Received: December 7, 2023. Revised: January 23, 2024.

390

Efficient Multi-signature and QR Code Integration for Document Authentication

Using EdDSA-based Algorithm

Mike Yuliana1* Wildan Dharma Walidaniy1

1
Department of Electrical Engineering, Politeknik Elektronika Negeri Surabaya, Surabaya, 60111, Indonesia
* Corresponding author’s Email: [email protected]

Abstract: The previous global pandemic has shifted many of these proposed needs to remote interactions, including
document authentication. The solution for remote document authentication is the use of digital signatures. However,
in reality, most implementations are inefficient and even weaken already secure methods. This study presents a novel
efficient digital signature method based on a multi-signature method that incorporates the edwards-curve digital
signature algorithm (EdDSA) algorithm, a quick-response (QR) code, and the cryptographically secure pseudorandom
number generator (CSPRNG) for key generation. This study assessed the performance efficiency of this method by
running it on a large number of executions in terms of generation, signature generation, and signature generation time
for three signers and a verifier. The results show that this proposed method outperforms all the existing methods that
this study compared across all parameters. In key generation performance assessments, this proposed method stands
out by employing CSPRNG that already been proven for its efficiency and security for cryptographic use. The usage
of the multi-signature method in the EdDSA algorithm has made this proposed method superior in signature generation
and signature verification performance assessment. On paper, EdDSA is the latest algorithm that has surpassed its
predecessors regarding security and efficiency. By using the multi-signature method, this study further improves the
signature generation and signature verification efficiency as signatures only need to be generated and verified once to
represent all signers. With the performance assessment done on this proposed device, this proposed method has an
average improvement of 76.27% across all parameters against the existing method. Additionally, using QR codes in
the method facilitates real-life signature verification by simplifying the scanning process for verifiers. This work
provides a secure and efficient solution to document file authentication.

Keywords: Document authentication, Digital signature, Multi-signature, CSPRNG, EdDSA, QR code.

adoption of digital signatures, especially in academic

1. Introduction institutions [1].
Despite the widespread use of digital signatures,
The global pandemic over the past three years has
their implementation has not been precise in practice.
necessitated interactions without physical contact and
There is a prevailing belief among some individuals
remote engagements. One notable consequence of
that a digital signature is simply a scanned image or
this phenomenon is the need for signing and verifying
text and QR code representation of a handwritten
documents. A signed document signifies its approval
signature [2]. Forgery is a significant risk associated
by the relevant parties. In traditional practice,
with such models. The inconsistency and
documents are signed in writing, often leading to the
susceptibility to fabrication of these items stem from
possibility of physical contact. In addition, the
the fact that their creation and verification processes
traditional signing process is inefficient as documents
both rely on human visual perception [3]. Real digital
must be physically delivered to the signer. Facing
signatures require cryptographic methods within
new situations during the pandemic has prompted a
them, moving beyond the representation of
shift to digital methods, such as the widespread
traditional handwritten signatures.

International Journal of Intelligent Engineering and Systems, Vol.17, No.2, 2024 DOI: 10.22266/ijies2024.0430.32
Received: December 7, 2023. Revised: January 23, 2024. 391

The essence of a digital signature lies in the same value as the digest in the digital signature, it
cryptographic algorithms used. These algorithms means the document has not been modified and can
safeguard the authenticity and integrity of the signed be considered authentic. The public key used also
documents. To accomplish this, the document is allows verification of whether the signature on the
encoded in a distinct and difficult-to-alter format document is indeed from an official entity because
(hash) using intricate mathematical operations [4]. the public key is a unique code for each individual
To better understand how digital signatures work, and is always paired with the privately held private
an understanding of cryptographic techniques is key [11].
necessary. Simply put, cryptography, which derives RSA, one of the first algorithms used in digital
its name from the Greek for "secret writing," is a signature systems, was introduced in 1977 [12]. The
method for generating secret messages. This RSA algorithm is based on the mathematical problem
technique ensures that messages or information can of logarithms, creating a one-way function that is
be securely transmitted without being understood by easy to calculate the result of an input but difficult to
third parties [5]. There are four main goals of reverse (compute the input value). DSA, which was
cryptography [6]: 1) Confidentiality: ensuring data introduced in 1993, also uses discrete logarithm
remains safe and private from unauthorized access; problems in its method but uses a different
2) Integrity: ensuring data is safe from modification mathematical approach. ECC, an algorithm
by unauthorized parties; 3) Authentication: ensuring renowned for its security and efficiency, was
that the sender or origin of data information is from introduced more recently. ECC is based on the
an official source; 4) Non-repudiation: ensuring that elliptic curve discrete logarithm (ECDLP), allowing
the sender of data information cannot deny their the use of smaller key sizes while maintaining
involvement in the transmission. security equivalent to its predecessors.
In a general sense, cryptography can be In real-world conditions, a document often needs
categorized into two distinct types: asymmetric to be signed by several entities before it is declared
cryptography and symmetric cryptography [7]. valid. In digital signatures, such a method is referred
Symmetric cryptography uses the same key for both to as a multi-party signature. Unlike common
encryption and decryption processes. Therefore, the methods where one entity generates one signature, a
sender and receiver of secret data must have the same multi-party signature creates one digital signature
information about the key used. In its implementation, from several entities [13]. There have been many
the channel used to share key information must be studies attempting to create secure and efficient
secure. Examples of symmetric cryptography include multi-signatures. For example, the latest study by
data encryption standard (DES), Blowfish, and Shankar et al. [14] proposed a method where
advanced encryption standard (AES). On the other biometric credentials such as Identity (ID) numbers
hand, asymmetric cryptography uses different keys and fingerprints are hashed to generate a private key.
for the encryption and decryption processes. These Then, the private keys of each user are used to
keys are the private key and the public key. The generate a master private key used to form a digital
private key is kept secret, while the public key is signature with the EdDSA, a family of ECC.
openly shared. digital signature algorithm (DSA), However, this method poses security vulnerabilities.
Rivest-Shamir-Adleman (RSA), and elliptic curve Because of the additional non-random information to
cryptography (ECC) are examples of asymmetric generate the private key, third parties can more easily
cryptography [8]. attack the system. This key generation method is also
Digital signatures themselves are part of considered inefficient due to unnecessary pre-
asymmetric cryptography. As explained earlier, processing.
asymmetric cryptography employs two keys. Since This proposed method aims to tackle this issue by
the digital signature aims for authentication, the using the dedicated random key generation process
private key of the signer is used to create the digital by using the CSPRNG algorithm so that the quality
signature [9]. The public key of the signer is then of randomness cannot be distinguished from pure
used for verification. The hash function also plays a random [15]. Because of this, the security issue can
crucial role in the digital signature. The hash function be fixed, so the attacker cannot predict how the
creates a digest of the document before signing to private key is generated. Using CSPRNG as key
ensure the document's integrity [10]. The document's generation also addresses the efficiency issue by
digest is checked during the verification process. The eliminating unnecessary preprocessing with existing
digest embedded in the digital signature is obtained methods. Additionally, this study proposes to
by decrypting it using the public key of the signer. If implement QR codes in the method to further
the document's digest during verification has the
International Journal of Intelligent Engineering and Systems, Vol.17, No.2, 2024 DOI: 10.22266/ijies2024.0430.32
Received: December 7, 2023. Revised: January 23, 2024. 392

Algorithm 1: Digital signature 2.1 Hash

1 Key Generation:
- Alice generates private key 𝐾𝑝𝑟 using a Hash functions are one of the most important
random generator and its pair public key aspects of authentication systems, especially digital
signatures. Hash is used to produce a fixed-length
𝐾𝑝𝑏
unique string using a one-way function that ensures
- Alice shares the 𝐾𝑝𝑏 publicly and keeps
authenticity [16]. The string commonly referred to as
the 𝐾𝑝𝑟 privately "digest" is utilized to compare messages obtained
2 Signature Generation: from another source with those obtained during the
- Alice creates a digest of message 𝑚 decryption procedure [17]. Secure hash algorithm
using hash function 𝐻(𝑚) (SHA)-2, specifically SHA-256, was used in the
- Alice encrypts the digest 𝐻(𝑚) using existing method to generate the private key of the
cryptographic function 𝐶 along with system [18]. SHA-2 is a family of hash algorithms
private key 𝐾𝑝𝑟 . The output is signature developed by the National Institute of Standards and
𝑆𝑚 that obtained by: Technology (NIST) in 2002 to update SHA-1 [19].
𝑆𝑚 = 𝐶(𝐾𝑝𝑟 , 𝐻(𝑚)) The family includes common hash algorithms such as
- Signed message 𝑚𝑠𝑖𝑔𝑛𝑒𝑑 is formed by SHA-256, SHA-384, and SHA-512. SHA-256
concatenating message 𝑚 and signature produces a digest value that is 256 bits long for each
𝑆𝑚 : input length, making it an appropriate choice for
𝑚𝑠𝑖𝑔𝑛𝑒𝑑 = 𝑚 ∥ 𝑆𝑚 generating the private key of the EdDSA algorithm
which requires a key that is 32 bytes (256 bits) long.
- 𝑚𝑠𝑖𝑔𝑛𝑒𝑑 is ready to be sent to verifier
3 Signature Verification: 2.2 Digital signature
- Bob receives the 𝑚𝑠𝑖𝑔𝑛𝑒𝑑
- Bob calculates his digest from the The main foundation of existing methods is
message using the same hash function digital signatures. Based on encryption methods in
𝐻(𝑚) asymmetric cryptography, digital signatures ensure
- Bob decrypts the received signature the authenticity of documents and provide proof that
using decryption function 𝐷. The result the signer is genuine [20]. In general, digital signature
is decrypted message 𝐷𝑆𝑚 algorithms involve three main operations: key
𝐷𝑆𝑚 = 𝐷(𝐾𝑝𝑏 , 𝑆𝑚 ) generation, signature generation, and signature
verification. Algorithm 1 below provides an example
- Bob compares the decrypted message
of how digital signatures generally work, with Alice
𝐷𝑆𝑚 that contains the digest with the acting as the signer and Bob as the verifier.
digest from his own received message
𝐻(𝑚) 2.3 ECC
- If 𝐷𝑆𝑚 equal to 𝐻(𝑚), the message is
valid In 1985, a new generation of public key
cryptography was introduced, namely ECC
improve efficiency in the signature verification developed by Victor Miller and Neal Koblitz [21].
process. With QR codes, signatures can be easily read ECC was a significant improvement over its
by verifiers. predecessors, such as RSA and DSA, as it provided
This study’s subsequent sections are as follows. better security with smaller key sizes. The smaller
The literature review of the current methodology is key size makes the cryptographic process less time-
detailed in section 2. The proposed method is consuming and more memory efficient [22]. An
described in section 3. The study's evaluation and elliptic curve is a curve with the general form of the
analysis are presented in section 4. Finally, the equation as Eq. (1):
conclusions of this study are in section 5.
𝑦 2 = 𝑥 3 + 𝑎𝑥 + 𝑏 (1)
2. Literature study
With the condition Eq. (2):
This section discusses the fundamental aspects of
existing digital signature method models. This 4𝑎3 + 27𝑏 2 ≠ 0 (2)
section covers cryptographic concepts such as hash,
digital signatures, ECC, multi-signature schemes, This cryptographic process utilizes a mathematical
and a comparison with existing methods. property of elliptic curves on a finite field 𝐺𝐹(𝑝),
International Journal of Intelligent Engineering and Systems, Vol.17, No.2, 2024 DOI: 10.22266/ijies2024.0430.32
Received: December 7, 2023. Revised: January 23, 2024. 393

which is difficult to invert. The general form of an overall security and efficiency. Bisheh-Niasar et al.
elliptic curve over GF(p) is defined as follows Eq. [26] proposed an optimized implementation of
(3): EdDSA. The researcher showcased significant
improvements in execution time and performance
𝑦 2 ≡ 𝑥 3 + 𝑎𝑥 + 𝑏(𝑚𝑜𝑑 𝑝) (3) while maintaining security levels comparable to
advanced encryption standard (AES)-128. The usage
In this case, 𝑝 is a prime number, and the elements in of EdDSA covers the security issue in ECDSA. The
𝐺𝐹(𝑝) are {0, 1, 2, … , 𝑝 − 1}. This property is known method employed a hash function for the key
as the elliptic curve discrete logarithm problem, generation which is the standard method. However, it
creating a one-way function that is both difficult to lacks multi-signature support which has the same
solve and efficient to implement. Therefore, ECC is limitations as the previous methods. The method also
rapidly replacing previous public key algorithms in does not specify the implementation of QR codes
communication networks and systems, including in within it system. G. Shankar et al. [14] propose an
the context of digital signatures. improved multi-signature scheme for ensuring the
authenticity of digital documents. The method
2.4 Multi-signature employed EdDSA for the cryptography algorithm
and includes support for multi-signature features. For
Multi-signature or multi-sig is a protocol in the key generation, the method used the user's
digital signature cryptography that allows multiple biometric credentials, derived from a hash of the
signers to produce one master signature that XOR operation on an Aadhaar number (ID number),
represents their validation of a data or message [23] fingerprint, and a random number to generate the
Verification is done by combining all signers' public private key defined as follows Eq. (4)
keys that have been given publicly. This approach
increases the security and efficiency of using digital 𝑘𝑝𝑟𝑖𝑣 = 𝐻𝑎𝑠ℎ(𝐹𝑝𝑟𝑖𝑛𝑡 ⨁𝐼𝐷⨁𝑅𝑁𝐺) (4)
signatures [24]. To crack a multi-sig, an attacker must
know the private keys of all authorized parties who
However, this method of key generation is not usual
signed the data, a very difficult task. The efficiency
which leads to efficiency and security issues. XOR-
of multi-sig is seen in the verification process. In
ing a biometric number with a random number may
traditional digital signatures, the verifier has to check
not be an optimal approach for key generation, as
all existing signatures with their respective owners.
directly utilizing the random number would suffice.
However, in the multi-sig method, the verifier only
Moreover, incorporating non-random information
needs to check one signature using the master public
into key generation might introduce vulnerabilities,
key.
potentially aiding attackers in key cracking.
2.5 Comparison to existing method Additionally, the paper does not specify support for
QR code implementation, which could be a valuable
Based on the fundamentals above, several papers aspect of document verification and user-friendly
have implemented the digital signature method using interactions.
various attributes. Wellem et al. [25], utilized the Our proposed method introduces several
standard elliptic curve digital signature algorithm improvements over the existing methods. Firstly, we
(ECDSA). single signature method and QR code to proposed to utilize the robust EdDSA digital
address the vulnerability of paper-based academic signature algorithm, known for its superiority over
documents by proposing a document authentication ECDSA [27], addressing security concerns and
system. The QR code embedded in printed improving efficiency. Secondly, our method
documents contains a digital signature, allowing for eliminates the insignificant pre-process in key
subsequent authentication through either document generation by employing CSPRNG for the key
upload or QR code scanning. However, this research generation process that is dedicated to the
has limited features that lead to security and cryptographic field, mitigating security issues present
efficiency issues. The ECDSA usage is already in previous methods and optimizing efficiency.
known to be vulnerable to side-channel analysis Thirdly, we introduce multi-signature support,
attacks. The research does not specify which key adding a layer to security and also the efficiency.
generation technique they use in the method. The Lastly, we propose the integration of QR code
inferior key generator method will lead to support to simplify the user experience during
vulnerability. Furthermore, the lack of support for signature verification. Table 1 provides a brief
multi-signature features will impact the method’s overview of previous research and its comparison
with our method.
International Journal of Intelligent Engineering and Systems, Vol.17, No.2, 2024 DOI: 10.22266/ijies2024.0430.32
Received: December 7, 2023. Revised: January 23, 2024. 394

Table 1. Attributes comparison of the previous method

Multi-
Cryptography
Method Key Generator Signature QR Code support
Algorithm
Support
Wellem et al. [25] ECDSA Not specified – ✓
Bisheh-Niasar et al. [26] ECDSA Hash function – –
The hash of XOR on the
Aadhaar number,
G. Shankar et al. [14] EdDSA ✓ –
fingerprint, and a random
number
Proposed Method EdDSA CSPRNG ✓ ✓

Algorithm 2: EdDSA algorithm

1 Key Generation: 3. Proposed method
Input: None This section describes the main components used
- Create random number for seed
in the efficient multi-signature method as an
(𝑘) using hash improvement over existing methods and its
- Split the hash of 𝑘 into two part, 𝑎 and 𝑏 implementation. This method utilizes EdDSA, which
half left of the seed as 𝑎, half right as 𝑏 is considered the most efficient and secure digital
- Private key (𝑑) = 𝑎 signature algorithm. To provide the key, this study
- Set base point (𝐺) utilizes the CSPRNG which provides an efficient and
- Public key (𝑄) = 𝑑𝐺 secure unpredictable random number. This study also
Output: private key (𝑑), public key (𝑄) integrates QR codes as a signature format to improve
2 Signature Generation: the efficiency of the verification process.
Input: Data (𝑀), private key (𝑑), public
key (𝑄), half right of seed (𝑏) 3.1 EdDSA
- Calculate the hash of 𝑏 and 𝑀
𝑟 = 𝐻𝑎𝑠ℎ(𝑑, 𝑀) EdDSA is one of the latest digital signature
- Find point in 𝑅 algorithms that still belongs to the ECC family.
𝑅 = 𝑟𝐺 EdDSA was developed by Bernstein in 2012 to
- Calculate the hash of Data (𝑀), public
overcome the weaknesses of its predecessor, ECDSA
[28]. EdDSA utilized Edwards curve that is defined
key (𝑄), and point 𝑅
as follows Eq. (5).
ℎ = 𝐻𝑎𝑠ℎ(𝑀, 𝑄, 𝑅)
- Find 𝑠 as component of the digital
𝑎𝑥 2 + 𝑦 2 = 1 + 𝑑𝑥 2 𝑦 2 (5)
signature
𝑠 = (𝑟 + ℎ ∗ 𝑑) 𝑚𝑜𝑑 𝑛
Where 𝑑 is a non-square constant. Edwards curves
Output: Digital signature (𝑅, 𝑠)
excel in terms of their superior efficiency in addition,
3 Signature Verification:
doubling, and tripling operations when compared to
Input: Public key (𝑄), Data (𝑀), and
the Weierstrass form of elliptic curves that are used
Digital signature (𝑅, 𝑠) in ECDSA. EdDSA offers a high increase in
- Get the hash value from the received data
efficiency without compromising the security level
(𝑀), public key (𝑄), and point 𝑅 [29]. ECDSA, its predecessor, was notoriously
ℎ = 𝐻𝑎𝑠ℎ(𝑀, 𝑄, 𝑅) vulnerable to side-channel analysis attacks because it
- Find point (𝑝1) = 𝑠𝐺 relied heavily on the quality of the random number
- Find point (𝑝2) using elliptic curve generator (RNG) algorithm for the cryptographic
calculation process. One of the advantages of EdDSA is that it
𝑝2 = 𝑅 + ℎ𝑄 overcomes this weakness by implementing a hash
- Compare point (𝑝1) and point (𝑝2), if function in the algorithm to replace the RNG [26].
equal the signature is valid. Otherwise, Based on the advantages offered, this research
the signature is invalid proposes using EdDSA as the cryptographic
Output: Valid or Invalid algorithm for the proposed digital signature method.
Algorithm 2 below explains how the EdDSA
cryptographic calculation process works.

International Journal of Intelligent Engineering and Systems, Vol.17, No.2, 2024 DOI: 10.22266/ijies2024.0430.32
Received: December 7, 2023. Revised: January 23, 2024. 395

3.2 CSPRNG first developed by an automotive company in Japan

in 1994 to facilitate tracking of the production
CSPRNG is a function specifically designed to process [31]. Data is read by scanning the QR Code
generate unpredictable random numbers, used for pattern, which is very efficient as it allows the data
security purposes in cryptographic processes. The input process by simply scanning the image.
numbers generated by CSPRNG have a high degree
of unpredictability and are indistinguishable from 3.4 Implementation
true random numbers. An attack involving the known
output of the CSPRNG generator has more than a This study has implemented the proposed method
50% chance of being predicted, making it on a device with the specifications described in Table
mathematically impossible to get an accurate 2. An illustration of the proposed method can be seen
prediction of the numbers to come [30]. Furthermore, in Fig. 1. In this study, the overall process is classified
even if an attacker knows the full CSPRNG algorithm into three main phases: key generation, signature
used, they cannot use backtracking to analyze generation, and signature verification. This method
previously generated numbers due to seed uses three signers as an example of a multi-signature
uncertainty. This makes it resistant to external and methodKey generation is the first step in executing
internal attacks. This study proposes the use of this method. This stage generates a key pair (private
CSPRNG for EdDSA key generation, taking into key and public key) for each signer CSPRNG is used
account the required security qualities. to perform number generation. EdDSA uses 32-byte-
long keys, so this study customized CSPRNG to
3.3 QR code generate values of that size. The random number
generated is used as the private key. The private key
QR Code is a 2D matrix image used as a data is provided to the EdDSA public key generator to
representation in the form of a pattern. QR Code was

Figure. 1 Proposed method

International Journal of Intelligent Engineering and Systems, Vol.17, No.2, 2024 DOI: 10.22266/ijies2024.0430.32
Received: December 7, 2023. Revised: January 23, 2024. 396

obtain the public key. The key pair is then stored by considered invalid, indicating that the signature or
each signer. These key pairs correspond to each other document may have been forged.
and operations performed with one key can only be
efficiently reversed by the other key. The example of 4. Evaluation and results
32 bytes private key generated from CSPRNG and its
In this section, this study evaluates the
public key pair from EdDSA in base64 format are
performance and security of the proposed method.
given as follows Eq. (6).
This evaluation aims to measure the efficiency and
security of the proposed method in comparison with
𝑘𝑝𝑟𝑖𝑣 = "𝑣𝑗𝑆𝐺𝑓𝑁𝐻𝑌9𝑛𝑄/𝐹𝑛𝑥𝑡𝑔𝑎𝑋𝑣𝑜/𝑤𝑦9
existing methods.
+ 𝐾𝑟𝑎𝑤𝑆00𝐶𝑏𝐼𝑙𝑑𝑝/ℎ𝑧4 = "
𝑘𝑝𝑢𝑏 = "5𝐴1𝐷𝑂4𝑅𝑜𝑡𝑌𝑘𝑋8𝑓𝑐𝑎0𝑘55𝑗𝑈𝐹 4.1 Device
𝑓𝑤𝐻𝑤𝑒𝑑𝐹4𝑢𝑅𝐿𝑔𝑑𝑏𝐷54𝑃7𝑤 = " (6)
The performance test was run on a hardware
In the signature generation phase, the objective is device with the specification Intel® Core™ i5-
to generate a digital signature of the document. The 8250U Processor, 8 GB RAM, NVIDIA® GeForce®
private keys of all signers are combined using the MX150 graphics card, and Windows 11 Home 64-Bit
AND operator to get a master private key that operating system (OS). The detailed specifications
represents all signers. This process is defined as can be found in Table 2 below.
follows Eq. (7).
4.2 Performance assessment
𝑚𝑎𝑠𝑡𝑒𝑟 𝑘𝑝𝑟𝑖𝑣 = 𝑘1𝑝𝑟𝑖𝑣 ∧ 𝑘2𝑝𝑟𝑖𝑣 ∧ 𝑘3𝑝𝑟𝑖𝑣 (7) This subsection aims to evaluate how well the
proposed method performs on a real-time application
The digest is calculated by hashing the document. by running the method on multiple execution counts
This digest value is passed to the EdDSA algorithm for 3 signers and a verifier. The findings provide an
along with the master private key to generate a digital understanding of the performance of the method on a
signature. To ease the process of future verification, larger scale in a practical implementation. The
the digital signature is encoded into a QR code format parameters observed in this study include key
and embedded into the signed document. generation, signature generation, and signature
Finally, at the signature verification stage, the generation time.
verifier will check the validity of the signature. The We conducted a performance assessment on our
resulting digest of the hashed document will be device with specifications outlined in Table 2.
compared with the digest of the digital signature. The Various methods were implemented to evaluate their
digest of the digital signature can be obtained by efficiency based on distinct cryptographic attributes,
decrypting the digital signature using the master such as the encryption algorithm used, key generation
public key from the combined public keys of all techniques, and the support for multi-signatures. The
signers. Master public key is obtained with Eq. (8). detailed differences between the various method can
be seen in Table 1.
𝑚𝑎𝑠𝑡𝑒𝑟 𝑘𝑝𝑢𝑏 = 𝑘1𝑝𝑢𝑏 ∧ 𝑘2𝑝𝑢𝑏 ∧ 𝑘3𝑝𝑢𝑏 (8) For comparison, this study compared the
proposed method with the methods proposed by
Verification is considered successful if both digests Wellem et al. [25], which uses the standard ECDSA
have the same value. Otherwise, the signature is single signature method, Bisheh-Niasar et al. [26],
with the EdDSA single signature method, and G.
Table 2. Device specifications Shankar et al. [14], with the EdDSA double signature
Specifications Details method but using Aadhar numbers, fingerprints, and
Model Acer Aspire E5-476G random numbers for the key generation process.
Processor Intel® Core™ i5-8250U In the assessment of key generation performance,
this study analyzes how efficient each method is in
RAM 8 GB DDR4 2400 MHz
generating key pairs, which is one of the fundamental
NVIDIA® GeForce®
Graphics
MX150
aspects in the cryptographic process. As can be seen
in Table 3 and Fig. 2 (a), the proposed method, which
Storage 128 GB M.2 SSD
uses CSPRNG for key generation, has the fastest
OS Windows 11 Home 64-Bit
execution time compared to the previous methods.
Programming Overall, for all execution counts, the
Python 3.9.7
language

International Journal of Intelligent Engineering and Systems, Vol.17, No.2, 2024 DOI: 10.22266/ijies2024.0430.32
Received: December 7, 2023. Revised: January 23, 2024. 397

Table 3. Performance comparison

Execution times
Method Parameter
10 50 100 200 500
Key
0.055086 0.118986 0.236023 0.464312 1.16126
Generation
Signature
0.0246723 0.127484 0.249966 0.506532 1.21894
Wellem et al. Generation
[25] Signature
0.0920871 0.474262 0.967783 2.06469 4.66371
Verification
Overall 0.1718454 0.720732 1.453772 3.035534 7.04391
Key
0.0023415 0.0076932 0.0154026 0.0306648 0.077323
Generation
Signature
0.0016377 0.0076458 0.0155037 0.0310347 0.076934
Bisheh-Niasar et Generation
al. [26] Signature
0.0038067 0.0182367 0.0365517 0.0768912 0.183661
Verification
Overall 0.0077859 0.0335757 0.067458 0.1385907 0.337918
Key
0.0219081 0.0923676 0.188202 0.377695 0.960287
Generation
Signature
0.001122 0.0052923 0.0106201 0.0211665 0.053014
G. Shankar et al. Generation
[14] Signature
0.0012437 0.006081 0.0121651 0.0245325 0.060815
Verification
Overall 0.0242738 0.1037409 0.2109872 0.423394 1.074116
Key
0.0019248 0.0075903 0.0151734 0.0304335 0.075952
Generation
Signature
0.001122 0.0053251 0.0105689 0.0211677 0.052826
Proposed Generation
method Signature
0.0012335 0.0060961 0.0121345 0.0245531 0.061249
Verification
Overall 0.0042803 0.0190115 0.0378768 0.0761543 0.190026

proposed method has an average of 91.79% better only need to generate one master signature
than the latest method [14] for key generation time. representing all three signers, while single-signature
On paper, EdDSA is already more efficient than methods have to generate three signatures for each
ECDSA. The method [14] has a longer execution signer. For all execution counts, the proposed
time because there is preprocessing that is not methods have an average of 95.71% and 92.25%
conventional for key generation, even though they better than [25] and [26], respectively. Both the
use the EdDSA algorithm. The proposed method uses proposed method and [14] have similar performance
CSPRNG which is proven to be very efficient and as they both use the same multi-signature EdDSA
secure for cryptographic processes. method.
For signature generation performance assessment, The final stage of the digital signature method
this study analyzes how efficiently each method process is testing the performance of signature
generates digital signatures. The proposed method, verification. This test analyzes how efficient each
which uses multi-signature EdDSA, has the fastest method is in verifying the pre-generated signatures.
execution time compared to the previous methods, as Table 3 and Fig. 2 (c) show similar results to the
shown in Table 3 and Fig. 2 (b). Again, EdDSA signature generation assessment, where the multiple
outperforms ECDSA in both single-signature and signature method surpasses the single signature
multi-signature methods. All multi-signature method in terms of efficiency. The multi-signature
methods outperform single-signature methods as they

International Journal of Intelligent Engineering and Systems, Vol.17, No.2, 2024 DOI: 10.22266/ijies2024.0430.32
Received: December 7, 2023. Revised: January 23, 2024. 398

1.4
1.2
Time (Second)

1
0.8
0.6
0.4
0.2
0
1 2 3 4 5
Execution Count

ECDSA single-sig [29] EdDSA single-sig [26] EdDSA multi-sig [13] Proposed method

(a)
1.4
1.2
Time (Second)

1
0.8
0.6
0.4
0.2
0
10 50 100 200 500
Execution Count

ECDSA single-sig [29] EdDSA single-sig [26] EdDSA multi-sig [13] Proposed method

(b)
5
4
Time (Second)

3
2
1
0
10 50 100 200 500
Execution Count

ECDSA single-sig [29] EdDSA single-sig [26] EdDSA multi-sig [13] Proposed method

(c)

8
Time (Second)

0
1 2 3 4 5
Execution Count

ECDSA single-sig [29] EdDSA single-sig [26] EdDSA multi-sig [13] Proposed method

(d)
Figure. 2 Performance comparison: (a) key generation, (b) signature generation, (c) signature verification, and (d) overall
performance

International Journal of Intelligent Engineering and Systems, Vol.17, No.2, 2024 DOI: 10.22266/ijies2024.0430.32
Received: December 7, 2023. Revised: January 23, 2024. 399

method only needs to verify one primary signature this study recommends the implementation of QR
that represents the signatures of all signers, while the codes in its method to facilitate the signature
single-signature method needs to verify all three verification process in real-world situations. With the
signatures. The proposed method overcomes all QR code, the signature can be easily read by the
previous methods with an average improvement of verifier through scanning the QR code. This study
98.72% and 67.16% compared to [25] and [26], evaluates the proposed method by performing an
respectively. [14] achieved similar results to this efficiency performance assessment through various
study as it used the same multi-signature EdDSA executions for 3 signers and a verifier. The observed
method used in this study. parameters include key generation, signature
In all three basic processes of digital signature, generation, and signature generation time. The
the proposed method that implements CSPRNG key proposed method can outperform all existing
generation in EdDSA multi-signature signature methods. Against the latest method [14], the
method, surpasses all existing methods in terms of proposed method shows an average improvement of
performance. This clearly shows that the use of 76.27% for all parameters in the performance
CSPRNG is the best choice for key generation as this assessment. This study makes a significant
algorithm is already efficient on paper and has been contribution to the field of document file
established for cryptographic use, in this case used in authentication by presenting an efficient and secure
digital signature methods. Multi-signature methods method.
outperform single signature methods in signature
generation and signature verification because they Nomenclature
only need one master signature to represent all
signers. Therefore, they only need to create and verify • 𝐾𝑝𝑟 : private key
signatures once. It should be noted that this study • 𝐾𝑝𝑏 : public key
previously conducted a performance assessment • 𝑚: the message that will be signed
against the DSA method and a commonly used • 𝐻(. ): hash function to generate digest
algorithm, namely RSA. However, the results were • digest: unique fixed-size representation of data
astronomically inefficient compared to the proposed • 𝐶(. ): encryption function
method. Therefore, this study decided not to include • 𝑆𝑚 : digital signature of the message
them in this proposed analysis, given that in Figure 2 • 𝑚𝑠𝑖𝑔𝑛𝑒𝑑 : signed message, combination of
it is difficult to distinguish between the two methods. message and digital signature
For the overall performance, the proposed method • 𝐷(. ): decryption function
achieved an average of 97.42% and 44.26% against
• 𝐷𝑆𝑚 : decrypted digital signature
[25] and [26], respectively. Meanwhile, against the
latest method [14], the proposed method shows an • 𝑘: random number used as a seed
improvement of 76.27%. This indicates that the • 𝑠𝑒𝑒𝑑: initial value
proposed method offers better efficiency compared to • 𝑎: the left half of the seed
the existing methods, especially in terms of lower • 𝑏: the right half of the seed
execution time on all steps in the digital signature • 𝑑: private key
method. • 𝐺: base point on the elliptic curve
• 𝑄: public key,
5. Conclusion • 𝑀: the message that will be signed
This study proposes an efficient double signature • 𝐻𝑎𝑠ℎ(. ): a hash function
method with QR code integration for document • 𝑟 : the result of hashing the concatenation of
authentication using an EdDSA-based algorithm. private key and message
This proposed method uses CSPRNG for • 𝑅: a point on the elliptic curve
unpredictable, robust, and efficient key generation. • ℎ : the result of hashing the concatenation of
EdDSA was chosen as the digital signature algorithm message, public key, and point R
as it is one of the latest algorithms that offers a high- • 𝑠: digital signature component
security level and superior efficiency compared to • 𝑛: the order of the elliptic curve
previous algorithms. The application of the multi- • 𝑝1: point 1 on elliptic curve
signature method in this study improves the overall • 𝑝2: point 2 on elliptic curve
efficiency, as it integrates all the signer's private and • 𝐺𝐹(. ): Galois Field
public keys. This allows signature generation and
signature verification to be done only once. Finally,

International Journal of Intelligent Engineering and Systems, Vol.17, No.2, 2024 DOI: 10.22266/ijies2024.0430.32
Received: December 7, 2023. Revised: January 23, 2024. 400

Conflicts of interest Scientific and Research Publications (IJSRP),

Vol. 9, No. 3, pp. 576-589, 2019.
The authors declare no conflict of interest [9] Z. A. Saputri, A. Sudarsono, and M. Yuliana,
“E-voting security system for the election of
Author contributions EEPIS BEM president”, In: Proc. of 2017
Conceptualization, M.Y. and W.D.W.; International Electronics Symposium on
methodology, M.Y. and W.D.W.; software, W.D.W; Knowledge Creation and Intelligent Computing
validation, M.Y.; formal analysis, M.Y. and W.D.W; (IES-KCIC), pp. 147–152, 2017.
supervision, M.Y.; funding acquisition, M.Y. [10] R. Kasodhan and N. Gupta, “A New Approach
of Digital Signature Verification based on
References BioGamal Algorithm”, In: Proc. of 3rd
International Conference on Computing
[1] A. A. Santosa, Y. T. Prasetyo, F. Alamsjah, A. Methodologies and Communication (ICCMC),
A. N. P. Redi, and I. Gunawan, “How the pp. 10–15, 2019.
COVID-19 Pandemic Affected the Sustainable [11] W. Lin, “Trends in Data Protection and
Adoption of Digital Signature: An Integrated Encryption Technologies”, Springer Nature
Factors Analysis Model”, Sustainability, Vol. 14, Switzerland, 2023.
No. 7, p. 4281, 2022. [12] R. Imam, Q. M. Areeb, A. Alturki, and F. Anwer,
[2] S. J. Basha, V. S. Veesam, T. Ammannamma, S. “Systematic and Critical Review of RSA Based
Navudu, and M. V. V. S. Subrahmanyam, Public Key Cryptographic Schemes: Past and
“Security enhancement of digital signatures for Present Status”, IEEE Access, Vol. 9, pp.
blockchain using EdDSA algorithm”, In: Proc. 155949–155976, 2021.
of the 3rd International Conference on [13] Q. Feng, K. Yang, M. Ma, and D. He, “Efficient
Intelligent Communication Technologies and Multi-Party EdDSA Signature With Identifiable
Virtual Mobile Networks, pp. 274–278, 2021. Aborts and its Applications to Blockchain”,
[3] G. Maulani, E. A. Nabila, and W. Y. Sari, IEEE Transactions on Information Forensics
“Digital Certificate Authority with Blockchain and Security, Vol. 18, pp. 1937–1950, 2023.
Cybersecurity in Education”, International [14] G. Shankar, L. H. Farhani, P. A. C. Angelin, P.
Journal of Cyber and IT Service Management Singh, A. Alqahtani, A. Singh, G. Kaur, and I. A.
(IJCITSM), Vol. 1, pp. 136-10, 2021. Samori, “Improved Multisignature Scheme for
[4] S. Pramanik and S. K. Bandyopadhyay, Authenticity of Digital Document in Digital
“Signature Image Hiding in Color Image using Forensics Using Edward-Curve Digital
Steganography and Cryptography based on Signature Algorithm”, Security and
Digital Signature Concepts”, In: Proc. of 2nd Communication Networks, Vol. 2023, pp. 1-18,
International Conference on Innovative 2023.
Mechanisms for Industry Applications (ICIMIA), [15] J. P. Arockiasamy, L. E. Benjamin, and R. U.
pp. 665-669, 2020. Vaidyanathan, “Beyond Statistical Analysis in
[5] A. M. Qadir and N. Varol, “A Review Paper on Chaos-Based CSPRNG Design”, Security and
Cryptography”, In: Proc. of 7th International Communication Networks, Vol. 2021, pp. 1-14,
Symposium on Digital Forensics and Security 2021.
(ISDFS), pp. 1–6, 2019. [16] P. P. Pittalia, “A Comparative Study of Hash
[6] S. A. Busafi and B. Kumar, “Review and Algorithms in Cryptography”, International
Analysis of Cryptography Techniques”, In: Proc. Journal of Computer Science and Mobile
of 9th International Conference System Computing (IJCSMC), Vol. 8, pp. 147-152,
Modeling and Advancement in Research Trends 2019.
(SMART), pp. 323–327, 2020. [17] W. D. Walidaniy, M. Yuliana, and H. Briantoro,
[7] M. T. Gençoğlu and M. T. Gençoğlu, “Improvement of PSNR by Using Shannon-
“Importance of Cryptography in Information Fano Compression Technique in AES-LSB
Security”, IOSR Journal of Computer StegoCrypto”, In: Proc. of 2022 International
Engineering (IOSR-JCE), Vol. 21, No. 1, pp. Electronics Symposium (IES), 2022.
65–68, 2019. [18] H. Bensalem, Y. Blaquière, and Y. Savaria,
[8] M. A. A. Shabi, “A Survey on Symmetric and “Acceleration of the Secure Hash Algorithm-
Asymmetric Cryptography Algorithms in 256 (SHA-256) on an FPGA-CPU Cluster Using
information Security”, International Journal of OpenCL”, In: Proc. of 2021 IEEE International

International Journal of Intelligent Engineering and Systems, Vol.17, No.2, 2024 DOI: 10.22266/ijies2024.0430.32
Received: December 7, 2023. Revised: January 23, 2024. 401

Symposium on Circuits and Systems (ISCAS), pp. Signatures”, In: Proc. of 2023 International
1–5, 2021. Electronics Symposium (IES), pp. 238–243,
[19] M. R. Anwar, D. Apriani, and I. R. Adianita, 2023.
“Hash Algorithm In Verification Of Certificate [30] P. Kietzmann, T. C. Schmidt, and M. Wählisch,
Data Integrity And Security”, Aptisi “A Guideline on Pseudorandom Number
Transactions on Technopreneurship (ATT), Vol. Generation (PRNG) in the IoT”, ACM Comput
3, No. 2, pp. 65–72, 2021. Surv, Vol. 54, No. 6, pp. 1–38, 2022.
[20] N. J. G. Saho, E. C. Ezin, B. Watson, E. Badouel, [31] E. Uçak, “Teaching Materials Developed Using
and O. Niang, “Comparative Study on the QR Code Technology in Science Classes”,
Performance of Elliptic Curve Cryptography International Journal of Progressive Education,
Algorithms with Cryptography through RSA Vol. 15, No. 4, pp. 215–228, 2019.
Algorithm”, Colloque Africain sur la Recherche
en Informatique et en Mathématiques Apliquées,
2020.
[21] A. H. Darrel and Menezes, “Elliptic Curve
Cryptography”, Encyclopedia of Cryptography,
Security and Privacy, pp. 1–2, 2019.
[22] L. Nayak and V. Jayalakshmi, “A Study of
Securing Healthcare Big Data using DNA
Encoding based ECC”, In: Proc. of 6th
International Conference on Inventive
Computation Technologies (ICICT), pp. 348–
352, 2021.
[23] G. Maxwell, A. Poelstra, Y. Seurin, and P.
Wuille, “Simple Schnorr multi-signatures with
applications to Bitcoin”, Des Codes Cryptogr,
Vol. 87, No. 9, pp. 2139–2164, 2019.
[24] Y. Xiao, P. Zhang, and Y. Liu, “Secure and
Efficient Multi-Signature Schemes for Fabric:
An Enterprise Blockchain Platform”, IEEE
Transactions on Information Forensics and
Security, Vol. 16, pp. 1782–1794, 2021.
[25] T. Wellem, Y. Nataliani, and A. Iriani,
“Academic Document Authentication using
Elliptic Curve Digital Signature Algorithm and
QR Code”, JOIV : International Journal on
Informatics Visualization, Vol. 6, No. 3, p. 667,
2022.
[26] M. B. Niasar, R. Azarderakhsh, and M. M.
Kermani, “Cryptographic Accelerators for
Digital Signature Based on Ed25519”, IEEE
Trans. Very Large Scale Integr VLSI Syst, Vol.
29, No. 7, pp. 1297–1305, 2021.
[27] G. J and S. Koppu, “An empirical study to
demonstrate that EdDSA can be used as a
performance improvement alternative to
ECDSA in Blockchain and IoT”, Informatica,
Vol. 46, No. 2, 2022.
[28] A. H. Darrel and Menezes, “Elliptic Curve
Signature Schemes”, Encyclopedia of
Cryptography, Security and Privacy, pp. 1–3,
2019.
[29] W. D. Walidaniy, M. Yuliana, and H. A.
Darwito, “Enhancing Document Authenticity
with QR Codes and ECC-Based Digital
International Journal of Intelligent Engineering and Systems, Vol.17, No.2, 2024 DOI: 10.22266/ijies2024.0430.32