Offline Handwritten Signature Forgery Detection using Deep Learning Methods

Abstract

Offline signature verification is one of the most challenging tasks in biometric authentication. Despite recent advances in this field using image recognition and deep learning, there are many remaining things to be explored. The most recent technique, which is Siamese Convolutional Neural Network, has been used a lot in this field and has achieved great results. In this thesis, we develop an architecture that combines the power of Siamese Triplet CNN and a stack Fully connected neural network for binary classification to automatically verify genuine and forgery signatures even if the forged signature is highly skilled. In the challenging public dataset for signature verification BHSig260, our model can achieve a low FAR = 13.66, which is slightly better than the SigNet model. Once the final model is trained, the one-shot learning should make it possible to determine if the input image is genuine or fraudulent just from one base image. Therefore, our model is expected to be extremely suitable for practical problems, such as banking systems or mobile authentication applications..., in which the amount of data for each identity is limited in quantity and variety.