Volume 10, Issue 3, March– 2025 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/25mar342
Signature Verification Using Deep
Learning and CNN
G. Abinesh1; Dr. V. Kavitha2; Prajith. J. V3
(Associate Professor)1
1,2,3
Department of Computer Science with Cognitive Systems, Sri Ramakrishna College of
Arts and Science, Coimbatore.
Publication Date: 2025/03/19
Abstract: Signature verification plays a crucial role in authentication and fraud detection across various domains such as
banking, legal documentation, and digital security. Traditional methods often struggle with intra-class variability, making deep
learning approaches, particularly Convolutional Neural Networks (CNNs), a promising alternative. This study presents a CNN-
based signature verification system that effectively distinguishes between genuine and forged signatures. The proposed model
extracts spatial features from handwritten signatures using multiple convolutional layers, enabling robust feature learning. A
Siamese network architecture is employed to compare signature pairs, utilizing contrastive or triplet loss to enhance verification
accuracy. The system is trained on publicly available signature datasets and evaluated using performance metrics such as
accuracy, precision, recall, and F1-score. Experimental results demonstrate that the CNN-based approach outperforms
traditional feature-based methods, providing improved generalization to unseen signatures. This research highlights the
potential of deep learning in enhancing signature verification reliability while reducing manual effort in forensic analysis. Index
terms: Signature Verification, Convolutional Neural Networks, Deep Learning, Siamese Network, Authentication.
How to Cite: G. Abinesh; Dr. V. Kavitha; Prajith. J. V (2025). Signature Verification Using Deep Learning and CNN. International
Journal of Innovative Science and Research Technology, 10(3), 374-381.
https://doi.org/10.38124/ijisrt/25mar342
I. INTRODUCTION pressure, signing speed, stroke order, and overall structure.
Signature verification systems are designed to analyze and
Signature verification is a crucial process used to compare these attributes to determine the authenticity of a
authenticate an individual's identity by analyzing and signature. The system can be broadly classified into two types:
confirming the legitimacy of their handwritten or electronic static (offline) verification and dynamic (online) verification.
signature. It plays a fundamental role in various sectors, Static verification analyzes scanned or printed signatures,
including banking, legal transactions, business agreements, while dynamic verification captures real-time signing
government records, and digital security, where ensuring the behavior, including pressure sensitivity and movement
authenticity of a signature is vital in preventing fraud and dynamics. These techniques help in distinguishing genuine
unauthorized access. The verification process can be signatures from forgeries, ensuring security and authenticity.
conducted manually by experts trained in handwriting
analysis or through automated systems that use advanced Data Retrieval
technologies like machine learning, artificial intelligence, and The process of signature verification involves multiple
biometric authentication to enhance accuracy and efficiency. stages, starting with data retrieval. Signature samples are
With the increasing reliance on digital platforms, signature extracted from image files or digital input devices, ensuring
verification has evolved significantly from traditional consistency and format standardization. Error handling
methods to modern electronic and biometric techniques. mechanisms are implemented to detect inconsistencies in
Traditional verification methods involve comparing a input data, such as distorted or incomplete signatures. Secure
signature against a known sample, assessing factors such as databases are utilized to store reference samples, providing a
stroke patterns, pressure, spacing, and overall consistency. robust foundation for comparison. By incorporating both
Expert forensic document examiners often perform this task offline and real-time verification methods, the system offers
in legal cases or financial disputes where signature forgery is a versatile approach to authentication.
suspected.
Data Processing
II. TYPES OF SIGNATURE VERIFICATION Data processing is another essential aspect of signature
verification. Once a signature is retrieved, it undergoes pre-
A signature is a unique and personal identifier that can processing, which includes conversion to grayscale, noise
be analyzed based on several key features, including stroke reduction, and image enhancement. These techniques
IJISRT25MAR342 www.ijisrt.com 374
Volume 10, Issue 3, March– 2025 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/25mar342
improve the clarity of the signature, making it easier to extract UI Layer: Facilitates user interaction, allowing users to
crucial features. Feature extraction methods such as Scale- input and retrieve verification results.
Invariant Feature Transform (SIFT), Histogram of Oriented Application Layer: Processes and analyzes signatures
Gradients (HOG), and Local Binary Patterns (LBP) are using machine learning models and feature extraction
applied to identify distinctive characteristics within the techniques.
signature. These features are then normalized to ensure Data Layer: Manages storage, retrieval, and security of
consistency across different samples and conditions. signature records.
User Authentication III. TESTING AND IMPLEMENTATION
User authentication is integrated into the system to
provide additional security. Before submitting a signature for Thorough testing and implementation strategies are
verification, users must be authenticated through secure login crucial for ensuring the reliability of the signature verification
mechanisms. This prevents unauthorized access and ensures system. Unit testing is performed on individual modules, such
that the submitted signature corresponds to a registered as data extraction, feature processing, and authentication
individual. Reference signatures are stored in encrypted algorithms, to validate their functionality. Integration testing
formats, adding an extra layer of security to the verification is conducted to ensure seamless interaction between different
process. components of the system. Additionally, datasets containing
both genuine and forged signatures are used to evaluate the
Verification Algorithms system’s accuracy and robustness. By refining algorithms and
The core of the signature verification system lies in its improving data processing techniques, the system continues
verification algorithms. Convolutional Neural Networks to evolve and improve over time.
(CNNs) and other machine learning models play a crucial role
in comparing signatures. These models are trained using large Impact of Artificial Intelligence on Signature Verification
datasets of genuine and forged signatures, allowing them to Artificial intelligence has revolutionized signature
learn patterns and detect discrepancies effectively. The verification by enhancing the precision and efficiency of
comparison process involves measuring similarity using verification systems. AI-driven models, such as
classifiers and distance metrics, which help determine the Convolutional Neural Networks (CNNs), employ deep
likelihood of a match. Integrating AI-driven techniques learning techniques to analyze vast amounts of signature data
significantly enhances the system’s ability to differentiate and identify discrepancies. These models continuously learn
between authentic and fraudulent signatures with high from new data, improving their accuracy over time. AI-
precision. powered verification systems can quickly distinguish
between genuine and fraudulent signatures, reducing manual
User Interface (Gui) workload and enhancing security.
To ensure ease of use, the system is equipped with a
graphical user interface (GUI) that allows users to upload Security Measures in Signature Verification
signatures and receive instant verification results. The Ensuring the security of signature verification systems
interface provides real-time feedback, highlighting key is crucial in preventing unauthorized access and fraud. Multi-
verification metrics and authentication confidence scores. factor authentication (MFA) is commonly integrated with
Users can also configure settings such as verification signature verification to provide an additional layer of
thresholds and preferred authentication methods, allowing for security. Encryption techniques safeguard stored signature
a customizable experience. The GUI is designed to be data, preventing unauthorized alterations. Furthermore,
intuitive and accessible, catering to both individual users and blockchain technology is being explored to create immutable
enterprise-level applications. signature records, ensuring authenticity and reducing the risk
of forgery.
Database Design
A well-structured database is essential for efficient Ethical and Legal Implications
signature verification. The system employs an optimized Signature verification must adhere to ethical and legal
database architecture that stores signature samples along with guidelines to protect user privacy and data integrity.
user information. Indexing techniques are used to enable Regulatory compliance with frameworks like GDPR and data
quick retrieval and matching, ensuring fast processing times. protection laws ensures that user data is handled securely and
Security measures, including encryption and access controls, transparently. Ethical concerns regarding AI biases and
are implemented to protect sensitive data and prevent signature data misuse must also be addressed through
unauthorized modifications. Logging mechanisms are also responsible AI practices and strict data governance policies.
integrated to maintain records of verification attempts,
providing valuable insights for auditing and analysis. Forecasting
Forecasting plays a significant role in improving
System Architecture verification accuracy by adapting to evolving signature
The system follows a layered architecture, consisting of variations. Signatures naturally change over time due to aging,
three primary components: health conditions, and other factors, which can lead to false
rejections in verification systems. Predictive analytics can
help adjust verification thresholds dynamically, ensuring that
IJISRT25MAR342 www.ijisrt.com 375
Volume 10, Issue 3, March– 2025 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/25mar342
genuine signature variations are accurately identified while troubleshooting, and compliance verification. One of the
maintaining strict security measures against forgeries. This primary components of documentation in signature
adaptability enhances user experience while ensuring the verification is algorithmic workflows. Detailed records of
integrity of the verification system. Another aspect of signature verification algorithms, including feature extraction
forecasting in signature verification involves anticipating techniques, classification models, and verification thresholds,
technological advancements and emerging threats. are essential for system maintenance and upgrades. These
Cybersecurity risks, including deepfake signature attacks and records help developers understand the underlying logic of
AI-generated forgeries, pose significant challenges to the verification system, enabling them to refine algorithms
traditional verification methods. and improve accuracy over time. Additionally, well-
documented workflows facilitate seamless integration with
By leveraging forecasting models, researchers and emerging technologies such as AI-powered fraud detection
developers can anticipate these threats and develop more and biometric authentication.
sophisticated defense mechanisms, such as multimodal
biometric authentication and blockchain-integrated Verification logs form another critical aspect of
verification systems. Predictive analysis also aids in resource documentation. Maintaining comprehensive logs of all
allocation, allowing organizations to optimize their verification attempts allows organizations to track
verification infrastructure based on anticipated demand and authentication history, detect anomalies, and analyze user
potential risks. behavior. These logs include timestamped records of
signature submissions, verification outcomes, confidence
IV. DOCUMENTATION scores, and flagged discrepancies. In case of disputes or
security breaches, verification logs serve as crucial evidence
Proper documentation in signature verification ensures for auditing and forensic investigations.
transparency, reliability, and compliance with security
standards. Comprehensive documentation serves as a Moreover, log data can be utilized for machine learning
reference guide for developers, auditors, security analysts, training, enabling continuous improvement of verification
and regulatory bodies, facilitating smooth system operation, models.
Fig 1 Training and Testing Phases of the CNN-Based Signature Verification System.
System Design Signature Verification: When the user selects the "Verify
Signature" option, they are prompted to upload a scanned
Command-Line Interface (CLI) Design: The system or digitally captured signature. The system then compares
design for signature verification includes a user-friendly this signature with stored reference samples, analyzing
command-line interface (CLI) that guides users through features such as stroke consistency, pressure sensitivity,
the verification process efficiently. and shape alignment to determine authenticity. The result
is displayed, indicating whether the signature is genuine
Main Menu: Upon launching the signature verification or potentially fraudulent.
system, users are presented with a main menu containing Upload Reference Signature: Users can upload reference
options such as: signatures to build a trusted signature database. The
system extracts essential features from the signature and
Verify Signature securely stores them for future verification. This step is
Upload Reference Signature crucial for organizations requiring multi-user
Set Verification Parameters authentication.
Exit
IJISRT25MAR342 www.ijisrt.com 376
Volume 10, Issue 3, March– 2025 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/25mar342
Verification Parameters: Users can configure settings match percentage and authenticity confidence level, is
such as sensitivity levels for matching, error tolerance, clearly presented with color-coded indicators to highlight
and authentication thresholds. Adjusting these parameters genuine and fraudulent results.
allows flexibility based on security requirements and Reference Signature Database Management: A dedicated
application domains. section in the GUI allows users to manage stored reference
Error Handling: The system is designed to handle various signatures. Users can add, update, or delete reference
error scenarios, including: signatures as needed. This functionality ensures that
organizations can maintain an up-to-date database for
Invalid or unreadable signature input improved verification accuracy.
Mismatched signature formats Verification Parameters and Customization: The settings
Database access errors panel provides users with customization options,
Threshold mismatches leading to ambiguous verification including the ability to adjust matching sensitivity, define
results acceptable signature variations, and enable advanced AI-
based verification models. This feature allows users to
V. GRAPHICAL USER INTERFACE tailor the verification process based on security
(GUI) DESIGN requirements and specific application scenarios.
User Feedback and Interaction: The GUI includes
Main Window: The GUI for signature verification is progress indicators during verification processes, ensuring
designed with a structured layout, allowing users to easily users remain informed about ongoing operations.
navigate through various functions. The main window Interactive elements such as tooltips, hover effects, and
contains sections for signature verification, uploading instructional prompts enhance the usability of the system,
reference signatures, and accessing system settings. Clear, reducing errors and improving the overall user experience.
labeled buttons enable smooth interaction, making the Error Handling and Notifications: The system is designed
process intuitive for users. to handle and display errors effectively. Notifications
Signature Upload and Verification Display: Users can appear in a dedicated message area, providing detailed
upload signatures using an interactive drag-and-drop descriptions of issues such as incorrect file formats,
feature or by selecting a file manually. Once uploaded, the unreadable signatures, or database access failures. Color-
system processes the signature and visually displays coded alerts indicate the severity of errors, guiding users
verification results. The verification outcome, including to take corrective actions promptly.
Fig 2 Software Interface for Model Configuration and Simulation
Signature Verification in the Database Context involves cryptographic techniques that allow verification of
Signature verification can be employed in the database the data's origin and prevent unauthorized modifications. In
to ensure data integrity, authenticity, and security. It the context of the Weather_Data table, signature verification
IJISRT25MAR342 www.ijisrt.com 377
Volume 10, Issue 3, March– 2025 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/25mar342
can ensure that recorded weather parameters remain stored data would be detectable, as the modified record’s
unaltered from their original entry. This is particularly hash would not match the original signed hash.
important when weather data is sourced from external APIs,
meteorological agencies, or IoT-based weather sensors, as Timestamp Integrity
unauthorized alterations could mislead decision-makers and The timestamp column in the Weather_Data table
researchers. plays a vital role in verifying the accuracy of stored data. By
ensuring that the timestamp is included in the signed data,
Implementing Digital Signatures signature verification can prevent attackers from modifying
Digital signatures can be integrated within the SQLite historical records without detection.
database design using hashing and asymmetric encryption
techniques. When a weather data record is inserted into the This is especially crucial in applications where real-time
database, a digital signature can be generated using a private weather tracking and predictions rely on accurate
key. This signature can be stored as an additional column in historical data.
the Weather_Data table. When a user retrieves or analyzes
data, the corresponding signature is verified using the Performance Considerations
associated public key to ensure that the data has not been Although digital signatures enhance security, they
tampered with. introduce computational overhead. The process of
A sample schema modification for signature storage might generating, storing, and verifying signatures requires
be: additional processing power and storage. However, SQLite
is designed to handle lightweight database applications
Alter Table Weather_Data Add Column Signature Text; efficiently, and implementing digital signatures with
optimized hashing algorithms ensures minimal performance
Database Operations in Signature Verification impact.
Database operations play a crucial role in
implementing and managing signature verification System Architecture Flow in Signature Verification
mechanisms. These operations include: The system architecture for signature verification
follows a structured flow that ensures the integrity,
Insertion of Signed Data: When a new weather record is authenticity, and security of stored data. The architecture
added, the system generates a digital signature for the consists of several key components that work together to
data using a private key and stores it alongside the record. generate, store, and verify digital signatures efficiently.
Retrieval and Verification: When retrieving weather data,
the system retrieves the stored digital signature and Data Acquisition Layer: The first step in the system
verifies it using the public key to ensure data integrity. architecture involves acquiring weather data from
Updating Records: If an update is necessary, a new trusted sources such as meteorological sensors, third-
signature must be generated to replace the previous one, party APIs, or manual entries by verified personnel. At
ensuring the integrity of modified records. this stage: The data is collected in real-time or batch
Deletion and Logging: Deleted records may be logged in mode. Each entry is formatted according to predefined
an archive along with their signatures to maintain an standards. The collected data undergoes preliminary
auditable history of data changes. validation to check for anomalies.
Batch Processing: Large datasets may require batch Signature Generation Module: Once the weather data is
verification processes to ensure efficient integrity validated, a digital signature is generated using a
checking without excessive computational overhead. cryptographic hashing algorithm combined with
asymmetric encryption. The steps include: Computing a
VI. ROLE OF PUBLIC AND PRIVATE KEYS hash value for the data record (excluding the signature
field). Encrypting the hash using a private key to produce
In asymmetric cryptography, private keys are used to a unique signature. Storing the signature alongside the
sign data, whereas public keys are used to verify the weather record in the database.
signatures. A meteorological agency or data provider could Database Storage Layer: The signed weather data is then
own the private key, ensuring that only their authorized stored securely in the SQLite database. The database
systems can generate valid signatures. On the other hand, operations for this layer include: Insertion of signed data
researchers, analysts, and decision-makers can use the to maintain data integrity. Updating records by
public key to verify that the received data remains unaltered. generating new signatures for modified data. Logging
This guarantees trust in the data's origin and integrity. deletion requests to prevent unauthorized removals.
Data Retrieval and Verification Layer: When users or
Preventing Unauthorized Modifications systems retrieve weather data, the signature verification
Without signature verification, unauthorized parties process is initiated to ensure authenticity. The
might alter weather records for various reasons, such as verification steps include: Retrieving the original data
financial gain (e.g., modifying weather conditions to support and its stored signature. Computing a new hash of the
fraudulent insurance claims) or misinformation campaigns. retrieved data. Decrypting the stored signature using the
By enforcing digital signatures, any attempt to manipulate corresponding public key. Comparing the newly
IJISRT25MAR342 www.ijisrt.com 378
Volume 10, Issue 3, March– 2025 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/25mar342
computed hash with the decrypted hash; if they match, be implemented to detect any unauthorized modifications to
the data is verified as authentic. stored data.
Access Control and Security Enforcement: To prevent
unauthorized modifications and access to sensitive data: Compliance with Regulatory Standards
Role-based access controls (RBAC) are implemented. Different industries have varying compliance
Only authorized personnel can modify data, and requirements regarding data integrity and security.
modifications trigger new signature generation. Logs of Signature verification must align with standards such as:
all access requests and modifications are maintained.
Audit and Compliance Layer: To meet regulatory and GDPR (General Data Protection Regulation)
security standards, the system includes: Regular integrity ISO/IEC 27001 (Information Security Management)
checks via batch signature verifications. Timestamp NIST (National Institute of Standards and Technology)
verification to ensure historical data remains intact. Guidelines
Secure logging of changes to maintain an audit trail.
VIII. SYSTEM TESTING AND
VII. CONSIDERATIONS IMPLEMENTATION
Implementing signature verification requires attention Testing Methodologies:
to several key considerations to ensure robust security and To ensure the effectiveness of signature verification,
efficiency: various testing methodologies should be employed:
Key Management Unit Testing: Individual components such as hashing,
Proper handling of cryptographic keys is essential. encryption, and database storage are tested in isolation.
Private keys must be stored securely to prevent unauthorized Integration Testing: The interaction between different
access, while public keys should be easily accessible for modules, including data acquisition, signature
verification. Key rotation policies should be implemented to generation, and verification, is validated.
mitigate risks of key compromise. Performance Testing: The system is tested for efficiency
under varying loads to ensure that signature verification
Performance Optimization does not introduce excessive delays.
Verifying digital signatures for large datasets can Security Testing: Penetration testing and vulnerability
introduce computational overhead. To optimize assessments are conducted to identify potential security
performance: weaknesses.
Batch verification techniques can be used for bulk data Implementation Strategies
processing. The successful implementation of the signature
Efficient hashing algorithms should be selected to verification system involves:
balance security and speed.
Indexing can be used in databases to improve retrieval Gradual Deployment: Rolling out the verification
times for signature verification. mechanism in phases to ensure smooth adoption.
Data Migration: If existing records need to be signed, a
Handling Data Updates migration strategy should be implemented to generate
Whenever data is modified, a new digital signature and store signatures for historical data.
must be generated to reflect the changes. This ensures that User Training: Ensuring that stakeholders understand the
altered records remain verifiable and trustworthy. However, verification process and can interpret verification results
excessive updates can lead to performance issues, requiring correctly.
optimized update strategies. Monitoring and Maintenance: Regularly monitoring
system performance and updating cryptographic
Ensuring Backward Compatibility techniques to counter emerging threats.
In cases where a signature verification system is added
to an existing database, backward compatibility must be Challenges and Mitigation Strategies
maintained. This can involve:
Computational Overhead: Optimization techniques such
Providing a fallback mechanism for older records as indexing and batch processing can improve
without signatures. performance.
Implementing progressive verification where unsigned Key Management: Secure storage and periodic rotation
records are gradually updated with signatures. of cryptographic keys are essential.
Backward Compatibility: Ensuring that older records
Security Against Replay and Tampering Attacks
without signatures can still be accessed and
Digital signatures should be timestamped to prevent progressively signed over time.
replay attacks where old signatures are reused to forge
authenticity. Additionally, cryptographic measures should
IJISRT25MAR342 www.ijisrt.com 379
Volume 10, Issue 3, March– 2025 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/25mar342
IX. CONCLUSION Adaptive Learning Models:
Implementing systems that learn from new data and
Technological Innovations: adapt to changes in an individual’s signature over time,
Recent advancements in machine learning and deep ensuring long-term reliability.
learning have refined feature extraction and pattern
recognition. These innovations enable systems to Real-Time Verification:
accommodate natural variations in handwriting while Enhancing the speed and efficiency of signature
accurately distinguishing genuine signatures from forgeries. verification systems for real-time applications in online
Integrating multimodal biometrics further strengthens banking, e-commerce, and secure transactions.
overall system reliability.
Cross-Platform Usability:
Persistent Challenges: Despite progress, issues such as Ensuring seamless signature verification across
intra-class variability and sophisticated forgery different devices and platforms, including smartphones,
techniques continue to test the limits of current systems tablets, and PCs.
Operational Challenges: Despite progress, issues such as
intra-class variability, sophisticated forgery techniques, Forgery Detection:
and environmental inconsistencies continue to challenge Developing more advanced techniques to detect
system accuracy. sophisticated signature forgeries, including the use of AI-
driven anomaly detection systems.
Continuous Research and Development Are Essential to
Overcome These Limitations. Cloud-Based Solutions:
Expanding cloud-based signature verification to
Pattern Recognition: Pattern recognition in signature enable easy scalability, remote access, and global
verification refers to the process of analyzing and applications.
distinguishing key characteristics from a person's
signature using advanced algorithms. These algorithms User Experience Improvements:
utilize machine learning and statistical techniques to Balancing security with a seamless, user-friendly
identify consistent features such as stroke sequences, experience, ensuring minimal friction in authentication
pressure variations, and overall structure. By examining processes.
these unique traits, the system can accurately identify
whether a signature is genuine or forged. Leveraging REFERENCES
pattern recognition allows for the detection of subtle
discrepancies that may be missed by traditional methods, [1]. Chatzisterkotis, Thomas (2015) An examination of
improving both accuracy and reliability. It plays a vital quantitative methods for Forensic Signature Analysis
role in creating secure and robust verification systems for and the admissibility of signature verification system
digital identity and transaction protection. as legal evidence. Master of Science by Research
Integrating multimodal biometrics: Integrating (MScRes) thesis, University of Kent,
multimodal biometrics involves combining signature https://kar.kent.ac.uk/id/eprint/54048.
authentication with other biometric methods, like [2]. Dr.V. Thangavel (2023) Use of Digital Signature
fingerprint or facial recognition, to strengthen security. Verification System (DSVS) in various Industries:
By using multiple biometric factors, the system cross- Security to protect against counterfeiting: Research.
checks various identifiers, reducing the risk of false Z-Global Banking eJournal Vol 2, Issue 15.
acceptance or rejection that could occur with a single [3]. Syed Zulkarnain Syed Idrus, Estelle Cherrier,
biometric trait. For example, if someone’s signature Christophe Rosenberger, Jean-Jacques
appears to match but their fingerprint or facial features Schwartzmann. (2013) A Review on Authentication
don’t, the system flags the discrepancy, offering more Methods. Australian Journal of Basic and Applied
reliable authentication. Sciences, 7 (5), pp.95-107. hal-00912435.
[4]. Abhishek Shende, Mahidhar Mullapudi and
Scope for Future Enhancement Narayana Challa, (2024) Enhancing Document
Verification Systems: A Review of Techniques,
Improved Algorithms: Challenges, and Practical Implementations,
Developing more robust and accurate algorithms, International Journal of Computer Engineering and
especially using deep learning, to handle variability in Technology (IJCET),15(1),16-25.
signatures and detect sophisticated forgeries. https://iaeme.com/Home/issue/IJCET?Volume=15&
Issue=1.
Multimodal Biometric Integration: [5]. N. Zaman, I. Karabey Aksakallı, and N. Bayğın,
Combining signature verification with other biometric (2023) “Digital Certificate Security: A Blockchain-
systems (e.g., fingerprints, facial recognition) to increase
based Approach for Fraud Prevention and
security and minimize fraud risk.
Verification”, Bitlis Eren Üniversitesi Fen Bilimleri
Dergisi, vol. 12, no. 4, pp. 1128–1138, doi:
IJISRT25MAR342 www.ijisrt.com 380
Volume 10, Issue 3, March– 2025 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/25mar342
10.17798/bitlisfen.1343747.Wang, C., Li, Q., &
Kim, S. (Eds.).
[6]. Fierrez-Aguilar, J., Krawczyk, S., Ortega-Garcia, J.,
Jain, A.K., 2005b. Fusion of local and regional
approaches for on-line signature verification. IWBRS
2005, 188–196.
[7]. Jain, A.K., Griess, F., Connell, S. (2002). On-line
signature verification. Pattern Recogn. 35, 2963–
2972.
[8]. Kashi, R.S., Hu, J., Nelson, W.L., Turin, W., (1997).
On-line handwritten signature verification using
Hidden Markov Model features. In: Proceedings of
the ICDAR, pp. 253–257.
[9]. Krawczyk, S., (2005). User authentication using on-
line signature and speech. Master’s Thesis, Michigan
State University, Department of Computer Science
and Engineering.
IJISRT25MAR342 www.ijisrt.com 381