Off-line Signature Verification and Forgery Detection Using Pixel Based Fuzzy Logic Technique

A simple, passive scheme to achieve the offline signature verification and forgery detection is proposed. This approach is based on the fuzzy technique of the Takagi-Sugeno model [12]. The ultimate goal of fuzzy logic is to deal with reasoning, inexact or fuzzy concepts. The parameters considered in this approach are the pixel comparison of the signatures, pixel area of the signature and the distance taken from the centre of the image. From these parameters, the efficiency rates are calculated. Based on these efficiency rates, the rule based fuzzy technique is framed to evaluate the testing process [1],[10]. The trained genuine specimen signatures are compared with the tested signatures, based on this algorithm; the signature is accepted or rejected. The experimental results show that the proposed scheme indentifies the two types of forgeries: unskilled and skilled forgeries effectively. The accuracy rate of our proposal is found to be about 85 to 89%.