A Fuzzy Approach for the Automatic Off-line PersianSignature Verification Problem

This paper proposes a fuzzy approach for Automaticoff-line signature verification. Automatic off-line SV (SignatureVerification) is an active area of research with numerousapplications such as bank check verification, identity recognition,authorization in legal documents, etc. Over the decades twodifferent types of computer based verification techniques arose:On-line and Off-line SV. Due to shape and size variability insignatures of the same writer and difficulty of verificationproblem, we present a fuzzy approach for SV. We use fuzzyapproach for measuring the similarity degree between thereference and test signature using Takagi–Sugeno FuzzyInference System (FIS). Four types of local features (distance,angle, proportion, and distance to Centre of Gravity) areextracted from control points on set of training signatures andthen these features fuzzified for training of FIS. According to theoutput of FIS, we make decision that test signature is forgery orgenuine. Our algorithm is compared with some related worksand the results of comparison are stated in section 7.