An Offline Signature Verification Using Deep Convolutional Neural Networks

Abstract

When a biometric authentication technology verifies a handwritten offline signature. Due to the time-variant character of signatures, signature verification is a difficult process. There are two primary forms of signatures, such as the dynamic signature used online. A static signature is one that is not currently active. The phrase intra-personal variability is used when an offline signature cannot be made in the same way even by a skilled signer. To prevent fraudulent signatures in this instance, we use a highly deep learning (DL) offline signature verification algorithm. The Convolution Neural Network (CNN) is a component of deep learning (DL). CNN was created and educated for two distinct. For two distinct models, such WI and WD, the writer-independent (WI) and writer-dependent (WD) approaches are the most crucial elements in the signature verification process.