SIGNATURE-BASED BIOMETRIC AUTHENTICATION: A DEEP DIVE INTO DEEP LEARNING APPROACHES
International Research Journal of Modernization in Engineering Technology and Science · 2024
All throughout the globe, biometrics is now being used extensively for identifying and verifying people and their signatures.When it comes to biometric authentication methods, signatures are among the most crucial.The primary uses for a person's distinctive signature are in establishing their identification and verifying the authenticity of important or legally binding documents.A key biometric for financial, administrative, and banking applications is the offline handwritten signature.The research used the CNN as its DL approach.In this study significant insights into signature verification through machine learning.Preprocessing techniques ensure the quality of signature images, while feature extraction via PCA enhances verification efficiency.Overall, this research lays a strong foundation for advancing signature verification systems with broad applications in authentication and fraud detection.Comparative analysis highlights CNN's exceptional accuracy of 99.14%, indicating its superiority over other models like VGG19 and Random Forest with accuracy of 84% and 97%, respectively.The technology can be utilised for trustworthy and effective signature verification in a variety of applications.