Abstract:
A wide variety of security applications employ Biometrics technology in order to identify or verify an individual based on their physiological or behavioral traits. Physiological refers to the human features such as iris, fingerprint etc., while Behavioral refers to the artificially created features such as voice and handwritten signature.
One of the above-mentioned traits, Handwritten Signature, has been used for various security-based applications such as accessing computer terminals, establishing authorization for performing automated transactions, access in the field of banking etc., As such since the task of identification and verification of identity is of the utmost priority in the field of information security, need for automation of the recognition and verification of the handwritten signature is much necessary.
Handwritten Signature verification is divided into two types. Online handwritten signature verification and Offline handwritten signature verification. The Online verification is referred to the process where the signature is acquired through graphic tablets or mobile devices whereas the Offline method is referred to process where the signature is acquired on a paper.
In this paper, the static geometric properties of handwritten signature acquired through the offline method are explored and various related features are extracted for the process of personal identification and verification. Those features are then fed to an Artificial Neural Network for proper training and classification purposes. The results are then compared to the recent and relevant advancements in the field of Offline Handwritten Signature Verification which clearly shows that while the proposed method falls behind the best achievements, it still clearly outperforms other proposed methods in the area.
