Abstract: In this project, Handwritten signature recognition is a crucial aspect of biometric authentication, widely used in banking, legal, and official documentation processes. This project focuses on developing an intelligent system capable of accurately recognizing and verifying handwritten signatures using machine learning techniques. The primary objective is to distinguish between genuine and forged signatures to ensure secure and reliable identity verification. The proposed model incorporates a CNN, effectively extracting hierarchical features from the signature images, while the Mobile Net architecture ensures the model's lightweight nature and adaptability to various devices, including those with limited computational resources. This paper introduces a robust system for handwritten signature classification employing Convolutional Neural Networks (CNN) and the lightweight Mobile Net architecture, optimizing the accuracy and efficiency of the signature verification process. Signature classification, a challenging task due to the inherent variations and idiosyncrasies in individuals’ handwriting styles, demands a technique that can understand and learn these nuanced differences. Furthermore, our approach leverages data augmentation and transfer learning techniques to enhance the model's generalization capabilities and performance on unseen data.

Keywords: Deep Learning, Convolutional Neural Networks (CNN), Mobile net, Image Processing.

| DOI: 10.17148/IARJSET.2025.125189