An intelligent system for Online Signature Verification

Online Signature Verification is one of the most feasible biometric authentication techniques and is being widely researched upon. In this paper, some of the popular approaches towards Online Signature Verification are reviewed, and an abstraction of the general process involved in these approaches has been proposed. The most critical area in this process has identified to be that of feature selection. This is because a vast number of features are available for online signature verification. An implementation of an online signature verification system has been done with a selection of features and has been evaluated with test data available from the first international signature verification competition. The comparison showed that the selected features with the thresholds applied performed reasonably fine. Several lines of experiments with the developed online verification system have been suggested to test for further improved performance.