A Survey: Handwritten Signature Verification Using Different Methods

Nowadays, with the development of correspondence via the Internet and the crises that occurred after 2019, many institutions in various fields, whether public or private, and banks and financial transactions, began to carry out most of them via the Internet and needed to sign various documents electronically. There are many problems with the signature images, documents, transactions, or files that make the verification process difficult or require preprocessing for signature. One important part of biometric authentication is signature verification, which tries to separate genuine signatures from fraud. Both online and offline signature verification techniques are examined in this survey. In offline systems, properties like size, shape, and texture are extracted for comparison from static representations of signatures. On the other hand, online systems collect dynamic data such as pen pressure, acceleration, and velocity. This paper gives a comprehensive assessment of the current studies and outcomes in the last 8 years in the subject of online and offline handwritten signature verification. It is possible to improve verification accuracy by creating a hybrid system that combines the advantages of both methods. This study investigates cutting-edge strategies to enhance the functionality of signature verification systems, such as machine learning algorithms and feature extraction techniques. The ultimate objective is to create dependable and strong systems that can successfully thwart intricate forgery efforts. Whereas the accuracy in most offline research may reach 100%, and this may be mostly due to the fact that the matching is of one-to-one type, while in online methods is one-to-many. Index Terms— Online signature verification, offline signature verification, genuine, forgeries.