SIGNATURE FORGERY RECOGNITION

SIGNATURE FORGERY RECOGNITION

¹Shriganesh Bhandari, ²Satyam Shukla, ³Omkar Koyande

¹Student, Computer Engineering, ²Student, Computer Engineering, ³Student, Computer Engineering

¹Prof. Srusthi Jadhav,

¹VPPCOE, Mumbai, India

Abstract:  Signatures are one of the most important techniques for biometric authentication. There are two kinds of signature nowadays, offline(static) and dynamic (online). There are greater distinctive characteristics of offline signatures, but there are less distinctive characteristics of offline signatures. Offline signatures are also harder to check. Furthermore, the most significant downside of offline signatures is that even the most talented signer does not sign the same way. This is called Intra-personal variability. Both checking the offline signatures is a difficult problem for researchers. In this research, we proposed an offline signature verification approach based on the Deep Learning to prevent signature fraud by malicious individuals.

Index Terms – Offline Signature Recognition, Siamese Neural Network, , Contrastive loss, Euclidean Distance.