Fusion of Deep Learning and Classical Learning for Offline Signature Verification

International Journal of Electronics and Communication Engineering

© 2026 by SSRG - IJECE Journal

Volume 13 Issue 3

Year of Publication : 2026

Authors : Bhavani S D, Bharathi R K

10.14445/23488549/IJECE-V13I3P113

How to Cite?

Bhavani S D, Bharathi R K, "Fusion of Deep Learning and Classical Learning for Offline Signature Verification," SSRG International Journal of Electronics and Communication Engineering, vol. 13,  no. 3, pp. 163-174, 2026. Crossref, https://doi.org/10.14445/23488549/IJECE-V13I3P113

Abstract:

As the technology is growing at a faster pace, authentication becomes an integral part of everyday activities to minimize fraudulent activities. Handwritten offline signature verification is considered one of the reliable methods for authenticating an individual’s identity, widely used in financial transactions, legal documents, business contracts, and so on. Handwritten Offline Signature Verification process involves three major phases: signature acquisition, feature extraction, and Verification. Within these stages, feature extraction plays a vital role because the outcome of the verification process is reliant on the quality of features extracted from the signatures. In this study, a hybrid approach based on a convolutional neural network (CNN) is introduced, where a lightweight CNN is employed to extract detailed and discriminative features from signature images. By convention, all deep neural networks extract high-dimensional features, which may contain features that do not contribute to the verification process. To overcome this issue, a correlation coefficient-based feature selection method is used to eliminate all non-contributing features, thereby reducing the computational complexity. In the final stage, the chosen discriminative features are supplied to the voting classifier for discriminating authentic and fake signatures. The five benchmark datasets, including CEDAR, BHSig260 (Bengali and Hindi), MCYT-75, and UTSig, are utilized to evaluate the proposed model. To enhance the model's transparency, the explainable AI technique LIME (Local Interpretable Model Agnostic Explanations) is employed to identify the features that influence the model’s decisions. Additionally, Grad-CAM (Gradient-Weighted Class Activation Mapping) is employed to visualize the specific parts of the signature image from which the CNN derives its features.

Keywords:

SSV, LIME, Grad-CAM, Voting Classifier, CNN, Correlation Coefficient.

References:

[1] Ansam A. Abdulhussien et al., “A Genetic Algorithm Based One Class Support Vector Machine Model for Arabic Skilled Forgery Signature Verification,” Journal of Imaging, vol. 9, no. 4, pp. 1-26, 2023.
[CrossRef] [Google Scholar] [Publisher Link]
[2] Fadi Mohammad Alsuhimat, and Fatma Susilawati Mohamad, “A Hybrid Method of Feature Extraction for Signatures Verification Using CNN and HOG a Multi-Classification Approach,” IEEE Access, vol. 11, pp. 21873-21882, 2023.
[CrossRef] [Google Scholar] [Publisher Link]
[3] Samy Bakheet, and Ayoub Al-Hamadi, “Automatic Detection of COVID-19 using Pruned GLCM-Based Texture Features and LDCRF Classification,” Computers in Biology and Medicine, vol. 137, pp. 1-10, 2021.
[CrossRef] [Google Scholar] [Publisher Link]
[4] Debanshu Banerjee et al., “A New Wrapper Feature Selection Method for Language-Invariant Offline Signature Verification,” Expert Systems with Applications, vol. 186, 2021.
[CrossRef] [Google Scholar] [Publisher Link]
[5] Faiza Eba Batool et al., “Offline Signature Verification System: A Novel Technique of Fusion of GLCM and Geometric Features using SVM,” Multimedia Tools and Applications, vol. 83, pp. 14959-14978, 2024.
[CrossRef] [Google Scholar] [Publisher Link]
[6] Aman Singla, and Ajay Mittal, “Exploring Offline Signature Verification Techniques: A Survey Based on Methods and Future Directions,” Multimedia Tools and Applications, vol. 84, pp. 2835-2875, 2025.
[CrossRef] [Google Scholar] [Publisher Link]
[7] Subhash Chandra, and Sushila Maheshkar, “Verification of Static Signature Pattern based on Random Subspace, REP Tree and Bagging,” Multimedia Tools and Applications, vol. 76, pp. 19139-19171, 2017.
[CrossRef] [Google Scholar] [Publisher Link]
[8] Moises Diaz et al., “A Perspective Analysis of Handwritten Signature Technology,” ACM Computing Surveys, vol. 51, no. 6, pp. 1-39, 2019.
[CrossRef] [Google Scholar] [Publisher Link]
[9] Luiz G. Hafemann, Robert Sabourin, and Luiz S. Oliveira, “Offline Handwritten Signature Verification—Literature Review,” 2017 Seventh International Conference on Image Processing Theory, Tools and Applications (IPTA), Montreal, QC, Canada, pp. 1-8, 2017.
[CrossRef] [Google Scholar] [Publisher Link]
[10] M. Muzaffar Hameed et al., “Machine Learning-Based Offline Signature Verification Systems: A Systematic Review,” Signal Processing: Image Communication, vol. 93, 2021.
[CrossRef] [Google Scholar] [Publisher Link]
[11] Maryam Houtinezhad, and Hamid Reza Ghaffary, “Writer-Independent Signature Verification based on Feature Extraction Fusion,” Multimedia Tools and Applications, vol. 79, pp. 6759-6779, 2020.
[CrossRef] [Google Scholar] [Publisher Link]
[12] Fu-Hsien Huang, and Hsin-Min Lu, “Multiscale Feature Learning Using Co-Tuplet Loss for Offline Handwritten Signature Verification,” arXiv preprint, pp. 1-16, 2025.
[CrossRef] [Google Scholar] [Publisher Link]
[13] Anamika Jain, Satish Kumar Singh, and Krishna Pratap Singh, “Signature Verification using Geometrical Features and Artificial Neural Network Classifier,” Neural Computing and Applications, vol. 33, pp. 6999-7010, 2021.
[CrossRef] [Google Scholar] [Publisher Link]
[14] Huan Li et al., “TransOSV: Offline Signature Verification with Transformers,” Pattern Recognition, vol. 145, 2024.
[CrossRef] [Google Scholar] [Publisher Link]
[15] Teressa Longjam, Dakshina Ranjan Kisku, and Phalguni Gupta, “Multi-Scripted Writer Independent Off-line Signature Verification using Convolutional Neural Network,” Multimedia Tools and Applications, vol. 82, pp. 5839-5856, 2023.
[CrossRef] [Google Scholar] [Publisher Link]
[16] Yusnur Muhtar et al., “FC-ResNet: A Multilingual Handwritten Signature Verification Model Using an Improved ResNet with CBAM,” Applied Sciences, vol. 13, no. 14, pp. 1-15, 2023.
[CrossRef] [Google Scholar] [Publisher Link]
[17] Ebrahim Parcham, Mahdi Ilbeygi, and Mohammad Amini, “CBCapsNet: A Novel Writer-Independent Offline Signature Verification Model using a CNN-based Architecture and Capsule Neural Networks,” Expert Systems with Applications, vol. 185, 2021.
[CrossRef] [Google Scholar] [Publisher Link]
[18] Sourodeep Roy et al., “Offline Signature Verification System: A Graph Neural Network based Approach,” Journal of Ambient Intelligence and Humanized Computing, vol. 14, pp. 8219-8229, 2023.
[CrossRef] [Google Scholar] [Publisher Link]
[19] Muhammad Sharif et al., “A Framework for Offline Signature Verification System: Best Features Selection Approach,” Pattern Recognition Letters, vol. 139, pp. 50-59, 2020.
[CrossRef] [Google Scholar] [Publisher Link]
[20] B.H. Shekar, Wincy Abraham, and Bharathi Pilar, “Offline Signature Verification using CNN and SVM Classifier,” 2022 IEEE 7th International Conference on Recent Advances and Innovations in Engineering (ICRAIE), Mangalore, India, pp. 304-307, 2022.
[CrossRef] [Google Scholar] [Publisher Link]
[21] B. Venkata Sivaiah et al., “Signatures Verification using CNN and HOG including Voting Classifier,” Proceedings of the International Conference on Computational Innovations and Emerging Trends, vol. 112, pp. 598-608, 2024.
[CrossRef] [Google Scholar] [Publisher Link]
[22] Yizhen Wang, Jianbin Zheng, and Yiwen Zhou, “An Efficient Offline Signature Verification Method Based on Improved Feature Extraction,” 2022 2nd International Conference on Computer Science, Electronic Information Engineering and Intelligent Control Technology (CEI), Nanjing, China, pp. 609-612, 2022.
[CrossRef] [Google Scholar] [Publisher Link]
[23] Zhaoya Wang et al., “Advances in Offline Handwritten Signature Recognition Research: A Review,” IEEE Access, vol. 11, pp. 120222 120236, 2023.
[CrossRef] [Google Scholar] [Publisher Link]
[24] Xiong Yu-Jie et al., “Past: Pairwise Attention Swin Transformer for Offline Signature Verification,” SSRN, pp. 1-14, 2024.
[CrossRef] [Google Scholar] [Publisher Link]
[25] Lidong Zheng et al., “HTCSigNet: A Hybrid Transformer and Convolution Signature Network for offline Signature Verification,” Pattern Recognition, vol. 159, 2025.
[CrossRef] [Google Scholar] [Publisher Link]
[26] Yiwen Zhou et al., “Handwritten Signature Verification Method Based on Improved Combined Features,” Applied Sciences, vol. 11, no. 13, pp. 1-14, 2021.
[CrossRef] [Google Scholar] [Publisher Link]
[27] R.S.A. Zneit, “On-Line Handwriting Signature Verification Based on Using Extreme Points Extraction,” Engineering, Technology & Applied Science Research, vol. 6, no. 4, pp. 1084-1088, 2016.
[CrossRef] [Google Scholar] [Publisher Link]
[28] Xingbiao Zhao et al., “Fusing Deep and Hand-Crafted Features by Deep Canonically Correlated Contractive Autoencoder for Offline Signature Verification,” Pattern Recognition, vol. 168, 2025.
[CrossRef] [Google Scholar] [Publisher Link]