Identification of Digital Signature Patterns Based on The CNN Method at Almas’udiyyah Islamic Boarding School

Abstract

This study aims to identify digital signature patterns using the Convolutional Neural Network (CNN) method at Al Mas’udiyyah Islamic Boarding School. Digital signatures are an essential form of authentication in electronic transactions. Using MATLAB, we developed a CNN model to classify signatures and evaluate its accuracy. The dataset comprises images of students' signatures. The research stages included collecting 60 signature images for training data and 30 signature images for testing data, which were then acquired using a scanner. The results show that the Convolutional Neural Network method can recognize each signature image with high accuracy during the testing process. Al Mas’udiyyah Islamic Boarding School frequently requires verification processes for administrative purposes, such as signing attendance sheets and documents. With a CNN-based digital signature verification system, the boarding school can ensure the security and authenticity of signatures automatically, reducing the risk of forged signatures and increasing efficiency. The CNN model developed in this study achieved an accuracy of 86% in identifying genuine and forged signatures.