaking the correlation between speech samples. The tech-
technique has been tested on WAV files of various sizes, and nique is built to withstand known-plaintext attacks and
the findings show that it is secure and resistant to attacks like provides a wide key space, making brute-force assaults
brute force and statistical threats. impossible.

VOLUME 11, 2023 63051

 K. M. Hosny et al.: Multimedia Security Using Encryption: A Survey

TABLE 7. Comparison between audio encryption methods.

This paper [101] proposes a new approach for selectively • DNA’s intricate complexity determines its security level,
encrypting encoded audio data utilizing a modified alternat- making it exceedingly challenging to perform computa-
ing step generator (ASG). The encrypted audio data maintains tions involving DNA and images.
standard compatibility and can withstand known plaintext • Many encryption methods have primarily concentrated
attacks using different keystreams to encrypt different audio on a limited number of performance factors and have
data frames. The ASG is a well-known clock-controlled not tackled the challenge of striking an optimal balance
generator that employs a combination of three linear feed- between conflicting aspects such as security and com-
back shift registers (LFSRs) and is vulnerable to enhanced plexity.
linear consistency assaults. To overcome this vulnerability, • Conventional encryption can significantly impair data
the authors propose incorporating nonlinearity in the stop/go availability since the original data becomes inaccessible
LFSRs of the ASG using nonlinear filter generators. The to anyone except the user who performed the encryption.
proposed method is tested using MP3 audio coding standards
and has low resource complexity. Based on past research, further advancements can be made in
multimedia encryption. In the future, it’s important to focus
on the following areas to continue research and progress in
this field:
VI. DISCUSSION
Recent security methods have been introduced for protecting • The tuning of key generators like chaotic maps can be
multimedia information. Different security measures, includ- improved by using metaheuristic techniques, machine
ing those outlined in Section II, have been used to test the learning [102], deep learning [103], deep belief net-
encryption methods and confirm their efficiency and secu- works, or deep transfer learning [104], [105].
rity. A comparison between the surveyed methods for image • Exploiting blockchain technology [106], [107] in pro-
encryption based on entropy, correlation, NPCR, UACI, and tecting multimedia data.
the key length is presented in Table 5. A comparison between • Current research in multimedia encryption doesn’t focus
the surveyed methods for video encryption based on entropy, on creating application-specific multimedia encryption
correlation, NPCR, UACI, and the key length is presented methods. In the future, research in application-aware
in Table 6. A comparison between the surveyed methods image encryption will be an important area of focus.
for audio encryption based on entropy, correlation, NPCR, • Encryption approaches for multispectral images and
UACI, and the key length is presented in Table 7. other multimedia data are an underdeveloped area. High-
The extensive analysis shows limitations to the practicality dimensional hyper-chaotic systems must be designed for
of the introduced encryption methods. The limitations can be these multidimensional forms of multimedia data.
summarized as follows: • With the rapid advancement of multimedia applica-
tions like medical and satellite imaging, high-resolution
• Cellular automata methods have limitations in their images are required, which leads to a need for com-
encryption/decryption algorithms due to their small rule putationally extensive multimedia encryption methods.
space and low diffusion power, which restricts the range Parallel multimedia encryption approaches can be used
of diversity achievable. to address this challenge.

63052 VOLUME 11, 2023

K. M. Hosny et al.: Multimedia Security Using Encryption: A Survey

• The integration of encryption with reversible data [13] R. S. Malladar and R. S. Kunte, ‘‘Selective video encryption using chaos
hiding/steganography is a growing area of research. for H.264 videos applicable to video on demand(VoD),’’ in Proc. 5th
Int. Conf. Commun. Electron. Syst. (ICCES), Jul. 2020, pp. 215–220, doi:
By combining both encryption and steganography, more 10.1109/ICCES48766.2020.9137861.
secure results can be obtained [14] A. Fitwi and Y. Chen, ‘‘Privacy-preserving selective video surveillance,’’
in Proc. 29th Int. Conf. Comput. Commun. Netw. (ICCCN), Aug. 2020,
pp. 1–10, doi: 10.1109/ICCCN49398.2020.9209688.
VII. CONCLUSION [15] S. M. U. Zia, M. McCartney, B. Scotney, J. Martinez, and A. Sajjad,
This paper conducts a comprehensive investigation of tech- ‘‘A novel image encryption technique using multi-coupled map lattice
system with generalized symmetric map and adaptive control parame-
niques for encrypting multimedia. The paper focuses on ter,’’ Social Netw. Comput. Sci., vol. 4, no. 1, pp. 1–16, Jan. 2023, doi:
encryption methods for image, video, and audio. Differ- 10.1007/S42979-022-01503-4.
ent multimedia encryption techniques were compared using [16] K. Kordov, ‘‘A novel audio encryption algorithm with permutation-
substitution architecture,’’ Electronics, vol. 8, no. 5, p. 530, May 2019, doi:
evaluation parameters to highlight their strengths and weak- 10.3390/electronics8050530.
nesses. It was found that the techniques need to have high [17] X. Wang and Y. Su, ‘‘An audio encryption algorithm based on DNA
levels of confusion, no correlation with the original multi- coding and chaotic system,’’ IEEE Access, vol. 8, pp. 9260–9270, 2020,
doi: 10.1109/ACCESS.2019.2963329.
media data, low computational complexity, and robustness [18] R. Wu, S. Gao, X. Wang, S. Liu, Q. Li, U. Erkan, and X. Tang, ‘‘AEA-
against cryptanalysis. Future research possibilities in multi- NCS: An audio encryption algorithm based on a nested chaotic system,’’
media encryption were explored. It was concluded that this Chaos, Solitons Fractals, vol. 165, Dec. 2022, Art. no. 112770, doi:
10.1016/J.CHAOS.2022.112770.
field still has room for improvement and growth. The paper [19] D. Wen, W. Jiao, X. Li, X. Wan, Y. Zhou, X. Dong, X. Lan, and W. Han,
aims to encourage researchers to explore the challenges in ‘‘The EEG signals encryption algorithm with K-sine-transform-based cou-
this area and to assist them in selecting the right approach pling chaotic system,’’ Inf. Sci., vol. 622, pp. 962–984, Apr. 2023, doi:
10.1016/J.INS.2022.12.001.
to develop new encryption methods that fit their specific [20] S. Jana, N. Dutta, A. K. Maji, and R. K. Pal, ‘‘A novel time-stamp-
applications, saving time. based audio encryption scheme using Sudoku puzzle,’’ in Proc. Int. Conf.
Frontiers Comput. Syst., Lecture Notes in Networks and Systems, vol. 404,
2023, pp. 159–169, doi: 10.1007/978-981-19-0105-8_16.
REFERENCES [21] A. Kumar and M. Dua, ‘‘Audio encryption using two chaotic map
[1] M. D. Abdulrahaman, N. Faruk, A. A. Oloyede, N. T. Surajudeen-Bakinde, based dynamic diffusion and double DNA encoding,’’ Appl. Acoust.,
L. A. Olawoyin, O. V. Mejabi, Y. O. Imam-Fulani, A. O. Fahm, and vol. 203, Feb. 2023, Art. no. 109196, doi: 10.1016/J.APACOUST.2022.
A. L. Azeez, ‘‘Multimedia tools in the teaching and learning processes: A 109196.
systematic review,’’ Heliyon, vol. 6, no. 11, Nov. 2020, Art. no. e05312, [22] Z. Dong, X. Wang, X. Zhang, M. Hu, and T. N. Dinh, ‘‘Global
doi: 10.1016/j.heliyon.2020.E05312. exponential synchronization of discrete-time high-order switched neural
[2] Z. Liang, Q. Qin, C. Zhou, and S. Xu, ‘‘Color image encryption networks and its application to multi-channel audio encryption,’’ Non-
algorithm based on four-dimensional multi-stable hyper chaotic system linear Anal., Hybrid Syst., vol. 47, Feb. 2023, Art. no. 101291, doi:
and DNA strand displacement,’’ J. Electron. Eng. Technol., vol. 18, no. 1, 10.1016/J.NAHS.2022.101291.
pp. 539–559, Jan. 2022, doi: 10.1007/S42835-022-01157-5. [23] R. I. Abdelfatah, ‘‘Audio encryption scheme using self-adaptive
[3] V. Vanitha and D. Akila, ‘‘Bio-medical image encryption using the mod- bit scrambling and two multi chaotic-based dynamic DNA
ified chaotic image encryption method,’’ in Artificial Intelligence on computations,’’ IEEE Access, vol. 8, pp. 69894–69907, 2020, doi:
Medical Data (Lecture Notes in Computational Vision and Biomechanics), 10.1109/ACCESS.2020.2987197.
vol. 37. Singapore: Springer, 2023, pp. 231–241, doi: 10.1007/978-981-19- [24] K. S. Babu, K. B. Raja, K. Kiran Kumar, T. H. Manjula Devi,
0151-5_20. K. R. Venugopal, and L. M. Patnaik, ‘‘Authentication of secret information
[4] C. Zhang and B. Du, ‘‘A fast piecewise image encryption scheme combin- in image steganography,’’ in Proc. IEEE Region 10 Conf., 2008, pp. 1–6,
ing NC1DNSM and P-box,’’ Integration, vol. 88, pp. 328–342, Jan. 2023, doi: 10.1109/TENCON.2008.4766581.
doi: 10.1016/J.VLSI.2022.10.003. [25] Q. Zhang, L. Liu, and X. Wei, ‘‘Improved algorithm for image encryp-
[5] Q. Lai, G. Hu, U. Erkan, and A. Toktas, ‘‘High-efficiency medical tion based on DNA encoding and multi-chaotic maps,’’ AEU Int.
image encryption method based on 2D logistic-Gaussian hyperchaotic J. Electron. Commun., vol. 68, no. 3, pp. 186–192, Mar. 2014, doi:
map,’’ Appl. Math. Comput., vol. 442, Apr. 2023, Art. no. 127738, doi: 10.1016/j.aeue.2013.08.007.
10.1016/J.AMC.2022.127738. [26] X.-Y. Wang, S.-X. Gu, and Y.-Q. Zhang, ‘‘Novel image encryption
[6] L. Liu and J. Wang, ‘‘A cluster of 1D quadratic chaotic map and its appli- algorithm based on cycle shift and chaotic system,’’ Opt. Lasers Eng.,
cations in image encryption,’’ Math. Comput. Simul., vol. 204, pp. 89–114, vol. 68, pp. 126–134, May 2015, doi: 10.1016/j.optlaseng.2014.12.025.
Feb. 2023, doi: 10.1016/J.MATCOM.2022.07.030. [27] M. Kaur, S. Singh, and M. Kaur, ‘‘Computational image encryption
[7] W. Song, C. Fu, Y. Zheng, M. Tie, J. Liu, and J. Chen, ‘‘A parallel techniques: A comprehensive review,’’ Math. Problems Eng., vol. 2021,
image encryption algorithm using intra bitplane scrambling,’’ pp. 1–17, Jul. 2021, doi: 10.1155/2021/5012496.
Math. Comput. Simul., vol. 204, pp. 71–88, Feb. 2023, doi: [28] B. Mondal, S. Singh, and P. Kumar, ‘‘A secure image encryption scheme
10.1016/J.MATCOM.2022.07.029. based on cellular automata and chaotic skew tent map,’’ J. Inf. Secur. Appl.,
[8] U. Erkan, A. Toktas, and Q. Lai, ‘‘2D hyperchaotic system based on Schaf- vol. 45, pp. 117–130, Apr. 2019, doi: 10.1016/j.jisa.2019.01.010.
fer function for image encryption,’’ Exp. Syst. Appl., vol. 213, Mar. 2023, [29] O. M. Al-hazaimeh, A. A. Abu-Ein, M. M. Al-Nawashi, and
Art. no. 119076, doi: 10.1016/J.ESWA.2022.119076. N. Y. Gharaibeh, ‘‘Chaotic based multimedia encryption: A survey
[9] Q. Lai, G. Hu, U. Erkan, and A. Toktas, ‘‘A novel pixel-split image for network and internet security,’’ Bull. Electr. Eng. Informat., vol. 11,
encryption scheme based on 2D Salomon map,’’ Exp. Syst. Appl., vol. 213, no. 4, pp. 2151–2159, Aug. 2022, doi: 10.11591/EEI.V11I4.3520.
Mar. 2023, Art. no. 118845, doi: 10.1016/J.ESWA.2022.118845. [30] B. Dhanalaxmi and S. Tadisetty, ‘‘Multimedia cryptography—A review,’’
[10] O. S. Faragallah, A. I. Sallam, and H. S. El-sayed, ‘‘Visual protection using in Proc. IEEE Int. Conf. Power, Control. Signals Instrum. Eng.
RC5 selective encryption in telemedicine,’’ Intell. Autom. Soft Comput., (ICPCSI), Sep. 2017, pp. 764–766, Jun. 2018, doi: 10.1109/ICPCSI.2017.
vol. 31, no. 1, pp. 177–190, 2022, doi: 10.32604/IASC.2022.019348. 8391817.
[11] M. Abomhara, O. Zakaria, O. O. Khalifa, A. A. Zaidan, and B. B. Zaidan, [31] H. Elkamchouchi, R. Anton, and Y. Abouelseoud, ‘‘Multimedia data
‘‘Enhancing selective encryption for H.264/AVC using advanced encryp- secure transmission: A review,’’ Int. J. Sci. Res. Manage., vol. 10, no. 12,
tion standard,’’ 2022, arXiv:2201.03391. pp. 949–971, Dec. 2022, doi: 10.18535/ijsrm/v10i12.ec01.
[12] J. He, Y. Xu, W. Luo, S. Tang, and J. Huang, ‘‘A novel selective encryp- [32] M. Magdy, K. M. Hosny, N. I. Ghali, and S. Ghoniemy, ‘‘Secu-
tion scheme for H.264/AVC video with improved visual security,’’ Signal rity of medical images for telemedicine: A systematic review,’’ Mul-
Process., Image Commun., vol. 89, Nov. 2020, Art. no. 115994, doi: timedia Tools Appl., vol. 81, no. 18, pp. 25101–25145, Mar. 2022,
10.1016/J.IMAGE.2020.115994. doi: 10.1007/S11042-022-11956-7.

VOLUME 11, 2023 63053

 K. M. Hosny et al.: Multimedia Security Using Encryption: A Survey

[33] S. Zhu and C. Zhu, ‘‘Plaintext-related image encryption algorithm based [54] K. M. Hosny, S. T. Kamal, and M. M. Darwish, ‘‘A novel color image
on block structure and five-dimensional chaotic map,’’ IEEE Access, vol. 7, encryption based on fractional shifted Gegenbauer moments and 2D
pp. 147106–147118, 2019, doi: 10.1109/ACCESS.2019.2946208. logistic-sine map,’’ Vis. Comput., vol. 39, pp. 1027–1044, 2023, doi:
[34] C. Zhu, G. Wang, and K. Sun, ‘‘Improved cryptanalysis and enhance- 10.1007/s00371-021-02382-1.
ments of an image encryption scheme using combined 1D chaotic maps,’’ [55] K. M. Hosny, S. T. Kamal, and M. M. Darwish, ‘‘A color
Entropy, vol. 20, no. 11, 2018, doi: 10.3390/e20110843. image encryption technique using block scrambling and chaos,’’
[35] Y. Zhang, B. Xu, and N. Zhou, ‘‘A novel image compression– Multimedia Tools Appl., vol. 81, no. 1, pp. 505–525, Jan. 2022,
encryption hybrid algorithm based on the analysis sparse repre- doi: 10.1007/s11042-021-11384-z.
sentation,’’ Opt. Commun., vol. 392, pp. 223–233, Jun. 2017, doi: [56] K. M. Hosny, S. T. Kamal, M. M. Darwish, and G. A. Papakostas,
10.1016/j.optcom.2017.01.061. ‘‘New image encryption algorithm using hyperchaotic system and
[36] Y. Xie, J. Yu, S. Guo, Q. Ding, and E. Wang, ‘‘Image encryption scheme Fibonacci Q-matrix,’’ Electronics, vol. 10, no. 9, p. 1066, Apr. 2021,
with compressed sensing based on new three-dimensional chaotic system,’’ doi: 10.3390/electronics10091066.
Entropy, vol. 21, no. 9, p. 819, Aug. 2019, doi: 10.3390/e21090819. [57] S. T. Kamal, K. M. Hosny, T. M. Elgindy, M. M. Darwish, and
[37] A. Ur Rehman, X. Liao, A. Kulsoom, and S. A. Abbas, ‘‘Selective encryp- M. M. Fouda, ‘‘A new image encryption algorithm for grey and color
tion for gray images based on chaos and DNA complementary rules,’’ medical images,’’ IEEE Access, vol. 9, pp. 37855–37865, 2021, doi:
Multimedia Tools Appl., vol. 74, no. 13, pp. 4655–4677, Jul. 2015, doi: 10.1109/ACCESS.2021.3063237.
10.1007/s11042-013-1828-7. [58] M. Yadollahi, R. Enayatifar, H. Nematzadeh, M. Lee, and J.-Y. Choi,
[38] E. Chen, L. Min, and G. Chen, ‘‘Discrete chaotic systems with one- ‘‘A novel image security technique based on nucleic acid con-
line equilibria and their application to image encryption,’’ Int. J. cepts,’’ J. Inf. Secur. Appl., vol. 53, Aug. 2020, Art. no. 102505, doi:
Bifurcation Chaos, vol. 27, no. 3, Mar. 2017, Art. no. 1750046, doi: 10.1016/j.jisa.2020.102505.
10.1142/s0218127417500468. [59] A. U. Rehman, X. Liao, R. Ashraf, S. Ullah, and H. Wang, ‘‘A color
[39] T. Sivakumar and R. Venkatesan, ‘‘A novel image encryption using callig- image encryption technique using exclusive-OR with DNA complemen-
raphy based scan method and random number,’’ KSII Trans. Internet Inf. tary rules based on chaos theory and SHA-2,’’ Optik, vol. 159, pp. 348–367,
Syst., vol. 9, no. 6, pp. 2317–2337, 2015, doi: 10.3837/tiis.2015.06.020. Apr. 2018, doi: 10.1016/j.ijleo.2018.01.064.
[40] D. Singh and V. Kumar, ‘‘A comprehensive review of computational [60] X. Wu, K. Wang, X. Wang, H. Kan, and J. Kurths, ‘‘Color image
dehazing techniques,’’ Arch. Comput. Methods Eng., vol. 26, no. 5, DNA encryption using NCA map-based CML and one-time keys,’’ Sig-
pp. 1395–1413, Nov. 2019, doi: 10.1007/S11831-018-9294-Z. nal Process., vol. 148, pp. 272–287, Jul. 2018, doi: 10.1016/j.sigpro.
[41] M. Kaur, D. Singh, and R. S. Uppal, ‘‘Parallel strength Pareto evolutionary 2018.02.028.
algorithm-II based image encryption,’’ IET Image Process., vol. 14, no. 6, [61] Y. Li, C. Wang, and H. Chen, ‘‘A hyper-chaos-based image encryp-
pp. 1015–1026, May 2020, doi: 10.1049/iet-ipr.2019.0587. tion algorithm using pixel-level permutation and bit-level permu-
[42] J. Qu, ‘‘Image encryption algorithm based on logistic chaotic tation,’’ Opt. Lasers Eng., vol. 90, pp. 238–246, Mar. 2017, doi:
scrambling system,’’ in Proc. IEEE 3rd Int. Conf. Inf. Syst. 10.1016/j.optlaseng.2016.10.020.
Comput. Aided Educ. (ICISCAE), Sep. 2020, pp. 519–522, [62] J. Wu, X. Liao, and B. Yang, ‘‘Color image encryption based on
doi: 10.1109/ICISCAE51034.2020.9236847. chaotic systems and elliptic curve ElGamal scheme,’’ Signal Pro-
[43] A. Kumar and M. Dua, ‘‘Novel pseudo random key & cosine transformed cess., vol. 141, pp. 109–124, Dec. 2017, doi: 10.1016/j.sigpro.2017.04.
chaotic maps based satellite image encryption,’’ Multimedia Tools Appl., 006.
vol. 80, no. 18, pp. 27785–27805, Jul. 2021, doi: 10.1007/s11042-021- [63] A. Y. Niyat, M. H. Moattar, and M. N. Torshiz, ‘‘Color image
10970-5. encryption based on hybrid hyper-chaotic system and cellular
[44] A. Kumar and M. Dua, ‘‘A GRU and chaos-based novel image encryption automata,’’ Opt. Lasers Eng., vol. 90, pp. 225–237, Mar. 2017, doi:
approach for transport images,’’ Multimed. Tools Appl., vol. 82, pp. 1–28, 10.1016/j.optlaseng.2016.10.019.
Oct. 2022, doi: 10.1007/s11042-022-13902-z. [64] H. Liu and A. Kadir, ‘‘Asymmetric color image encryption scheme using
[45] A. Belazi, A. A. Abd El-Latif, and S. Belghith, ‘‘A novel image 2D discrete-time map,’’ Signal Process., vol. 113, pp. 104–112, Aug. 2015,
encryption scheme based on substitution-permutation network and doi: 10.1016/j.sigpro.2015.01.016.
chaos,’’ Signal Process., vol. 128, pp. 155–170, Nov. 2016, doi: [65] H. Liu, A. Kadir, and Y. Niu, ‘‘Chaos-based color image block encryption
10.1016/j.sigpro.2016.03.021. scheme using S-box,’’ AEU Int. J. Electron. Commun., vol. 68, no. 7,
[46] M. Kaur and V. Kumar, ‘‘A comprehensive review on image encryption pp. 676–686, Jul. 2014, doi: 10.1016/j.aeue.2014.02.002.
techniques,’’ Arch. Comput. Methods Eng., vol. 27, no. 1, pp. 15–43, [66] J. Karmakar and M. K. Mandal, ‘‘Chaos-based image encryption
Jan. 2020, doi: 10.1007/S11831-018-9298-8. using integer wavelet transform,’’ in Proc. 7th Int. Conf.
[47] Q. Cun, X. Tong, Z. Wang, and M. Zhang, ‘‘A new chaotic Signal Process. Integr. Networks, Feb. 2020, pp. 756–760, doi:
image encryption algorithm based on dynamic DNA coding and 10.1109/SPIN48934.2020.9071316.
RNA computing,’’ Vis. Comput., vol. 2023, pp. 1–20, Jan. 2023, [67] Q. Yao et al., ‘‘Color image encryption based on discrete trinion Fourier
doi: 10.1007/S00371-022-02750-5. transform and random-multiresolution singular value decomposition,’’
[48] S. Bhowmik and S. Acharyya, ‘‘Image encryption approach using Multimedia Tools Appl., vol. 79, nos. 37–38, pp. 27555–27581, Oct. 2020,
improved chaotic system incorporated with differential evolution and doi: 10.1007/S11042-020-09296-5.
genetic algorithm,’’ J. Inf. Secur. Appl., vol. 72, Feb. 2023, Art. no. 103391, [68] M. Guan, X. Yang, and W. Hu, ‘‘Chaotic image encryption algorithm
doi: 10.1016/j.jisa.2022.103391. using frequency-domain DNA encoding,’’ IET Image Process.,
[49] D. Zareai, M. Balafar, and M. FeiziDerakhshi, ‘‘EGPIECLMAC: Effi- vol. 13, no. 9, pp. 1535–1539, Jul. 2019, doi: 10.1049/iet-ipr.2019.
cient grayscale privacy image encryption with chaos logistics maps 0051.
and Arnold Cat,’’ Evolving Syst., vol. 2023, pp. 1–31, Jan. 2023, [69] X. Wu, D. Wang, J. Kurths, and H. Kan, ‘‘A novel lossless color
doi: 10.1007/s12530-022-09482-w. image encryption scheme using 2D DWT and 6D hyperchaotic sys-
[50] S. Benaissi, N. Chikouche, and R. Hamza, ‘‘A novel image encryp- tem,’’ Inf. Sci., vols. 349–350, pp. 137–153, Jul. 2016, doi: 10.1016/j.ins.
tion algorithm based on hybrid chaotic maps using a key image,’’ 2016.02.041.
Optik, vol. 272, Feb. 2023, Art. no. 170316, doi: 10.1016/j.ijleo.2022. [70] H. Singh, A. K. Yadav, S. Vashisth, and K. Singh, ‘‘Optical image
170316. encryption using Devil’s vortex toroidal lens in the Fresnel transform
[51] J. Wen, X. Xu, K. Sun, Z. Jiang, and X. Wang, ‘‘Triple-image bit- domain,’’ Int. J. Opt., vol. 2015, pp. 1–13, Dec. 2015, doi: 10.1155/2015/
level encryption algorithm based on double cross 2D hyperchaotic 926135.
map,’’ Nonlinear Dyn., vol. 111, no. 7, pp. 6813–6838, Apr. 2023, doi: [71] E. A. Naeem, M. M. A. Elnaby, N. F. Soliman, A. M. Abbas,
10.1007/S11071-022-08158-Z. O. S. Faragallah, N. Semary, M. M. Hadhoud, S. A. Alshebeili, and
[52] H. Zang, M. Tai, and X. Wei, ‘‘Image encryption schemes based on a class F. E. A. El-Samie, ‘‘Efficient implementation of chaotic image encryption
of uniformly distributed chaotic systems,’’ Mathematics, vol. 10, no. 7, in transform domains,’’ J. Syst. Softw., vol. 97, pp. 118–127, Nov. 2014,
p. 1027, Mar. 2022, doi: 10.3390/MATH10071027. doi: 10.1016/j.jss.2014.07.026.
[53] K. M. Hosny, S. T. Kamal, and M. M. Darwish, ‘‘Novel encryption for [72] K. M. Hosny, M. A. Zaki, N. A. Lashin, and H. M. Hamza,
color images using fractional-order hyperchaotic system,’’ J. Ambient ‘‘Fast colored video encryption using block scrambling and multi-
Intell. Humaniz. Comput., vol. 13, no. 2, pp. 973–988, Feb. 2022, doi: key generation,’’ Vis. Comput., vol. 2022, pp. 1–32, Nov. 2022,
10.1007/s12652-021-03675-y. doi: 10.1007/s00371-022-02711-y.

63054 VOLUME 11, 2023

K. M. Hosny et al.: Multimedia Security Using Encryption: A Survey

[73] K. M. Hosny, M. A. Zaki, H. M. Hamza, M. M. Fouda, and N. A. Lashin, [93] S. Adhikari and S. Karforma, ‘‘A novel audio encryption method
‘‘Privacy protection in surveillance videos using block scrambling-based using Henon–Tent chaotic pseudo-random number sequence,’’
encryption and DCNN-based face detection,’’ IEEE Access, vol. 10, Int. J. Inf. Technol., vol. 13, no. 4, pp. 1463–1471, Aug. 2021,
pp. 106750–106769, 2022, doi: 10.1109/ACCESS.2022.3211657. doi: 10.1007/s41870-021-00714-x.
[74] A. T. Maolood, E. K. Gbashi, and E. S. Mahmood, ‘‘Novel lightweight [94] I. Yasser, M. A. Mohamed, A. S. Samra, and F. Khalifa, ‘‘A chaotic-
video encryption method based on ChaCha20 stream cipher and hybrid based encryption/decryption framework for secure multimedia com-
chaotic map,’’ Int. J. Electr. Comput. Eng., vol. 12, no. 5, pp. 4988–5000, munications,’’ Entropy, vol. 22, no. 11, p. 1253, Nov. 2020, doi:
Oct. 2022, doi: 10.11591/ijece.v12i5.pp4988-5000. 10.3390/E22111253.
[75] J. Sethi, J. Bhaumik, and A. S. Chowdhury, Chaos-Based Uncompressed [95] F. Peng, X.-Q. Gong, M. Long, and X.-M. Sun, ‘‘A selective encryption
Frame Level Video Encryption. Singapore: Springer, 2022, pp. 201–217. scheme for protecting H.264/AVC video in multimedia social network,’’
[76] A. Hafsa, M. Fradi, A. Sghaier, J. Malek, and M. Machhout, ‘‘Real-time Multimedia Tools Appl., vol. 76, no. 3, pp. 3235–3253, 2017, doi:
video security system using chaos-improved advanced encryption standard 10.1007/s11042-016-3710-x.
(IAES),’’ Multimedia Tools Appl., vol. 81, no. 2, pp. 2275–2298, Jan. 2022, [96] D. Shah, T. Shah, and S. S. Jamal, ‘‘Digital audio signals encryption by
doi: 10.1007/s11042-021-11668-4. Mobius transformation and Hénon map,’’ Multimed. Syst., vol. 26, no. 2,
[77] J. Karmakar, A. Pathak, D. Nandi, and M. K. Mandal, ‘‘Sparse represen- pp. 235–245, Apr. 2020, doi: 10.1007/s00530-019-00640-w.
tation based compressive video encryption using hyper-chaos and DNA [97] N. Sasikaladevi, K. Geetha, and K. N. Venkata Srinivas, ‘‘A multi-tier secu-
coding,’’ Digit. Signal Process., vol. 117, Oct. 2021, Art. no. 103143, doi: rity system (SAIL) for protecting audio signals from malicious exploits,’’
10.1016/j.dsp.2021.103143. Int. J. Speech Technol., vol. 21, no. 2, pp. 319–332, Jun. 2018, doi:
[78] X. Li, H. Yu, H. Zhang, X. Jin, H. Sun, and J. Liu, ‘‘Video encryption based 10.1007/S10772-018-9510-0.
on hyperchaotic system,’’ Multimedia Tools Appl., vol. 79, nos. 33–34, [98] E. A. Albahrani, ‘‘A new audio encryption algorithm based on chaotic
pp. 23995–24011, Sep. 2020, doi: 10.1007/S11042-020-09200-1. block cipher,’’ in Proc. Annu. Conf. New Trends Inf. Commun. Technol.
[79] A. Alarifi, S. Sankar, T. Altameem, K. C. Jithin, M. Amoon, Appl. (NTICT), Jul. 2017, pp. 22–27, doi: 10.1109/NTICT.2017.7976129.
and W. El-Shafai, ‘‘A novel hybrid cryptosystem for secure stream- [99] J. B. Lima and E. F. Da Silva Neto, ‘‘Audio encryption based on the
ing of high efficiency H.265 compressed videos in IoT multimedia cosine number transform,’’ Multimedia Tools Appl., vol. 75, no. 14,
applications,’’ IEEE Access, vol. 8, pp. 128548–128573, 2020, doi: pp. 8403–8418, Jul. 2016, doi: 10.1007/S11042-015-2755-6.
10.1109/ACCESS.2020.3008644. [100] E. Hato and D. Shihab, ‘‘Lorenz and Rossler chaotic system for speech
[80] H. Elkamchouchi, W. M. Salama, and Y. Abouelseoud, ‘‘New video signal encryption,’’ Int. J. Comput. Appl., vol. 128, no. 11, pp. 25–33,
encryption schemes based on chaotic maps,’’ IET Image Process., vol. 14, Oct. 2015, doi: 10.5120/ijca2015906670.
no. 2, pp. 397–406, Feb. 2020, doi: 10.1049/iet-ipr.2018.5250. [101] S. P. James, S. N. George, and P. P. Deepthi, ‘‘An audio encryption
[81] R. R. kumar, D. Ganeshkumar, A. Suresh, and K. Manigandan, ‘‘A new technique based on LFSR based alternating step generator,’’ in Proc. IEEE
one round video encryption scheme based on 1D chaotic maps,’’ in Int. Conf. Electron., Comput. Commun. Technol. (CONECCT), Jan. 2014,
Proc. 5th Int. Conf. Adv. Comput. Commun. Syst. (ICACCS), Mar. 2019, pp. 1–6, doi: 10.1109/CONECCT.2014.6740185.
pp. 439–444, doi: 10.1109/ICACCS.2019.8728443. [102] G. Qi, H. Wang, M. Haner, C. Weng, S. Chen, and Z. Zhu, ‘‘Convolutional
[82] D. Valli and K. Ganesan, ‘‘Chaos based video encryption using maps and neural network based detection and judgment of environmental obstacle in
Ikeda time delay system,’’ Eur. Phys. J. Plus, vol. 132, no. 12, pp. 1–18, vehicle operation,’’ Wiley Online Libr., vol. 4, no. 2, pp. 80–91, Jun. 2019,
Dec. 2017, doi: 10.1140/epjp/i2017-11819-7. doi: 10.1049/trit.2018.1045.
[83] S. Kotel, M. Zeghid, A. Baganne, T. Saidani, I. Daradkeh, and T. Rached, [103] I. Lahsaini, M. El Habib Daho, and M. A. Chikh, ‘‘Deep transfer
’’FPGA-based real-time implementation of AES algorithm for video learning based classification model for COVID-19 using chest CT-
encryption,’’ Recent Adv. Telecommun., Inform. Educ. Technol., pp. 17–18, scans,’’ Pattern Recognit. Lett., vol. 152, pp. 122–128, Dec. 2021, doi:
2014. 10.1016/j.patrec.2021.08.035.
[84] O. S. Faragallah, H. S. El-Sayed, and W. El-Shafai, ‘‘Efficient opto [104] Y. Tingting, W. Junqian, W. Lintai, and X. Yong, ‘‘Three-stage net-
MVC/HEVC cybersecurity framework based on Arnold map and discrete work for age estimation,’’ CAAI Trans. Intell. Technol., vol. 4, no. 2,
cosine transform,’’ J. Ambient Intell. Humans. Comput., vol. 1, pp. 1–16, pp. 122–126, Jun. 2019, doi: 10.1049/trit.2019.0017.
Jul. 2021, doi: 10.1007/s12652-021-03382-8. [105] H. S. Basavegowda and G. Dagnew, ‘‘Deep learning approach for
[85] M. Huang, C. Yang, and Y. Zhang, ‘‘Selective encryption of microarray cancer data classification,’’ CAAI Trans. Intell. Technol., vol. 5,
H.264/AVC based on block weight model,’’ in Proc. IEEE 18th Int. no. 1, pp. 22–33, Mar. 2020, doi: 10.1049/trit.2019.0028.
Conf. Commun. Technol. (ICCT), Oct. 2018, pp. 1368–1373, doi: [106] A. I. Awad, M. M. Fouda, M. M. Khashaba, E. R. Mohamed,
10.1109/ICCT.2018.8599959. and K. M. Hosny, ‘‘Utilization of mobile edge computing on the
[86] A. I. Sallam, E.-S. M. El-Rabaie, and O. S. Faragallah, ‘‘Efficient HEVC Internet of Medical Things: A survey,’’ ICT Exp., May 2022, doi:
selective stream encryption using chaotic logistic map,’’ Multimedia Syst., 10.1016/J.ICTE.2022.05.006.
vol. 24, no. 4, pp. 419–437, 2018, doi: 10.1007/s00530-017-0568-3. [107] Z. Feixiang, L. Mingzhe, W. Kun, and Z. Hong, ‘‘Color image encryption
[87] M. Altaf, A. Ahmad, F. A. Khan, Z. Uddin, and X. Yang, ‘‘Computation- via Hénon–Zigzag map and chaotic restricted Boltzmann machine over
ally efficient selective video encryption with chaos-based block cipher,’’ blockchain,’’ Opt. Laser Technol., vol. 135, Mar. 2021, Art. no. 106610,
Multimedia Tools Appl., vol. 77, no. 21, pp. 27981–27995, Nov. 2018, doi: doi: 10.1016/J.OPTLASTEC.2020.106610.
10.1007/s11042-018-6022-5.
[88] Y. Tew, K. Wong, and R. C.-W. Phan, ‘‘Region-of-interest
encryption in HEVC compressed video,’’ in Proc. IEEE Int.
Conf. Consum. Electron.-Taiwan (ICCE-TW), May 2016, pp. 1–2,
doi: 10.1109/ICCE-TW.2016.7520709.
[89] Y. Tew, K. Wong, and R. C.-W. Phan, ‘‘Joint selective encryption and data KHALID M. HOSNY (Senior Member, IEEE) was
embedding technique in HEVC video,’’ in Proc. Asia–Pacific Signal Inf. born in Zagazig, Egypt, in 1966. He received the
Process. Assoc. Annu. Summit Conf. (APSIPA), Dec. 2016, pp. 1–5, doi: B.Sc., M.Sc., and Ph.D. degrees from Zagazig
10.1109/APSIPA.2016.7820912. University, Egypt, in 1988, 1994, and 2000. He is
[90] M. N. Asghar, M. Ghanbari, M. Fleury, and M. J. Reed, ‘‘Sufficient currently a Professor in information technology
encryption based on entropy coding syntax elements of H.264/SVC,’’ with the Faculty of Computers and Informatics,
Multimedia Tools Appl., vol. 74, no. 23, pp. 10215–10241, 2015,
Zagazig University. From 1997 to 1999, he was
doi: 10.1007/s11042-014-2160-6.
[91] S. S. Giradkar and A. Bhattacharya, ‘‘Securing compressed video a Visiting Scholar with the University of Michi-
streams using RC4 encryption scheme,’’ in Proc. Global Conf. Com- gan, Ann Arbor, and the University of Cincinnati,
mun. Technol. (GCCT), Apr. 2015, pp. 640–644, doi: 10.1109/GCCT.2015. Cincinnati, USA. He is a Senior Member of ACM.
7342740. He has published four edited books and more than 140 papers in inter-
[92] H. Shen, L. Zhuo, and Y. Zhao, ‘‘An efficient motion reference struc- national journals. His research interests include image processing, pattern
ture based selective encryption algorithm for H.264 videos,’’ IET Inf. recognition, multimedia, and computer vision. He is an editor and a scientific
Secur., vol. 8, no. 3, pp. 199–206, May 2014, doi: 10.1049/iet-ifs.2012. reviewer of more than 70 international journals. He was among the top 2%
0349. of scientists according to the Stanford ranking (2020–2022).

VOLUME 11, 2023 63055

 K. M. Hosny et al.: Multimedia Security Using Encryption: A Survey

MOHAMED A. ZAKI was born in Abu Kabir, MOSTAFA M. FOUDA (Senior Member, IEEE)
Egypt, in 1995. He received the B.Sc. degree in received the Ph.D. degree in information sciences
information technology from the Faculty of Com- from Tohoku University, Japan, in 2011. He is cur-
puters and Informatics, Zagazig University, Egypt, rently an Assistant Professor with the Department
in 2017. He is currently a Teaching Assistant with of Electrical and Computer Engineering, Idaho
the Information Technology Department, Faculty State University, ID, USA. He also holds the posi-
of Computers and Informatics, Zagazig University. tion of a Full Professor with Benha University,
Egypt. His research interests include cyber secu-
rity, machine learning, the IoT, and 6G networks.
He has served on the technical committees for sev-
eral IEEE conferences. He is also a reviewer in several IEEE TRANSACTIONS
and magazines. He is an Editor of IEEE TRANSACTIONS ON VEHICULAR
TECHNOLOGY and an Associate Editor of IEEE ACCESS.

NABIL A. LASHIN received the B.Sc. degree in HANAA M. HAMZA received the B.Sc., M.Sc.,
communication and electronics engineering from and Ph.D. degrees from the Faculty of Computers
Zagazig University, Egypt, in 1993, the M.Sc. and Informatics, Zagazig University, Egypt. She is
degree in communication and electronics engi- currently a Lecturer with the Faculty of Computers
neering from Cairo University, in 1999, and the and Informatics, Zagazig University.
Ph.D. degree in electrical engineering and com-
puter science from the Technical University of
Berlin, Germany, in 2005. He is currently an Assis-
tant Professor in information technology with
Zagazig University.

63056 VOLUME 11, 2023