Offline Signature Recognition Using Machine Learning

Offline Signature Recognition Using Machine Learning

Mohammad Ikhsan, Bin Zakaria
Gunay, Karli

Biometric behavior can be recognized through the signature behavior of a person. It is mostly used for authorization and authentication in legal documentation papers. Signature recognition has two ways of verification, dynamic or online recognition and static or offline recognition. In this paper we use offline recognition to analyze signature images using Artificial Neural Network. We used mark minutia masking as the feature extraction. We proposed offline signature recognition using machine learning with supervised learning algorithm. The aim of using artificial neural network is to automatically find signatures that match to the owners of the signatures. Based on our evaluation, after we compared feed forward backpropagation and other supervised learning network such cascade-forward network, it revealed cascade-forward shown the highest accuracy100 % with low mean square error 0. Keywords: biometric, offline signature, machine learning

Conference or Workshop Item
PeerReviewed

2012-05-31

1142