Professional Documents
Culture Documents
Article
A Finite State Machine-Based Improved
Cryptographic Technique
Mohammad Mazyad Hazzazi 1, * , Raja Rao Budaraju 2 , Zaid Bassfar 3 , Ashwag Albakri 4
and Sanjay Mishra 5
1 Department of Mathematics, College of Science, King Khalid University, Abha 61421, Saudi Arabia
2 Senior Member of Technical Staff, Oracle, 3990 Scottfield Street, Dublin, CA 94568, USA
3 Department of Information Technology, University of Tabuk, Tabuk 71491, Saudi Arabia
4 Department of Computer Science, College of Computer Science & Information Technology,
Jazan 45142, Saudi Arabia
5 Department of USCS, Uttaranchal University, Dehradun 248007, India
* Correspondence: mmhazzazi@kku.edu.sa
Abstract: With the advent of several new means of communication, safeguarding the confidentiality
of messages has become more crucial. Financial institutions, virtual currencies, and government
organizations are all examples of high-risk contexts where information exchanges need particular
care. The importance of data security in preventing unauthorized access to data is emphasized.
Several cryptographic methods for protecting the secrecy and integrity of data were compared. In
this research, the proposed work includes a new Turbo Code-based encryption algorithm. The
Turbo encoder’s puncturing process is controlled by a secret key, and a typical random sequence
is generated to encrypt the data and fix any mistakes. Key generation utilizing pre-existing data
eliminates the requirement for sending keys over a secure channel. Using recurrence relations and
the Lower–Upper (LU) decomposition method, the presented study suggests a novel approach to
message encryption and decryption. The resulting encrypted grayscale image has a very high level
of security, with an entropy of 7.999, a variation from perfection of 0.0245, and a correlation of 0.0092
along the diagonal, 0.0009 along the horizontal, and −0.0015 along the vertical. Directly decrypted
Citation: Hazzazi, M.M.; Budaraju,
pictures have a Peak Signal-to-Noise Ratio (PSNR) of 56.22 dB, but the suggested approach only
R.R.; Bassfar, Z.; Albakri, A.; Mishra, manages an embedding capacity of 0.5 bpp (bits per pixel). This may be achieved by decreasing the
S. A Finite State Machine-Based size of the location map by only 0.02 bpp.
Improved Cryptographic Technique.
Mathematics 2023, 11, 2225. Keywords: Modified Cryptographic Turbo Code; Peak Signal-to-Noise Ratio; horizontal correlation;
https://doi.org/10.3390/ Finite State Machine
math11102225
MSC: 94A60
Academic Editors: Raúl M. Falcón
and Lingfeng Liu
Figure 1. Basic Data Hiding Technique.
Figure 1. Basic Data Hiding Technique.
The channel is used to send incoming messages to the source encoder with the best
feasible data rate and accuracy [3]. The first encoder may read the symbols as code words.
There is a unique coded message associated with every symbol. Before a source decoder
can convert the output of a binary channel decoder back into a symbol sequence, it must
first perform the inverse mapping of the source encoder’s code words. Incorrect binary
sequences may have been sent due to interference in the signal’s path to its destination.
Channel coding [4] methods are useful for eliminating this kind of mistake. Channel
encoding involves an encoder at the transmitter end adding additional bits to the output
of the source encoder using an al-algorithm. Information encoded in channel 1 might be
decoded by a channel coding block decoder at the receiving end. Many error-correcting
methods are used in channel encoding to ensure that the original bit sequence may be
recovered in its entirety.
Copyright protection and preventing malicious manipulation are the most pressing
and difficult issues. Research suggests that keeping secrets might help alleviate these
sorts of concerns. Covert data storage in media may be useful for a variety of purposes,
such as facilitating the identification of images, safeguarding intellectual property, and
regulating access. The integrity of the message would be compromised if information were
distorted in a way that was not immediately obvious to the human eye. When dealing
with overly sensitive information or medical imaging, even a little bit of inaccuracy might
have serious consequences. Reverse Data Hiding (RDH) [5] provides information on how
to inject data into digital media without any quality degradation. After data has been
extracted from digital media, the whole storage device may be retrieved. The owner of the
picture has little reason to trust the cloud service providers, since they will almost certainly
encrypt the shot before uploading it. The picture restoration and analysis processes must
be carefully separated for usage in medical and defense applications. The original photo
should be encrypted using the owner’s public key. The data hider uses a data-hiding key
to include the information into the encrypted signal. After receiving the information, it
may be decoded from the image. If the sender and receiver both have the data-hiding key,
the recipient may decode the sent signal without knowing what is in the encrypted data.
Therefore, even if the re-receiver has the cypher [6], they cannot decode the data in the
image without also obtaining the corresponding cypher. His cover photo might be obtained
via public decoding of the encrypted signal he was allegedly sent.
The use of the Modified Cryptographic Turbo Code Finite State Machine (MCTC-
FSM) is expanding into new domains, such as privacy protection [7], encrypted image
authentication, and authorship verification. One major drawback of current methods is
Mathematics 2023, 11, 2225 3 of 25
that the encoded information can only be accessed either before or after the image has been
decoded. Thus, the MCTC-FSM method has been validated as a practical choice for both
secret extraction and image recovery. Most TC methods may be an improvement over what
has come before, but none of them can ensure complete success when it comes to extracting
data or restoring images. Due to their prohibitive computational cost, slow embedding rate,
and inadequate security [8], these approaches are not well-suited for usage in real-time
applications. The proposed work includes a ground-breaking MCTC-FSM technique with
robust security, complete cover and secret reversibility, and a large embedding capacity
at a reasonable cost. In this research, we provide a unique MCTC-FSM method based
on two-level embedding. The cypher text may be incorporated into the cypher image
using this method, even if the cover and secret are unknown. Only those in possession
of the corresponding decryption keys will be able to read the encrypted information. The
suggested methodology improves the quality of both embedding and encryption, with
correlations tending towards zero and histogram values becoming uniform. The final
image should have an SSIM of 1, a BER of 0, and a PSNR of infinity.
1.1. Motivation
Error-control coding is a method used to improve a channel’s resistance to interference.
The proper unity between mathematical modelling and efficient algorithms, and more
critically the realization of this principle in actual systems, is essential for data transfer that
is both quick and secure. To code channels in a way that comes close to the Shannon limit,
Turbo Code (TC) is a major contender. The bit error rate (BER) performance of currently
available Turbo Codes is vastly enhanced, either in the waterfall zone or the error-floor
region, or both. Yet, the flattening effect is an issue for most forward-correcting codes. Yet,
the development of communication technology calls for an overly complex coding scheme
with great potential to address this problem. As a result, the development of better TC
designs has been a hot topic in the scientific community during the last several decades.
However, from the early nineteenth century, a large amount of study has been con-
ducted in the field of computing intelligence approaches. Scientists have successfully
implemented these algorithms to improve the efficiency of modern communication net-
works. Existing linear programming methodologies, such as Pentagrid’s method, Bellman’s
theory, etc., cannot simplify the complexities of real-world engineering challenges and
cannot produce stable optimum values for multi-optima-based rough surfaces. Considering
this, the development of strong meta-heuristic algorithms inspired by nature and their use
in a wide range of engineering issues has arisen as a new and significant topic of study.
The literature has sparked some fresh ideas for creating more effective TC. As adding
another dimension to a Turbo Code (TC) significantly boosts the system’s BER perfor-
mance, this study tries to build both enhanced Three-Dimensional (3D) and unique Four-
Dimensional (4D) TCs. A new modulation method called Superposition Modulation (SM)
has been included into this setting. In addition, the unique Simple Augmented State
Diagram (SASD) method has been used to assess the suggested 4D-performance TCs. More-
over, utilizing the finite length rule and the asymptotic spectral shape function, an in-depth
investigation of the asymptotic behavior of the innovative hybrid interleave-based 4D-TC
has been performed. By using the time-slicing and interleave-gain technique, we have
also analyzed the worst-case upper limit of the suggested structure. Improved versions of
previously developed algorithms inspired by nature have also been successfully utilized
in the effective design of the TC. Lastly, the suggested TCs’ BER performances based on
existing communication standards such as WiMAX and WIFI are discussed.
1.2. Significance
During the last several decades, the efficient design of Turbo Codes (TCs) for cutting-
edge communication systems has been a hotbed of innovative study. Scientists have
proposed numerous helpful methods, which can be roughly divided into five categories:
TC structural analysis, TC ensemble analysis, TC energy allocation strategies, TC design
Mathematics 2023, 11, 2225 4 of 25
2. Literature Survey
Steganography [9] entails encrypting secret information inside a seemingly benign
picture. Heavy embedding, on the other hand, drastically alters the cover picture, which
flags the presence of embedded data. Thus, steganography is typically insufficient for
securing massive volumes of disguised data. Data security may be improved by combining
encryption with other methods, such as steganography. Encryption relies on a mathematical
function to encrypt data, making it unreadable by anyone except the intended recipient.
The data is encrypted, but it may be snooped on if it is transferred through a wireless
channel. When combined with encryption, steganography is one of the most effective
techniques to conceal sensitive data [10]. To be effective, software must be able to restore
both the secret information and the cover media without tampering with them.
Reversible Data Hiding (RDH) may be used to decode and decrypt both the cover
picture and the secret data since it is a lossless embedding mechanism. Researchers are now
focusing on RDH in encrypted domains, also known as RDH in Encrypted Domain (RDH-
ED). In the RDH-ED framework, there are two categories, separable and non-separable,
which reflect whether the hidden information can be recovered. Independent RDH-ED
allows for data decryption without disclosing any sensitive keys. Nevertheless, in the case
of non-separable RDH-ED, the cover picture must be decrypted before the hidden data
can be accessed. Puech initially presented the non-separable RDH-ED cypher based on the
original cover art’s Advanced Encryption Standard (RDH) [11] encryption.
Before decrypting the encrypted image and analyzing its local standard deviation for
the hidden image, we first mark it. Obtaining the secret information is impossible without
first decoding the cover art, making this technique inseparable from the cover art itself.
This technique is simple to develop; however, it has a low embedding capacity of only
0.0625 bpp (bits per pixel).
Mathematics 2023, 11, 2225 5 of 25
Zhang proposes a another RDH-ED approach in [12] that includes encrypting the
original cover picture using an exclusive-or (XOR) operation. By carefully examining each
individual component, we were able to put together the whole image. A line was drawn
through the middle of each cube to separate the spaces. To conceal the information, we
need to switch the three Least Significant Bits (LSB) of each pixel in a particular section of
the block. This technique may simultaneously reveal the cover art and the secret data to
the recipient. If the block size is off, extracting data or restoring images will fail. Using
a new estimate equation and side-matching algorithm, we enhance embedding capacity
and drastically reduce the error rate between neighboring blocks in approach [13]. Because
data recovery requires decryption of the picture, this technique is a non-separable RDH-ED,
developing differentiated histograms of data and adjusting preexisting ones [14].
The security and dependability of a system are compromised when error-correcting
codes are used with other security mechanisms. The necessity to enhance the current
ME technique [15] with multi-bit error correction and a larger embedding capacity and
reversibility inspired the proposed study.
Further secret information may be added to a steganographic [16] picture by giving
it the illusion of encryption. The cover image may be reversed if you reveal the hidden
words on the FSM’s map of its whereabouts. The cover image’s components have new
homes after being embedded, and the map showing where they will be kept is called
the location map. At the receiver, a position map may be extracted from the codewords
to reconstruct the parts of the cover picture that were untouched by the changes made
during embedding. Integrating ME with FSM, which employs a cryptographic FSM, would
allow for the embedding of secret data and a cover picture with improved quality, security,
and reliability.
Table 1. Cont.
The suggested Cryptographic Turbo Code is used to encrypt and encode the stegano-
graphic image (an embedded cover picture). It is recommended in the Modified Crypto-
graphic Turbo Code (MCTC) that a secret-key FSM should replace the regular FSM, thereby
making it a variant of the FSM. With the secret key, a secret-key FSM encrypts the data and
corrects transmission problems by shuffling the bits of the input into a new sequence [29].
As an alternative, we suggest using a cryptographic FSM based on elliptic curves to produce
random bitstream noise (MCTC-FSM) [24]. The sensitive data is protected on two fronts by
the joint efforts of the Modified Minimum Encryption and Minimal Common Transceiver
Complexity. This may be achieved by using a steganographic image to covertly conceal the
information and then encrypting the whole thing [25]. The cover photo’s security and em-
beddability are both improved by using a steganographic picture for encryption. Recovery
quality suffers as embedding capacity rises because more advanced embedding methods
change more of the cover image’s bits. To get around this issue, the MCTC authors suggest
relaying the in-codeword locations of the changed bits through syndrome embedding.
Hence, the receiver can decode both the altered bits of the cover image and the hidden
data included inside the steganographic picture. Data embedding might be conducted
backwards [30].
Due to their reliance on a physical map, current RDH-ED systems can locate the appro-
priate cover image, but they are unable to insert it. To fill up the discovered knowledge gap,
it will be necessary to enhance embedding capabilities via reliable recovery of the cover
art. When it comes to correcting bugs, Turbo code is unparalleled. On the downside, it
does not safeguard information during transmission. The bit error rate could be decreased
by using TC in conjunction with encryption. The present MCTC-FSM is quite effective
at detecting and repairing errors. Yet, they might be built with cryptographically safe
encryption and portable data storage in mind. There is a known void in the literature about
the development of a cryptographically sound FSM that can create a random interleaving
pattern at an affordable computing cost.
3. Proposed Methodology
The classic method accomplishes security and reliability via a cascade implementation
of the traditional cryptography system and the error-correcting coding scheme. To ensure
the safety of the sent data, it is first encrypted and then the transmission channels are
encoded. Information received must be decoded and encrypted at the receiving end before
it can be used to reconstruct the original message. The Advanced Encryption Standard
(AES) is an algorithm for secure data transmission that has been recommended for use.
Mathematics 2023, 11, 2225 7 of 25
For the goal of data security, it is often used in systems with constrained means, such
as satellites.
The performance of a convolution coding unit connected in parallel mode can be
assessed from the Input Redundancy Weight Enumerating Function (IRWEF) of the con-
verged constituent recursive convolution code C as given by
where the coefficient ACv,t implies the coded strings number-devising parity check weight
spawned by an input bit stream of weight w. Here, W and Z suggest two dummy variables.
The Conditional Weight Enumerating Functions (CWEFs) Aw,C1 (w, Z ) and Aw,C2 (w, Z )
of the encoding modules C1 and C2 can be specified in Equation (2) and
Equation (3), respectively.
Aw,C1 (w, Z ) = ∑ Aw,Z
C1
Zz (2)
z
∑ Aw2
c Z
Aw,C2 (w, Z ) = Zz (3)
z
C
Here, Aw,z1
and AC 2
w,s signify the coefficients of the weight enumerating function of the
encoding modules C1 and C2 , respectively. These coefficients denote coded strings with
parity check weight z produced by input bit string of weight w. Now, by picking up the
appropriate value of λ which is the input factor and N is the pixel range, the fractional
part of the parity sequences of length λN and λK from C1 and C2 are amassed by a P/S
multiplexing unit to create the coded bit sequence of length λ( N + K ). Therefore, the
CWEF [17] of the multiplexing unit which accrues the fractional sequence approaching
from convolution encoders connected in a parallel fashion and detached by an interleaver
of size l1 is specified by:
The authenticity of Equation (4) has been assessed by considering the following norm:
It states that the interleaver π1 of length l1 should map a set of incoming
bits of weight w
l1
into all of its dissimilar P permutation with identical probability . Now, the output
w
of the multiplexing unit is permuted by the interleaving unit π2 with a length of λ( N + K ).
Therefore, the CWEF of the interleaving unit π2 is given by Equation (5).
Where w5 and w6 characterize the input weights of the tracks P2 and P5 , respectively,
which are connected to the puncturing module. In this context, wa and wb represent output
weights of the bit strings coming from the puncturing unit. The weight of the output code
sequence impending from the post-encoding unit is signified by (h − w − wa − wb ) = w3 .
00
Additionally, w10 and w1 characterize the input weights of the tracks P3 and P4 , respectively,
which are connected to the puncturing module.
The straitened higher bound on the BEP for the ML soft decoding technique under the
Additive White Gaussian Noise (AWGN) channel condition can be evaluated by the Input
Output Weight Enumerator Function (IOWEF) coefficients Bw,k s . The straitened higher
where RC signifies the code rate. A flowchart of the proposed work is shown in Figure 2.
1 𝐸
𝑃 𝑤 ∗ 𝑤 ∗𝐵 , ∗𝑄 2∗𝑅 ∗ ∗ℎ∗ (5)
𝑁 𝑁
Figure 2. Flowchart of the proposed work.
Figure 2. Flowchart of the proposed work.
To provide security and dependability, RDH and the FSM combine to provide excel-
To provide security and dependability, RDH and the FSM combine to provide excellent
lent security but with a drop in dependability compared to the original implementation
security but with a drop in dependability compared to the original implementation of FSM.
of FSM. As the RDH-128, RDH-192, and RDH-256 all use keys of length 128, 192, or 256
As the RDH-128, RDH-192, and RDH-256 all use keys of length 128, 192, or 256 bits to
bits to encrypt and decode 128-bit data blocks, this makes sense. This limits the 4-FSM’s
encrypt and decode 128-bit data blocks, this makes sense. This limits the 4-FSM’s error-
error-correcting capability by limiting the maximum frame length of the FSM to 128 bits.
correcting capability by limiting the maximum frame length of the FSM to 128 bits. A
A longer input bit sequence may improve the MCTC-error-correcting FSM’s performance.
longer input bit sequence may improve the MCTC-error-correcting FSM’s performance.
Nevertheless, encoding a bit sequence longer than 128 bits necessitates many iterations of
Nevertheless, encoding a bit sequence longer than 128 bits necessitates many iterations of
MCTC, which increases the time and energy needed to process the data.
MCTC, which increases the time and energy needed to process the data.
In this scheme, the m length-based input sequence is passed through the Serial-to-
Parallel (S/P) unit. Thereafter, symbols impending from the S/P unit are moved through
the Binary Phase-Shift Keying (BPSK) segment. Next, the amplitudes are assigned to each
of these symbols. Finally, these symbols are linearly superimposed to generate a finite-
alphabet output symbol Y. This whole modulation arrangement can be mathematically
articulated by the expression
m1 m1 m1
Y= ∑ Ci1 = ∑ β i1 di1 = ∑ β i1 (1 − 2bi1 ) (6)
i =1 i =1 i =1
where β i1 indicates the magnitude of the i1th binary chip and bi1 ∈ {0, 1}. Three power
distribution techniques, namely Equal Power Allocation (EPA), Unequal Power Allocation
(UPA), and Grouped Power Allocation (GPA), have been successfully used for the design
of the efficient 3D-TC.
Instead, one might modify the existing FSM implementation in such a manner that it
provides security and error correction at a reduced computational cost. The only function
that the conventional FSM can achieve is to correct errors in the data during transmission;
this is a limitation. However, the random shuffling that the random FSM conducts may be
used for encryption as well, so it is not only for fixing mistakes.
g0 ( E ) = 1 + E 2 + E 3 g1 ( E ) = 1 + E + E 3 (8)
Figure 1 illustrates the shift register implementation of the transfer function of each
RSC encoder. The delay in the shift register is denoted by the E in this equation.
An MCTC sits in between the two RSC encoders to provide separation. The input bit
sequence A, which is delivered to the MCTC, is given to the first RSC encoder, which then
encodes the sequence.
It is sufficient to send either the A or the B systematic bits since they are both equiva-
lently represented by scrambled copies of the other. However, to prevent the data from
being stolen, the MCTC will send out systematic bits, B. In addition, to prevent the confi-
dential data from being guessed by a user who is not permitted to access them, the code
word of the first encoder is interleaved with MCTC-2 to produce the letter E. The MCTC’s
code word is referred to as MCTC, and it is provided as:
CCTC = { B, E, F } (9)
The term “code rate” refers to the proportion of the number of information bits to the
number of encoded bits. The amount of information bits included in the MCTC is L, while
the number of parity bits is 2L. This combination yields a coding rate of:
M 1
RCTC = = (10)
Mathematics 2023, 11, x FOR PEER REVIEW M+M+M 3M 10 of 26
On the other hand, this results in an increase in the system’s bandwidth. An increase
in the code rate and a decrease in the code’s bandwidth cost may be achieved by puncturing
puncturing the parity bits or removing some of them. A decrease in the performance of
the parity bits or removing some of them. A decrease in the performance of the code’s
the code’s error-correcting mechanism is brought on by puncturing; however, a compe-
error-correcting mechanism is brought on by puncturing; however, a competent puncturing
tent puncturing technique may keep this decrease in performance to a minimum. In the
technique may keep this decrease in performance to a minimum. In the field of cellular
field of cellular transmission, the rate of 1/2-pierced FSM is used. The secret word for the
transmission, the rate of 1/2-pierced FSM is used. The secret word for the rate of 1/2 MCTC
rate of 1/2 MCTC is as shown in Figure 3.
is as shown in Figure 3.
Figure 3. Cryptographic Turbo Code (CTC) encoder using MCTC FSM.
Figure 3. Cryptographic Turbo Code (CTC) encoder using MCTC FSM.
When it is time to decode the information, the decoder will substitute false bits with
When it is time to decode the information, the decoder will substitute false bits with
aa value of zero for the bits that have been punctured. This will allow the information to be
value of zero for the bits that have been punctured. This will allow the information to
decoded. When the RSC encoders have completed encoding all the bits in the input bit
be decoded. When the RSC encoders have completed encoding all the bits in the input bit
sequence, they must reset to an all-zero state before going on to the next input sequence.
sequence, they must reset to an all-zero state before going on to the next input sequence.
This ensures that the encoders are ready for the next input sequence. To reset all the shift
This ensures that the encoders are ready for the next input sequence. To reset all the shift
registers to their original state, which is all zeroes, the input bit sequence is extended with
what are known as termination bits, which are also called tail bits. Termination bits are
introduced into a code at the end of each code word. This is performed to ensure that the
error-correcting performance of the code does not deteriorate because of the introduction
of the termination bits.
When it is time to decode the information, the decoder will substitute false bits with
a value of zero for the bits that have been punctured. This will allow the information to be
decoded. When the RSC encoders have completed encoding all the bits in the input bit
sequence, they must reset to an all-zero state before going on to the next input sequence.
This ensures that the encoders are ready for the next input sequence. To reset all the shift
Mathematics 2023, 11, 2225 10 of 25
registers to their original state, which is all zeroes, the input bit sequence is extended with
what are known as termination bits, which are also called tail bits. Termination bits are
introduced into a code at the end of each code word. This is performed to ensure that the
registers to their original state, which is all zeroes, the input bit sequence is extended with
error-correcting performance of the code does not deteriorate because of the introduction
what are known as termination bits, which are also called tail bits. Termination bits are
of the termination bits.
introduced into a code at the end of each code word. This is performed to ensure that the
3.3. Decryption of The Cryptographic Turbo Code
error-correcting performance of the code does not deteriorate because of the introduction
of theFigure 4 provides a visual representation of the structure of a CTC decoder. The noisy
termination bits.
bits that are received are “C-C-T-C,” which equals “B-E-F.” The MCTC de-FSM first per-
3.3. Decryption of The Cryptographic Turbo Code
forms a de-interleaving operation on the received bits B to produce an A using the keys
km, ks, P, and G that are shared by the sender and the receiver. This process is repeated
Figure 4 provides a visual representation of the structure of a CTC decoder. The
until the desired output is achieved.
noisy bits that are received are “C-C-T-C,” which equals “B-E-F.” The MCTC de-FSM first
The iterative procedure will continue until the estimates produced by the two decod-
performs a de-interleaving operation on the received bits B to produce an A using the keys
ers tend to converge on one another. The iterative decoding technique that is employed in
km, ks, P, and G that are shared by the sender and the receiver. This process is repeated
the TC enhances the performance of the error correction with each subsequent iteration.
until the desired output is achieved.
Figure 4. Cryptographic Turbo Code (CTC) decoder using MCTC de-FSM.
Figure 4. Cryptographic Turbo Code (CTC) decoder using MCTC de-FSM.
The iterative procedure will continue until the estimates produced by the two decoders
tend to converge on one another. The iterative decoding technique that is employed in the
TC enhances the performance of the error correction with each subsequent iteration.
Equation (11) shows the data-hiding text-using Finite State Machine with the Likeli-
hood Ratio and message.
Pr di = +1 | R2
L2ext and Lo (di ) = L1ext (di ) + L2ext (di ) + L a A− L
( di ) = log
Pr (di = −1 | R2 ) (12)
dˆt = 1 i f ( L o ( d i ) > 0)
(13)
dˆ1 = 0 i f ( Lo (di ) < 0)
If the receiver uses the incorrect key to the FSM, the decoders receive incorrect bit
sequences R1 and R2 , thereby recovering the incorrect information shown in Equation (12).
Y ( v ) = X ( u ), v = π ( u ) (14)
where (u) is the mapping function.
A secret number is used in conjunction with a complicated procedure on a computer
to encrypt and decode data. To prevent unauthorized parties from understanding sensitive
information sent over public networks, it is necessary to encrypt the message before
Mathematics 2023, 11, 2225 11 of 25
sending it. Only the receiver to whom they are addressed should be able to decode them.
A communication encrypted with a key should be decipherable only by the owner of the
same key.
In traditional TCs, the input data are interleaved with a well-known permutation
pattern that is intended to increase the efficiency of the coding process. This makes it
possible for a random FSM to be employed. However, to store the interleaving pattern,
the random FSM needs an additional memory and, to send data to the receiver, it needs a
wider bandwidth (Table 2).
are not permitted to do so. The interleaving sequence that is produced by the MCTC is
overly sensitive to the fact that the MCTC has a secret key. When the secret key is altered,
the interleaving pattern also undergoes considerable revision, which makes it impossible
for an unauthorized user to deduce the order in which the inputs will be sent.
Elliptic curve cryptography is a typical encryption procedure that offers a high degree
of protection to the data that is sent from a sender to a receiver over a public network.
This level of protection is provided by the elliptic curve. The elliptic curve equation must
be satisfied by a cloud of points for the elliptic curve arithmetic to be performed, since
this is the basis for elliptic curve encryption. The positive integers a and b satisfying the
curve equation that defines the points on the elliptic curve Ep(a, b) over a finite field Fp is
as follows:
y2 = (x3 + ax + b) mod p. (15)
The domain parameters, also known as the parameters of the elliptic curve, a, b, and
p, describe the point (x, y) on the elliptic curve in such a manner that both x and y are
contained inside a finite field, Fp. These domain parameters are also known as the elliptic
curve parameters.
This is the lowest positive integer that might possibly fulfil the requirements of the
condition. The security of the cryptosystem may be ensured by choosing the value of p that
will be a large prime integer. This will produce an elliptic curve of high order that contains
many points, which will satisfy the requirements for safety.
The creation of a top-secret lookup table is the first step in the implementation of the
MCTC that has been suggested. This step is subsequently followed by the development
of interleaving locations. Let us assume that the point G on the elliptic curve of prime
order Ep is a concealed generator (a, b). The most top-secret lookup table is created using
computational techniques
Qd = (km × Pd ), d = 0, 1, 2, 3, . . . , N, (17)
where Qd is the collection of elliptic points that are concealed inside the index t of the
lookup table.
The key km may have a single bit altered, which will cause a whole new lookup table
to be generated, as shown in Table 3. This protects the private lookup table from being read
by a third party who is not authorized to examine it in any capacity. Both the transmitter
and the receiver make use of two hidden parameters that they refer to as G and km . These
parameters are only known to the transmitter and the receiver. These parameters are
accountable for ensuring that the confidential lookup table’s privacy is always preserved.
In addition, even if the adversary is successful in learning one of the secret parameters, G
or km , the ECDLP that is used in elliptic curve arithmetic makes it exceedingly difficult for
them to uncover the other secret parameter.
The MCTC oversees creating the interleaved bit locations v for each input bit location
u that it receives, and it is accountable for doing so. The MCTC is the one responsible for
producing the interleaved bit location v, which serves as the index of the secret lookup
table. This index is associated with the elliptic point Pu , which can be determined using
the formula
Pu = (u × Q), u = 1, 2, . . . , N − 1
(18)
and Q = (ks × P),
where ks is the secret key and P is a secret elliptic curve point given to MCTC.
Mathematics 2023, 11, 2225 13 of 25
Table 3 provides an illustration of the case that may be used to explain the production
of an interleaved bit location for the mask key km = 2,853,234. Case 1 and case 2 demonstrate
the positions of the interleaved bits for the same secret key ks but for different secret points,
P. Even if the keys km and ks are the same, the interleaving sequence is completely different
because of a change in P, as can be seen in the results for cases 1 and 2. Cases 2 and
3 demonstrate that even though P remains unchanged, the generation of a totally new
interleaving sequence is brought about by a single bit shift in the secret key, ks . According
to the findings, the user must be in possession of practically all the secret key parameters to
successfully recover the right interleaving pattern. An entirely new sequence is generated
if there is a modification to even one of the parameters.
Every FSM has a matching de-FSM that works on the interleaved sequence to construct
the sequence in its natural order. This order is maintained throughout the process. To
de-interleave the sequence that has been interleaved, the receiver will need to construct an
interleaving sequence that is awfully close to the one that was sent. It is possible for the
receiver to produce a comparable interleaving sequence by secretly broadcasting the keys
“km , ks , G, P.” Because of this, there is no longer a need for additional bandwidth, which
was necessary to share the whole table between the sender and the receiver.
While using a secure system, an adversary is rendered incapable, or at best exceedingly
improbable, of regaining access to the original data. Because of this, one of the most
important criteria for a secure system is the ability to generate and transmit random data.
Since the random output is hard to anticipate, it makes it difficult for an adversary to launch
an attack on the system. Because the output that is created is more random, the data that is
collected is also more unexpected, which gives the system an increased level of safety.
The amplitude, direction, and form of the link between the input and output may be
seen with the use of an MCTC scatterplot. When looking at the correlation coefficient, a
result that is high is indicative of an extraordinarily strong link. A pattern in which there is
no obvious center and a practically null CC would be a great depiction of a random Finite
State Machine (FSM). Dispersion, on the other hand, is a statistical metric that is used to
evaluate how much of a stretch there is in a certain distribution. If the dispersion value is
made higher, the locations of the bits that are interleaved will become more spread apart.
NIST FIPS-140-2, the industry-standard statistical test suite, is used to provide further
verification on the proposed MCTC’s statistical characteristics. In the following, you will
find a detailed explanation of the different examinations.
A scatter plot may be used to illustrate the unpredictability of the output sequence. It
is desirable for the scatter plot to contain points that are distributed in a way that is both
random and consistent. If the scatter plot displays a pattern that can be predicted, the
parameters that were used to construct the sequence may be uncovered, which would leave
the cryptosystem vulnerable to attack.
The bit locations in the input and the bit positions in the output have absolutely no
link with one another, as is seen from the picture. Altering the values of ks also results in
a striking transformation of the pattern. Because of this, the scatterplot ensures that the
random numbers that are generated are one-of-a-kind for each user and are founded on
the MCTC keys that they have. The link between the MCTC’s input and output is one that
is overly sensitive to the way its key is configured. An opponent who does not possess
the relevant keys will be unable to discern the connection between the bits that are input
and bits that are output because of this. This time, the test was conducted between two
interleaved bit location sequences known as vi and vi+1 . Both sequences were produced by
the MCTC using the ith and (i + 1)th secret key and an interleaving length of 20,000 bits. The
interleaving sequence was generated with the use of an elliptic curve with the parameters
E19991(3,1), km = 49853, G = (52,17), and P = (62,2).
As a result of the points being evenly distributed around the graph, it is hard to draw
any conclusions about the nature of the cypher or their connection to one another that
are relevant.
This indicates that the association is not extraordinarily strong. Because of this, it is
exceedingly difficult for an unauthorized user to anticipate the interleaving pattern due to
the low value of the correlation (Table 4).
Table 4. Correlation coefficient between the input and the output bit locations of MCTC for different
values of ks .
The Quadrature Permutation Polynomial, also known as QPP; the random FSM,
sometimes known as RI; and the deterministic FSM, often known as DI, are all frequently
used by the TC. It was determined whether the MCTC, QPP, RI, and DI had all been
generated in a random fashion. In Figure 4, a comparison is made between the scatter
plots of the RI, QPP, DI, and MCTC. The output sequence can be seen in the figure because
QPP and DI are interleaved in such a way that makes it visible; as a result, the output
sequence is predictable and open to attack. However, the interleaved bit sequence may be
Mathematics 2023, 11, 2225 15 of 25
11, x FOR PEER REVIEW easily determined by an authorized user based on the original bit sequence, even
16 of 26 though
both the RI and the suggested MCTC demonstrate a large degree of unpredictability in the
pattern (Figure 5).
(b) QPP FSM
(a) Random FSM
(d) Proposed FSM
(c) Deterministic FSM
Figure 5. The relationship between the original bit locations and the interleaved bit locations for
Figure 5. The relationship between the original bit locations and the interleaved bit locations for both
both the present and planned FSM.
the present and planned FSM.
To calculate the normalized dispersion and spreading factor of the proposed MCTC,
To calculate the normalized dispersion and spreading factor of the proposed MCTC,
simulations were run using an input bit sequence of lengths 133, 499, and 2019, cor-
simulations were run using an input bit sequence of lengths 133, 499, and 2019, corre-
respondingly. This allowed for the determination of the normalized dispersion and
spondingly. This allowed for the determination of the normalized dispersion and spread-
spreading factor.
ing factor. According to the information shown in Table 5, the value of the RI standard deviation
According to the information shown in Table 5, the value of the RI standard deviation
is somewhere around 0.8. Because the suggested MCTC has a dispersion value that is close
to that of the RI, it is possible to use it as a method that interleaves bits in a way that is both
is somewhere around 0.8. Because the suggested MCTC has a dispersion value that is close
effective and efficient. In addition to this, it comes as quite a shock. Despite this, the DI
to that of the RI, it is possible to use it as a method that interleaves bits in a way that is
and the QPP both exhibit a little degree of dispersion to varying degrees. This is a direct
both effective and efficient. In addition to this, it comes as quite a shock. Despite this, the
consequence of the fact that the DI and QPP generate a predictable interleaving pattern for
DI and the QPP both exhibit a little degree of dispersion to varying degrees. This is a direct
the data they process (Table 6).
consequence of the fact that the DI and QPP generate a predictable interleaving pattern
for the data they process (Table 6).
Table 5. The dispersion of each of the FSMs is compared here.
L RI DI
Table 5. The dispersion of each of the FSMs is compared here. QPP MCTC
γ sp γ sp γ sp γ sp
133 L 0.8245 RI 2 0.0300 DI 10 QPP
0.0455 16 MCTC
0.8216 2
499 0.8130 γ 2 sp 0.0075
γ sp 18 γ 0.0210 sp 32 γ 0.8124 sp 2
2019 0.8132 2 0.0019 27 0.0103 64 0.8131 2
133 0.8245 2 0.0300 10 0.0455 16 0.8216 2
499 0.8130 2 0.0075 18 0.0210 32 0.8124 2
2019 0.8132 2 0.0019 27 0.0103 64 0.8131 2
Table 6. Comparison of dispersion for MCTC with different values of secret key 𝑘𝑠.
Mathematics 2023, 11, 2225 16 of 25
Table 6. Comparison of dispersion for MCTC with different values of secret key ks .
16 16
5000 i∑
Tp = aF (i )2 − 5000 (19)
=1
2. Poker test:
The bitstream that is going to be analyzed has been sectioned up into subcomponents
that each comprise four bits, considering that F(i) stands for the count of each of the
four-bit values.
3. Run test:
A run is the term used to describe an unbroken string of bits that are all the same.
With the run test, a tally is taken of the number of runs of ones and zeros of varied lengths
that are present in the input sequence. These runs may be of varying lengths. The value
of the test parameter known as Tr is determined by the number of ones and zeros that are
present in a certain bitstream.
4. Long run test:
Mathematics 2023, 11, 2225 17 of 25
The value of the test parameter Tl must be more than or equal to 26, and it is deter-
mined by counting the number of consecutive ones or zeros that are not broken up in any
way. If the bitstream does not include any runs of this sort, then the test may be regarded
to have been successful.
Estimates were made for the test parameters for each of the four tests based on
around one-hundred different binary sequences that were generated by the MCTC that
was recommended. During the generation process, a one-of-a-kind private key was used
to construct each binary sequence of length 20,000 utilizing that sequence. The binary
sequence was created by converting each of the random numbers that were produced by
the MCTC into a bit, bi, by utilizing the binary representation of the numbers that were
obtained by calculating
bi = 1, if ν > N/2
(20)
= 0, if ν < N/2
The results of three of the examples are shown in Table 7, which demonstrate that
the test value is located within the appropriate interval for each of the four statistical tests
required by the FIPS 140-2.
Required
Statistical Test Case 1 Case 2 Case 3
Interval
Mono bit test Input 1 9998 10,001 10,019
Poker test Input 2 10.9696 11.4560 15.4880
Example 1 Input 3 1785 1720 1470
Example 2 Input 4 1789 1789 1247
Long run test Input 5 1 2 3
Final Result success success success
of N, where N is an integer that is greater than zero. If all these conditions are met, then the
sequence of numbers has a period of N
νi = νi+N , f or i ≥ 0, (21)
where the letter “i” denotes a particular location inside the integer v. The periodicity of the
proposed MCTC is largely governed by several different parameters, the most significant of
which is the number of points on an elliptic curve that are created by the generator point G.
The suggested MCTC is mostly determined by these factors. It is possible to define
Mathematics 2023, 11, x FOR PEER REVIEW a cyclic
19 of 26
group using an elliptic curve of prime order. The building of a cyclic group results in the
production of all the points on an elliptic curve. This is accomplished by selecting a single
point from the cyclic group to serve as the beginning point in the process. One point from
selecting a single point from the cyclic group to serve as the beginning point in the process.
the collection may be chosen at random to serve as an example of this. An example of a
One point from the collection may be chosen at random to serve as an example of this. An
cyclic group is shown in each of the Figure 6a–c. These groups were generated by an elliptic
example of a cyclic group is shown in each of the Figure 6a–c. These groups were gener-
curve with the order 19 and the generator points (10,11), (7,6), and (16,4), respectively. Each
ated by an elliptic curve with the order 19 and the generator points (10,11), (7,6), and (16,4),
of these figures is shown in its own separate figure. An elliptic curve served as the basis
respectively. Each of these figures is shown in its own separate figure. An elliptic curve
for the construction of each of these groups. In the field of mathematics, an example of an
served as the basis for the construction of each of these groups. In the field of mathematics,
elliptic curve can be found denoted by the notation E17(2,2), which corresponds to this
an example of an elliptic curve can be found denoted by the notation E17(2,2), which cor-
cyclic group. The equation can generate all the points on the elliptic curve, even when
responds to this cyclic group. The equation can generate all the points on the elliptic curve,
the value of G is altered in some way. This is something that can be verified by personal
even when the value of G is altered in some way. This is something that can be verified by
observation.
personal The value
observation. ofvalue
The G, which is a variable
of G, which that canthat
is a variable be can
changed, is whatis determines
be changed, what
determines the specific sequence that is used to create the point each time. In addition, the
the specific sequence that is used to create the point each time. In addition, the point is
point is created by following this sequence.
created by following this sequence.
Figure 6. Comparison of computational cost of MCTC and proposed MCTC.
Figure 6. Comparison of computational cost of MCTC and proposed MCTC.
4.3. Bit Error Rate of Cryptographic Turbo Code
Bits that are sent across a digital communication system incur the danger of getting
corrupted if the channel through which they travel is impaired in some way, such as by
noise, distortion, interference, fading, or other similar phenomena. It is necessary for a
communication system to be able to recognize and rectify problems such as these for it to
Mathematics 2023, 11, 2225 19 of 25
Table 8. PSNR measurements taken from pictures recovered by CTC using both the right and
wrong keys.
Figure 7. BER performance of the uncoded MCTC, MCTC with TC, and proposed MCTC and TC.
Figure 7. BER performance of the uncoded MCTC, MCTC with TC, and proposed MCTC and TC.
The code’s total performance is heavily dependent on how well the MCTC algorithm
The BER plot of MCTC is shown in Figure 7; compared to TC, MCTC loses just 0.2 dB
is implemented. Many factors affect its performance in term of error correction: (i) The
of coding at a BER of 10-4, while gaining 0.8 dB when compared to MCTC-TC. The diagram
MCTC’s architecture; (ii) The number of decoding rounds conducted by the CTC decoder;
displays this data. The fact that MCTC’s random shuffling is controlled by a secret key also
(iii) the CTC’s frame length.
helps to ensure the privacy of any sensitive data it transmits.
Figure 7. BER performance of the uncoded MCTC, MCTC with TC, and proposed MCTC and TC.
We analyze how adjusting these parameters impacts the CTC’s performance. The bit
The code’s total performance is heavily dependent on how well the MCTC algorithm
error rate (BER) is calculated by encoding and decoding a random bit sequence of 65,536
is implemented. Many factors affect its performance in term of error correction: (i) The
The code’s total performance is heavily dependent on how well the MCTC algorithm
bits at a rate of 1/2 CTC, sending it via an AWGN channel, and then measuring the number
MCTC’s architecture; (ii) The number of decoding rounds conducted by the CTC decoder;
is implemented. Many factors affect its performance in term of error correction: (i) The
of errors in the decoded version eight times.
(iii) the CTC’s frame length.
MCTC’s architecture; (ii) The number of decoding rounds conducted by the CTC decoder;
(i) We Effect of the design of MCTC
analyze how adjusting these parameters impacts the CTC’s performance. The
(iii) the CTC’s frame length.
CTC analysis suggested MCTC for error correction. Figure 8 shows BER versus SNR
bit error rate (BER) is calculated by encoding and decoding a random bit sequence of
We analyze how adjusting these parameters impacts the CTC’s performance. The bit
scatter plots for each interval. CTC with MCTC has 104 BER at 0.4 dB SNR, 0.2 dB better
65,536 bits at a rate of 1/2 CTC, sending it via an AWGN channel, and then measuring the
error rate (BER) is calculated by encoding and decoding a random bit sequence of 65,536
than CTC with RI. Like the RI, the MCTC randomly interleaves with a dispersion of 0.8.
bits at a rate of 1/2 CTC, sending it via an AWGN channel, and then measuring the number
number of errors in the decoded version eight times.
Yet, deterministic FSMs such as the QPPI and SKI FSM use FSM spread to guarantee that
of errors in the decoded version eight times.
(i) Effect of the design of MCTC
permuted bits are no more than a specified distance apart. SKI created both FSMs. A quad-
(i) Effect of the design of MCTC
ratic permutation polynomial generates interleaved bit positions with a broad spread to
CTC analysis suggested MCTC for error correction. Figure 8 shows BER versus SNR
CTC analysis suggested MCTC for error correction. Figure 8 shows BER versus SNR
provide the QPPI error-correcting performance equivalent to the RI. Error correction is
scatter plots for each interval. CTC with MCTC has 104 BER at 0.4 dB SNR, 0.2 dB better
scatter plots for each interval. CTC with MCTC has 104 BER at 0.4 dB SNR, 0.2 dB better
achieved by transferring error patterns across decoders. Nevertheless, the SKI length and
than CTC with RI. Like the RI, the MCTC randomly interleaves with a dispersion of 0.8.
than CTC with RI. Like the RI, the MCTC randomly interleaves with a dispersion of 0.8.
key affect the spread between interleaved bit positions; thus, a big spread is not always
Yet, deterministic FSMs such as the QPPI and SKI FSM use FSM spread to guarantee that
Yet, deterministic FSMs such as the QPPI and SKI FSM use FSM spread to guarantee that
guaranteed. QPPI and SKI feature interleaved bit locations that can take several values,
permuted bits are no more than a specified distance apart. SKI created both FSMs. A quad-
permuted bits are no more than a specified distance apart. SKI created both FSMs. A
but their patterns are uniformly spaced, making them simple to predict and prone to se-
ratic permutation polynomial generates interleaved bit positions with a broad spread to
quadratic permutation polynomial generates interleaved bit positions with a broad spread
curity breaches. Correlations from the interleaved pattern’s regularity lower the BER
provide
to the QPPI error-correcting
provide the QPPI error-correcting performance equivalent to the RI. Error correction is
performance equivalent to the RI. Error correction is
curve’s error floor.
achieved by transferring error patterns across decoders. Nevertheless, the SKI length and
achieved by transferring error patterns across decoders. Nevertheless, the SKI length and
key affect the spread between interleaved bit positions; thus, a big spread is not always
key affect the spread between interleaved bit positions; thus, a big spread is not always
guaranteed. QPPI and SKI feature interleaved bit locations that can take several values,
guaranteed. QPPI and SKI feature interleaved bit locations that can take several values, but
but their patterns are uniformly spaced, making them simple to predict and prone to se-
their patterns are uniformly spaced, making them simple to predict and prone to security
curity breaches. Correlations from the interleaved pattern’s regularity lower the BER
breaches. Correlations from the interleaved pattern’s regularity lower the BER curve’s
curve’s error floor.
error floor.
Figure 8. Effect of FSM on the BER performance of CTC.
(ii) Effect of the number of decoding iterations of CTC decoder
CTC decoding uses iterative decoding to exchange data between the protocol’s two
decoder modules. Figure 9 shows a BER for the CTC’s efficiency during decoder iterations
Figure 8. Effect of FSM on the BER performance of CTC.
Figure 8. Effect of FSM on the BER performance of CTC.
(ii) Effect of the number of decoding iterations of CTC decoder
(ii) Effect of the number of decoding iterations of CTC decoder
CTC decoding uses iterative decoding to exchange data between the protocol’s two
decoder modules. Figure 9 shows a BER for the CTC’s efficiency during decoder iterations
Mathematics 2023, 11, x FOR PEER REVIEW 22 of 26
Figure 9. Complexity of iterative decoding grows with the number of iterations.
Figure 9. Complexity of iterative decoding grows with the number of iterations.
(iii) TC’s
Effect of frame length of the CTC
error-correcting efficiency rises from one to two iterations. The decoders commu-
nicatedWhat followed was a spectacular error-correcting performance [26] from TC findings.
and shared a lot of information throughout training. Hence, the receiver receives
Conclusions are drawn for frame lengths greater than 65,536 bits. Yet, frame lengths of
relevant input bits to accurately assess the situation. After a few cycles, the decoders may
around 1000 bits are required for video transmission. However, far shorter ones may be
learn the input bits to anticipate. More iterations do not enhance decoder information
Figure 9. Complexity of iterative decoding grows with the number of iterations.
sufficient for uses such as voice transmission. Figure 10 shows how well CTC works for
sharing. Because of this, the BER performance only improved after the ninth decode cycle.
error correction over a variety of frame durations. This process yielded the following sets
(iii)
(iii) Effect of frame length of and
of Effect of frame length of the CTC
digits: 256, 508, 1020, 2048, the CTC
65,536. Increasing the frame length improves the bit
What followed was a spectacular error-correcting performance [26] from TC findings.
error rate performance, as seen in the Figure 10, and this is particularly apparent at higher
What followed was a spectacular error-correcting performance [26] from TC findings.
Conclusions are drawn for frame lengths greater than 65,536 bits. Yet, frame lengths of
Signal-to-Noise Ratios. Increasing the frame length allows the CTC to generate more un-
Conclusions are drawn for frame lengths greater than 65,536 bits. Yet, frame lengths of
around 1000 bits are required for video transmission. However, far shorter ones may be
correlated codewords and carry out efficient decoding due to the increased randomness
around 1000 bits are required for video transmission. However, far shorter ones may be
sufficient for uses such as voice transmission. Figure 10 shows how well CTC works for
in the interleaved sequence.
sufficient for uses such as voice transmission. Figure 10 shows how well CTC works for
error correction over a variety of frame durations. This process yielded the following sets
error correction over a variety of frame durations. This process yielded the following
of digits: 256, 508, 1020, 2048, and 65,536. Increasing the frame length improves the bit
sets of digits: 256, 508, 1020, 2048, and 65,536. Increasing the frame length improves the
error rate performance, as seen in the Figure 10, and this is particularly apparent at higher
bit error rate performance, as seen in the Figure 10, and this is particularly apparent at
Signal-to-Noise Ratios. Increasing the frame length allows the CTC to generate more un-
higher Signal-to-Noise Ratios. Increasing the frame length allows the CTC to generate more
correlated codewords and carry out efficient decoding due to the increased randomness
uncorrelated codewords and carry out efficient decoding due to the increased randomness
in the interleaved sequence.
in the interleaved sequence.
Figure 10. Effect of frame length on the BER performance of CTC.
4.5. Effect of Signal‐to‐Noise Ratio on Image Quality
As a result of channel interference, the sent data is garbled. If there was severe bit
corruption, the quality of the restored picture may be severely reduced. Moreover, if many
bits are corrupted, it is possible that the correct secret information may be obtained
Figure 10. Effect of frame length on the BER performance of CTC.
Figure 10. Effect of frame length on the BER performance of CTC.
Mathematics 2023, 11, x FOR PEER REVIEW
Mathematics 2023, 11, x FOR PEER REVIEW 23 of 26 23 of 26
incorrectly.
incorrectly. Hence,
Hence, we we can examine
can examine the of
the impact impact of channel
channel noise on noise on the quality
the picture picture quality
achieved achieved
with CTC with
by CTC by graphing
graphing the CC
the CC and the and
PSNR the PSNR
with the with the Signal-to-Noise
Signal-to-Noise Ratio Ratio
achieved with CTC by graphing the CC and the PSNR with the Signal-to-Noise Ratio (SNR).
(SNR). The Signal-to-Noise Ratio (SNR) was improved from 0.0 to 1.0. Figure 11a presents
(SNR). The Signal-to-Noise Ratio (SNR) was improved from 0.0 to 1.0. Figure 11a presents
The Signal-to-Noise Ratio (SNR) was improved from 0.0 to 1.0. Figure 11a presents the
the results of plotting PSNR against SNR and CC against SNR. A picture of good quality
the results of plotting PSNR against SNR and CC against SNR. A picture of good quality
results of plotting PSNR against SNR and CC against SNR. A picture of good quality will
will have a CC of at least 0.8 and a PSNR of at least 50 dB. Figure 11b shows that at SNR
will have a CC of at least 0.8 and a PSNR of at least 50 dB. Figure 11b shows that at SNR
have a CC of at least 0.8 and a PSNR of at least 50 dB. Figure 11b shows that at SNR values
values larger than 0.5 dB, the suggested CTC yields a CC of 1 and a PSNR of more than
values larger than 0.5 dB, the suggested CTC yields a CC of 1 and a PSNR of more than
50 dB. The research shows that the CTC can restore a high-quality picture to one that was
larger than 0.5 dB, the suggested CTC yields a CC of 1 and a PSNR of more than 50 dB. The
50 dB. The research shows that the CTC can restore a high-quality picture to one that was
previously distorted by noise of the same strength as the original signal (Table 9, Figure
research shows that the CTC can restore a high-quality picture to one that was previously
previously distorted by noise of the same strength as the original signal (Table 9, Figure
12). 12).
distorted by noise of the same strength as the original signal (Table 9, Figure 12).
Figure 11. Image quality of image recovered by proposed CTC: (a) Correlation Coefficient versus
Figure 11. Image quality of image recovered by proposed CTC: (a) Correlation Coefficient versus
Figure 11. Image quality of image recovered by proposed CTC: (a) Correlation Coefficient versus
SNR; (b) PSNR versus SNR.
SNR; (b) PSNR versus SNR.
SNR; (b) PSNR versus SNR.
Table 9. Simulation Parameters for steam cipher using proposed MCTC.
Table 9. Simulation Parameters for steam cipher using proposed MCTC.
Table 9. Simulation Parameters for steam cipher using proposed MCTC.
Simulation Parameter
Simulation Parameter Type/Value Type/Value
Simulation Parameter Type/Value
FSM Encoder
FSM Encoder With modified Turbo coding
With modified Turbo coding
FSM Encoder
Matrix Generation
Matrix Generation With
3 × 3 modified
3 × 3 Turbo coding
Matrix Generation
Input rate of data
Input rate of data 256 256 3 × 3
Input rate of data
Overall length of the frame
Overall length of the frame 256
16,789 bits
16,789 bits
Overall length of the frame Quadrature Phase Shift Keying (QPSK)
Transduction
Transduction 16,789 bits
Quadrature Phase Shift Keying (QPSK)
Transduction
Configuration Options for the Universal Quadrature Phase Shift Keying (QPSK)
Configuration Options for the Universal
Configuration Options for the Universal a = 5, b = 2, p = 26,741
a = 5, b = 2, p = 26,741
Random Number Generator (EC-URNG)
Random Number Generator (EC-URNG) a = 5, b = 2, p = 26,741
Random Number Generator (EC-URNG)
Code for a hidden transmitter (Ka)
Code for a hidden transmitter (Ka) 45,671 45,671
Code for a hidden transmitter (Ka) 45,671
Hidden code for the receiver (Kb)
Hidden code for the receiver (Kb) 7894 7894
Hidden code for the receiver (Kb) 7894
G = (15,647,15,689), Km =45,679, K =753, P =
G = (15,647,15,689), Km =45,679, K =753, P =
parameters kept under wraps (G, Km, K, P)
parameters kept under wraps (G, Km, K, P) G = (15,647,15,689), Km = 45,679, K = 753,
parameters kept under wraps (G, Km, K, P) (867,159,786).
(867,159,786).
P = (867,159,786).
(a) (a) (b) (b) (c) (c)
(d) (d) (e) (e) (f) (f)
Figure 12. Stream cypher images with their corresponding histograms, from left to right: (a) The
original picture; (b) The decrypted image by an authorized user; (c) The encrypted image; (d) The
histogram of the original image; (e) The histogram of the decrypted image; (f) The histogram of the
encrypted image.
Mathematics 2023, 11, 2225 23 of 25
Mean Square
Secret key (k) Input Feature Coefficient PSNR
Error
Private Key 1 6.5471 2.2238 1.1582 × 123 6.5266
Private Key 2 6.3258 −2.2252 1.1633 × 123 6.5268
Private Key 3 6.2478 −2.2233 1.1631 × 123 6.525
Private Key 4 6.1023 −2.2232 1.1656 × 123 6.3992
Private Key 5 6.7894 −2.2233 1.1536 × 123 6.5326
5. Conclusions
There are four factors that are crucial to the success of the proposed methods: key
length, key order, block size, and intermediate cypher generation. We use a larger key
than the methods proposed by previous studies, with a key length of 128–512 bits, a key
matrix order of 4 × 4 to 6 × 6, and block sizes of 16 bytes to 36 bytes. Compared to existing
solutions, the proposed approaches are superior since their encryption and decryption
times scale linearly with the key size. Guessing the Plaintext takes a lot longer and becomes
almost impossible without knowing the intermediate encryption. Based on the comparative
analysis provided in the previous chapter, the following is provided. The Plaintext can
be concealed using the proposed techniques; however, recovering the Plaintext without
knowledge of the intermediate encryption and key is incredibly challenging due to the
large key size, the high order of the key matrix, and the large block size of the Plaintext.
Future work on image encryption applications may benefit from the frequent binary to
decimal translation, which increases the strength of the technique. Over time, we want to
enhance this algorithm’s protection of text as well as audio, video, and still photos. Other
statistical methods will be used to assess the keys’ degree of unpredictability.
make use of a Finite State Machine, a recurrence matrix, and the LU decomposition of a
randomly selected non-singular matrix, together with key matrix orders of 44 and 55, key
sizes of 128 and 200 bits, and block sizes of 16 and 25 bytes. By increasing the matrix size
and the number of rounds, further research may make it possible to send more data safely
in parallel. The resulting algorithms are more secure because of this and might potentially
keep sensitive information hidden. The suggested works provide a novel encryption and
decryption approach based on the Genetic Algorithm, making it exceedingly difficult to
guess the key by combining a 128-bit key, a 4-by-4 key matrix, and a Plaintext block size of
16 bytes. More time will be spent theorizing on the Plaintext as a result of this.
Author Contributions: Validation, R.R.B. and Z.B.; Investigation, M.M.H.; Writing–original draft,
A.A.; Writing–review & editing, S.M. All authors have read and agreed to the published version of
the manuscript.
Funding: The authors extend their appreciation to the Deanship of Scientific Research at King Khalid
University for funding this work through a research program under grant number R. G. P. 2/109/43.
Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.
Data Availability Statement: Available online: https://sipi.usc.edu/database/database.php?volume=
misc.
Acknowledgments: The authors extend their gratitude to the deanship of scientific research at King
Khalid University for funding this work through the research groups program under grant number
R. G. P. 2/109/43.
Conflicts of Interest: The authors declare no conflict of interest.
References
1. Ray, P.K.; Dila, G.K.; Patel, B.K. Application of some recurrence relations to cryptography using finite state machine. Int. J.
Comput. Sci. Electron. Eng. 2014, 2, 220–223.
2. Almakhour, M.; Sliman, L.; Samhat, A.E.; Mellouk, A. A formal verification approach for composite smart contracts security
using FSM. J. King Saud Univ. Comput. Inf. Sci. 2023, 35, 70–86. [CrossRef]
3. Alawida, M.; Teh, J.S.; Alshoura, W.H. A New Image Encryption Algorithm Based on DNA State Machine for UAV Data
Encryption. Drones 2023, 7, 38. [CrossRef]
4. He, P.; Tu, Y.; Bao, T.; Sousa, L.; Xie, J. COPMA: Compact and Optimized Polynomial Multiplier Accelerator for High-Performance
Implementation of LWR-Based PQC. IEEE Tran. VLSI Syst. 2023, 31, 596–600. [CrossRef]
5. Polese, S. Strength Evaluation of Cryptographic Primitives to Linear, Differential And Algebraic Attacks. Ph.D. Thesis, Department
of Computer Science Giovanni Degli Antoni, University of Milan, Milan, Italy, 2022.
6. Rashid, M.; Sonbul, O.S.; Zia, M.Y.I.; Kafi, N.; Sinky, M.H.; Arif, M. Large Field-Size Elliptic Curve Processor for Area-Constrained
Applications. Appl. Sci. 2023, 13, 1240. [CrossRef]
7. Hasija, T.; Kaur, A.; Ramkumar, K.R.; Sharma, S.; Mittal, S.; Singh, B. A Survey on Performance Analysis of Different Architectures
of AES Algorithm on FPGA. Mod. Electr. Dev. Commun. Syst. 2023, 948, 39–54.
8. Kumar, A.; Singh, P.; Patro, K.A.K.; Acharya, B. High-throughput and area-efficient architectures for image encryption using
PRINCE cipher. Integration 2023, 90, 224–235. [CrossRef]
9. Alharbi, A.R.; Tariq, H.; Aljaedi, A.; Aljuhni, A. Latency-Aware Accelerator of SIMECK Lightweight Block Cipher. Appl. Sci. 2023,
13, 161. [CrossRef]
10. Abebe, A.T. Lightweight and Efficient Architecture for AES Algorithm based on FPGA. ISEL Acad. J. Electr. Telecommun. Comp.
2023, 8, 8.
11. Nair, M.; Sadhukhan, R.; Mukhopadhyay, D. Generating Secure Hardware using ChatGPT Resistant to CWEs, Cryptol. ePrint
Arch. 2023. (Preprint). Available online: https://eprint.iacr.org/2023/212 (accessed on 20 February 2023).
12. Alhomoud, A.; Jamal, S.S.; Altowaijri, S.M.; Ayari, M.; Alharbi, A.R.; Aljaedi, A. Large Field-Size Throughput/Area Accelerator
for Elliptic-Curve Point Multiplication on FPGA. Appl. Sci. 2023, 13, 869. [CrossRef]
13. Huang, X.; Dong, Y.; Ye, G.; Shi, Y. Meaningful image encryption algorithm based on compressive sensing and integer wavelet
transform. Front. Comp. Sci. 2023, 17, 173804. [CrossRef]
14. Kamil Khudhair, S.; Sahu, M.; KR, R.; Sahu, A.K. Secure Reversible Data Hiding Using Block-Wise Histogram Shifting. Electronics
2023, 12, 1222. [CrossRef]
15. Shi, Y.Q.; Li, X.; Zhang, X.; Wu, H.T.; Ma, B. Reversible data hiding: Advances in the past two decades. IEEE Access 2016, 4,
3210–3237. [CrossRef]
Mathematics 2023, 11, 2225 25 of 25
16. Shi, Y.Q. Reversible Data Hiding. In Digital Watermarking: Third International Workshop, Proceedings of the IWDW 2004, Seoul,
Republic of Korea, 30 October–1 November 2004; Cox, I.J., Kalker, T., Lee, H.K., Eds.; Springer: Berlin/Heidelberg, Germany, 2005.
17. Zhang, X. Reversible data hiding in encrypted image. IEEE Signal Proc. Lett. 2011, 18, 255–258. [CrossRef]
18. Peng, F.; Li, X.; Yang, B. Improved PVO-based reversible data hiding. Digit. Signal Proc. 2014, 25, 255–265. [CrossRef]
19. Dhanda, S.S.; Brahmjit, S.; Poonam, J. Lightweight cryptography: A solution to secure IoT. Wirel. Pers. Commun. 2020, 112,
1947–1980. [CrossRef]
20. Tseng, H.W.; Hsieh, C.P. Prediction-based reversible data hiding. Infor. Sci. 2009, 179, 2460–2469. [CrossRef]
21. Zhang, X. Reversible data hiding with optimal value transfer. IEEE Trans. Multi. 2012, 15, 316–325. [CrossRef]
22. Alshudukhi, K.S.; Khemakhem, M.A.; Eassa, F.E.; Jambi, K.M. An Interoperable Blockchain Security Frameworks Based on
Microservices and Smart Contract in IoT Environment. Electronics 2023, 12, 776. [CrossRef]
23. Tai, W.L.; Yeh, C.M.; Chang, C.C. Reversible data hiding based on histogram modification of pixel differences. IEEE Trans. Circ.
Syst. Vid. Technol. 2009, 19, 906–910.
24. Li, X.; Li, B.; Yang, B.; Zeng, T. General framework to histogram-shifting-based reversible data hiding. IEEE Trans. Image Proc.
2013, 22, 2181–2191. [CrossRef]
25. Kalker, T.O.N.; Willems, F.M. Capacity bounds and constructions for reversible data-hiding. In Proceedings of the 2002 14th
International Conference on Digital Signal Processing Proceedings, DSP 2002, Santorini, Greece, 1–3 July 2002; IEEE: New York, NY,
USA, 2003.
26. Wang, X.; Zhao, M.; Feng, S.; Chen, X. An image encryption scheme using bit-plane cross-diffusion and spatiotemporal chaos
system with nonlinear perturbation. Soft Comput. 2023, 27, 1–18. [CrossRef]
27. Wu, S.T. A Key-Based Multi-Mode Clock-Controlled Stream Cipher for Real-Time Secure Communications of IoT. Electronics
2023, 12, 1076. [CrossRef]
28. Makhloufi, A.E.; Adib, S.E.; Raissouni, N. Highly Efficient Security Level Implementation in Radiation-Tolerance FPGA Using a
Combination of AES Algorithm and Hamming Code: LST-SW Case. Int. J. Electr. Electron. Eng. Telecommun. 2022, 1–12.
29. Almuzaini, K.K.; Shalini, S.; Sindhu, P.M.; Sandeep, K.; Stephen, O.; Prashant, K.S.; Piyush, K.P.; Piyush, K.S. Design and
analysis of energy aware interior gateway routing algorithm with particle swarm optimization. Int. J. Commun. Syst. 2023, e5466.
[CrossRef]
30. Shukla, P.K.; Amer, A.; Piyush, K.P.; Adel, R.A.; Sajjad, S. AES Based White Box Cryptography in Digital Signature Verification.
Sensors 2022, 23, 9444. [CrossRef]
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.