Online and Offline Signature Verification: A Combined Approach

Abstract

Handwritten signature verification is an emerging area. In this paper, an automatic signature verification system has been proposed. This work focuses on both online and offine features of handwritten signatures and aims at combining their results to verify the signature. Signatures are collected for both online and offine. Online data collected is the signing process captured using a webcam and offine data collected are the scanned signatures. Initially both data undergoes appropriate preprocessing steps. Then feature extraction is done where features based on pen tip tracking are used in case of online and gradient and projection based features are used in case of offine method. Later the online and offine method verifies the signature separately and finally their results are combined and the signature is verified using SVM. Paper also compares the results of online, offine and combined approach.