SOMU SRIKANTH REDDY, UPPALAPATI SRIKANTH, SUNKARI RAKESH, Mr.A.T.BARANI VIJAYA KUMAR
Abstract
Signature verification is crucial for authenticating documents in banking, legal, and educational sectors. The paper outlines an automated system that uses machine learning and image processing to identify genuine signatures and detect forgeries. By analyzing unique signature features, the system ensures accurate verification, enhancing security and efficiency in document authentication processes. The proposed method employs Histogram of Oriented Gradients (HOG) to capture distinctive features reflecting the shape and orientation of elements in signature images. A Random Forest (RF) classifier, utilizing an ensemble approach, is trained on these features to improve accuracy and mitigate overfitting. Comprehensive preprocessing is conducted, including converting images to grayscale and resizing them for consistency. Hyperparameter tuning via GridSearchCV optimizes the model’s performance, yielding accuracies of 99.70% on the CEDAR dataset, 99.59% on the BHSig-B dataset, and 98.94% on the BHSig-H dataset. Precision, recall, and F1-scores surpass 94% for CEDAR and BHSigB, and 98% for BHSig-H. The results highlight the model’s effectiveness in identifying skilled forgeries, with strong generalization and computational efficiency across these datasets. This approach is chosen for its reduced computational demand compared to prior methods, driven by HOG’s efficient feature extraction and RF’s optimized ensemble framework, which together minimize processing time while delivering high accuracy.
Keywords
Signature verification, machine learning, image processing, Histogram of Oriented Gradients (HOG), Random Forest, GridSearchCV, biometric authentication, forgery detection, digital security
Conclusion
This study demonstrates the effectiveness of a machine learning-based pipeline for offline signature verification. By leveraging Histogram of Oriented Gradients (HOG) for feature extraction and Random Forest for classification, the system achieves strong accuracy in distinguishing between genuine and forged signatures. The integration of preprocessing, class balancing using SMOTE, and hyperparameter tuning with GridSearchCV ensures a consistent and optimized workflow.
The model’s performance is validated through key evaluation metrics and confusion matrix analysis, confirming its generalization to unseen data. With its modular design and joblib-based serialization, the system is scalable and deployable for real-time applications in domains like banking, legal verification, and digital identity management.
References
[1]Ji, Longcheng, Wang, Hong, Hou, Junxu, Chen, Zhouping, & Li, Ziyang. (2025). Signature authenticity verification using a Cross-Path Four-Stream Network for preventing disguising frauds. Computers and Electrical
Engineering, 122, 109998. https://doi.org/10.1016/j.compeleceng.2024.109998 [2]Zheng, Lidong, Wu, Da, Xu, Shengjie, & Zheng, Yuchen. (2025). HTCSigNet: A Hybrid Transformer and Convolution Signature Network for offline signature verification. Pattern Recognition, 159, 111146.
https://doi.org/10.1016/j.patcog.2024.111146 [3]Ren, Jian-Xin, Xiong, Yu-Jie, Zhan, Hongjian, & Huang, Bo. (2023). 2C2S: A two-channel and two-stream transformer based framework for offline signature verification. Engineering Applications of Artificial Intelligence, 118, 105639.
https://doi.org/10.1016/j.engappai.2022.105639 [4]B. H. Shekar, W. Abraham and B. Pilar, “Offline Signature verification using CNN and SVM classifier,” 2022 IEEE 7th International Conference on Recent Advances and Innovations in Engineering (ICRAIE),
MANGALORE, India, 2022, pp. 304-307, doi: 10.1109/ICRAIE56454.2022.10054336.
https://ieeexplore.ieee.org/document/10054336 [5]Li, Huan, Wei, Ping, Ma, Zeyu, Li, Changkai, & Zheng, Nanning. (2024). TransOSV: Offline Signature Verification with Transformers. Pattern
Recognition, 145, 109882. https://doi.org/10.1016/j.patcog.2023.109882 [6]J. Li, Y. Wen and L. He, “SCConv: Spatial and Channel Reconstruction Convolution for Feature Redundancy,” 2023
IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Vancouver, BC, Canada, 2023, pp. 6153-6162, doi: 10.1109/CVPR52729.2023.00596.
https://ieeexplore.ieee.org/document/10204928 [7]P. Wei, H. Li and P. Hu, “Inverse Discriminative Networks for Handwritten Signature Verification,” 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Long Beach, CA, USA, 2019, pp. 5757-5765, doi: 10.1109/CVPR.2019.00591.
https://ieeexplore.ieee.org/document/8954001 [8]Z. Zeng, “Multi-scale Attention-based Individual Character Network for Handwritten Signature Verification,” 2022 3rd International Conference on
Computer Vision, Image and Deep Learning & International Conference on Computer Engineering and Applications (CVIDL & ICCEA), Changchun, China, 2022, pp. 1-5, doi: 10.1109/CVIDLICCEA56201.2022.9824261.
https://ieeexplore.ieee.org/document/9824261 [9]Amjad, Haadia, Goeller, Kilian, Seitz, Steffen, Knoll, Carsten, Bajwa, Naseer, Tetzlaff, Ronald, & Malik, Muhammad Imran. (2024). Block Induced Signature Generative Adversarial Network (BISGAN): Signature
Spoofing Using GANs and Their Evaluation. arXiv. https://doi.org/10.48550/arXiv.2410.06041 [10]Ishfaq, Muhammad, Saadia, Ayesha, Alserhani, Faeiz M., & Gul, Ammara. (2024). Enhancing Security: Infused Hybrid Vision Transformer for Signature Verification. IEEE Access, 12, 137504– 137520.
https://doi.org/10.1109/ACCESS.2024.3447083 [11]Dey, Sounak, Dutta, Anjan, Toledo, J. Ignacio, Ghosh, Suman K., Llados, Josep, & Pal, Umapada. (2017). SigNet: Convolutional Siamese Network for Writer Independent Offline Signature Verification. arXiv. https://doi.org/10.48550/arXiv.1707.02131 [12]Parcham, E., Ilbeygi, M., & Amini, M. (2021). CBCapsNet: A novel writer-independent offline signature verification model using a CNN- based architecture and capsule neural networks. Expert Systems with Applications, 185, 115649. https://doi.org/10.1016/j.eswa.2021.115649 [13]Lai, Songxuan, Jin, Lianwen, & Yang, Weixin. (2017). Toward highperformance online HCCR: A CNN approach with DropDistortion, path signature and spatial stochastic max-pooling. Pattern Recognition Letters, 89, 60–66. https://doi.org/10.1016/j.patrec.2017.02.011 [14]C. Li, F. Lin, Z. Wang, G. Yu, L. Yuan and H. Wang, “DeepHSV: UserIndependent Offline Signature Verification Using Two-Channel CNN,” 2019 International Conference on Document Analysis and Recognition
(ICDAR), Sydney, NSW, Australia, 2019, pp. 166-171, doi: 10.1109/ICDAR.2019.00035.
https://ieeexplore.ieee.org/document/8977952 [15]Xiong, Yu-Jie, & Cheng, Song-Yang. (2021). Attention Based Multiple Siamese Network for Offline Signature Verification. In Document Analysis and Recognition – ICDAR 2021 (pp. 337–349). Lecture Notes in Computer Science, vol 12823. Springer, Cham. https://doi.org/10.1007/978-3-03086334-0_22 [16]X. Cairang et al., “Learning Generalisable Representations for Offline Signature Verification,” 2022 International Joint Conference on Neural Networks (IJCNN), Padua, Italy, 2022, pp. 1-7, doi:
10.1109/IJCNN55064.2022.9892224.
https://ieeexplore.ieee.org/document/9892224 [17]C. Ren, J. Zhang, H. Wang and S. Shen, “Vision Graph Convolutional Network for Writer-Independent Offline Signature Verification,” 2023 International Joint Conference on Neural Networks (IJCNN), Gold Coast, Australia, 2023, pp. 1-7, doi: 10.1109/IJCNN54540.2023.10192006.
https://ieeexplore.ieee.org/abstract/document/10192006 [18]Prajapati, Prakash Ratna, Poudel, Samiksha, Baduwal, Madan, Burlakoti, Subritt, & Panday, Sanjeeb Prasad. (2021). Signature verification using convolutional neuralnetwork and autoencoder. Journal of the Institute of Engineering, 16(1), 33–42.
https://api.semanticscholar.org/CorpusID:237250627 [19]Dosovitskiy, Alexey, Beyer, Lucas, Kolesnikov, Alexander, Weissenborn, Dirk, Zhai, Xiaohua, Unterthiner, Thomas, Dehghani, Mostafa, Minderer, Matthias, Heigold, Georg, Gelly, Sylvain, Uszkoreit, Jakob, & Houlsby, Neil. (2021). An image is worth 16×16 words: Transformers for image recognition at scale. arXiv.
https://doi.org/10.48550/arXiv.2010.11929 [20]Z. Wojna et al., “Attention-Based Extraction of Structured Information from Street View Imagery,” 2017 14th IAPR International Conference on Document Analysis and Recognition (ICDAR), Kyoto, Japan, 2017,
pp. 844-
850, doi: 10.1109/ICDAR.2017.143.
https://ieeexplore.ieee.org/document/8270074 [21]Li, Haoyang, Li, Heng, Zhang, Hansong, & Yuan, Wei. (2021). Blackbox attack against handwritten signature verification with region- restricted adversarial perturbations. Pattern Recognition, 111, 107689.
https://doi.org/10.1016/j.patcog.2020.107689
How to Cite This Paper
SOMU SRIKANTH REDDY, UPPALAPATI SRIKANTH, SUNKARI RAKESH, Mr.A.T.BARANI VIJAYA KUMAR (2026). SIGNATURE VERIFICATION USING IMAGE PROCESSING. International Journal of Computer Techniques, 13(2). ISSN: 2394-2231.
