MULTISCRIPT PATTERN RECOGNITION AND HANDWRITTEN SIGNATURE VERIFICATION SYSTEM FOR FORENSIC DOCUMENT EXAMINATION
Keywords:
Multiscript Pattern Recognition, Handwritten Signature Verification, Forensic Document Examination, Machine Learning, Neural Networks, Support Vector MachinesAbstract
Background: In the realm of forensic document examination, the accurate recognition and verification of handwritten signatures across multiple scripts present significant challenges. Multiscript pattern recognition is crucial for authenticating documents in diverse linguistic contexts. This study explores the development of a multiscript pattern recognition and handwritten signature verification system tailored for forensic applications. Methods: The proposed system integrates advanced image processing techniques, machine learning algorithms, and a multiscript database to analyze and verify handwritten signatures. The preprocessing phase involves noise reduction, normalization, and feature extraction. The system employs a combination of neural networks and support vector machines (SVM) for pattern recognition, capable of handling different scripts and languages. The signature verification process includes dynamic and static feature analysis, ensuring comprehensive coverage of variations in handwriting. Results: The system was evaluated on a dataset comprising signatures from multiple scripts, including Latin, Devanagari, Arabic, and others. The pattern recognition module achieved an accuracy rate of 92%, while the signature verification module demonstrated an accuracy of 89% in distinguishing genuine signatures from forgeries. The multiscript capability of the system was particularly effective in environments with diverse linguistic representations, enhancing the reliability of forensic examinations. Conclusion: The developed multiscript pattern recognition and handwritten signature verification system shows promise for forensic document examination, offering a robust tool for verifying signatures across various scripts. The integration of machine learning techniques and multiscript databases significantly improves the accuracy and reliability of forensic analyses, making it a valuable asset in legal and investigative contexts
References
Choi, S., & Kim, H. (2022). Multiscript handwritten signature verification using deep convolutional neural networks. Pattern Recognition Letters, 159, 148-155. https://doi.org/10.1016/j.patrec.2022.05.011
Gupta, R., & Sharma, S. (2021). A survey of multiscript text recognition and handwriting verification systems. Journal of Forensic Sciences, 66(4), 1324-1337. https://doi.org/10.1111/1556-4029.14887
Verma, S., & Sharma, R. (2023). An overview of handwritten signature verification methods: A forensic perspective. International Journal of Forensic Sciences, 9(2), 77-89. https://doi.org/10.1016/j.ijfs.2023.100250
Zhao, Y., & Zhang, H. (2021). Handwritten signature verification using deep feature extraction and classification techniques. Computer Vision and Image Understanding, 207, 103199. https://doi.org/10.1016/j.cviu.2020.103199
Pereira, J., & Fonseca, D. (2020). Multiscript handwritten document analysis and verification: A review. Artificial Intelligence Review, 53(3), 1691-1720. https://doi.org/10.1007/s10462-019-09718-5
Joshi, S., & Agrawal, S. (2022). A comparative study of signature verification techniques using feature extraction and classification. Journal of Computer Vision and Image Processing, 5(1), 43-59. https://doi.org/10.1016/j.jcvip.2021.12.004
Kumar, P., & Jha, P. (2022). Signature verification in multiscript environments: Advances and challenges. Forensic Science International: Digital Investigation, 40, 301-314. https://doi.org/10.1016/j.fsidi.2022.300167
Bhavani, S. D., & Bharathi, R. K. (2024). A multi-dimensional review on handwritten signature verification: strengths and gaps. Multimedia Tools and Applications, 83(1), 2853-2894.
Diaz, M., Ferrer, M. A., Impedovo, D., Malik, M. I., Pirlo, G., & Plamondon, R. (2019). A perspective analysis of handwritten signature technology. Acm Computing Surveys (Csur), 51(6), 1-39.
Longjam, T., Kisku, D. R., & Gupta, P. (2023). Writer independent handwritten signature verification on multi-scripted signatures using hybrid CNN-BiLSTM: A novel approach. Expert Systems with Applications, 214, 119111.
Downloads
Published
Issue
Section
License
Copyright (c) 2024 Ankit Singh, Vaibhav Saran, Meenakshi Mahajan (Author)
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
