Abstract: Handwritten signature verification is an essential aspect of biometric authentication, commonly applied in sectors like banking, legal affairs, and official documentation. This project presents the development of an intelligent system designed to accurately identify and verify handwritten signatures using machine learning. The system's primary aim is to detect and distinguish between authentic and forged signatures, thereby enhancing identity security.
Our approach integrates a Convolutional Neural Network (CNN) for extracting detailed features from signature images, combined with the MobileNet architecture to maintain a lightweight model that performs well even on devices with limited processing power. Given the natural variation in individual handwriting, signature verification is inherently complex. Our model is built to learn and adapt to these subtle variations effectively.
To further strengthen the system’s performance, we employ data augmentation techniques and leverage transfer learning. These strategies help improve the model’s generalization capabilities, enabling it to perform reliably on previously unseen data. Overall, this work proposes a high-performance, resource-efficient solution for handwritten signature classification and verification.
Keywords: Deep Learning, Convolutional Neural Networks (CNN), Mobile net, Image Processing.
|
DOI:
10.17148/IARJSET.2025.125189