Deep learning approaches for online signature authentication: a comparative study of pre-trained CNN models

Deep learning approaches for online signature authentication: a comparative study of pre-trained CNN models M Ranga Swamy http://orcid.org/0009-0004-3741-2268 Vijayalakshmi P V Rajendran Abstract

Authorization is essential for handling document assurance and security. Nowadays, it constitutes one of the top responsibilities for securing information and effectiveness in every domain. Technological advances have made interactions with machinery more effortless. As a result, the demand for authentication for various legitimate causes is growing rapidly. Therefore, biometric-based identification has dramatically accelerated. This is an improvement over the other approaches. The present work is intended to apply convolutional neural networks for mining features and supervised machine-learning techniques to verify handwritten signatures. Raw images of signatures were used to train the CNN models for feature extraction and data augmentation. In the present work, pre-trained CNN models, such as VGG16, Inception-v3, ResNet50, and Xception, were used to separate authentic from fake signatures. Supervised learning methods, including Logistic Regression and SVM, were used to classify features. The test data were obtained from the ICDAR 2011 Signature Dataset. The results obtained from the present work showed a clear improvement over traditional methods over 69 different signatures. VGG-16 with RMSProp achieved an impressive validation accuracy of 83%, demonstrating robustness with minimal overfitting. Compared with existing techniques, the proposed deep learning approach proved to be more accurate and reliable for signature verification.
01 21 2025 03 31 2025 015230 10.1088/crossmark-policy iopscience.iop.org Deep learning approaches for online signature authentication: a comparative study of pre-trained CNN models Engineering Research Express paper © 2025 IOP Publishing Ltd. All rights, including for text and data mining, AI training, and similar technologies, are reserved. 2024-09-27 2025-01-09 2025-01-21 https://iopscience.iop.org/page/copyright https://iopscience.iop.org/info/page/text-and-data-mining 10.1088/2631-8695/ada86d https://iopscience.iop.org/article/10.1088/2631-8695/ada86d https://iopscience.iop.org/article/10.1088/2631-8695/ada86d/pdf https://iopscience.iop.org/article/10.1088/2631-8695/ada86d/pdf https://iopscience.iop.org/article/10.1088/2631-8695/ada86d/pdf https://iopscience.iop.org/article/10.1088/2631-8695/ada86d/pdf https://iopscience.iop.org/article/10.1088/2631-8695/ada86d https://iopscience.iop.org/article/10.1088/2631-8695/ada86d/pdf https://iopscience.iop.org/article/10.1088/2631-8695/ada86d https://iopscience.iop.org/article/10.1088/2631-8695/ada86d/pdf Sensors Roszczewska 10.3390/s24113524 24 2024 Online signature biometrics for mobile devices M 10.1109/ICECAA58104.2023.10212410 866 2023 Online digital cheque signature verification using deep learning approach Egypt. Informatics J. Agrawal 10.1016/j.eij.2022.10.004 24 1 2023 Classification and comparison of ad hoc networks: a review 2024 IEEE Symposium on Security and Privacy (SP) Schorlemmer 1160 2024 10.1109/SP54263.2024.00215 Signing in four public software package registries: quantity, quality, and influencing factors Multimed. Tools Appl. Bhavani 10.1007/s11042-023-15357-2 83 2853 2024 A multi-dimensional review on handwritten signature verification: strengths and gaps Leghari 465 2023 Energies Jose APL 2022 7611 2022 Offline handwritten signature verification using deep neural networks Neural Comput. Appl. Salturk 10.1007/s00521-024-09690-2 36 11311 2024 Deep learning-powered multimodal biometric authentication: integrating dynamic signatures and facial data for enhanced online security Appl. Sci. Kao 10.3390/app10113716 10 11 2020 An offline signature verification and forgery detection method based on a single known sample and an explainable deep learning approach Girshick 10.1109/CVPR.2014.81 580 2014 Rich feature hierarchies for accurate object detection and semantic segmentation Lect. Notes Comput. Sci. (including Subser. Lect. Notes Artif. Intell. Lect. Notes Bioinformatics) Scherer 10.1007/978-3-642-15825-4_10 6354 92 2010 Evaluation of pooling operations in convolutional architectures for object recognition He 10.1109/CVPR.2016.90 770 2016 Deep residual learning for image recognition Zhang 10.1109/ISCID.2016.2033 2 103 2016 Multi-phase offline signature verification system using deep convolutional generative adversarial networks Procedia Comput. Sci. Jahandad 10.1016/j.procs.2019.11.147 161 475 2019 Offline signature verification using deep learning convolutional neural network (CNN) architectures GoogLeNet inception-v1 and inception-v3 Diagnostics Rustam 10.3390/diagnostics12061474 12 1 2022 Incorporating CNN features for optimizing performance of ensemble classifier for cardiovascular disease prediction J. Phys. Conf. Ser. Sharma 10.1088/1742-6596/1969/1/012044 1969 2021 A comprehensive study on offline signature verification Int. J. Acad. Multidiscip. Res. Alajrami 3 39 2019 Verification of handwritten signature using deep learning Mohsen 88 2015 A novel approach for finger vein verification based on self-taught learning International Symposium on Artificial Intelligence and Signal Processing Fayyaz 2015 211 2015 Online signature verification based on feature representation Expert Syst. Appl. Ghosh 10.1016/j.eswa.2020.114249 168 2021 A recurrent neural network based deep learning model for offline signature verification and recognition system Kim 2016 Online signature verification using deep convolutional neural network J. Ambient Intell. Humaniz. Comput. Kim 9 191 2018 Online signature verification using deep neural network with hidden Markov model Sudharshan 10.1109/ICDSIS55133.2022.9915833 2022 39 2022 Handwritten signature verification system using deep learning Procedia Comput. Sci. Poddar 10.1016/j.procs.2020.03.133 170 610 2020 Offline Signature recognition and forgery detection using deep learning IEEE Trans. Inf. Forensics Secur. Menotti 10.1109/TIFS.2015.2398817 10 864 2015 Deep representations for iris, face, and fingerprint spoofing detection African J. Comput. ICT Oladele 7 11 2014 Forged signature detection using artificial neural network Zhang 2016 Signature verification using convolutional neural network Sadkhan 241 2022 Analysis of different types of digital signature Artif. Intell. Rev. Minaee 10.1007/s10462-022-10237-x 56 8647 2023 Biometrics recognition using deep learning: a survey in China Communications Yang 10.23919/JCC.2021.01.014 18 161 2021 Digital signature based on ISRSAC IEEE Access Malik 10.1109/ACCESS.2020.3033848 8 195832 2020 Deepairsig: end-to-end deep learning based in-air signature verification J. Electron. Imaging Sharma 10.1117/1.JEI.31.4.041210 31 41210 2022 Offline signature verification using deep neural network with application to computer vision Appl. Sci. Hoang 10.3390/app11052092 11 1 2021 Improvement for convolutional neural networks in image classification using long skip connection Electron Khan 13 10 2024 FireXplainNet: optimizing convolution block architecture for enhanced wildfire detection and interpretability Int. J. Intell. Syst. Appl. Eng. Swamy 12 3928 2024 Intelligent systems and applications in engineering online signature authentication using pre-trained optimization techniques Liwicki 10.1109/ICDAR.2011.294 1480 2011 Signature verification competition for online and offline skilled forgeries (SigComp2011) Pattern Recognit. Hafemann 10.1016/j.patcog.2017.05.012 70 163 2017 Learning features for offline handwritten signature verification using deep convolutional neural networks ACM Comput. Surv. Diaz 10.1145/3274658 51 1 2019 A perspective analysis of handwritten signature technology Comput. Vis. Image Underst. Deng 10.1006/cviu.1999.0799 76 173 1999 Wavelet-based off-line handwritten signature verification Pal 10.1109/DAS.2016.48 72 2016 Performance of an off-line signature verification method based on texture features on a large indic-script signature dataset Int. J. Adv. Comput. Sci. Appl. Rum 10.14569/IJACSA.2021.0120312 12 102 2021 FishDeTec: a fish identification application using image recognition approach Electron Fathimathul 11 1 2022 A novel method for the classification of butterfly species using pre-trained CNN models CEUR Workshop Proc. Malik 768 26 2011 Evaluation of local and global features for offline signature verification Appl. Sci. Zhou 10.3390/app11135867 11 2021 Handwritten signature verification method based on improved combined features