Abstract
Signature verification is one of the biometric techniques frequently used for personal identification. In many commercial scenarios, such as bank check payment, the signature verification process is based on human examination of a single known sample. Although there is extensive research on automatic signature verification, yet few attempts have been made to perform the verification based on a single reference sample. In this paper, we propose an off-line handwritten signature verification method based on an explainable deep learning method (deep convolutional neural network, DCNN) and unique local feature extraction approach. We use the open-source dataset, Document Analysis and Recognition (ICDAR) 2011 SigComp, to train our system and verify a questioned signature as genuine or a forgery. All samples used in our testing process are collected from a new author whose signatures are not present in the training or other stages. From the experimental results, we get the accuracy between 94.37% and 99.96%, false rejection rate (FRR) between 5.88% and 0%, false acceptance rate (FAR) between 0.22% and 5.34% in our testing dataset.
Highlights
In forensic labs, handwriting examination is mainly performed by trained experts, and the result is primarily based on their experience and domain knowledge
We propose an off-line handwritten signature verification method using a single known genuine signature
Deep learning-based methods have demonstrated their great capability for signature verification, most of these classifiers still rely on large-scale datasets to process learning tasks
Summary
In forensic labs, handwriting examination is mainly performed by trained experts, and the result is primarily based on their experience and domain knowledge. Compared to other biometric approaches such as fingerprints, handwritten signatures have relatively high intra-class variability (a high variability between the signatures of a specific person) as well as low inter-class variability (skilled forgeries can be very similar to the genuine one) For these reasons, signature verification has been widely used in the field of personal authentication, it is one of the most challenging issues in biometric technology, especially when we perform signature verification based on a single known sample. Signature verification has been widely used in the field of personal authentication, it is one of the most challenging issues in biometric technology, especially when we perform signature verification based on a single known sample It is because we don’t have enough samples to exclude the high-variability factor. Deep learning-based signature verification methods have achieved great breakthrough in recent years, most of them still require several (more than one) genuine reference signature samples for training the networks [1,2,3,4]
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
