Genuine Forgery Signature Detection using Radon Transform and K-Nearest Neighbour

Authentication is very much essential in managing security. In modern times, it is one in all priorities. With the advent of technology, dialogue with machines becomes automatic. As a result, the need for authentication for a variety of security purposes is rapidly increasing. For this reason, biometrics-based certification is gaining dramatic momentum. The proposed method describes an off-line Genuine/ Forgery signature classification system using radon transform and K-Nearest Neighbour classifier. Every signature features are extracted by radon transform and they are aligned to get the statistic information of his signature. To align the two signatures, the algorithm used is Extreme Points Warping. Many forged and genuine signatures are selected in K-Nearest Neighbour classifier training. By aligning the test signature with each and every reference signatures of the user, verification of test signature is done. Then the signature can be found whether it is genuine or forgery. A K-Nearest Neighbour is used for classification for the different datasets. The result determines how the proposed procedure is exceeds the current state-of-the-art technology. Approximately, the proposed system's performance is 90 % in signature verification system.