Online Signature Verification with Neural Networks Classifier and Fuzzy Inference

Compared to physiologically based biometric systems such as fingerprint, face, palm-vein and retina, behavioral based biometric systems such as signature, voice, gait, etc. are less popular and many are still in their infancy. A major problem is due to inconsistencies in human behavior which require more robust algorithms in their developments. In this paper, an online signature verification system is proposed based on neural networks classifier and fuzzy inference. The software has been developed with a robust validation module based on Pearson's cot-relation algorithm in which more consistent sets Of user's signature are enrolled. In this way, more consistent sets of training patterns are used to train the neural network modules based on the popular back-propagation algorithm. To increase the robustness not only the neural network threshold is used for the verification, the time and length of the signature are also calculated. A fuzzy inference module is then set lip to infer the three thresholds for human-like decision outputs. The signature verification system shows better consistency and is more robust than previous designs.